THREE ESSAYS ON HEALTH AND EDUCATION IN GUYANA. 

 

BY 

 

MS. NICHOLA LATOYA WILLIAMS 

 

 

 

 

 

 

 

A thesis submitted to the Newcastle University  

for the degree of  

DOCTOR OF PHILOSOPHY 

 

Department of Economics 

Newcastle University Business School 

Newcastle University 

 

September 2023 

 



ii 
 

Abbreviations 

 

BCG  Bacillus Calmette-Guérin vaccine 

CAPE  Caribbean Advanced Proficiency Examination  

CARICOM Caribbean Community and Common Market 

COVID-19 Coronavirus disease 

CHS  Community High Schools 

CPCE  Cyril Potter College of Education 

CSEC   Caribbean Secondary Education 

CSHE   Carnegie School of Home Economics  

DES  Digest of Education Statistics  

DPT  Diphtheria, Pertussis (whooping cough), and tetanus 

EDs  Enumeration districts  

GBOS  Guyana Bureau of Statistics  

GDHS  Guyana Demographic and Health Survey 

GITC   Guyana Industrial Training Centre             

GLFS  Guyana Labour Force Survey 

GoG  Government of Guyana 

GSSs  General Secondary Schools 

GTI   Government Technical Institute 

HAZ  Height-for-age z-scores 

IPCC  Intergovernmental Panel on Climate Change 

MICS  Multiple Indicator Cluster Survey 

MoE  Ministry of Education 

NATI  New Amsterdam Technical Institute  

OLS  Ordinary Least Square  

PIAAC Program for the International Assessment for Adult Competencies 

PSU  Primary Sampling Unit 

UN   United Nations 

UNICEF United Nations International Children's Emergency Fund  

USAID United States Agency for International Development 

USE  Universal Secondary Education 

WHO  World Health Organisation 

SD  Standard Deviation 

SSEE  Secondary School Entrance Examination              

SSPE  Secondary School Proficiency Examination  

 

 

 

 

 

 

 

  



iii 
 

Abstract 

 

This dissertation, through a set of three substantive chapters, examines the effects of weather 

shocks, educational reform and birth order on health and education-related outcomes in 

Guyana. In doing so, this dissertation brings a wealth of new literature for Guyana, which 

remains an underexplored context. Findings from this dissertation have implications for 

enhancing the economic and social well-being of the nation.  

The first chapter examines the effects of in-utero exposure to extreme weather events on early 

childhood health in flood-prone Guyana. Using three waves of the Multiple Indicator Cluster 

Surveys, we find that prenatal exposure to higher-than-average rainfall events or positive 

rainfall shocks (i.e., floods) negatively affects the height-for-age z score and increases the 

likelihood of stunting and severe stunting. Given the importance of exposure to weather shocks 

in early gestation, we examine the impact by trimesters and find that positive rainfall shocks in 

the first trimester matter the most.  

The second chapter examines the effects on educational attainment and labour market 

outcomes of a Universal Secondary Education (USE) policy during 2003-07 for the conversion 

of schools from a pre-vocational track into an academic track. Using data from the Guyana 

Labour Force Surveys and the Digest of Education Statistics, we find that greater exposure to 

the USE leads to an increase in the completion of secondary or tertiary education in Guyana. 

Our findings suggest that the exposure to the USE policy reform had positive effects on the 

completion of secondary or tertiary education and the potential to earn higher wages in Guyana. 

Overall, we find that the USE (2003-2007) policy reform did not close the achievement gaps 

in education across ethnicities. 

The final chapter examines the effects of birth order on early childhood health in Guyana and 

finds that the height-for-age z-score increases more for second-and third-born male children 

when compared to first-born children. We also find that mothers both breastfed and invested 

more in vaccinations for children whose birth order is higher than one. In addition, we find that 

third-born and fourth-born and higher Muslim children tend to have higher birth weights when 

compared to firstborn children.  

 

 



iv 
 

Acknowledgements 

 

It is a genuine pleasure to express my deep gratitude to my supervisors, Dr Sara Maioli and Dr 

Smriti Sharma, who both made completing this dissertation possible. Their continued guidance 

and advice assisted me in completing the various stages of my project. Their timely scholarly 

advice, meticulous scrutiny, and empirical approaches have helped me accomplish this task. I 

would also like to thank my PhD panel, Dr Nils Braakman and Dr John Wildman, who have 

both encouraged me throughout the years.  

I would also like to give social thanks to my family and friends, who all believed in me and 

were my constant support throughout the years. Your prayer, encouragement and support 

sustained me throughout this journey. I am also grateful to my cohort members, especially my 

office mates, for their moral support.  

Warmest thanks to my aunt (Mrs Kim Sundack-Hinds), mom (Ms Camille Sundack), 

grandmother (Ms Francena Cato) and dad (Mr Ian Williams) who have all encouraged, 

supported, and prayed for me throughout this journey.  

Special thanks to Ms Naomi Bonté, Dr Loria-Mae Heywood, Ms Tabitha Akande, Dr Clydon 

Thierens, Dr Shannon Leakey, Dr Tabitha Sindani, Mrs Yemi Oyedeji and Mrs Toshia James 

who have all encouraged and inspired me to keep working hard throughout the years.  

To my late grandfather (Mr George Elmer Sundack) and late uncle (Mr Troy Sundack), the 

memories of your love, support and provision have motivated me to complete this journey. I 

know you are celebrating in heaven.  

Finally, I would like to thank God for being my constant help throughout this PhD journey. I 

have experienced your guidance, love, and provision during this journey. You are the one that 

provided wisdom, health, and strength to sustain me. This PhD is all for your Glory. I will keep 

on trusting you for my future.  

  



v 
 

CONTENTS 

 

List of Figures           viii 

 

List of Tables              ix 

 

 

CHAPTER 1: - Introduction ............................................................................................ 11 

1.1 Background and Motivation .......................................................................................... 11 

1.2 Research Context .......................................................................................................... 13 

1.3 Guyana’s Colonial History and Independence ............................................................... 15 

1.4 Economic Growth and Country Comparison.................................................................. 16 

1.5 Guyana’s Health and Education Sectors ........................................................................ 20 

1.6 Layout of the Dissertation ............................................................................................. 22 

 

CHAPTER 2 The effects of weather shocks on early childhood health in Guyana         23 

2.1 Introduction ................................................................................................................... 23 

2.2 Background on Guyana and its Climate ......................................................................... 28 

2.3 Data Sources and Descriptive Statistics ......................................................................... 30 

2.3.1 Multiple Indicator Cluster Surveys (MICS) ......................................................................... 30 

2.3.2 Rainfall Data ....................................................................................................................... 31 

2.3.3 Summary Statistics.............................................................................................................. 32 

2.4 Empirical strategy ......................................................................................................... 34 

2.5 Results………………………..………...……………..………………………………….37 

2.5.1 Regional Rainfall Condition (1995-2014) and HAZ scores...................................... 37 

2.5.2 The effect of rainfall shocks on postnatal investments ............................................. 43 

2.5.3 Heterogeneous Effects ............................................................................................ 46 

2.5.4 Robustness Checks ................................................................................................. 49 

2.5.5 Placebo Test............................................................................................................ 52 

2.6 Conclusion .................................................................................................................... 53 

References……………………………………………………………………………………55 

 



vi 
 

CHAPTER 3 Effects of a School Conversion Policy on Educational Attainment and 

Labour Market Outcomes in Guyana         61 

3.1 Introduction ................................................................................................................... 61 

3.2 Context.......................................................................................................................... 66 

3.2.1 Guyana’s Education System .................................................................................... 66 

3.2.2 Description of the changes caused by the USE policy reform. ................................. 70 

3.2.3 Guyana’s Labour Market ........................................................................................ 73 

3.3 Data sources and descriptive statistics ........................................................................... 74 

3.3.1 The Guyana Labour Force Survey (GLFS) .............................................................. 74 

3.3.2 Digest of Education Statistics (DES) ....................................................................... 76 

3.4 Empirical Strategy ......................................................................................................... 76 

3.5 Result ............................................................................................................................ 78 

3.5.1 Summary Statistics ................................................................................................. 78 

3.5.2 The effect of the USE (2003-2007) policy reform on schooling in Guyana .............. 80 

3.5.3 The effect of the USE (2003-2007) on labour market outcomes in Guyana.............. 84 

3.5.4 Robustness Checks ................................................................................................. 87 

3.6 Conclusion .................................................................................................................. 104 

References………………………………………………………………………………......106 

 

CHAPTER 4 Birth Order Effects on Early Childhood Health in Guyana  110 

4.1 Introduction ................................................................................................................. 110 

4.2 Literature Review ........................................................................................................ 112 

4.3 Data sources and descriptive statistics ......................................................................... 114 

4.3.1 The 2009 Guyana Demographic and Health Survey (GDHS) ................................ 114 

    4.3.2 Summary Statistics…………………………………………………………………116 

4.4 Empirical Strategy ....................................................................................................... 118 

4.5 Results and Discussions............................................................................................... 119 

4.5.1 Postnatal Investments ............................................................................................ 121 

4.5.2 Robustness Check ................................................................................................. 124 

4.5.3 Heterogeneous effects ........................................................................................... 127 

4.5.4 Policy Implications ............................................................................................... 136 



vii 
 

4.6 Conclusion .................................................................................................................. 137 

References…………………………………………………………………………………..138 

 

CHAPTER 5 Conclusion          141 

5.1 Introduction ................................................................................................................. 141 

5.2 Research Highlights .................................................................................................... 141 

5.3 Needs for Further Research ......................................................................................... 144 

5.4 Summary ..................................................................................................................... 145 

Appendix A The effects of weather shocks on early childhood health in Guyana 147 

Appendix A....................................................................................................................... 147 

Appendix 1 Further details on data and variable description........................................... 147 

 

Appendix B Effects of a School Conversion Policy on Education Attainment and Labour 

Market Outcomes in Guyana        149 

Appendix B ....................................................................................................................... 149 

Appendix 1 Further details on data and variable description .............................................. 149 

Appendix B.1: The USE Policy reform and Teacher's qualification……………………….150 

 

Appendix C Birth Order Effects on Early Childhood Health in Guyana  154 

Appendix C ....................................................................................................................... 154 

Appendix 1 Further details on data and variable description........................................... 154 

Appendix 2: Recommended vaccination schedule in Guyana ......................................... 157 

 

Bibliography            158 

                     

  



viii 
 

List of Figures 

 

Figure 1.1: Map of Administrative Regions and Population Density of Guyana. .................. 14 

Figure 1.2: GDP by Economic Activity at Current Prices (G$’M) 20006-2021. ................... 17 

Figure 1.3: Employment by Economic Sectors (2017-2021). ............................................... 17 

Figure 1.4: The GDP growth rate comparison for Guyana, its neighbouring counties and 11 

CARICOM countries. ......................................................................................................... 18 

Figure 1.5: The GDP growth rate comparison for Guyana, its neighbouring counties and 11 

CARICOM countries. ......................................................................................................... 19 

Figure 1.6: Guyana’s yearly Budget Allocation to its Health and Education Sectors (2007-

2022) .................................................................................................................................. 20 

 

Figure 2.1: Map of the average annual rainfall by regions of Guyana .................................. 29 

 

Figure 3.1: The structure and organisation of the education system prior to the USE (2003-

2007) policy reform............................................................................................................. 68 

Figure 3.2: A comparison of the number of institutions and share of institutions for the period 

1999-2019. .......................................................................................................................... 70 

Figure 3.3: The number of students in secondary institutions between pre-and-post-policy 

periods. ............................................................................................................................... 71 

Figure 3.4: The year-on-year percentage change for students in secondary institutions for the 

period (1999-2019). ............................................................................................................ 72 

Figure 3.5: The dropout rates in secondary institutions over time (1999-2019). ................... 73 

  

Figure B1: The share of students in secondary educational institutions by type during (1999-

2019) ……………………………………………………………………………..…………151 

Figure B2: The share of trained teachers for types of secondary educational institutions during 

(1999-2019) ………………………………………………………...………………………152 

Figure B3: The share of untrained teachers for the period (1999-2019)……………………152 

Figure B4: The share of unqualified teachers in the secondary institutions for the period (1999-

2019)………………………………………………………………………………..………153 

 

 



ix 
 

List of Tables 

Table 2.1: Summary statistics .............................................................................................. 33 

Table 2.2: The effects of in-utero rainfall shocks on child health ......................................... 39 

Table 2.3: The trimester-specific effects of in-utero rainfall shocks on child health ............. 40 

Table 2.4: The effects of prenatal and postnatal rainfall shocks on child health.................... 42 

Table 2.5: The effects of in-utero rainfall shocks on postnatal investments .......................... 44 

Table 2.6: The trimester-specific effects of in-utero rainfall shocks on postnatal investments

 ........................................................................................................................................... 45 

Table 2.7: Heterogeneous effects by gender of the child ...................................................... 47 

Table 2.8: Heterogeneous effects by location ...................................................................... 48 

Table 2.9: Heterogeneous effects by mother’s education ..................................................... 49 

Table 2.10: The effects of in-utero rainfall shocks on child health (90th, 10th percentile 

measure) ............................................................................................................................. 50 

Table 2.11: The trimester-specific effects of in-utero rainfall shocks on child health (90th, 10th 

percentile measure) ............................................................................................................. 51 

Table 2.12: The effects of in-utero rainfall shocks on child health ....................................... 52 

 

Table 3.1: Summary statistics .............................................................................................. 79 

Table 3.2: Effect of the USE on Educational Attainment ..................................................... 81 

Table 3.3: Educational Attainment - Restricted regressions (Ethnic groups: Africans, Mixed 

and Amerindians)…………………………………………………………………………….83 

Table 3.4: Educational Attainment - Restricted regressions (Ethnic groups: Other and East 

Indian – reference group)…………………………………………………………….……….83 

Table 3.5: Effect of the USE on the Labour Market in Guyana……………………………...85 

Table 3.6: Labour Market Outcomes - Restricted regressions (Ethnic groups: Africans and 

Mixed)………………………………………………………………………………………..86 

Table 3.7: Labour Market Outcomes - Restricted regressions (Ethnic groups: Africans and 

Mixed)………………………………………………………………………………………..86 

Table 3.8: Robustness Check - Effect of the USE on Educational Attainment in Guyana…..87 

Table 3.9: Robustness Check - Effect of the USE on the Labour Market in Guyana…….….90 

Table 3.10: Robustness Check - Effect of the USE on Educational Attainment………….…91 

Table 3.11: Robustness Check - Effect of the USE on the Labour Market in Guyana……....92 

Table 3.12: Summary statistics respondents who have not migrated vs those who migrated 93 

Table 3.13: Robustness Check: Effect of the USE on Educational attainment in Guyana…..94 

Table 3.14: Robustness Check: Effect of the USE on the Labour Market in 

Guyana……………………………………………………………………………………….96 



x 
 

Table 3.15: Robustness Check: Effect of the USE on Educational Attainment in Guyana (Share 

of non-GSS institutions…………………………………………………………………….. 98 

Table 3.16: Robustness Check: Effect of the USE on the Labour Market (Share of non-GSS 

institutions) ……………………………………………………………………………...….99 

Table 3.17: Robustness Check (Baseline specification): Multiple Hypothesis Testing (Westfall-

Young, Bonferroni and Sidak-Holm Corrections) ………………………………………….100 

Table 3.18: Robustness Check (DID-IV: 2SLS 

Regression)…….………………………………………………………………………......  101 

Table 3.19: Placebo test - Effect of the USE on Educational Attainment…….…………….102 

Table 3.20: Table 3.20: Placebo test - Effect of the USE on the Labour Market in Guyana 103 

 

Table 4.1: Summary statistics ............................................................................................ 117 

Table 4.2: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level 

(OLS Regressions) ............................................................................................................ 120 

Table 4.3: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level 

(OLS Regressions) ............................................................................................................ 123 

Table 4. 4: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at PSUs level 

(OLS Regressions) ............................................................................................................ 125 

Table 4.5: DHS Guyana 2009 Birth Order Effects on Child Health (Postnatal Investments) - 

Clustered at PSUs level (OLS Regressions) ....................................................................... 126 

Table 4.6: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level 

(OLS Regressions) – Heterogeneous Effects (ETHNICITY) ............................................. 129 

Table 4.7: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level 

(OLS Regressions) – Heterogeneous Effects (RELIGION) ................................................ 130 

Table 4.8: Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) 

– Heterogeneous Effects (EDUCATION) .......................................................................... 132 

Table 4.9: Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) 

– Heterogeneous Effects (GENDER)................................................................................. 133 

Table 4.10: DHS Guyana 2009 Birth Order Effects on Child Health – Heterogeneous effects 

(Mother’s age at each birth subgroups) …………………………………………………….137 

Table 4.11: DHS Guyana 2009 Birth Order Effects on Child Health – Heterogeneous effects 

(Mother’s age at first birth) …...……………………………………………………………138 

 

Table C.1: Recommended Vaccination Schedule in Guyana .............................................. 157 



11 
 

CHAPTER 1: - Introduction 
 

1.1 Background and Motivation 

 

In the literature, health and education are seen as the two main types of investments in human 

capital that predict the economic well-being of every society at large. The two topics are 

interconnected in various ways, i.e., they are substantial returns to education when improved 

health exists, and health complements education in rising the returns to labour (Hunt-McCool 

and Bishop, 1998). Through three distinct chapters, this dissertation sheds light on these topics 

of immense relevance for a country like Guyana. Guyana is a suitable context to study these 

questions because ninety percent of its inhabitants live on its coastal plain, which is below sea 

level, making Guyana one of the most endangered countries in the world and its country’s 

capital prone to extreme weather conditions and flooding (World Bank, 2022). Furthermore, 

Guyana’s educational sector has undergone a Universal Secondary Education (USE) policy 

reform that has not been empirically analysed to determine its success. For these reasons, 

coupled with the fact that Guyana has rich sources of data, which can be analysed to determine 

whether the government has reasons to be concerned with the country’s human capital and 

overall economic well-being. Furthermore, the results of this dissertation have possible policy 

implications which can be used by policymakers to make informed decisions to improve the 

economic well-being of Guyana.  

In the 21st century, global warming has been the focus of every country across the globe, and 

Guyana is no exception to the harmful effects of global warming on its economy. Furthermore, 

variations and changes in rainfall and temperature, and the frequency and intensity of extreme 

weather events such as floods and droughts can negatively affect child health outcomes such 

as stunting of growth, wasting and severe stunting (Glewwe and King, 2001; Alderman et al., 

2006 Victora et al., 2008; Ogasawara and Yumitori, 2018; Aguilar and Vicarelli, 2022). 

Consequently, over the years, Guyana has witnessed a rise in stunting (i.e., height-for-age z-

score < -2) among children under 5 (Kaieteur News, 2014; UNICEF, 2019). Furthermore, 

Chapter one of this dissertation investigates weather shock's effects on Guyana's childhood 

health.  

Guyana’s secondary institutions suffered from high levels of dropout in the 1990s due to its 

three-tier secondary education structure that offered unequal learning opportunities among the 



12 
 

secondary departments of primary schools or Primary Tops and Community High Schools, 

which offered a more pre-vocational curriculum than General Secondary Schools who were 

more academic in nature (Jennings and Miller 1999; Gafar 2003; World Data on Education, 

Guyana 2006/07). However, the Universal Secondary Education (2003-2007) policy reform 

was created with the aim of converting secondary classes in Primary Tops and Community 

High Schools into General Secondary Schools (World Bank, 2013; Ministry of Education, 

2022). More important, the USE policy reform made secondary education universally available 

to all secondary school students in Guyana. Consequently, all secondary school students can 

have equal access to a more academic curriculum rather than a pre-vocational one. A study has 

shown that academic qualifications attract higher wage premia than vocational qualifications 

(Dearden et al., 2002). Studies have also shown that educational policy reforms tend to increase 

educational attainment (Riphahn, 2012; Duflo et al., 2021; Brudevold-Newman, 2021). As 

such, Chapter two of this dissertation examined the effects of a school conversion policy on 

educational attainment and labour market outcomes in Guyana.  

As it relates to birth order and childhood health in Guyana, existing literature finds a strong 

positive correlation between birth order and the health of babies at birth/newborns, i.e., birth 

weight tends to increase among later-born when compared to first-borns (Björkegren and 

Svaleryd 2016; Brenøe and Molitor, 2017; Pruckner et al., 2019). Literature has also shown 

that there can be policy implications of birth order effects when mothers spend considerable 

time breastfeeding and on postnatal investments (Björkegren and Svaleryd, 2016). According 

to Guyana Demographic and Health Survey (2009), the average number of children per family 

is 2.8 children per family. In recent years, Guyana has also witnessed a rise in single-parent 

households headed by women. Furthermore, 30 percent of Guyana’s households are headed by 

women (Office of the United Nations High Commissioner for Human Rights, 2020). 

Consequently, the third chapter of this dissertation examined the birth order effects and early 

childhood health in Guyana.  

To the researcher’s knowledge, to date, there has been no previous study on the effects of 

weather shocks, birth order and the impact of the Universal Secondary Education (2003-2007) 

policy reform in Guyana. Consequently, this research is both novel and timely given the reports 

in the media on extreme weather events that are caused by the changing climatic conditions, 

the need for the reduction of child mortality and improve maternal health and the Ministry of 

Education’s goal of achieving quality USE for all secondary school children in Guyana. 

Moreover, the socioeconomic implications for children, mothers and the country compel a 



13 
 

closer examination of the problems highlighted in this research. Therefore, this study aims to 

examine the effects of weather shocks, birth order and educational reform on health and 

education-related outcomes in Guyana. To achieve this objective, the following research 

questions are proposed and addressed by this research: 

 

• How do weather shocks affect childhood health in Guyana? 

• How has the USE (2003-2007) policy reform impacted Guyana's educational 

attainment and labour market outcomes?  

• How can birth order affect childhood health in Guyana? 

Furthermore, the rest of the chapter will provide information on the research context, Guyana’s 

colonial history and independence, economic growth and country comparison, Guyana’s health 

and education sectors, and the chapter concludes with the dissertation layout.  

 

1.2  Research Context 
 

The geographic focus of this research is the Cooperative Republic of Guyana. Guyana, 

formerly known as British Guiana, is the only English-speaking country in South America. It 

is in the north-eastern corner of South America between 20N and 80N Latitude and shares 

borders with Suriname to the east, Brazil to the south and southwest, Venezuela to the west, 

and the Atlantic Ocean to the north. Guyana is the third-smallest sovereign state in South 

America after Uruguay and Suriname. Guyana has a land area of 83,000 square miles and a 

population of approximately 800,000 inhabitants (World Bank, 2022). 

 



14 
 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

Figure 1.1: Map of Administrative Regions and Population Density of Guyana. 

Source: Author’s illustration based on Guyana’s Population Density (2012 Census) provided by Bureau of 

Statistics, Guyana. 

 

Guyana is a word derived from one of the native Amerindian languages meaning “land of many 

waters.” This name was given due to the numerous rivers that traverse long distances in the 

country, with the main rivers being Essequibo, Demerara, Berbice and Corentyne. As Figure 

1.1 shows, Guyana comprises ten administrative regions and nine towns. The capital city of 

Georgetown is situated in region 4, and it is the largest city in Guyana.  Georgetown is also the 

most densely populated region and home to 80% of Guyana’s population (World Bank, 2022).  

 



15 
 

1.3 Guyana’s Colonial History and Independence  
 

The history of Guyana can be traced back to its first people who were said to have travelled 

from Eurasia way before the 12th Century. These are known as the Amerindians or Indigenous 

people of Guyana (Sanders, 1995). In the colonial era of the 15th Century, Guyana’s 

government was defined by successive policies of the Spanish, French, Dutch, and British 

settlers. In the mid-17th century during the period of the Transatlantic Slave trade, Guyana’s 

second largest ethnic group (the Africans) were brought as slaves by Dutch settlers (Khemraj, 

2016). The British first occupied Guyana during the Napoleonic Wars in 1814 and in 1831, 

Guyana was declared a British colony. In 1834 slavery was abolished in Guyana, and in 1838, 

the then British Guiana commenced the Indian Indentureship system, where East Indians 

(Guyana’s largest ethnic group to date) were brought to Guyana to work on sugar plantations. 

In 1838, the Portuguese also arrived in Guyana from Madeira to work on the sugar plantations 

under the indentureship system (Khemraj, 2016). In 1851, Chinese indentured labourers arrived 

in Guyana to serve the estates.  

Guyana’s colonial past created a rich multi-ethnical and multi-cultural economy (Khemraj, 

2016). Consequently, Guyana has six ethnic groups: Amerindian, East Indian (Indo-Guyanese), 

African (Afro-Guyanese), Chinese, Portuguese, and Mixed race.  In terms of employment, the 

Indo-Guyanese (East Indians) dominated both employment and income distribution categories 

in Guyana during 2006-2016 when compared to Afro-Guyanese (Africans) (Constantine, 

2021). Furthermore, in terms of education, the East Indians and Africans excel comparably in 

both primary and secondary levels of examinations in Guyana (Guyana Chronicle, 2011). 

In 1966, British Guiana became an independent nation under the new name Guyana. On 

February 23rd, 1970, Guyana was proclaimed a cooperative republic within the British 

Commonwealth and was named ‘The Co-operative Republic of Guyana or Guyana.’ In 1970, 

Guyana also welcomed its first President (of Chinese descent), Mr Raymond Arthur Chung, 

and Mr Forbes Burnham (of African descent) was declared the Prime Minister. In Guyana, 

executive power is exercised by the President, advised by a cabinet.  Furthermore, in the late 

1970s, Guyana was declared a socialist nation by then-President Forbes Burnham.  

In 1973 Guyana, along with three other Caribbean countries (Jamaica, Barbados, and Trinidad 

and Tobago), formed an alliance known as CARICOM (or ‘The Caribbean Community’) to 

allow better trade agreements and movement of people between countries. At present, 

CARICOM consists of fifteen Caribbean nations and dependencies (CARICOM, 2022).  In the 



16 
 

early 1980s, Guyana followed a series of inward-looking policies, and in 1985, the country's 

economic situation started to deteriorate (Chandisingh, 1983; Gafar, 1996: Khemraj, 2016). In 

response to rescuing the country from its economic downturn, in 1988, International Monetary 

Fund (IMF) and Commonwealth were invited to review the economy. In 1989, Guyana 

embraced the structural reforms instituted by the IMF and embarked on a new era of outward-

looking policies to better stimulate the economy.  

Guyana’s political climate is one marked by racial tension due to ethnicity-based politics 

among the two most predominant races, i.e., the East Indians and Africans (Ryan, 2020). In 

2001, the People’s Progressive Party (PPP) (a predominantly East Indian party) won and went 

on to lead the country. However, in 2015 the opposition party ‘A Partnership for National 

Unity’ (APNU), a coalition of the main opposition party and other smaller parties headed by 

former President David Granger (of African descent) won the elections and led the country up 

until 2020. In 2020, the PPP won the general and regional elections in Guyana and President 

Irfaan Ali was appointed as President.  

 

1.4 Economic Growth and Country Comparison 
 

Guyana is richly endowed with vast deposits of minerals such as gold, diamond, bauxite, and 

manganese which are all major exports. The government of Guyana has invested heavily in 

mechanising the country’s gold and bauxite mining sectors to improve production (World 

Atlas, 2022).  

In 2015, the government of Guyana estimated that 83.95 percent of the country’s total land 

area was covered with forests. Guyana has positioned itself to become ‘the Breadbasket of the 

Caribbean’ as its rich soil and climate make it ideal for agricultural exports of its traditional 

crops such as rice, sugar, and non-traditional crops such as coconut and vegetables. Guyana 

also has a large fishery and livestock sector.1  

 

With a GDP per capita of US$9,374.8 in 2021, Guyana is considered as an upper middle-

income country (World Bank, 2022). However, it is among the poorest countries in South 

 
1 The Breadbasket of the Caribbean – Kaieteur News (kaieteurnewsonline.com) 

https://www.kaieteurnewsonline.com/2020/08/12/the-breadbasket-of-the-caribbean/


17 
 

America due to 43.4 percent of the population living on less than US$5.5 per person a day in 

2011 Purchasing Power Parity (World Bank Factsheet, 2020).  

 

Figure 1.2: GDP by Economic Activity at Current Prices (G$’M) 20006-2021. 

Source: Bureau of Statistics, Guyana, 2022. 

 

Figure 1.2 shows the Gross Domestic Product by Economic Activity at Current prices for the 

period 2006-2021. Over the period 2006-2021, Guyana has witnessed a steady rise in the 

service sector’s contribution to the Gross Domestic Product of the country. However, in recent 

times, Guyana’s mining and quarrying sectors have witnessed a stack increase, owing to the 

exploration of oil in the country.2 The Agriculture and Manufacturing sectors have witnessed 

a steady growth in GDP contribution averaging at GY$46 million and GY$136 million 

respectively.  

 

Figure 1.3: Employment by Economic Sectors (2017-2021). 

Source: Guyana Labour Force Survey, Bureau of Statistics, Guyana, 2022 

 
2 Exxon makes three new oil discoveries in Guyana and boosts reserves | Reuters 

0

100000

200000

300000

400000

500000

600000

700000

800000

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

m
il

li
o
n

s 
G

$

Years

Agriculture, forestry and fishing Mining and quarrying Manufacturing Services

18.9 17.5
13.3 12.1 13.8

4.2 3.9 4.1 3.4 5
8.4 10 12 10.2 8.8

68.5 68.6 70.6
74.3 72.4

0

10

20

30

40

50

60

70

80

2017 2018 2019 2020 2021

p
er

ce
n

ta
g
e 

(%
)

Years

Agriculture, forestry and fishing Mining and quarrying Manufacturing Services

https://www.reuters.com/business/energy/exxon-makes-three-new-discoveries-guyana-increases-oil-reserves-2022-04-26/


18 
 

Figure 1.3 shows the employment by economic sectors for the period 2017-2021. Guyana’s 

service sector is responsible for an average of 71 percent of its employment, while the 

agriculture, forestry and fishery employ approximately 15 percent of its labour force, followed 

by manufacturing and mining and quarrying sectors with an average of 4 percent and 10 percent 

respectively.   

In May 2015, a US-based company Exxon Mobil announced the discovery of a significant 

amount of oil in the Stabroek Block, Georgetown, 120 miles offshore Guyana in the North 

Atlantic Ocean.3 In recent times, Exxon Mobil has announced several new oil discoveries in 

the Stabroek Block, with an anticipated recoverable oil exceeding eight billion barrels.4 To 

draw a comparison, Venezuela, Guyana’s neighbour hosts the world’s largest oil reserves, i.e., 

more than 300 billion barrels.5 Moreover, these recent oil discoveries place Guyana in the list 

of the top 20 countries with the largest proven oil reserves.6 Consequently, Guyana is poised 

to experience a major boost in its GDP by 300-1000 percent by the end of 2025.7  

 

Figure 1.4: The GDP growth rate comparison for Guyana, its neighbouring counties and 11 CARICOM 

countries. 

Source: World Bank (2022) 

 
3 ExxonMobil announces significant oil discovery offshore Guyana 
4 Guyana project overview | ExxonMobil 
5 International - U.S. Energy Information Administration (EIA) 
6 Guyana is number 17 for largest oil reserves in the world | OilNOW 
7 Will Guyana soon be the richest country in the world? - BBC News 

-40

-30

-20

-10

0

10

20

30

40

50

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

G
D

P
 G

ro
w

th
 r

at
e 

%

Years
Guyana Suriname
Brazil Venezuela
Jamaica Barbados
Belize Dominica
Grenada Antigua and Barbuda
Saint Lucia The Bahamas
Saint Vincent and the Grenadines Trinidad and Tobago

https://corporate.exxonmobil.com/Locations/Guyana/News-releases/ExxonMobil-announces-significant-oil-discovery-offshore-Guyana
https://corporate.exxonmobil.com/Locations/Guyana/Guyana-project-overview
https://www.eia.gov/international/overview/country/VEN
https://oilnow.gy/featured/worlds-largest-oil-reserves-by-country/
https://www.bbc.co.uk/news/world-latin-america-48185246


19 
 

Figure 1.4 shows the GDP growth rate comparison between Guyana, its neighbouring countries 

and 11 CARICOM countries. Over the period 2006-2019, Guyana and the CARICOM 

countries had relatively similar growth rates, however in 2020, Guyana recorded a GDP growth 

rate 43.5 percent, this represents a 38.1 percent increase over 2019.  

 

Figure 1.5: The GDP growth rate comparison for Guyana, its neighbouring counties and 11 CARICOM countries. 

Source: World Bank (2022) 

 

Figure 1.5 shows the GDP growth rate comparison for Guyana, its neighbouring countries and 

11 CARICOM countries. Over the years, Guyana’s GDP per capita has increase, however in 

2020 Guyana witnessed a sharp increase owing to oil extraction from Guyana’s recent oil 

discovery. For the period 2006-2021, St Kitts and Nevis, Barbados, Suriname and Antigua and 

Barbados are the countries with the highest GDP per capita. While Venezuela has witnessed a 

sharp decline in their GDP per capita.  

As it relates to climate change, Guyana is extremely susceptible to the effects of climate change 

due to its low-lying coastline which is prone to damage due to rising sea levels (World Bank 

2022; Lakhan, 2014).8 This and the other effects of climate change pose a major threat to 

economic growth 

 
8 World Bank Document 

0

5000

10000

15000

20000

25000

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

G
D

P
 p

er
 c

ap
it

a 
U

S
 (

$
)

Years
Guyana Suriname
Brazil Venezuela
Jamaica Barbados
Belize Dominica
Grenada Antigua and Barbuda
Saint Lucia The Bahamas
Saint Vincent and the Grenadines Trinidad and Tobago
Haiti St Kitts and Nevis

https://documents1.worldbank.org/curated/en/099307104052251719/pdf/IDU0468c78c10092804f1c082d004de6dfcd66cc.pdf


20 
 

1.5 Guyana’s Health and Education Sectors 
 

The Government of Guyana (GoG) continues to invest heavily in Guyana’s health and 

education sectors. In 2022 GoG allocated 26.7% of its GY$552.9 billion budget to the 

education (GY$74.4 billion) and health (GY$73.2 billion) sectors (Guyana Budget Bulletin, 

2022). In Guyana, an average of 28% of its yearly budget is allocated towards its Health and 

Educational Sectors (Ministry of Finance, Guyana, Budget Speech, 2022). Figure 6 below 

shows the yearly budget by the government of Guyana to the Health and Educational Sectors 

(2007-2022). The Health and Education sectors remain top priorities for the GoG, as such based 

on Figure 1.6, the budget expenditure has steadily increased over the period 2007-2022. 

However, the stark increase in the budget expenditure for the health and education sectors in 

2021-2022 is due mainly to controlling the Covid-19 pandemic.  

 

 

Figure 1.6: Guyana’s yearly Budget Allocation to its Health and Education Sectors (2007-2022) 

Source: Ministry of Finance, Guyana 

 

According to the World Bank Data, Guyana’s government expenditure, total (% of GDP) on 

education and health sectors is approximately 9 percent for the period 2007-2018 while its 

larger neighbouring country Brazil is 14.3 percent. Furthermore, with respect to the CARICOM 

countries, the average government expenditure as a total % of GDP on these sectors is 10 

percent (World Bank Data, 2022).   

 

10.6 11 12.8 13.4 14
17 19.2 21.5 23.2

28
31.2 33.3

35.9

51.7 53.5

73.2

17.1 18.5 20.3
24.3

21.8
26.5 28.7

32.3 31.8

40.3
43.1 41

52.5 52

60.7

74.4

0

10

20

30

40

50

60

70

80

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

E
x

p
en

d
it

u
re

 -
G

yd
 $

 b
il

li
o
n

Years

Health Education



21 
 

Guyana operates a universal healthcare system where primary healthcare facilities and services 

provided by the Georgetown Public Hospital Corporation (GPHC) are free to access to every 

citizen and resident of Guyana (Ministry of Health, Guyana, 2022). Furthermore, healthcare, 

and medical facilities in Guyana are provided by both the public and private sectors. Although 

primary healthcare is free of cost, complex procedures are partially subsidized.  

 

Although public healthcare is free, the early-life health of children remains a major concern. 

According to the World Bank, the (1971-2014) records on the prevalence of stunting (i.e., an 

height-for-age z score < -2) for children under 5, Guyana witnessed an increase in stunting 

from 14 percent in 1998 to 19.5 percent at the end of 2009, however it witnessed a record low 

in 2014 of 12 percent.9  Furthermore, in 2016 UNICEF disclosed that 1 in every 4 Amerindian 

child in the remote interior areas of Guyana is considered as suffering from stunted growth.10 

In 2020, it was noted that 28 percent of children in hinterland regions experienced stunted 

growth. It was also noted that 21 percent of mothers in these regions practice exclusive 

breastfeeding, which is below the South American average of 57.3 percent.11  Consequently, 

childhood health remains a major concern for the Government of Guyana (UN Network, 2020). 

 

In 2022,  the GoG with the support of the Inter-American Development Bank (IADB) continues 

to advance efforts to improve the standard of healthcare, and this is reflected in increased life 

expectancy (which was 69.91 years as of  2019); reduction in maternal and child mortality; 

decreased incidence, prevalence and mortality from communicable diseases; high levels of 

immunization coverage; greater awareness of environmental health issues; and improved water 

and sanitation facilities (Ministry of Health, Guyana, 2022). 

 

In 1961, the GoG established the Ministry of Education (MoE).12 After gaining independence 

from Britain in 1966, Guyana inherited a well-established educational system with a curriculum 

and educational aims patterned after the British system. During the 1976 - mid-1990s, under 

the market socialism regime, education was free from nursery to university (UNWOMEN, 

 
9https://www.ceicdata.com/en/guyana/health-statistics/gy-prevalence-of-stunting-height-for-age--of-children-

under-5  
10 https://guyanatimesgy.com/1-in-4-amerindian-children-considered-stunted-unicef/  
11https://www.unnetworkforsun.org/news/fueling-gains-nutrition-harnessing-guyanas-oil-boom-invest-human-

development  
12 The statutory age for beginning compulsory education in Guyana is 5 years 6 months to 15 years.  

https://www.ceicdata.com/en/guyana/health-statistics/gy-prevalence-of-stunting-height-for-age--of-children-under-5
https://www.ceicdata.com/en/guyana/health-statistics/gy-prevalence-of-stunting-height-for-age--of-children-under-5
https://guyanatimesgy.com/1-in-4-amerindian-children-considered-stunted-unicef/
https://www.unnetworkforsun.org/news/fueling-gains-nutrition-harnessing-guyanas-oil-boom-invest-human-development
https://www.unnetworkforsun.org/news/fueling-gains-nutrition-harnessing-guyanas-oil-boom-invest-human-development


22 
 

2021). However, in the early 1990s, under the IMF structural reform period, access to education 

became free from nursery to secondary and this reform exists in modern-day Guyana.13   

Guyana’s secondary institutions have also experienced major reforms under the Universal 

Secondary Education (USE) (2003-2007) policy reform. Prior to 2003, secondary education in 

Guyana operated under a tier system where they were three types of secondary schools, namely, 

secondary departments of primary schools (popularly known as “Primary Tops”), Community 

High schools (CHSs) and General Secondary Schools (GSSs). The USE policy reform sought 

to make secondary education universally available to all students. Although the Ministry of 

Education has successfully converted the CHSs into GSSs, as of 2020, there are some Primary 

Tops that are still to be converted. Consequently, the MoE has set the attainment of quality 

USE as a major objective in its current Education Strategic Plan (Ministry of Education, 2022).  

 

 1.6 Layout of the Dissertation 
 

In addition to this initial chapter, this study comprises of five chapters. Chapter 2 presents the 

effects of weather shocks on childhood health in Guyana. Chapter 3 explores the effects of a 

school conversion policy (namely the Universal Secondary Education, 2003-2007 policy) on 

educational attainment and labour market outcomes in Guyana. Chapter 4 presents the effects 

of birth order on child health in Guyana. Chapter 5 summarises the key findings and explores 

the overall policy implications and opportunities of future research emerging from this 

dissertation. 

 

 
13 Guidelines and procedures for the placement of students at the various levels of the Education System. 

https://education.gov.gy/en/index.php/parents/information-parents/school-placement/1403-guidelines-and-procedures-for-the-placement-of-students-at-the-various-levels-of-the-education-system#:~:text=The%20Government%20of%20Guyana%20provides,six%20months%20to%20fifteen%20years.


23 
 

CHAPTER 2  

 

The effects of weather shocks on early childhood health in Guyana. 

 

2.1 Introduction 
 

Over time, the world has witnessed a rise in global temperatures and an increase in the 

frequency and severity of natural disasters such as earthquakes, bush fires, tsunamis, floods, 

droughts, and hurricanes, all largely attributed to climate change (López et al., 2015; IPCC, 

2022). These catastrophic events have been shown to have negative effects on countries’ GDP, 

investments, employment and earnings, education, and health (Botzen et al., 2019). Moreover, 

lower-income countries are especially affected by climate change, and this hinders them from 

attaining the Sustainable Development Goals (United Nations, 2019). Consequently, it is 

important for governments in developing countries to understand the effects of such shocks 

and accordingly devise effective strategies to mitigate the adverse effects of weather shocks. 

 

In this study, we analyse the effects of in-utero weather shocks on early childhood health in 

Guyana. This is the first study to address this question in the context of Guyana. We use rich 

anthropometric data for children under age 5, from three rounds of the Guyana Multiple 

Indicator Cluster Survey (MICS) and combine that with rainfall shocks in the region and year 

of birth, constructed using geo-spatial precipitation data. These large-scale data are 

underutilized in the Guyanese context and allow us to shed light on the health impacts of 

weather shocks in a middle-income flood-prone country. In doing so, we provide evidence 

related to the ‘fetal origins hypothesis’.1 This hypothesis states that the intra-uterine 

environment determines the long-run health outcomes of children and their human capital 

achievements. We focus on three measures of early childhood health: height-for-age z-score 

(hereafter, HAZ), and indicators for stunting and severe stunting. A child’s HAZ is an 

important measure that captures lingering effects of any growth disturbances that might have 

occurred earlier in their life (Leroy et al., 2015). 2 Stunting, or low height-for-age, is the most 

prevalent measure of child malnutrition, and it usually reflects insufficient nutrient intake 

during the early stages of development. Stunting generally occurs before age two and once 

established, is usually permanent, and most children do not regain the height lost or achieve a 

 
1 Almond and Currie (2011) and Almond et al. (2018) provide overviews of this literature. 
2 While the weight-for-age z-score which captures short term illness, HAZ captures chronic illnesses (Victora et 

al., 2008) 



24 
 

normal body height. Some consequences may be extremely severe: stunted growth in childhood 

due to incomplete development of vital organs may lead to premature death later in life. Less 

extreme effects also include delayed development, impaired cognitive function, and poor 

school performance (UNICEF, 2007; WHO, 2015). 

 

Our results show that prenatal exposure to positive rainfall shocks (i.e., where rainfall is 

significantly above the long-term regional average) in the nine months preceding birth 

increases the likelihood of being stunted and severely stunted while negative rainfall shocks 

(i.e., where rainfall is significantly below the long-term regional average) do not affect growth 

significantly. We also consider the timing of these shocks in terms of the trimester of exposure. 

We find that the adverse effects of positive rainfall shocks on HAZ and stunted and severely 

stunted growth are greater during the first trimester. We also find that negative rainfall shocks 

in the third trimester positively impact the HAZ. Moreover, when examining heterogeneous 

effects, we find that positive rainfall shock increases height-for-age z-scores in urban areas as 

compared to rural areas. However, we do not find any differential effects of rainfall shocks on 

health outcomes based on child gender and maternal education.  

This study and its findings contribute to the broader literature on shocks and their impact on 

childhood health. Studies have empirically shown that exposure to adverse conditions in early 

childhood – disasters, conflict, diseases, famine, and pollution – has significant negative 

impacts on health outcomes such as HAZ and stunting (e.g., Glewwe and King, 2001; 

Alderman et al., 2006; Victora et al., 2008; Ogasawara and Yumitori, 2018; Aguilar and 

Vicarelli, 2022). Furthermore, exposure to shocks in-utero also leads to growth faltering i.e., a 

slower rate of weight gain in childhood than normal for age and sex (Almond et al., 2009; 

Banerjee et al., 2010; Dercon and Porter, 2014). This is important as children who experience 

slow height growth are found to perform worse in school, score poorly on tests of cognitive 

function, and have poorer psychomotor development and fine motor skills. They also tend to 

have lower activity levels, interact less frequently in their environments, and fail to acquire 

skills at normal rates (Salm and Schunk, 2012). Rainfall shocks during the in-utero period led 

to long-term negative impacts such as lower overall total years of schooling and lower wages 

in adulthood (Almond et al., 2018; Rosales-Rueda, 2018). Further, negative rainfall 

fluctuations in-utero are associated with lower birth weight, premature labour, high infant 

mortality rate, lower academic test scores, higher incidence of delayed enrolment and higher 

dropouts and grade repetition (Rocha and Soares, 2011). Consequently, environmental, and 



25 
 

economic shocks experienced in-utero and in early childhood may have long-lasting impacts 

on one’s health and economic outcomes in adulthood (Currie and Vogl, 2013; Strauss and 

Thomas, 2008). 

Our findings also speak to the literature on whether positive or negative weather shocks affect 

household welfare and health outcomes more. The literature studies both positive rainfall 

shocks (floods) and negative rainfall shocks (droughts) on health and human development 

outcomes and the results are not always consistent. For instance, Leight et al. (2015) find that 

both positive and negative rainfall shocks in-utero had significant negative effects on children’s 

HAZ in China, but only positive rainfall shocks negatively impacted their cognitive skills in 

primary school. Carrillo (2020) finds that negative effects of prenatal droughts on adult 

outcomes were smaller relative to the negative effects of floods. 

Maccini and Yang (2009) find that positive rainfall in one’s birth year and birth district have a 

positive relationship with adult height, completed years of education, and wealth in adulthood 

for women in Indonesia. On the contrary, Skoufias et al. (2011) studied the impacts of rainfall 

shocks on the welfare of households in rural Mexico and revealed that positive rainfall shocks 

are associated with lower HAZ while negative rainfall shocks had different impacts based on 

regions and socioeconomic characteristics. Rocha and Soares (2015) studied water scarcity and 

birth outcomes in Brazil’s semiarid regions and found that negative rainfall shocks during the 

in-utero periods are correlated with higher infant mortality, lower birth weight, and shorter 

gestation periods. Kumar et al. (2016) also documents that exposure to drought in-utero is 

associated with lower weight-for-age z-scores and increased probabilities of being underweight 

for children in India.  

Our findings also contribute to an understanding of critical periods where shocks can have the 

greatest health impacts. Some studies find that birth outcomes are most sensitive to maternal 

stress during the first trimester of pregnancy (Camacho, 2008; Currie and Rossin-Slater, 2013; 

Bozzoli and Quintana-Domeque, 2014). However, Kramer (1987) and Stein et al. (2000) posit 

that nutritional deficit impact birth weight mostly in the third trimester of pregnancy. Rosales-

Rueda (2018) finds that children who had prenatal exposure to severe floods in Ecuador during 

the 1997-1998 El Niño phenomenon, particularly during the third trimester of pregnancy, 

scored relatively lower on cognitive tests and were shorter five and seven years later. Carrillo 

(2020) finds that the long-run effects on educational and health outcomes from exposure to 

floods and droughts occurred mainly in the first trimester of gestation. Moreover, Randell et 



26 
 

al. (2020) finds that greater than average rainfall in Ethiopia during the third trimester in utero 

is positively associated with severe stunting, however, a child’s HAZ increases with greater 

exposure to early-life rainfall. Le and Nguyen (2021) find that exposure to excess rainfall in 

the first trimester in fifty-five low- and middle-income countries led to a reduction in children’s 

HAZ by 0.14 standard deviations and an increase in stunting by 0.03 percentage points. They 

also find that exposure to excess rainfall in the second and third trimesters lead to a reduction 

in children’s HAZ by 0.14 and 0.14 standard deviations respectively. Moreover, when 

examining the lack of rainfall, they also found similar results.  

 

There are several mechanisms through which in-utero weather shocks have been shown to 

affect child health. Firstly, through poverty and loss of income, weather shocks can have 

serious welfare consequences both in the short and long run (Edoka, 2013). Further, it can lead 

to intergenerational effects which can lead to poverty traps particularly for the disadvantaged 

groups (Almond et al., 2018). Secondly, this can lead to maternal stress and malnutrition for 

the mother which can exacerbate the incidence of negative health impacts on the developing 

foetus. Kramer (1987) posits that the increase in the likelihood of low birth weight may be 

driven by intrauterine growth restriction, which is largely determined by poor maternal 

nutrition. Aguilar and Vicarelli (2022) showed that negative effects of floods on income, food 

consumption, and diet composition during early childhood appear to be key mechanisms 

behind the impacts on children’s outcomes. Thirdly, the incidence of positive rainfall shocks 

can lead to a disease environment through water-borne diseases such as leptospirosis, malaria, 

yellow fever, dengue fever among others (Maccini and Yang, 2009).  Finally, there can be 

difficulties in access to health facilities during periods of excessive rainfall (Bonjean et al., 

2012; Dimitrova and Bora, 2020) due to damaged road infrastructure and shortage of staff 

availability and this is another mechanism that contributes to the adverse effects of weather 

shocks on child health.  

 

Over the period, there are several reports of flooding in Guyana in both urban/rural or 

coastal/inland areas. According to Pelling (1999), ninety percent of the Guyanese population 

are at risk from contemporary flood hazard, the potential impacts of climate change and sea-

level rise. For instance, in 2005, Guyana experienced high levels of rainfall which resulted in 

regions 3, 4 and 5 being flooded. Regions 1, 2 and 6 were also affected. According to the Word 

Bank (2005) report, the 2005 flood adversely affected sectors such as health, education, water 



27 
 

and sanitation, food, agriculture, and infrastructure. The flood results in health centers and 

clinics being closed due to equipment and drug supplies being damaged. As such, public health 

programs in the affected areas been interrupted. In addition, water treatment plants were out of 

service due damages.3 The 2005 flood also resulted in economic impacts that were estimated 

to be around US$465 million (59% of the Guyana’s then GDP). Majority of the losses were 

seen in Guyana’s Rice sector, thus resulting in thousands of acres of rice and other sources of 

economic activities that were concentrated in the low-lying coastal plains and most fertile 

agricultural regions 4 and 5.4  

Moreover, in January 2006, Guyana experienced yet another episode of flooding regions 1, 2, 

3, 5, 6 and 9, where regions 2 and 5 were officially declared disaster areas.5 In January 2012, 

heavy rainfall resulted in several communities being flooded in regions 1, 2, 3, 4, 5, and 6.6  

Moreover, studies have shown that urban areas are better able to deal with weather shocks as 

they are better equipped in all ways to withstand and recover from shocks when compared to 

rural areas (Burgess, 2014; Hallegate et al., 2020). Furthermore, based on the heterogeneity by 

urban/rural areas, it is evident that in the case of Guyana, urban areas were better able to cope 

with weather shock impacts on child health outcomes than in rural areas. Consequently, it is 

expected that the mechanism listed above might have affected in-utero exposure to weather 

shocks.  

Furthermore, the estimates in this chapter are thus a combination of the effect on actual affected 

people and the proportion of people affected by a shock. We argue that the proportion of people 

affected by a shock is close to 100% at the regional level.  

 

The rest of the paper is organised as follows. Section 2.2 provides a background on Guyana 

and its climate. Section 2.3 describes the data sources and descriptive statistics. Section 2.4 

introduces the empirical strategy. Section 2.5 presents the results and Section 2.6 concludes.  

 

 

 
3 https://www.gfdrr.org/sites/default/files/publication/pda-2005-guyana.pdf  
4 https://www.gfdrr.org/en/floodpreventioninguyana   
5 file:///Users/nichola/Downloads/gyfl03.pdf  
6 https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-guyana-situation-

report--3&catid=125:guyana  

https://www.gfdrr.org/sites/default/files/publication/pda-2005-guyana.pdf
https://www.gfdrr.org/en/floodpreventioninguyana
file:///C:/Users/nichola/Downloads/gyfl03.pdf
https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-guyana-situation-report--3&catid=125:guyana
https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-guyana-situation-report--3&catid=125:guyana


28 
 

2.2 Background on Guyana and its Climate 
 

The Co-operative Republic of Guyana is the third-smallest sovereign state in mainland South 

America with an area of 214,970 km2. It is classified as an upper-middle-income country 

(World Bank, 2020) with 41.2 percent of its people living below the USD 5.50 a day poverty 

line (Inter-American Development Bank, 2017). Approximately, ninety percent of Guyana’s 

population inhabits flood-prone areas along the coastal plains which is half to one metre below 

sea level. Further, the coastal plains are also home to seventy five percent of the country’s 

economic activities. 

 

Guyana has an equatorial climate that is characterised by a high level of rainfall variability and 

the seasons and climate are determined by this variability. Consequently, there is high humidity 

and small variations in temperature owing to it being located just above the equator (Guyana 

Second National Communication to the United Nations Framework Convention on Climate 

Change, 2012). The mean temperature in Guyana is 25-27.5oC throughout the year in most 

regions except for the upland regions in the west, where the mean temperature is 20-23oC 

(World Bank Climate Change Knowledge Portal, 2020).  

 

In general, most of Guyana has two rainy and two dry seasons. However, in the southernmost 

parts there is one peak period of rainfall. The rainy seasons vary by location. The first one runs 

from December to early February (the small rainy season) and affects mainly the northern, 

coastal regions which receive approximately 6-12 inches of rainfall per month. The second 

begins in late April and lasts until early July (the main rainy season). Consequently, most of 

the country receives an average monthly rainfall of 10-18 inches between May and July. 

However, the periods in between are drier, but not extremely dry, since, during the two drier 

periods (February – April and July – November), the average precipitation per month based on 

month variability lies between 2.3 – 7 inches of rainfall. 

 

Studies conducted by Douglas (1995) and Smith et al. (1999) indicate “that the sea level on the 

region of Guyana is increasing at a rate in excess of 10mm/year – or 2 to 5 times faster than 

the global estimate.” (Guyana Second National Communication to the United Nations 

Framework Convention on Climate Change, 2012, pp. 23). Therefore, changes in the sea level 

increase the severity of individual rainfall events, and variations in annual rainfall levels pose 

significant threats to Guyana's East Demerara Water Conservancy system, a 150-year-old 



29 
 

storage system providing irrigation and safe water to ninety percent of people living in the 

coastal plains. The rise in sea level also affects the drainage infrastructure and future 

agricultural production (see Figure 2.1 below for a map of Guyana showing the average annual 

rainfall by regions). For instance, in 2011 coastal farmers in Guyana were hit by the rising sea 

levels and failing seawalls which resulted in agricultural losses (Thomson Reuters Foundation 

News, 2011).  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.1: Map of the average annual rainfall by regions of Guyana. 
Source: Author’s illustration based on rainfall data provided by Guyana Hydrometeorological Service and the 

University of Delaware (UDEL). 

 

According to the projections of land inundation during minimum and maximum storm surges, 

Guyana is expected to experience increases in temperature and increased rainfall in the rainy 

season and less precipitation in months where there are already water deficits. In Guyana, it is 

evident that the topography makes it susceptible to flooding, which has health impacts on the 

populace.  



30 
 

2.3 Data Sources and Descriptive Statistics 
 

2.3.1 Multiple Indicator Cluster Surveys (MICS) 

This paper uses individual-level data from the Multiple Indicator Cluster Surveys (MICS) for 

Guyana. The three rounds of Guyana MICS in 2000, 2006-07 and 2014 were carried out by the 

Guyana Bureau of Statistics in collaboration with UNICEF. MICS is a nationally representative 

sample survey of households and was designed to provide estimates on a variety of indicators 

about children and women across the country. It gathers information from women aged 15-49 

on demographic topics such as birth history, anthropometric data, health service utilisation, and 

nutritional status of mothers and young children under the age of five.  

The Guyana MICS uses the 2002 and 2012 Population and Housing Censuses to divide the 

regions into 300 enumeration districts (EDs). The sample sizes for the women who were 

successfully interviewed in each round of the MICS 2000, 2006-07 and 2014 were 4,801, 5,035 

and 5,076 respectively, resulting in a total of 14,912 women. Each survey aimed at targeting 

approximately 6,000 households in 300 EDs, i.e., 20 households per ED. The total number of 

households was then divided by the number of sample households per ED and 300 clusters 

were allocated to two domains (coastal/urban and interior/rural) across the 10 regions.   

Our outcome variables are height-for-age z-scores (HAZ), and dummy variables for stunted 

and severely stunted growth.  In this study, the HAZ was taken directly from the MICS 

surveys.7 In essence, the z-score is calculated by standardising a child’s height, given age and 

sex, against an international standard of well-nourished children. ‘Stunted’ is a dummy variable 

that takes a value of 1 if the child’s height is -2 SD below the age-sex standardised height of a 

healthy reference population, and 0 otherwise. Similarly, ‘severely stunted’ is a dummy 

variable that takes a value of 1 if the child’s height is more than -3 SD below the mean on the 

WHO Child Growth Standards and 0 otherwise.  

All rounds of the MICS collected detailed information on the month and year of birth of the 

child, region of birth, and height/length of children that were measured by MICS enumerators 

for children aged 0-4. While the overall sample across all 3 rounds of MICS comprises 8,720 

 
7The HAZ is calculated by comparing the child’s height/length with the median value in the reference population. 

The difference is divided by the standard deviation of the reference population as illustrated by the formula: Z-

score = (Individual value of the child – median value of children in the reference population)/Standard deviation 

of the reference population. 



31 
 

children aged 0-4, the HAZ scores are available for 7,636 children, and therefore, we limit the 

sample to those children for this study. 

Control variables used in this analysis is a dummy for the gender of the child (takes the value 

1 if male and 0 if female), a dummy for maternal higher education (which takes the value 1 if 

mother has higher education and 0 otherwise), a dummy for the location of the household (takes 

the value 1 for urban areas and 0 for rural), month of birth, year of birth, and region of birth 

fixed effects, and MICS survey wave fixed effects.  

 2.3.2 Rainfall Data 

This paper uses monthly rainfall data from the Guyana Hydrometeorological Service 

(HydroMet) for 1980-2014 for 68 weather stations spread across the ten regions. As there were 

approximately 15 percent missing station data in the HydroMet, we conducted correlation tests 

between the University of Delaware (UDel) gridded high-resolution station (land) data for 

precipitation and HydroMet data for those stations with complete data in both datasets and 

found that the two datasets were highly correlated.8 Consequently, we used the UDel 1980 -

2014 precipitation data to fill the small gaps in the Guyana HydroMet service rainfall records. 

We use the monthly station-level data to calculate monthly averages for each region. The 

advantage of taking this approach is to have accurate station-level data within the ten regions 

used in the analysis.  

To construct measures for rainfall shocks, we first need to establish historical means and 

standard deviations. The historical mean and standard deviation are calculated for each region 

and month over 1980-2014 (e.g., mean historical rainfall for region 1 for January is calculated 

as the average of rainfall in January across all years spanning 1980-2014; similarly for standard 

deviation). Accordingly, we define a given month as having ‘normal rainfall’ if rainfall fell 

within plus/minus 1 standard deviation of the historical monthly mean for that region. 

Similarly, we can construct indicator variables for ‘positive rainfall shocks’ and ‘negative 

rainfall shocks’ for a given region and month if the rainfall is more than plus 1 or less than 

minus 1 standard deviation with respect to the historical monthly mean of that region 

respectively. Based on the above, we define in-utero positive, negative, and normal rainfall 

 
8 Bovolo et al., 2011 paper shows the gaps in the Guyana Hydromet Service Data in 1999-2001. See: Finescale 

regional climate patterns in the Guianas, tropical South America, based on observations and reanalysis data 

(ncl.ac.uk) 

https://www.staff.ncl.ac.uk/isabella.bovolo/2011-Bovolo-et-al-IJC-Climate-of-the-Guianas.pdf
https://www.staff.ncl.ac.uk/isabella.bovolo/2011-Bovolo-et-al-IJC-Climate-of-the-Guianas.pdf
https://www.staff.ncl.ac.uk/isabella.bovolo/2011-Bovolo-et-al-IJC-Climate-of-the-Guianas.pdf


32 
 

shocks as the fraction of months during the 9 months preceding the month of birth that the 

positive, negative, and normal rainfall indicators equal one. For example, if an individual was 

born in September, then in-utero exposure to positive (negative) rainfall is computed as the 

share of positive (negative) rainfall months during the 9 months spanning January-September.  

In addition, to gain an understanding of which periods of pregnancy are most critical in terms 

of impact of shocks, we also calculate the shocks based on trimesters of birth, i.e., first trimester 

(embryonic), second trimester (foetal) and third trimesters (perinatal).  We construct variables 

for positive and negative rainfall shocks in each trimester of birth. The first trimester positive 

(negative) rainfall is the fraction of months during the 6-8 months before birth that the positive 

(negative) rain indicator equals one. Similarly, the second and third trimesters are defined as 

3-5 months and 0-2 months preceding birth respectively.  

It is important to note that using this approach does not provide any absolute measure of the 

rainfall shocks, as it varies by years, months, and regions. That is, for each month, there is one 

value for the climatical average or historical mean. For instance, for the month of January, the 

average rainfall is the mean for Januarys between the years (1980-2014) for all the stations 

within the regions and constitutes the historical mean or climatological average. Hence, each 

regional result is relative to its own normal rainfall conditions (every region and month would 

have their own historical means based on the (1980-2014) period) as rainfall patterns and 

amounts are considerably varied both spatially and temporally across the country (See Figure 

2.1). The historical mean was calculated by taking several years for the same month. Therefore, 

if the rainfall in that year and month is greater than the historical mean for that month and year 

plus 1 SD standard deviation then it is classified as a positive rainfall shock. This rainfall shock 

is not indicative of flood conditions, merely wetter than normal conditions as there are no 

available flood records for the country. 

 

2.3.3 Summary Statistics 
 

Table 2.1 below provides descriptive statistics of the key variables used in our analysis. The 

HAZ for children ranges from -9.62 to 6.58 with a mean of -0.655 and a standard deviation of 

1.54. On average 16 percent of the children have stunted growth while 5 percent are severely 

stunted. 52 percent of the children are male. On average, 11 percent of the prenatal period of 9 

months was characterised by positive rainfall shocks, 5 percent by negative rainfall shocks and 



33 
 

84 percent by normal rainfall. These figures are similar across the different trimester-specific 

rainfall shock variables. About 2 percent of the mothers have no education, 23 percentage have 

primary education and 75 percent have higher education (classified as those with secondary or 

university education). 74 percent of the children reside in urban locations.  

 

 

Table 2.1: Summary statistics 

Variable  Mean Standard 

deviation 

Min Max N 

Rainfall shock-related variables 

Fraction of Positive rainfall (in-utero) 

Fraction of Negative rainfall (in-utero) 

Fraction of Normal rainfall (in-utero) 

Fraction positive rain – 1st trimester 

Fraction positive rain – 2nd trimester 

Fraction positive rain – 3rd trimester 

Fraction negative rain – 1st trimester 

Fraction negative rain – 2nd trimester 

Fraction negative rain – 3rd trimester 

Fraction normal rain – 1st trimester 

Fraction normal rain – 2nd trimester 

Fraction normal rain – 3rd trimester 

Fraction positive rain – 12 months after 

Fraction negative rain – 12 months after 

Fraction of positive rain (90th10th percentile) 

Fraction positive rain (90th 10th) – 1st trimester 

 

Children Characteristics 

Male 

Height-for-age (z-score) 

Stunted 

Severely Stunted 

 

Postnatal investments 

BCG vaccination 

DPT vaccination 

Measles vaccination 

Ever breastfed 

 

Mother Characteristics 

No education 

Maternal primary education 

Maternal higher education 

Urban 

 

0.108 

0.053 

0.839 

0.112 

0.103 

0.109 

0.055 

0.055 

0.050 

0.834 

0.842 

0.841 

0.107 

0.046 

0.088 

0.089 

 

 

0.515 

-0.641 

0.157 

0.051 

 

 

0.993 

0.927 

0.855 

0.928 

 

 

0.023 

0.220 

0.756 

0.739 

 

0.112 

0.089 

0.131 

0.189 

0.181 

0.188 

0.141 

0.141 

0.135 

0.218 

0.213 

0.215 

0.101 

0.074 

0.095 

0.169 

 

 

0.499 

1.521 

0.364 

0.220 

 

 

0.082 

0.261 

0.355 

0.260 

 

 

0.148 

0.415 

0.429 

0.439 

 

0 

0 

0.333 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

 

 

0 

-5.99 

0 

0 

 

 

0 

0 

0 

0 

 

 

0 

0 

0 

0 

 

0.556 

0.556 

1 

0.667 

0.667 

0.667 

1 

1 

1 

1 

1 

1 

0.5 

0.5 

0.444 

1 

 

 

1 

5.91 

1 

1 

 

 

1 

1 

1 

1 

 

 

1 

1 

1 

1 

 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

7626 

 

 

7626 

7626 

7626 

7626 

 

 

5161 

4375 

4366 

6433 

 

 

7626 

7626 

7626 

7626 

Notes: The sample is restricted to the MICS 2000-2014 on individuals born between 1995 and 2014 and to observations 

where height-for-age z scores are available. Higher education comprises both Secondary and tertiary education. Rainfall 



34 
 

shocks were constructed using data from the Guyana Hydrometeorological Service and the University of Delaware 

Precipitation V5.01. 

 

 

2.4 Empirical strategy 
 

To measure the relationship between in-utero rainfall shocks and early childhood health, we 

employ linear regressions using the following specification:  

 

𝐻𝑖𝑗𝑚𝑡 =   +  
1

𝑅1𝑖𝑗𝑚𝑡 + 
2

𝑅2𝑖𝑗𝑚𝑡 + 
3

𝑍𝑖𝑗𝑚𝑡  +𝑗 + 𝑡 +  
𝑚

+  
𝑤

+ 
𝑖𝑗𝑚𝑡

    (1) 

 

where  𝐻𝑖𝑗𝑚𝑡 is the dependent variable of interest and it represents either HAZ for child i born 

in region j in month m and year t, dummy for being stunted (i.e., HAZ < -2 SD) or dummy for 

being severely stunted (i.e., HAZ < -3 SD). The coefficients of interest 
1
 and 

2
 measures the 

effects of prenatal exposure to rainfall shocks on childhood health. 𝑅1𝑖𝑗𝑚𝑡 is the fraction of 

positive rainfall months during the 9 months preceding birth while 𝑅2𝑖𝑗𝑚𝑡 is the fraction of 

negative rainfall months during the 9 months preceding birth.   𝑍𝑖𝑗𝑚𝑡 is a vector of controls 

which includes a dummy for the gender of child (takes the value 1 if male, 0 if female), a 

dummy for location (takes the value 1 if urban, 0 if rural), and a dummy for maternal education 

(takes the value 1 for higher education, 0 if otherwise). We also include region-of-birth fixed 

effects (𝑗), which absorb any unobservable time-invariant region-level determinants of child 

health, including geography, infrastructure such as roads and area-specific risks of diseases. 

The year-of-birth fixed effects  (
𝑡
)  and month-of-birth fixed effects (

𝑚
) capture any climatic 

or socioeconomic conditions or aggregate shocks impacting the entire country in a certain year 

and month and secular trends in health outcomes. 
𝑤

 is the vector of the MICS survey wave 

fixed effects. 𝑖𝑗𝑚𝑡 is the stochastic error term clustered at enumeration district (ED) level from 

MICS.  

 

Additionally, to understand the effect of trimester-specific positive and negative rainfall 

shocks, we estimate an equation below similar to equation 1, but where the key variables of 

interest (𝑖𝑗𝑚𝑡) are the vectors of trimester-specific shocks.  

 

𝐻𝑖𝑗𝑚𝑡 =   +  ϒ1𝑖𝑗𝑚𝑡 +  ϒ2𝑍𝑖𝑗𝑚𝑡  + 𝑗 + 𝑡 +  
𝑚

+ 
𝑤

+ 
𝑖𝑗𝑚𝑡

                                    (2) 

 



35 
 

In Section 4, we examine factors that might differentially influence the effects of rainfall shocks 

on childhood health. In particular, we examine the heterogeneous effects based on child’s 

gender, maternal higher education, as well as the location of households (rural or urban). The 

following specifications are estimated building on equation 1: 

 

𝐻𝑖𝑗𝑚𝑡 =   +  1𝑅1𝑖𝑗𝑚𝑡 + 2𝑅2𝑖𝑗𝑚𝑡 + 3𝑍𝑖𝑗𝑚𝑡 +  4𝑋𝑖𝑗𝑚𝑡 + 5𝑀𝑖𝑗𝑚𝑡 +  
6

𝑍𝑖𝑗𝑚𝑡  x  𝑅1𝑖𝑗𝑚𝑡 +

7𝑍𝑖𝑗𝑚𝑡  x  𝑅2𝑖𝑗𝑚𝑡 + 𝑗 + 𝑡 +  
𝑚

+ 
𝑤

+ 𝑖𝑗𝑚𝑡               (3)

                                       

𝐻𝑖𝑗𝑚𝑡 =   +  𝑥1𝑅1𝑖𝑗𝑚𝑡 + 𝑥2𝑅2𝑖𝑗𝑚𝑡 + 𝑥3𝑍𝑖𝑗𝑚𝑡 +  𝑥4𝑋𝑖𝑗𝑚𝑡 + 𝑥5𝑀𝑖𝑗𝑚𝑡 + 𝑥6𝑀𝑖𝑗𝑚𝑡  x 𝑅1𝑖𝑗𝑚𝑡 +

𝑥7 𝑀𝑖𝑗𝑚𝑡  x 𝑅2𝑖𝑗𝑚𝑡 + 𝑗 + 𝑡 +  
𝑚

+ 
𝑤

+ 𝑖𝑗𝑚𝑡                          (4) 

 

𝐻𝑖𝑗𝑚𝑡 =   +  Ф1𝑅1𝑖𝑗𝑚𝑡 + Ф2𝑅2𝑖𝑗𝑚𝑡 + Ф3𝑍𝑖𝑗𝑚𝑡 +  Ф4𝑋𝑖𝑗𝑚𝑡 + Ф5𝑀𝑖𝑗𝑚𝑡 + Ф6𝑋𝑖𝑗𝑚𝑡  x 𝑅1𝑖𝑗𝑚𝑡 +

Ф7𝑋𝑖𝑗𝑚𝑡  x 𝑅2𝑖𝑗𝑚𝑡 + 𝑗 + 𝑡 +  
𝑚

+ 
𝑤

+ 𝑖𝑗𝑚𝑡              (5) 

    

where 1, 2, 𝑥1, 𝑥2, Ф1, Ф2 in equations 3-5 represents the main effect of in-utero exposure to 

positive (negative) rainfall shocks on early childhood health. For instance, if 1 (2) in 

equation 3 is negative, this would suggest that the exposure to positive (negative) rainfall 

shocks had a negative impact on early childhood health and if the coefficients are positive the 

opposite is true. 6 − 7  in equations 3 are the coefficients on interactions of prenatal rainfall 

with the child’s gender dummy and urban location dummy. 6 𝑎𝑛𝑑 7 gives the heterogeneous 

effects, i.e., it gives the differential effects of the child’s gender, urban location and the 

mother’s education of those children who had experienced positive (negative) rainfall shocks 

while in-utero. Note, that this analogy would also hold in the case of equations 4 and 5.  

 

Moreover, if 6 and 7 are positive, this would suggest that the positive shocks have 

differentially more positive effect on childhood health for male, while negative 6  and 7 

would suggest the opposite.  Likewise, this analogy would also hold when examining maternal 

higher education and urban/rural location in equations 4 and 5 respectively. 

 

Given that the sample is based on surviving children only, the problem of selective mortality 

(either during pregnancy or in early infancy) may be of concern. Studies have shown that most 

miscarriages occur in the first trimester (Maconochie et al., 2006; Yakusheva and Fletcher, 

2015). Consequently, if the exposure to rainfall shocks affects this selection process, then the 



36 
 

estimated impacts after birth would need to be a combination of selection and a direct treatment 

effect. However, using this selected sample would bias the estimates towards zero. Therefore, 

the estimates in this paper should then be taken to the lower bound of the true effect.  

 

It is also important to note that between the years (2000-2014), Guyana’s under-five mortality 

rate declined sharply from 47 in 2000 to 34 per 1,000 live births in 2014 respectively (UNICEF, 

2023; Guyana MICS 2014). As such, based on the available data provided by the World Bank 

and UNICEF, it is evident that child mortality rates in Guyana have been decreasing in recent 

years. Furthermore, according to the Ministry of Health, there were no known weather shock 

related deaths of children under the age of five years old in Guyana during the 1995-2014 

period. In addition, the MICS 2006-07 and MICS 2014 reported 41 and 34 per thousand birth 

under-five mortality respectively. Thus, representing lower under five mortality rates (see 

MICS 2006-07 report and MICS 2014 report).  

 

In addition, when comparing child mortality between Guyana and other developing countries 

in the low-middle income bracket based on the 2000-2014 period, Bangladesh recorded 58 and 

40 per thousand birth under-five mortality rates respectively in the year 2007 and 2014, while 

Bolivia a country in South America recorded 50 and 33 per thousand birth under-five mortality 

rates respectively in the year 2007 and 2014(World Bank Data, 2023). Furthermore, the 

average under-five mortality rate for all low-middle countries for the year 2000, 2007 and 2014 

stood at 84, 64 and 48 per thousand births respectively (World Bank Data, 2023). Therefore, 

within the 2000-2014 period, the under-five child mortality rate in Guyana was lower than the 

average for countries in the same bracket. 

 

 

 

 

 

 

 

 

 

 



37 
 

 

2.5 Results 

2.5.1 Regional Rainfall Condition (1995-2014) and HAZ scores 

Figure 2.2: Showing Guyana’s Regional Rainfall Condition for the period (1995-2014) 



38 
 

Figure 2.2 above shows the rainfall condition at regional levels based on positive, negative, 

and normal rainfall shocks. The rainfall data is at the regional level. The regions are large areas 

with low population density and centralised health services. The smaller coastal regions of 3, 

4 and 6 has densely populated small areas along the coastal stirp with better transportation and 

health services infrastructure. Considering that the finest resolution of the MICS data is 

regional scale, the rainfall shocks were calculated at the regional level. Figure 2.2 provides a 

depiction of the impact of the positive and negative shocks by regions for children born within 

the nineteen-year period (1995-2014) based on the MICS surveys done in Guyana. It is 

important to make the distinction that the study scales the rainfall between what are defined as 

positive and negative shocks and not floods as individual flood events and extents are not 

available for the regions. 

Based on Figure 2.2, the regions that were most affected are regions 3, region 4, region 6, 

region 8 and region 10. However, in the MICS 2000, 2006-07 and 2014, the greater sample 

which comprised of 3589 respondents were found in regions 3, 4 and 6 (MICS 2000, 2006-07 

and 2014 Reports). According to the 2012 Census, Guyana’s region population distribution in 

regions 3, 4 and 6 were 107,416, 313, 429 and 109,431of Guyana’s 746,955 population 

respectively. These three areas make up 71 percent of Guyana’s population. Consequently, 

these three regions are the most density populated regions in Guyana (Guyana National Land 

Use Plan, 2013; BOS, 2023).91 Moreover, according to the MICS 2000, 2006-07 and 2014, 

regions 3, 4 and 6 comprised of an average of 16 percent, 45 percent, and 18 percent of the 

sample population respectively (MICS 2000, 2006-07 and 2014 Reports). As such, regions 3, 

4 and 6 are not only populous but also flat, so the excessive rainfall (as measured by a positive 

shock) is likely to make more damage there compared to a hilly region (World Bank Climate 

Change Knowledge Portal, 2023). On the other hand, Regions 8 and 10 are inland regions that 

are relatively hilly and are less flood prone and comprised of approximately 3 percent and 5 

percent respectively of the population sample (MICS 2000, 2006-07 and 2014 Reports). 

2.5.2 The effect of rainfall shocks during the gestational period 

 

Tables 2.2 and 2.3 below present the main results where we regress our main outcome variables 

of HAZ score, and indicator variables for stunted and severely stunted growth on prenatal 

 
9https://observatorioplanificacion.cepal.org/sites/default/files/plan/files/2013%20NLUP%20Guyana.pdf; 

https://statisticsguyana.gov.gy/wp-content/uploads/2019/10/2012_Preliminary_Report.pdf  

https://observatorioplanificacion.cepal.org/sites/default/files/plan/files/2013%20NLUP%20Guyana.pdf
https://statisticsguyana.gov.gy/wp-content/uploads/2019/10/2012_Preliminary_Report.pdf


39 
 

rainfall shocks. Table 2.2 considers separately the fraction of positive and negative rainfall 

shocks during the 9 months preceding birth as the primary variables of interest using equation 

1 while Table 2.3 presents the rainfall shocks variables in the first, second and third trimester 

of gestation. 

Table 2.2: The effects of in-utero rainfall shocks on child health 

 (1) (2) (3) 

 Height for age z  

score 

Stunted Severely  

Stunted 

Positive Rainfall -0.248  0.116**    0.075** 

 (0.192) (0.053) (0.031) 

Negative Rainfall 0.172 -0.034 -0.032 

 (0.225) (0.066) (0.036) 

Male    -0.100*** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Urban -0.008 -0.020 0.000 

 (0.086) (0.021) (0.010) 

Maternal higher education  0.258*** -0.059*** -0.023*** 

 (0.041) (0.011) (0.006) 

Constant -1.460*** 0.313***   0.088*** 

 (0.161) (0.046) (0.025) 

Month of birth FE Yes Yes Yes 

Year of birth FE Yes Yes Yes 

Region of birth FE Yes Yes Yes 

Survey wave FE Yes Yes Yes 

Controls Yes Yes Yes 

Observations 7626 7626 7626 

R2 0.095 0.058 0.023 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children 

born between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 
9 months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks 

are defined as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region 

calculated over 1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware 

Precipitation V5.01. Standard errors in parentheses are clustered at the enumeration district level. * p < 0.10, **p < 

0.05, *** p < 0.01. 

 

In Table 2.2, the coefficient on the positive rainfall shock for HAZ score in column 1 is negative 

indicating that higher exposure to positive rainfall shocks during the 9 months in-utero 

adversely affects child height, but it is not statistically significant. However, when examining 

stunted and severely stunted growth in columns 2 and 3 in Table 2.2, we find that a one standard 

deviation increase in positive rainfall in the 9 months before birth increases the likelihood of 

being both stunted and severely stunted by 1.30 and 0.84 percentage points respectively.102 On 

the other hand, the coefficients of the fraction of months with negative rainfall shocks are not 

 
10 NB:// The percentage point for height-for-age z-score, stunted, and severely stunted are calculated by using 

their coefficient values x the standard deviation of positive (negative) rainfall in the 9 months (trimesters (1st, 2nd 

and 3rd)) before birth x 100.  



40 
 

statistically significant. The total number of children under 5 in Guyana for the period 1995-

2014 was approximately 132,719 (Bureau of Statistics, Guyana, 2016). Therefore, a 1.30 

percentage point increase in stunted growth (or 0.013/0.157 = 8 percent increase, when taking 

into consideration the mean of stunted) would increase Guyana’s overall stunted growth by 

approximately 0.013 x 132,719 = 1,725 children. 

Similarly, a 0.84 percentage point increase in severely stunted growth (or 0.0084/0.051 = 16.5 

percent increase when taking into consideration the mean of stunted) would increase Guyana’s 

overall severely stunted growth by approximately 0.0084 x 132,719 = 1,115 children. 

However, these are non-trivial numbers. 

 

Table 2.3: The trimester-specific effects of in-utero rainfall shocks on child health 

 (1) (2) (3) 

 Height for 

age z score 

Stunted Severely  

Stunted 

1st trimester- positive    -0.245*** 0.048*   0.030* 

 (0.093) (0.026) (0.016) 

2nd trimester - positive -0.025 0.044*  0.026 

  (0.106) (0.027) (0.016) 

3rd trimester - positive 0.017 0.023 0.020 

   (0.105) (0.026) (0.015) 

1st trimester - negative 0.063 0.000 0.002 

  (0.139) (0.034) (0.021) 

2nd trimester- negative -0.124 -0.022 -0.001 

  (0.123) (0.036) (0.020) 

3rd trimester - negative  0.264** -0.011 -0.034* 

 (0.126) (0.036) (0.019) 

Male -0.100*** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Urban -0.008 -0.020 0.000 

 (0.086) (0.021) (0.010) 

Maternal higher education   0.256*** -0.059*** -0.023*** 

 (0.041) (0.011) (0.006) 

Constant -1.474*** 

(0.164) 

0.310*** 

(0.046) 

 0.087*** 

(0.025) 

Month of birth FE Yes Yes Yes 

Year of birth FE Yes Yes Yes 

Region of birth FE Yes Yes Yes 

Survey wave FE            Yes                       Yes              Yes 

Controls          Yes                       Yes              Yes 

Observations         7626                     7626             7626 

R2         0.096                     0.058            0.023 

Notes: Sample restricted to 2000, 2006-07 and 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born 

between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall in Panel A is the fraction of months during the 9 

months before birth that the positive shock (negative shock) indicator equals one. 1st trimester positive (negative) rainfall 
is the fraction of months during the 6-8 months before birth that the positive shock (negative shock) indicator equals one. 



41 
 

2nd trimester positive (negative) rainfall is the fraction of months during the 3-5 months before birth that the positive shock 

(negative shock) indicator equals one. 3rd trimester positive (negative) rainfall is the fraction of months during the 0-2 

months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are defined 

as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region calculated over 

1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware Precipitation V5.01. 
Standard errors in parentheses are clustered at the enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

Examining the role of trimester-specific shocks in Table 2.3, we find that positive rainfall 

shocks in the first trimester negatively impact HAZ (column 1). Therefore, on average children 

who experienced a 1 SD increase in positive rainfall in the first trimester in-utero exhibit a 

decline in their HAZ by 0.046 SD. Further, we find that a one standard deviation increase in 

positive rainfall during the first trimester increases the likelihood of stunted and severely 

stunted by 0.91 and 0.57 percentage points respectively. This finding is consistent with 

previous findings. For instance, Le and Nguyen (2021) who studied 55 lower- and middle-

income countries and found that exposure to excess rainfall in the first trimester leads to a 

reduction in HAZ and an increase in stunting. Carrillo (2020) found exposure in the first 

trimester had long-run effects on health outcomes in Brazil. We also find that a one standard 

deviation increase in positive rainfall in the second trimester increases the likelihood of stunted 

growth by 0.80 percentage points respectively. However, when looking at negative rainfall 

shocks, we find that a one standard deviation increase in negative rainfall shocks in the third 

trimester increases the HAZ by 0.037 SD. This result is comparable to Skoufias et al. (2011), 

who found that negative rainfall shocks resulted in increases in HAZ in some regions of 

Mexico. The results in Table 2.3 highlight the value of considering trimester-specific exposure 

to shocks.  

Studies have found that often it is not just in-utero shocks but also those in the first year after 

birth that matter for determining child health (Rosales-Rueda, 2018; Duque et al., 2018; 

Carrillo, 2020). To examine that, we also compute positive and negative rainfall shocks during 

the first twelve months after the child’s birth in addition to the rainfall shocks during the 9 

months of gestation.113  

 

Table 2.4 presents the results where we control for positive and negative rainfall shocks in the 

12 months after birth as well as during each of the trimesters in the gestation period. The results 

indicate that greater exposure to positive rainfall shocks during the first year of life is associated 

with lower HAZ though the coefficients are not statistically significant. However, when 

 
11 The sample size is now reduced, as it now considers children who were at least 1 year old (i.e., 1-4 years old) 

at the time of the survey. 

 



42 
 

looking at negative rainfall exposure, we find that exposure to negative rainfall shocks in the 

12 months after birth negatively affect height for age z scores by 0.051 SD and positively 

impacted severely stunted growth by 1.17 percentage point.  Rosales-Rueda (2018) also finds 

floods in the first year of life to negatively affect height, however, her point estimates were not 

statistically significant. Results show that exposure to negative rainfall shocks during the third 

trimester is associated with an increase in the HAZ by 0.045 SD.  

Table 2.4: The effects of prenatal and postnatal rainfall shocks on child health 

 (1) (2) (3) 

 Height for 

age z score 

Stunted Severely  

Stunted 

Pos 1-12 months after birth -0.003 -0.032 -0.069 

 (0.287) (0.070) (0.045) 

Neg 1-12 months after birth -0.689* 0.142 0.158** 

 (0.362) (0.098) (0.063) 

1st trimester – Positive rainfall -0.166 0.048 0.027 

 (0.108) (0.029) (0.018) 

2nd trimester – Positive rainfall 0.001 0.045 0.020 

 (0.113) (0.031) (0.019) 

3rd trimester – Positive rainfall 0.136 0.004 0.011 

 (0.114) (0.028) (0.017) 

1st trimester – Negative rainfall 0.096 -0.006 -0.004 

 (0.158) (0.040) (0.022) 

2nd trimester – Negative rainfall -0.079 -0.023 -0.001 

 (0.133) (0.039) (0.021) 

3rd trimester – Negative rainfall  0.335** -0.022 -0.040** 

 (0.134) (0.038) (0.020) 

Male -0.069* 0.013 0.001 

 (0.035) (0.009) (0.005) 

Urban 0.096 -0.044* -0.013 

 (0.091) (0.023) (0.011) 

Maternal higher education  0.289*** -0.066*** -0.023** 

 (0.044) (0.013) (0.007) 

Constant -1.405*** 

(0.180) 

0.324*** 

(0.054) 

0.101*** 

(0.029) 

Month of birth FE    Yes Yes Yes 

Year of birth FE    Yes Yes Yes 

Region of birth FE    Yes Yes Yes 

Survey wave FE              Yes                    Yes    Yes 

Controls             Yes                    Yes    Yes 

Observations            6180 6180   6180 

R2           0.099 0.069   0.030 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for individuals 

born between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 9 

months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are 
defined as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region 

calculated over 1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware 



43 
 

Precipitation V5. Standard errors in parentheses are clustered at the enumeration districts level. * p < 0.10, **p < 0.05, 
*** p < 0.01. 

 

2.5.2 The effect of rainfall shocks on postnatal investments 
 

In this section, we examine whether parents make investments depending on the conditions 

faced while the child was in-utero. Studies have found evidence that parents often try to 

equalise (Breining et al., 2015), reinforce (Almond & Mazumder, 2013; Adhvaryu and 

Nyshadham, 2016), or undertake compensatory investments for weaker children because of 

altruism and to afford them an equal opportunity to be productive to compete with their peers 

in their adult life (Behrman et al., 1982; Bhalotra, 2004; Aizer and Cunha, 2014). Studies have 

also shown that childhood vaccination and breastfeeding among others are all fundamental 

actions that protect children from communicable disease outbreaks after positive rainfall 

shocks (Mach et al., 2009; Gribble et al., 2011; Greenwood, 2014).  Consequently, given that 

access to health facilities is limited in Guyana, particularly to those living in the rural areas 

(The Pan American Health Organisation, 2001); vaccinations are seen as an important health 

variable. We focus primarily on children whose parent would have reported that their children 

have some of the WHO’s recommended dose of BCG, DPT and measles vaccinations.124 

Similarly, breastfeeding is also an input because it is an important source of nutrition for 

infants, particularly for those in vulnerable environments where access to safe drinking water 

is inadequate. WHO recommends that mothers exclusively breastfeed infants for the first six 

months of their life. In addition, breastfeeding is of critical importance during and after 

negative rainfall shocks (Quinn et al., 2008; Mach et al., 2009) and there is some evidence to 

suggest that it is also associated with better childhood cognitive development (Kramer et al., 

2008; Del Bono et al., 2012).  

 

The Guyana MICS Survey provides detailed information on postnatal investments such as 

vaccinations (measles, DPT and BCG) and whether the child was breastfed (but not the 

duration of breastfeeding). Consequently, we create dummy variables for measles, DPT, BCG 

and breastfed.  In Table 2.5, we find that greater exposure to positive rainfall shocks by 1 

standard deviation during the 9 months in-utero leads to lower likelihood of getting the measles 

and DPT vaccinations by 1.22 and 1.85 percentage points respectively and an increase in 

 
12 In Guyana, the recommended vaccination schedule is BCG within weeks after birth, DPT at two months, four 

months and six months, and measles at 9-11 months after birth (Guyana, PAHO/WHO, 2018). 

 



44 
 

breastfeeding practices by 1.23 percentage points. It is expected that mothers will invest more 

in postnatal vaccinations, as they try to reduce the risk of their infant contracting infectious 

diseases from the amplification of the disease environment during periods of positive rainfall 

shocks while in-utero. However, given that measles is the vaccination with the lowest 

population coverage (85.5 percent), such a finding seems plausible. Moreover, DPT and BCG 

had a vaccination coverage rate of 92.6 percent and 99.3 percent respectively. We find that 

BCG was the least responsive to weather shocks, as almost the entire population was 

vaccinated. Further, there is essentially a zero effect of negative in-utero rainfall shocks on 

postnatal investments.   

 

Table 2.5: The effects of in-utero rainfall shocks on postnatal investments 

 (1) (2) (3) (4) 

 Child ever 

breastfed? 

Measles 

Vaccination 

DPT 

Vaccination 

BCG 

Vaccination 

Positive Rainfall  0.110*** -0.109***  -0.165*** -0.015 

 (0.037) (0.033) (0.035) (0.011) 

Negative Rainfall -0.050 0.021 0.016 -0.015 

 (0.044) (0.042) (0.037) (0.016) 

Male 0.002 -0.001 -0.002 0.006** 

 (0.007) (0.005) (0.005) (0.002) 

Urban -0.001 -0.008 -0.021** -0.004 

 (0.014) (0.011) (0.012) (0.004) 

Maternal higher education 0.011 0.016** 0.013* 0.000 

 (0.009) (0.007) (0.007) (0.003) 

Constant 0.927*** 

(0.034) 

1.141*** 

(0.023) 

1.054*** 

(0.028) 

1.003*** 

(0.007) 

Month of birth FE 

Year of birth FE 

Region of birth FE 

Survey wave FE 

Controls 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Observations 6433 4366 4375 5161 

R2 0.028 0.733 0.531 0.023 

Notes: Sample restricted to 2000, 2006-07 and 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born between 1995 

and 2014 (age 0-4 years old). Outcome variables are all dummy variables. Positive (negative) rainfall in Panel A is the fraction of months 
during the 9 months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are defined 

as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region calculated over 1980-2014 using 

data from Guyana Hydrometeorological Service and the University of Delaware Precipitation V.501. Standard errors in parentheses are 

clustered at the enumeration district level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

Upon examining postnatal investment responses to trimester-specific shocks in Table 2.6, we 

find that children with greater exposure to positive rainfall shocks in the first and second 

trimesters have a higher probability of being breastfed. Greater exposure to first trimester 



45 
 

positive shocks and second trimester negative rainfall shocks leads to a higher probability of 

children being vaccinated for measles, however the results were not significant. On the other 

hand, positive shocks in the second and third trimesters reduces the probability of being 

vaccinated for measles. Overall, we find that prenatal exposure to positive and negative shocks 

leads to mostly ambiguous results when it relates to postnatal investments.  

Table 2.6: The trimester-specific effects of in-utero rainfall shocks on postnatal investments 

 (1) (2) (3) (4) 

 Child ever 

breastfed? 

Measles 

Vaccination 

DPT 

Vaccination 

BCG 

Vaccination 

1st trimester – positive rainfall 0.033* 0.014 -0.071*** -0.005 

 (0.019) (0.019) (0.019) (0.007) 

2nd trimester – positive rainfall  0.060** -0.048** -0.067*** 0.005 

 (0.020) (0.016) (0.018) (0.007) 

3rd trimester – positive rainfall 0.022 -0.077*** -0.025 -0.015** 

 (0.018) (0.019) (0.017) (0.007) 

1st trimester – negative rainfall -0.008 -0.037 -0.040* -0.011 

 (0.026) (0.026) (0.024) (0.011) 

2nd trimester – negative rainfall 0.014 0.023  0.037 -0.007 

 (0.027) (0.019) (0.023) (0.010) 

3rd trimester – negative rainfall -0.053** 0.029 0.008 0.002 

  (0.026) (0.024) (0.022) (0.007) 

Male 0.002 -0.002 -0.002 0.006*** 

 (0.007) (0.005) (0.005) (0.002) 

Urban -0.001 -0.007 -0.021** -0.004 

 (0.014) (0.012) (0.012) (0.004) 

Maternal higher education 0.011 0.016** 0.013* 0.000 

  (0.010) (0.007) (0.007) (0.003) 

Constant  0.925*** 

(0.034) 

1.147*** 

(0.023) 

1.064*** 

(0.029) 

1.002*** 

(0.007) 

Month of birth FE 

Year of birth FE 

Region of birth FE 

Survey wave FE 

Controls 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Observations 6433 4366 4375 5161 

R2 0.029 0.734 0.532 0.024 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born between 

1995 and 2014 (age 0-4 years old). Outcome variables are all dummy variables. Positive (negative) rainfall in Panel A is the 

fraction of months during the 9 months before birth that the positive shock (negative shock) indicator equals one. 1st trimester 

positive (negative) rainfall is the fraction of months during the 6-8 months before birth that the positive shock (negative shock) 

indicator equals one. 2nd trimester positive (negative) rainfall is the fraction of months during the 3-5 months before birth that 
the positive shock (negative shock) indicator equals one. 3rd trimester positive (negative) rainfall is the fraction of months during 

the 0-2 months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are defined 

as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region calculated over 1980-

2014 using data from Guyana Hydrometeorological Service and the University of Delaware Precipitation V.501. Standard errors 

in parentheses are clustered at the enumeration district level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 



46 
 

2.5.3 Heterogeneous Effects  
 

In this section, we investigate whether there are heterogeneous effects of rainfall shocks on 

child health based on (i) the gender of the child; (ii) whether the child belongs to an urban 

household; (iii) and the education level of the mother. To analyse this, split the samples. 

 

We first examine the sex of the child. For the past twenty years, the under-five mortality rate 

in Guyana is higher for male children as compared to females (UNICEF, 2020). In addition, 

research also shows that in times of trouble (i.e., severe weather conditions, earthquakes, 

natural disasters etc), maternal stress caused by the external environment around the time of 

conception exacerbates the difference, resulting in a reduction in the male to female sex ratio, 

thus suggesting that the male embryo is more vulnerable than that of a female (Kraemer, 2000).  

As such, in general, it would be a rational decision for households to direct more postnatal 

investments towards boys rather than girls in periods of adverse weather shocks. 

 

Results presented in Table 2.7 (column 1), show that the interaction term for males and the 

positive and negative rainfall variables are positive but not statistically significant when 

examining the height-for-age z score. Therefore, there is no evidence to suggest that the effects 

of exposure to positive and negative rainfall shock significantly varies by gender. Moreover, 

we found similar results for stunting.  

 

Weather shocks may have affected urban and rural households differently since access to 

medical facilities in urban and rural areas may be different due to greater road developments 

in urban areas (Rose and Corbin, 2017). Maternal education can be seen as an important 

determinant for the health of the children in-utero. Studies have shown that negative exposure 

to weather shocks are attenuated during the first trimester if the mother is more educated 

(Rosales-Rueda, 2018).  

 

 

 

 

 

 

 



47 
 

Table 2.7: Heterogeneous effects by gender of the child 

 (1) (2) (3) 

 Height for age z  

score 

Stunted Severely  

Stunted 

Positive rainfall shock -0.182 0.091   0.082* 

 (0.249) (0.063) (0.042) 

Negative rainfall shock 0.229 -0.048 -0.043 

 (0.271) (0.083) (0.045) 

Male -0.081 0.011 0.005 

 (0.052) (0.013) (0.008) 

Urban -0.008 -0.020 0.000 

 (0.086) (0.021) (0.010) 

Maternal higher education  0.257***   -0.059*** -0.023*** 

 (0.041) (0.011) (0.006) 

Positive rain x Male -0.127 0.047 -0.015 

 (0.295) (0.072) (0.048) 

Negative rain x Male -0.107 0.027 0.023 

 (0.359) (0.088) (0.051) 

Constant -1.470*** 

(0.162) 

   0.316*** 

(0.047) 

   0.088*** 

(0.025) 

Month of birth FE   Yes Yes Yes 

Year of birth FE   Yes Yes Yes 

Region of birth FE   Yes Yes Yes 

Survey wave FE Yes            Yes     Yes 

Controls Yes            Yes     Yes 

Observations 7626             7626    7626 

R2 0.095              0.058    0.023 
Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born 

between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 9 months 

before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are defined as 

+/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region calculated over 

1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware Precipitation V5.01. 

Standard errors in parentheses are clustered at the enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

Second, we examine the role of urban versus rural location in Table 2.8. Socioeconomic factors 

and other variables are important when persons choose where they reside, hence, the 

importance of analysing the heterogeneity by location. Residential location can also influence 

childhood health outcomes through access to adequate health facilities in rural areas etc (Hailu 

et al., 2020). In Table 2.8, column 1 below, we find that positive rainfall shock increases height-

for-age z-scores in urban areas (coefficient is 1.096 and significant at 1%) as compared to rural 

areas. The reason for the positive impact in urban areas is due to better access to medical 

services, infrastructure, disaster relief resources and resilience (Rözer, 2022). Whereas the rural 

areas have more sparse infrastructure and public service provision.  

 

  



48 
 

Table 2.8: Heterogeneous effects by location 

 (1) (2) (3) 

 Height for age 

z score 

Stunted Severely  

Stunted 

Positive rainfall shock -1.075*** 0.208* 0.063 

 (0.323) (0.110) (0.056) 

Negative rainfall shock 0.436 -0.198 -0.092 

 (0.488) (0.149) (0.063) 

Male -0.101*** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Urban -0.088 -0.022 -0.006 

 (0.101) (0.027) (0.012) 

Maternal higher education 0.259*** -0.060*** -0.023*** 

 (0.041) (0.011) (0.006) 

Positive rainfall shock x Urban 1.096*** -0.120 0.017 

 (0.374) (0.118) (0.063) 

Negative rainfall shock x Urban -0.330 0.221 0.083 

 (0.551) (0.160) (0.065) 

Constant -1.413*** 

(0.163) 

   0.317*** 

(0.048) 

  0.093*** 

(0.026) 

Month of birth FE          Yes Yes Yes 

Year of birth FE          Yes Yes Yes 

Region of birth FE          Yes Yes Yes 

Survey wave FE            Yes                  Yes         Yes 

Controls              Yes        Yes Yes 

Observations            7626        7626 6693 

R2            0.096         0.061 0.022 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born 

between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 9 months 
before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks are defined as 

+/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region calculated over 

1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware Precipitation V5.01. 

Standard errors in parentheses are clustered at the enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

To examine the differential effects of maternal education on the effects of rainfall shocks on 

child health, we create binary variables that take the value of one if the mother had completed 

‘secondary education’ or ‘tertiary education’, and zero otherwise (i.e., mothers with primary 

and no education). Studies have shown differential responses as it relates to maternal education, 

i.e., maternal education can play a significant role in mitigating the negative impacts of weather 

shocks on childhood health (Rosales-Rueda, 2018; Dimitrova and Bora, 2020). As such 

maternal education is an important variable to consider for heterogeneity. Results presented in 

Table 2.9 show that higher education of the mother positively affects the height-for-age z score 

and negatively affects stunted and severely stunted; however, when interacting mother’s 

education with the rainfall shock variables in (columns 1-3), we find that there are no 



49 
 

heterogeneous effects of mother’s education and the in-utero exposure to positive and negative 

rainfall shocks.  

 

Table 2.9: Heterogeneous effects by mother’s education 

 (1) (2) (3) 

 Height for age 

z score 

Stunted Severely stunted 

Positive rainfall shock -0.241 0.058 0.043 

 (0.356) (0.104) (0.057) 

Negative rainfall shock 0.068 -0.017 0.001 

 (0.370) (0.104) (0.066) 

Male -0.100*** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Maternal higher education 0.250*** -0.065*** -0.024** 

 (0.065) (0.017) (0.010) 

Positive rain shock x Higher -0.010 0.075 0.041 

 (0.381) (0.105) (0.061) 

Negative rain shock x Higher 0.148 -0.026 -0.047 

 (0.383) (0.103) (0.065) 

Urban -0.008 -0.020 0.000 

 (0.086) (0.021) (0.010) 

Constant  -1.453*** 

(0.165) 

  0.317*** 

(0.047) 

  0.088*** 

(0.025) 

Month of birth FE        Yes Yes               Yes 

Year of birth FE        Yes Yes  Yes 

Region of birth FE        Yes Yes  Yes 

Survey wave FE            Yes    Yes      Yes 

Controls            Yes    Yes      Yes 

Observations          7626      7626      7626 

R2         0.095       0.058     0.023 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born 

between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 9 months 

before birth that the positive shock (negative shock) indicator equals one. Higher education is a dummy variable that 
takes the value 1 for mothers who completed secondary and/or tertiary education and 0 otherwise (no education and 

primary education). Positive and negative shocks are defined as +/- 1 standard deviation respectively with respect to the 

historical monthly mean of each birth region calculated over 1980-2014 using data from Guyana Hydrometeorological 

Service and the University of Delaware Precipitation V5.01. Standard errors in parentheses are clustered at the 

enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

 

2.5.4 Robustness Checks  

 

For robustness, we consider alternative measures of rainfall shocks. More specifically, 

following Jayachandran (2006), we define a region-month-year as having a positive rainfall 

shock which takes a value of one if the rainfall in that month exceeds the ninetieth percentile 

of the region’s historical monthly average, and 0 otherwise. A negative rainfall shock takes a 

value of one if the rainfall in that month is below the tenth percentile of the region’s historical 



50 
 

monthly average, and 0 otherwise. Similarly, normal rainfall lies within the ninetieth and tenth 

percentiles of the historical regional mean. We then compute the fraction of months in the 9 

months preceding birth that is characterized by positive shocks, negative shocks, and normal 

rainfall, and we also split it into trimesters. 

In Table 2.10 we find that the results for stunted, and severely stunted growth remain robust, 

i.e., positive rainfall shocks significantly increase the likelihood of being stunted and severely 

stunted by 1.74 and 0.97 percentage points respectively.  

 

Table 2.10: The effects of in-utero rainfall shocks on child health (90th, 10th percentile measure)  

 (1) (2) (3) 

 Height for age  

z score 

Stunted Severely  

Stunted 

Positive Rainfall -0.357 0.155** 0.087** 

 (0.221) (0.057) (0.033) 

Negative Rainfall 0.018 -0.030 -0.006 

 (0.177) (0.049) (0.024) 

Male -0.101** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Urban -0.130 -0.019 0.000 

 (0.087) (0.021) (0.010) 

Maternal higher education    0.257*** -0.059*** -0.023** 

 (0.042) (0.011) (0.006) 

Constant -1.412*** 

(0.163) 

0.303*** 

(0.046) 

0.078** 

(0.025) 

Month of birth FE     Yes          Yes    Yes 

Year of birth FE     Yes          Yes    Yes 

Region of birth FE Yes          Yes    Yes 

Survey wave FE     Yes                      Yes Yes 

Controls     Yes                      Yes Yes 

Observations 7626 7626 7626 

R2 0.095 0.058 0.023 
Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born 

between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 9 months before 

birth that the positive shock (negative shock) indicator equals one. Standard errors in parentheses are clustered at the 

enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

Table 2.11 presents the results for the impact of positive and negative rainfall shocks on each 

trimester in-utero. The estimates in column 1 suggest that on average children who experienced 

a standard deviation increase in positive rainfall in the first trimester in-utero exhibited a 

decline in their height-for-age z score by 0.044 SD. This result is quantitatively analogous to 

our main result in Table 2.3.  In columns 2 and 3, we find that a one standard deviation increase 

in positive rainfall shocks in the first trimester also increases the likelihood of being stunted 



51 
 

and severely stunted by 1.21 and 0.87 percentage points respectively. However, when looking 

at the impacts of negative rainfall shocks, the results are not very robust.  

 

Table 2.11: The trimester-specific effects of in-utero rainfall shocks on child health (90th, 10th 

percentile measure) 

 (1) (2) (3) 

 Height for age z 

score 

Stunted Severely  

Stunted 

1st trimester – positive rainfall  -0.233**    0.064**    0.046** 

 (0.106) (0.028) (0.016) 

2nd trimester – positive rainfall -0.020 0.052* 0.022 

 (0.118) (0.030) (0.017) 

3rd trimester – positive rainfall -0.108 0.038 0.021 

 (0.114) (0.029) (0.017) 

1st trimester – negative rainfall -0.024 -0.004 0.017 

 (0.098) (0.026) (0.015) 

2nd trimester – negative rainfall -0.027 -0.029 -0.010 

 (0.099) (0.026) (0.016) 

3rd trimester – negative rainfall 0.075 0.002 -0.013 

 (0.097) (0.027) (0.013) 

Male -0.100*** 0.018** 0.004 

 (0.034) (0.008) (0.005) 

Urban -0.010 -0.019 0.000 

 (0.087) (0.021) (0.010) 

Maternal higher education   0.256*** -0.059*** -0.023** 

 (0.041) (0.011) (0.006) 

Constant -1.419*** 

(0.163) 

0.300*** 

(0.046) 

0.076** 

(0.025) 

Month of birth FE        Yes    Yes               Yes 

Year of birth FE        Yes    Yes  Yes 

Region of birth FE        Yes    Yes  Yes 

Survey wave FE            Yes       Yes      Yes 

Controls                Yes       Yes      Yes 

Observations 7626 7626 7626 

R2 0.095 0.058 0.023 
   Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children born between 

1995 and 2014 (age 0-4 years old). The 1st trimester positive (negative) rainfall is the fraction of months during the 6-8 months 

before birth that the positive shock (negative shock) indicator equal one. 2nd trimester positive (negative) rainfall is the fraction 

of months during the 3-5 months before birth that the positive shock (negative shock) indicator equals one. 3rd trimester positive 

(negative) rainfall is the fraction of months during the 0-2 months before birth that the positive shock (negative shock) indicator 

equals one. Positive (negative) rainfall shock=1 and 0 otherwise, if for a given region and month rainfall was above the 90th 

(below the 10th) percentile of the region’s historical monthly average calculated over 1980-2014 using data from Guyana 

Hydrometeorological Service and the University of Delaware Precipitation V5.01. Standard errors in parentheses are clustered 

at the enumeration districts level. * p < 0.10, **p < 0.05, *** p < 0.01. 
 

 

 

 



52 
 

2.5.5 Placebo Test 
 

Table 2.112: The effects of in-utero rainfall shocks on child health 

 (1) (2) (3) 

 Height for age z  

score 

Stunted Severely  

Stunted 

Positive Rainfall -0.248  0.116**    0.075** 

 (0.192) (0.053) (0.031) 
 [0.103] [0.000] [0.001] 

Negative Rainfall 0.172 -0.034 -0.032 

 (0.225) (0.066) (0.036) 

 [0.337] [0.457] [0.275] 

Month of birth FE Yes Yes Yes 

Year of birth FE Yes Yes Yes 

Region of birth FE Yes Yes Yes 

Survey wave FE Yes Yes Yes 

Controls Yes Yes Yes 

Observations 7626 7626 7626 

R2 0.095 0.058 0.023 

Notes: Sample restricted to 2000, 2006-07 & 2014 Guyana Multiple Indicator Cluster Survey (MICS) for children 

born between 1995 and 2014 (age 0-4 years old). Positive (negative) rainfall is the fraction of months during the 

9 months before birth that the positive shock (negative shock) indicator equals one. Positive and negative shocks 
are defined as +/- 1 standard deviation respectively with respect to the historical monthly mean of each birth region 

calculated over 1980-2014 using data from Guyana Hydrometeorological Service and the University of Delaware 

Precipitation V5.01. The results for positive and negative rainfall shocks were run separately. All regressions 

include controls for male, urban and maternal higher education. Standard errors in parentheses are clustered at the 

enumeration district level. * p < 0.10, **p < 0.05, *** p < 0.01. 

 

When performing randomization inference using the ritest command in STATA on the 

outcome variables. The results in Table 2.12 show that exposure to positive rainfall shocks in 

the 9 months preceding birth for the outcome variables stunted and severe stunted growth are 

significant, with p=c/n being 0.000 and 0.001 respectively. These results are also significant in 

our main regressions, thus indicating that results are robust when performing placebo tests.  

However, when looking that the negative rainfall shocks the results were not significant. This 

is also like our results in our baseline equations.  

 

 

 

 

 



53 
 

2.6 Conclusion  
 

This study uses historical weather data from the Guyana Hydrometeorology Service combined 

with the Guyana Multiple Indicator Cluster Survey microdata for 2000, 2006-07 and 2014 to 

generate insights into the effects of in-utero weather shocks on early childhood health as 

measured by HAZ and stunting of children. This study contributes to a growing body of 

economic evidence on the adverse effects of extreme weather events experienced in early 

childhood, by providing evidence of the effects of early-life exposure to positive and negative 

rainfall shocks in Guyana.    

 

The findings reveal that a one standard deviation increase in the exposure to positive rainfall 

shocks during in-utero significantly increases the likelihood of being stunted and severely 

stunted. Findings also show that in-utero exposure to positive rainfall shocks in the first 

trimester of gestation negatively impacts the HAZ and increases the likelihood of stunted and 

severely stunted growth. We find some mixed evidence that parents are more likely to 

undertake postnatal investments in the form of vaccinations and breastfeeding for children who 

experienced more shocks during gestation. Our findings show that climatic variability also 

appears to have substantial heterogeneous impacts based on location, such that negative rainfall 

shocks have a greater effect on the probability of stunting in urban areas as compared to rural 

areas. Although this study looks at the early childhood impacts of weather shocks in Guyana, 

it is likely that this will lead to long-term impacts on adult health, education, and socioeconomic 

outcomes as highlighted by previous studies.  

 

However, there have been several responses made by the Government of Guyana to mitigate 

the problem of increased flooding in Guyana. For instance, in 2010 the Ministry of Agriculture 

opened 4,500 acres of new land to allow traditional coastal farmers to start moving their homes 

and farms inland (Stabroek News, 2010). Moreover, the projected adaptation costs for building 

and reinforcing levees and seawalls to flood-proofing health clinics was approximately USD 

$1 billion which was equivalent to nearly 20 percent of Guyana’s GDP in 2010. Furthermore, 

in 2013, the Hope Canal (East Demerara Water Conservancy-Northern Relief Channel) was 

constructed in response to the 2005 flood that affected regions 3, 4 and 5. The Hope Canal was 

constructed to mitigate the effects of flooding on more than 300 communities in region 4. 

Furthermore, in June 2021, the government of Guyana declared a disaster as flooding affected 

all ten regions. Consequently, the Government of Guyana had proposed to build three outfall 



54 
 

channels for regions 3, 5 and 6 to mitigate Guyana’s yearly flood phenomenon (Stabroek News, 

2021).  

 

Although policy has been enacted to mitigate the occurrence of floods in Guyana, one potential 

policy recommendation emerging from this study is that Guyana’s Ministry of Health should 

consider collaborating with a multidisciplinary research team to specifically assess the 

medium-and long-term effects of adverse weather shocks on child and maternal health. 

Nutrition, and medical outreach programmes for pregnant mothers, in regions that are more 

susceptible to flooding should be considered. Furthermore, adding climatologists to this team 

to help monitor and predict the onset of these shocks would better prepare to put provisions in 

place to improve maternal and early childhood health in the aftermath of shocks.  

 

Furthermore, this paper contributes to our understanding of the effects of such shocks on child 

health outcomes. Since we believe that the severity of climate change effects is likely to 

increase in the future, it is important to provide evidence to Guyana’s authorities about the 

impact of rainfall shocks on child health, which is fundamental for the long-term health 

outcomes and, ultimately, for Guyana’s economic prosperity. 

  



55 
 

References 

 

Adhvaryu, A. and Nyshadham, A., (2016). Endowments at birth and parents' investments in 

children. The Economic Journal, 126(593), pp.781-820. 

 

Aguilar, A. and Vicarelli, M. (2022). El Nino and children: medium-term effects of early-life 

weather shocks on cognitive and health outcomes. World Development, 150, 105690. 

 

Aizer, A. and Cunha, F., (2014). Child endowments, parental investments, and the production 

of human capital. National Bureau for Economic Research Working paper, 18429. 

Alderman, H., Hoddinott, J., and Kinsey, B. (2006). Long Term Consequences of Early 

Childhood Malnutrition. Oxford Economic Papers, 53(4), pp. 450-474.  

Almond, D., Currie, J. and Duque, V., (2018). Childhood circumstances and adult outcomes: 

Act II. Journal of Economic Literature, 56(4), pp.1360-1446. 

Almond, D., Edlund, L. and Palme, M., (2009). Chernobyl's subclinical legacy: prenatal 

exposure to radioactive fallout and school outcomes in Sweden. The Quarterly Journal of 

Economics, 124(4), pp. 1729-1772.  

Almond, D. and Mazumder, B., (2013). Foetal origins and parental responses. Annual. Rev. 

Econ., 5(1), pp.37-56. 

Altshuler, K., M. Berg, L.M. Frazier, J. Laurenson, J. Longstreth, W.Mendez, and C. A 

Molgaard., (2003). Critical periods in development. OCHP Paper Series on Childrens Health 

and the Environment, 3:1–31 

Agüero, J., (2014). Long-term effect of climate change on health: Evidence from heat waves 

in Mexico, IDB-WP-481. 

Banerjee, Abhijit, Duflo, Esther, Postel‐Vinay, Gilles, and Watts, Tim., (2010). “Long run 

health impacts of income shocks: Wine and phylloxera in 19th century France,” The Review 

of Economics and Statistics 92(4): 714– 728. 

Barker, D.J., (1995). Fetal origins of coronary heart disease. British Medical 

Journal, 311(6998), pp.171-174. 

Behrman, J.R., Pollak, R.A. and Taubman, P., (1982). Parental preferences and provision for 

progeny. Journal of political economy, 90(1), pp.52-73. 

Beuermann, D. and Pecha, C., (2020). The effects of weather shocks on early childhood 

development: Evidence from 25 years of tropical storms in Jamaica. Economics & Human 

Biology, 37, p. 100851. 

Bhalotra, S., (2004). Early childhood investments in human capital: parental resources and 

preferences. University of Bristol Discussion Paper, 4, p.561. 



56 
 

Bonjean, C., Brunelin, S. and Simonet, C., (2012). Impact of climate related shocks on child’s 

health in Burkina Faso (No. 201232). CERDI. 

Botzen, W.W., Deschenes, O. and Sanders, M., (2019). The economic impacts of natural 

disasters: A review of models and empirical studies. Review of Environmental Economics and 

Policy, 13(2), pp.167-188. 

 

Bovolo, C.I., Pereira, R., Parkin, G., Kilsby, C. and Wagner, T., (2012). Fine‐scale regional 

climate patterns in the Guianas, tropical South America, based on observations and reanalysis 

data. International Journal of climatology, 32(11), pp.1665-1689. 

 

Bozzoli, C. and Quintana-Domeque, C., (2014). The weight of the crisis: Evidence from 

newborns in Argentina. Review of Economics and Statistics, 96(3), pp.550-562. 

 

Breining, S., Daysal, N.M., Simonsen, M. and Trandafir, M., (2015). Spillover effects of 

early-life medical interventions. 

 

Burgess, R., Deschenes, O., Donaldson, D. and Greenstone, M., (2014). The unequal effects of 

weather and climate change: Evidence from mortality in India. Cambridge, United States: 

Massachusetts Institute of Technology, Department of Economics. Manuscript. 

Camacho, A., (2008). Stress and birth weight: evidence from terrorist attacks. American 

Economic Review, 98(2), pp. 511-15.  

 

Carrillo, B., (2020). Early Rainfall Shocks and Later-Life Outcomes: Evidence from 

Colombia. The World Bank Economic Review, 34(1), pp.179-209. 

Climate Change., (2007). Impacts, Adaptation and Vulnerability. Ipcc.ch. (2007). Available at: 

https://www.ipcc.ch/site/assets/uploads/2018/03/ar4_wg2_full_report.pdf [Accessed 12th Apr. 

2020].  

Climate Change Legislation in Guyana., (2015). Available at 

http://www.lse.ac.uk/GranthamInstitute/wp-content/uploads/2015/05/GUYANA.pdf 

[Accessed: 5th Apr. 2020] 

Currie, J. and Rossin-Slater, M., (2013). Weathering the storm: Hurricanes and birth 

outcomes. Journal of Health Economics, 32(3), pp.487-503. 

Currie, J. and Vogl, T., (2013). Early-Life Health and Adult Circumstance in Developing 

Countries. Annual Review of Economics, 5(1), pp. 1-36.  

Del Bono, E., Ermisch, J. and Francesconi, M., (2012). Intrafamily resource allocations: a 

dynamic structural model of birth weight. Journal of Labour Economics, 30(3), pp.657-706. 

Dercon, S. and Porter, C., (2014). Live Aid Revisted: Long-term impacts of the 1984 Ethiopian 

Famine on Children. Journal of the European Economic Association, 12(4), pp. 927-948.  

Dimitrova, A. and Bora, J.K., (2020). Monsoon weather and early childhood health in 

India. PloS one, 15(4), p.e0231479. 

http://www.lse.ac.uk/GranthamInstitute/wp-content/uploads/2015/05/GUYANA.pdf


57 
 

Douglas, B., (1995). 'Global Sea level change: Determination and interpretation', 

Revs.Geophys. (supp.), pp. 1425-1432. 

Duque, V., Rosales-Rueda, M. and Sanchez, F., (2018), May. How do early-life shocks interact 

with subsequent human-capital investments? Evidence from administrative data. In IZA World 

of Labor Conference. 

Edoka, I.O., (2013). Weather shocks and nutritional status of disadvantaged children in 

Vietnam. Health Econometrics and Data Group, The University of York, WP, 13(10). 

Glewwe, P. and King, E., (2001). The Impact of Early Childhood Nutritional Status on 

Cognitive Development: Does the Timing of Malnutrition Matter? The World Bank Economic 

Review, 15(1), pp. 81-113.  

Greenwood, B., (2014). The contribution of vaccination to global health: past, present and 

future. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1645), 

p.20130433. 

 

Gribble, K.D., McGrath, M., MacLaine, A. and Lhotska, L., (2011). Supporting breastfeeding 

in emergencies: protecting women's reproductive rights and maternal and infant 

health. Disasters, 35(4), pp.720-738. 

 

Guyana, Gov., (2012). Guyana Second National Communication to the United Nations 

Framework Convention on Climate Change, pp. 1-437 

 

Guyana., (2000, 2006-07 and 2014). Multiple Indicator Cluster Survey. Available at: 

https://www.unicef.org/statistics/index_24302.html [Accessed 9th May 2020] 

 

Hallegatte, S., Vogt-Schilb, A., Rozenberg, J., Bangalore, M. and Beaudet, C., 2020. From 

poverty to disaster and back: A review of the literature. Economics of Disasters and Climate 

Change, 4, pp.223-247. 

Hoddinott, J. and Kinsey, B., (2001). Child Growth in the Time of Drought. Oxford Bulletin of 

Economics and Statistics, 63(4), pp. 409-436.  

Jayachandran, S., (2006). Selling Labor Low: Wage Responses to Productivity Shocks in 

Developing Countries. Journal of Political Economy, 114(3), pp. 538-575. 

Jayachandran, S. and Pande, R., 2017. Why are Indian children so short? The role of birth order 

and son preference. American Economic Review, 107(9), pp.2600-2629. 

Kramer, M., (1987). Determinants of low birth weight: methodological assessment and meta-

analysis. Bulletin of the world health organization, 65(5), pp. 663. 

Kramer, M., Aboud, F., Mironova, E., Vanilovich, I., Platt, R.W., Matush, L., Igumnov, S., 

Fombonne, E., Bogdanovich, N., Ducruet, T. and Collet, J., (2008). Breastfeeding and child 

cognitive development: new evidence from a large randomized trial. Archives of general 

psychiatry, 65(5), pp. 578-584. 

https://www.unicef.org/statistics/index_24302.html


58 
 

Kumar, S., Molitor, R. and Vollmer, S., (2016). Drought and early child health in rural 

India. Population and Development Review, pp.53-68. 

 

Lakhan, V.C., (2014). Use of geospatial techniques to relocate flood-prone coastal 

communities in Guyana. Contract Report No. 11. Caribbean Research Solutions, Guyana and 

Toronto 

 

Le, K. and Nguyen, M., (2021). In-utero exposure to rainfall variability and early childhood 

health. World Development, 144, p.105485. 

Lee, B. E., E. H. Ha, H. S. Park, Y. J. Kim, Y. C. Hong, H. Kim, and J. T. Lee., (2003). 

“Exposure to Air Pollution during Different Gestational Phases Contributes to Risks of Low 

Birth Weight.” Human Reproduction 18 (3): 638–43 

Leight, J., Glewwe, P. and Park, A., (2015). The impact of early childhood rainfall shocks on 

the evolution of cognitive and non-cognitive skills.  

Leroy, J.L., Ruel, M., Habicht, J.P. and Frongillo, E.A., (2015). Using height-for-age 

differences (HAD) instead of height-for-age z-scores (HAZ) for the meaningful measurement 

of population-level catch-up in linear growth in children less than 5 years of age. BMC 

pediatrics, 15(1), pp.1-11. 

 

López, R.E., Thomas, V. and Troncoso, P., (2015). Climate change and natural 

disasters. Santiago. 

Maccini, S. and Yang, D., (2009). Under the Weather: Health, Schooling, and Economic 

Consequences of Early-Life Rainfall. American Economic Review, 99(3), pp. 1006-1026.  

Mach, O., Lu, L., Creek, T., Bowen, A., Arvelo, W., Smit, M., Masunge, J., Brennan, M. and 

Handzel, T., (2009). Population-based study of a widespread outbreak of diarrhea associated 

with increased mortality and malnutrition in Botswana, Jan–Mar 2006. The American journal 

of tropical medicine and hygiene, 80(5), pp.812-818. 

Ministry of Communities., (2016). Capital Towns and Economic Development. Guyana: 

communities. Available at: https://moc.gov.gy/capital-towns-and-economic-development 

[Accessed 10th April 2020]. 

National Institute for Health Care Excellence., (2017). Available at: https://www.nice.org.uk/ 

[Accessed: 1st May 2020] 

 

Ogasawara, K. and Yumitori, M., (2019). Early-life exposure to weather shocks and child 

height: Evidence from industrializing Japan. SSM-population health, 7, p.100317 

 

Pörtner, H.O., Roberts, D.C., Adams, H., Adler, C., Aldunce, P., Ali, E., Begum, R.A., Betts, 

R., Kerr, R.B., Biesbroek, R. and Birkmann, J., (2022). Climate change 2022: Impacts, 

adaptation and vulnerability. IPCC Sixth Assessment Report. 

 

Quinn, D., Lavigne, S.V., Chambers, C., Wolfe, L., Chipman, H., Cragan, J.D. and Rasmussen, 

S.A., (2008). Addressing concerns of pregnant and lactating women after the 2005 hurricanes: 

https://moc.gov.gy/capital-towns-and-economic-development
https://www.nice.org.uk/


59 
 

the OTIS response. MCN: The American Journal of Maternal/Child Nursing, 33(4), pp.235-

241. 

 

Randell, H., Gray, C. and Grace, K., (2020). Stunted from the start: Early life weather 

conditions and child undernutrition in Ethiopia. Social Science & Medicine, p.113234. 

Rocha, R. and Soares, R., (2015). Water scarcity and birth outcomes in the Brazilian 

semiarid. Journal of Development Economics, 112, pp. 72-91. 

Rocha, R. and Soares, R., (2011). September. Early life health and educational performance: 

evidence from rainfall fluctuations in the Brazilian northeast. In 33º Meeting of the Brazilian 

Econometric Society. 

Rosales-Rueda, M., (2018). The impact of early life shocks on human capital formation: 

evidence from El Niño floods in Ecuador. Journal of Health Economics, 62, pp. 13-44.  

Rose, S. and Corbin, H., (2017). Prospects and constraints of rehabilitating and expanding 

Guyana’s road network, ISSN 15169111, Papers Donaea, 371. 

Salm, M. and Schunk, D., (2012). The relationship between child health and developmental 

gaps, and parental education: Evidence from administrative data. Journal of the European 

Economic Association, 10(6), pp. 1425-1449. 

Skoufias, E., Vinha, K. and Conroy, H.V., (2011). The impacts of climate variability on welfare 

in rural Mexico. The World Bank. 

Smith, R.C., Ainley, D., Baker, K., Domack, E., Emslie, S., Fraser, B., Kennett, J., Leventer, 

A., Mosley-Thompson, E., Stammerjohn, S. and Vernet, M., (1999). Marine ecosystem 

sensitivity to climate change: Historical observations and paleoecological records reveal 

ecological transitions in the Antarctic Peninsula region. BioScience, 49(5), pp. 393-404. 

Stein, A.D. and Lumey, L.H., (2000). The relationship between maternal and offspring birth 

weights after maternal prenatal famine exposure: the Dutch Famine Birth Cohort 

Study. Human biology, pp.641-654. 

Thomson Reuters Foundation News., (2011). Available at 

https://news.trust.org/item/20110601141800-sh9ro  [Accessed: 15th May 2020] 

 

United Nations., (2019), Unprecedented Impacts of Climate Change Disproportionately 

Burdening Developing Countries, Delegate Stresses, as Second Committee Concludes General 

Debate, https://www.un.org/press/en/2019/gaef3516.doc.htm [ Accessed: 10th Jan 2021] 

United Nation International Children’ Emergency Fund (UNICEF)., (2007). “Child & 

Development 2005 Survey (Multiple Indicators Cluster Survey 3) Final Report”, Ulaanbaatar.  

United Nation International Children’ Emergency Fund (UNICEF)., (2020). 

https://data.unicef.org/country/guy/ [Accessed: 1st Sept 2020] 

 

https://news.trust.org/item/20110601141800-sh9ro
https://www.un.org/press/en/2019/gaef3516.doc.htm
https://data.unicef.org/country/guy/


60 
 

Victora C.G., Adair L., Fall C. Maternal and child undernutrition: Consequences for adult 

health and human capital. Lancet. (2008);371:340–357. [PMC free article] [PubMed] [Google 

Scholar] 

 

World Bank., (2020), Available at: https://data.worldbank.org/country/GY [Accessed: 10th 

May 2020] 

 

World Bank Climate Change Knowledge Portal., (2020). Available at: 

https://climateknowledgeportal.worldbank.org/ [Accessed: 10th June 2020] 

 

World Bank Data., (2023), Available at:  

https://data.worldbank.org/indicator/SH.DYN.MORT?end=2014&locations=XO&start=2000 

[Accessed: 5 June 2023] 

World Health Organisation., (2020). Global Database on Child Growth and Malnutrition. 

Who.int. Available at: https://www.who.int/nutgrowthdb/about/introduction/en/index5.html. 

[Accessed date: 6th April. 2020]. 

Yakusheva, O. and Fletcher, J., (2015). Learning from teen childbearing experiences of close 

friends: Evidence using miscarriages as a natural experiment. Review of Economics and 

Statistics, 97(1),pp.29-43.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2258311/
https://pubmed.ncbi.nlm.nih.gov/18206223
https://scholar.google.com/scholar_lookup?journal=Lancet&title=Maternal+and+child+undernutrition:+Consequences+for+adult+health+and+human+capital&author=C.G.+Victora&author=L.+Adair&author=C.+Fall&volume=371&publication_year=2008&pages=340-357&pmid=18206223&
https://scholar.google.com/scholar_lookup?journal=Lancet&title=Maternal+and+child+undernutrition:+Consequences+for+adult+health+and+human+capital&author=C.G.+Victora&author=L.+Adair&author=C.+Fall&volume=371&publication_year=2008&pages=340-357&pmid=18206223&
https://data.worldbank.org/country/GY
https://climateknowledgeportal.worldbank.org/
https://data.worldbank.org/indicator/SH.DYN.MORT?end=2014&locations=XO&start=2000
https://www.who.int/nutgrowthdb/about/introduction/en/index5.html


61 

CHAPTER 3 

 

Effects of a School Conversion Policy on Educational Attainment and Labour Market 

Outcomes in Guyana 

 

3.1 Introduction  
 

The fundamental role that education plays in the lives of individuals and modern labour markets 

is widely acknowledged. Education generates positive externalities, given that the social returns 

to education tend to exceed the sum of the individual returns to education (Sianesi, 2002; 

Psacharopoulos and Patrinos, 2018). It is also an important policy instrument to fight against 

poverty and promote the general well-being of individuals. Furthermore, studies in different 

countries and periods have posited that highly educated individuals earn higher wages, experience 

less unemployment and work in more prominent professions than their less-educated counterparts 

(Shavit and Müller, 2000; Enu et al., 2014; Kampelmann et al., 2018). Besides affecting wages 

and earnings, education may influence several other labour market outcomes, namely: the time to 

the first stable job; employment/unemployment spells; worker productivity; hours worked; nature 

of work; worker’s health; and fringe benefits (Card, 1999; Heckman et al., 2006; Ye Chen, 2021).  

 

In this study, we aim to shed light on the effects of a policy that enhances school completion and 

investigate its effects on educational attainment and labour market outcomes. We do this by 

exploiting a key policy change in Guyana: in 2002, Guyana adopted a five-year plan from 2003-

2007 to achieve Universal Secondary Education (USE). The USE (2003-2007) reform policy was 

piloted through one main donor-funded project called the Secondary School Reform Project 

(SSRP). The SSRP was specifically designed for school conversions within all the regions of 

Guyana (Ministry of Education, Guyana, 2022).  Prior to the USE (2003-2007) policy reform, 

Guyana’s secondary education system operated under a differentiated three-tier system which 

consisted of three different types of secondary schools, namely: General Secondary Schools 

(GSSs), Community High Schools (CHSs), and the secondary department of primary schools, also 

known as Primary Tops. The latter two secondary schools were geared towards a pre-vocational 

curriculum, while the GSSs provided an academic curriculum (World Bank, 2013). This resulted 



62 

in the yearly dropout rates being over 60 percent for children in Community High Schools (CHSs) 

and Primary Tops (Williams, 1993; Jennings, 2000; Gill-Marshall, 2000). Moreover, students who 

attended CHSs and Primary Tops had little to no job prospects because the certificates received 

from these institutions were based on a pre-vocational curriculum rather than an academic one 

(World Data on Education, Guyana, 2006/07; Jennings, 2000; London, 2018). Despite these 

certificates being used to enrol students into technical or vocational post-secondary school 

programs. Guyana’s youth unemployment rate was among the highest in the Caribbean. Between 

2001 and 2002, Guyana witnessed its highest youth unemployment rate, an alarming 50 percent; 

which was far above the World’s average (Kaieteur News, 2015; Caribbean Development Bank, 

2015).  

 

In addition, Amerindian students were at a further disadvantage compared to the other ethnic 

groups because they resided in areas where access to general secondary education was limited and 

difficult (UNICEF, 2017). Consequently, this research also seeks to examine whether the USE 

(2003-2007) policy reform positively impacted secondary or tertiary education completion 

amongst Amerindian students. Furthermore, this research will also examine the heterogeneity 

across the various ethnic subgroups to determine the impact of the USE policy on the various 

groups, as they can be differences in economic outcomes and policy impacts by ethnicity, 

especially among ethnic minorities (Driessen, 2000).  

 

The USE (2003-2007) policy reform was done through converting secondary classes in primary 

schools (“Primary Tops”) and grades 7-10 in Community High Schools (CHSs) into General 

Secondary Schools (GSSs).   

The World Bank funded the SSRP and provided a total of US$17.3 million to the USE (2003-

2007) policy reform with the objectives of improving the quality, relevance, equity, and efficiency 

of secondary education in Guyana through improving the school environment by supporting 

rehabilitation and repairs, building extensions and construction of multi-purpose science and 

computer laboratories. In addition, the SSRP also developed measures to improve the quality and 

efficiency of lower secondary education by providing new curricula designed to improve the 

provision of 5 core subjects (namely, English, Mathematics, Science, Social Studies, and reading) 

supported by the distribution of textbooks and the training of teachers (World Bank, 



63 

2005).Therefore, the government expected that this conversion of schools from a pre-vocational 

track into an academic track would create a unified academic curriculum for all secondary school 

students, reduce secondary school drop-out rates, increase the number of qualified teachers in the 

secondary system and encourage an increase in the number of students who complete GSSs (World 

Data on Education, Guyana, 2006/07). Our study is the first to analyse the effects of this policy.  

 

According to the Ministry of education (2023), there is no official age for entry into secondary 

school, however on a normal basis given the entry age of age 5 or 6 for the six-year programme 

for primary schools, students usually enter the secondary level at approximately 12 years old. 

Consequently, we defined the treated as age 11, as the Ministry of education confirms that children 

can also attend secondary school at 11 plus (Ministry of education, 2023).  

According to the Ministry of education (2023), a child can be in a general secondary school as old 

as age 17-18 age old, depending on the age in which they have started secondary school. However, 

the occurrence of an 18-year-old student being in a GSS is very rare. Therefore, we have assumed 

that an individual would have been in general secondary education between the ages of 11-17 

based on the Guyana Labour Force Survey (GLFS). 

Consequently, this study utilises data from the Digest of Education Statistics (DES) for the period 

1999-2019 provided by the Ministry of Education of Guyana, as well as the Guyana Labour Force 

Survey (GLFS) quarterly data for the period (2018-2019). These two rich datasets are underutilised 

and allow us to investigate the impact of the USE (2003-2007) policy reform on the completion of 

secondary or tertiary education and on labour market outcomes. As such, we focus on three major 

outcome variables for educational attainment: completion of secondary or tertiary education, 

completion of tertiary education, attendance of secondary or tertiary education. In addition, we 

consider labour markets outcomes such as wages and employment. We use a difference-in-

difference specification which focuses on respondents who were aged 11-17 years in 2003 (the 

treated group) and 18-24 years old in 2003 (the control group) where 2003 is the year when the 

USE was introduced. We compute the intensity of exposure to the USE (2003-2007) policy using 

the share of non-GSSs institutions out of all secondary schools in each region in 2002, based on 

the DES. The findings from our study show that greater exposure to the Universal Secondary 

Education policy led to an increase in the completion of secondary or tertiary education in Guyana. 



64 

We also find that exposure to the policy led to an increase in the wages of treated cohorts. 

Furthermore, our results are robust when applying a range of alternative specifications.  

This study and its findings contribute to the literature on supply-side policies aimed at increasing 

the supply of schooling and the conversion of vocational or technical education into an academic 

track. Several studies have shown that educational reform, such as school construction and 

secondary education fee-reduction programs led to an increase in the completion of education and 

higher earnings (Duflo, 2001; Barrera-Osorio et al., 2013, Garlick, 2013; Duflo et al., 2021). For 

instance, Duflo (2001) studies the effects of the 1970s Indonesian’s government school 

construction program on men’s education and earnings and finds that it led to an increase in 

education and earnings in Indonesia. Studies have also shown that educational interventions 

through scholarships and mentoring in secondary schools were used as an incentive to encourage 

students. Secondary school scholarships in Ghana increased educational attainment, knowledge, 

skills, and preventative health behaviour, while reducing female fertility (Duflo et al., 2021).  

 

The literature has also shown that there are beneficial results of abolishing or reducing school fees 

or introducing education subsidies for school attendance and enrolment (Alderman et al., 2001; 

Deininger, 2003; Schultz, 2004; İşcan et al., 2015; Kan and Klasen, 2021; Brudevold-Newman, 

2021). Government policies in Kenya, such as fee reduction and capacity expansion through 

increasing class sizes and the number of classes, increased the proportion of students continuing 

from primary to secondary education (Brudevold-Newman, 2021). Studies have shown that, in 

general, free education policies result in positive impacts on educational attainment and financial 

inclusion. For instance, the Free Primary Education (FPE) policy in Kenya was associated with an 

increase in educational attainment and increased use of formal financial services (Ajayi and Ross, 

2020). In addition, fee abolition in West German secondary schools resulted in an increase in 

educational attainment in upper secondary schools (Riphahn, 2012).  

In general, studies have shown that formal education increases competency in basic skills in adults 

(OECD, 2013a; Massing and Schneider, 2017). For instance, Brunello and Rocco (2017) in their 

study using the Program for the International Assessment for Adult Competencies (PIAAC) 

dataset shows that persons who pursued an academic path had more basic skills and were more 

https://largescaleassessmentsineducation.springeropen.com/articles/10.1186/s40536-017-0041-y#ref-CR33


65 

competent when it came to literacy and numeracy skills relative to those who had vocational 

education.   

There are several studies that have shown that having academic qualifications increases the social 

returns, wages, and employment opportunities/typology of jobs of individuals. Psacharopoulos 

(1994) estimates the returns to vocational and academic qualifications from 32 studies and revealed 

that social returns to academic qualifications are approximately 15.5% per annum while the social 

returns for vocational qualifications are lower (at 11.7% per annum). The equivalent private returns 

are 10.6% and 10.5% respectively. Dearden et al. (2002) find that the wage premia for academic 

qualifications are higher than from vocational qualifications at the same level in Great Britain.  

Malamud and Pop-Eleches (2010) find that men in Romania who were affected by the 1973 

educational reform requiring students in vocational education to attend two additional years of 

general education were less likely to work as manual workers and craftsmen than their counterparts 

who were born too early to be affected by the policy.  

Studies have also found intergenerational impacts of education that are transmitted to future 

generations. For instance, Akresh et al. (2018), examined the long-term and intergenerational 

impacts of one of the largest primary school construction programs in Indonesia and finds that 

educational benefits persist for both men and women 43 years after the program. Furthermore, 

mother’s program exposure led to education benefits that were transmitted to the next generation, 

with the largest effects in upper secondary and tertiary education when compared to father’s 

exposure. Moreover, men who were exposed to the program were more likely to be formal 

workers, work outside agriculture and migrate.  

Furthermore, there are many studies that provide evidence that education and earnings are 

positively correlated with each other; as such, higher education increases one’s earning abilities or 

return to schooling (Ashenfelter and Krueger, 1994; Harmon et al., 2003: Heckman et al., 2006; 

Peet et al., 2015). 

 

Education reforms via the voucher system also led to positive impacts on education attainments 

and labour market outcomes. Consequently, a Universal Voucher System reform which provided 

subsidised school funding in Chile resulted in both an increase in the graduation rate among 

secondary school children and an increase in the percentage of students completing at least two 



66 

years of college. In addition, the Universal Voucher System reform also increased lifetime utility 

and reduced earning inequality in Chile (Bravo et al., 2010). Murnane et al. (2017) also 

investigated the consequences of educational voucher reform in Chile and found that on average, 

students’ test scores increased and income-based gaps in those scores declined by one-third in the 

five years after the reform. Psacharopoulos and Patrinos (2018), in their study on returns to 

investment in education, find that women experience higher average rates of return to schooling 

and the returns are higher in low-income countries. They also find that private sector employment 

had higher returns, thus supporting the productive value of education. Furthermore, studies that 

evaluate the impacts of formal education on occupational choice found that, for men, education 

increases the probability of wage work (Duflo, 2004) and decreases the probability of self-

employment (Ozier, 2018). 

Consequently, this study is the first-ever study to analyse a USE policy reform that converts 

secondary schools with a vocational curriculum into academic curriculums. The study also 

contributes to the methodological and economic literature for Guyana, as it is the first study for 

Guyana to investigate the effects of a secondary education policy reform on the completion of 

secondary or tertiary education and labour market outcomes. The rest of the paper is organised as 

follows. Section 3.2 provides context on Guyana’s education system. Section 3.3 describes the 

data sources, and Section 3.4 introduces the empirical strategy. Section 3.5 presents the results, 

and Section 3.6 concludes. 

3.2 Context 
 

This section provides background on Guyana’s education system and the labour market context 

throughout our study period, ending before the onset of the COVID-19 pandemic crisis.  

3.2.1 Guyana’s Education System 

Education in Guyana is largely free from nursey to secondary levels. However, there are some 

private institutions where one must pay tuition.  Although not compulsory, education in Guyana 

begins with two years of nursery or kindergarten education, followed by six years of mandatory 

primary education (grades 1-6).  



67 

Furthermore, as it relates to secondary education, based on the performance at the SSEE 

examinations, students were rewarded either a senior general secondary (i.e., an optional seven-

year curriculum for the top 5% performers – Grades 7-13) or a junior general secondary school 

(i.e., five-year curriculum – Grades 7-11) (Gafar, 2003). Grades 12 and 13 are classified as sixth 

form or post-secondary (see Figure 3.1 for the structure and organisation of Guyana’s educational 

system prior to the USE (2003-2007) policy reform).1  

Prior to 2003, secondary education operated under a tier system, where it comprised of a secondary 

department of a primary school (formally known as ‘Primary Tops’), Community High Schools 

(CHSs) and General Secondary School (GSSs).  

 

According to Gafar (2003), access to GSSs was based both on the completion of the Secondary 

School Entrance Examination (SSEE or the United Kingdom’s Standardised Assessment Tests 

(SATs) equivalent exam) or (11+) exam, and the number of places available in the GSSs. 

Following the SSEE, the top 55 percent were placed in the more prestigious GSSs whilst the 

remaining students with low scores were placed in CHSs, or Primary Tops.  

 

Furthermore, Jennings and Miller (1999) argued that CHSs classrooms were poorly equipped and 

did not foster an environment conducive to learning. They also reported that a 1990 survey done 

on 71 CHSs and Primary Tops revealed that only 38.5 percent of the students graduated, and 61.2 

percent dropped out in their final year without proper certification. The reason for the high dropout 

rate was due to shortage of textbooks, lack of qualified teachers and poor facilities. The courses 

offered in Primary Tops and CHSs were more tailored towards prevocational training while those 

offered in GSSs were more academic in nature. Some courses provided in CHSs, and Primary 

Tops were woodworking, technical drawing, home economics, health and physical education, 

music, art and craft, agriculture, and industrial arts and other vocational courses.  

 

 

 

 

 
1Thirteen public senior secondary schools offer the sixth form in Guyana (Ministry of Education, 2022). 



68 

Nursery (2 years) 

 

Primary (6 years) 

                                                                    SSEE 

                 (≥50 %)                                    (49-40 %)                           (≤ 39 %) 

 

         CHS                                             Primary Tops  

                       (3 years)       (4 years) 

              SSPE 1         SSPE 11 

              GSS    Pass     

                        CSEC (5 years)                Fail                       

                                  Vocational Training  

  CAPE (2 years)             CPCE Primary/Secondary 

                    Teacher Training (2/3 years)                                   

                                                                           GTI/NATI     GTIC          CSHE           

            University 

  (4-6 years)   

    Undergraduate + Masters 

 

 

Figure 3.1: The structure and organisation of the education system prior to the USE (2003-2007) policy reform. 

Source: Author’s illustration based on Digest of Education Statistics Reports (MOE). 

 

 

 

Furthermore, the Primary Tops and CHSs provided a four-year curriculum which also prepared 

students to sit the SSPE in year three (Grade 9) and year four (Grade 10) (World Data on 

Education, Guyana, 2010/11). The SSPE curriculum in the CHSs covered four core subject areas 

(English, Mathematics, Social Studies, and Integrated Science), written in two parts (Parts 1 and 

11). Students who passed Part 1 in their third year of study were able to transfer to a junior general 

secondary school, to pursue a more academic curriculum. Part 2 was administered at the end of 

the fourth year (Grade 10) and was designed to give the students who had failed Part 1 a second 

chance of attending a GSS (World Data on Education, 2006/07). Consequently, this examination 

THE WORLD OF WORK 



69 

once passed allowed students the opportunity to be accepted into a junior GSS and prepare for the 

CSEC examination or the United Kingdom’s General Certificate of Secondary Education (GCSE). 

On the other hand, the student that did not pass Part 1 nor Part 2 of the SSPE went onto to graduate 

(in Grade 10) from the CHS schools was awarded with a SSPE results certificate to either apply 

for a job or to enrol into one of the technical or vocational after-school programs such as Guyana 

Industrial Training Centre (GITC) programs in furniture making, electrical installation and 

automotive mechanics etc. However, it has been noted that employers held this certificate in low 

esteem, and it was very difficult for graduates to gain employment (World Data on Education, 

Guyana, 2010/11; London, 2018).2 On the other hand, the GSSs have always offered a five-year 

programme that is largely academic in nature, i.e., in the first three years it is compulsory for 

students to study English Language, English Literature, Mathematics, Social Studies, Integrated 

Science, Agricultural Science along with other core subjects. However, more teaching periods are 

given to English Language and Mathematics which are the two core courses that are required for 

both tertiary education and job search.  

 

After completing the fifth form in GSSs, students can either go onto the sixth form at a senior 

secondary school to complete the Caribbean Advance Proficiency Examination (CAPE) or the 

London General Certificate of Education (GCE) A-Levels equivalent. They can also apply directly 

to the tertiary institutions in Guyana, i.e., the University of Guyana, Turkeyen or Berbice Campus 

and teacher training at the Cyril Potter College of Education (CPCE) (Gafar, 2003). 

The University of Guyana requires students with a minimum of five CSEC General Proficiency 

(Grades 1 and 2; and 3 from 1998 onwards) or a minimum of five passes at CAPE, including the 

English Language, which is applicable for the pursuit of all majors. However, passes in 

Mathematics is applicable only for designated programmes such as Medicine, Chemistry, Biology, 

Medical Technology, Engineering, Computer Science, Economics and Business Management, 

Accounting and Finance among others (University of Guyana, 2022). Consequently, given that the 

USE (2003-2007) seeks to provide access to GSSs education universally to all students, it may 

 
2 See the link below for the explanation of the Primary Tops and CHSs SSPE examinations: 

http://www.educoas.org/portal/bdigital/contenido/interamer/bkiacd/interamer/interamerhtml/millerhtml/mil_jen.htm, 

Secondary School | K12 Academics 

http://www.educoas.org/portal/bdigital/contenido/interamer/bkiacd/interamer/interamerhtml/millerhtml/mil_jen.htm
https://www.k12academics.com/Education%20Worldwide/Education%20in%20Guyana/secondary-school


70 

increase students' access to tertiary education in Guyana, thus leading to improved labour market 

outcomes. 

 

3.2.2 Description of the changes caused by the USE policy reform. 
 

In this section, we present some descriptive analysis below using the DES (1999-2018) to illustrate 

what Guyana’s education system looked like pre-reform as it relates to the number of secondary 

institutions, the number of students in secondary institutions, the year-on-year percentage change 

for students in secondary institutions and the dropout rate. 

 

 
NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE) 

Figure 3.2: A comparison of the number of institutions and share of institutions for the period 1999-2019. 

Source: Author’s illustration based on Digest of Education Statistics (1999-2019). 

 

 

Furthermore, the USE (2003-2007) policy reform resulted in a fall in the number of CHSs and 

Primary Tops between the pre-policy years (1999-2002) and the post-policy periods (2003 

onwards. Furthermore, we examined the share of secondary institutions for the period (1999-2019) 

and find that the share of GSSs increased whilst the share of Primary Tops and CHSs declined 



71 

over time (see Figure 3.2 above for the comparison of the number of institutions and the share of 

institutions between the pre-and-post-policy periods). This result is consistent with the goal of the 

USE (2003-2007) policy reform which sought to convert Primary Tops and CHSs from a 

vocational track to an academic track.   

 

Moreover, as of 2018, there remain some Primary Tops to be converted to GSSs in Guyana. 

However, the Ministry of Education in collaboration with the World Bank conducted the 2008 

Geoff Howse report which outlined what needed to be done to fully achieve Universal Secondary 

Education in Guyana.3 This report was updated in 2021 and was accepted by Guyana’s Minister 

of education. The Ministry of Education (MOE) has subsequently begun the implementation 

process (Chief Planning Officer, MOE, 2022).   

 

We can also see that the policy led to an increase in GSSs in the post-policy period, as expected. 

Furthermore, when looking at the number of students in secondary institutions for the period 1999-

2019, we see a steady increase in GSSs students in the post policy period, a reduction of students 

in Primary Tops, as well as a complete phase out of CHSs in the year 2011 (see Figure 3.3 for the 

number of students between the pre-and-post-policy periods).  

 

NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE) 

Figure 3.3: The number of students in secondary institutions between pre-and-post-policy periods. 
Source: Digest of Education Statistics (1999-2019). 

 
3 See World Bank Document for the Geoff Howse Report, 2008.  

https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941eport0by0Geoff0Howse.pdf?sequence=1&isAllowed=y


72 

When looking at the year-on-year percentage change in the total number of students in Primary 

Tops, CHSs and GSSs for the period 1999-2019 (see Figure 3.4 for the year-on-year percentage 

change of students in secondary institutions during pre-and-post-policy periods), we see decreases 

in year-on-year percentage change in the number of students in CHSs and Primary Tops. However, 

for GSSs, there is a general increase in the post-policy period compared to the pre-policy period.  

NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 
Figure 3.4: The year-on-year percentage change for students in secondary institutions for the period (1999-2019). 

Source: Digest of Education Statistics (1999-2019). 

 

Overall, the results from Figures 3.2 and 3.3 reveal that the USE (2003-2007) policy reform 

reduced the total number of secondary institutions and the year-on-year percentage change in the 

number of students in Primary Tops and CHSs but increased it for GSS institutions. 

 

Moreover, we also examined the share of students across types of secondary institutions for the 

period (1999-2019) and find that the share of students in GSSs increased whilst it declined in 

Primary Tops and CHSs (see Figure 3.5 below for the share of students in secondary institution 

for the period (1999-2019)).  

When examining the dropout rate over time, we find that the dropout rates were higher for Primary 

Tops and CHSs in both the post-and-pre policy periods. Furthermore, we find that as CHSs and 



73 

Primary Tops are converted to GSSs, the dropout rates for GSSs declined significantly (see Figure 

3.6 below for the dropout rate in secondary institutions for the period (1999-2019)). 

 

NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 

Figure 3.5: The dropout rates in secondary institutions over time (1999-2019). 

Source: Digest of Education Statistics (1999-2019). 

 

The USE policy also impacted the number of trained, untrained and unqualified teachers in 

secondary institutions in Guyana (see Appendix B.1 for further discussion). 

  

3.2.3 Guyana’s Labour Market  

 

In this research, we examine the impact of the USE (2003-2007) policy reform on labour outcomes 

such as employment, wage employed, formal wage worker, work hours and log wage.   

In Guyana, the statutory minimum age of employment is fifteen years old. An individual is 

therefore deemed employed if he/she is of working age and is engaged in any activity to produce 

or provide services for pay (or without pay). Consequently, majority of employees in Guyana are 

formally employed based on a permanent or fixed-term contract (ILO, 2022). However, there are 

many who are employed without written contract. In general, the work hours in Guyana are forty 



74 

hours per week or 8 hours per day (ILO, 2022). In 2019, the public sector monthly minimum wage 

in Guyana was $70,000 Gyd/336.57 US$ per month, $3,230.27 Gyd/15.53 US$ per day and 

$403.79 Gyd/1.94 US$ per day (Bureau of Statistics, Guyana 2022).4  

 

Guyana population over the last 20 years has been an average of 750,000. Guyana had a total 

population of 563,167 aged 15 and above in 2019. There were more women than men within the 

working-age population. The labour force participation rate was 49 percent, with men and women 

recording 61 and 38 percent respectively. Furthermore, the participation rates in rural and urban 

areas were 48 and 51 percent respectively. Guyana’s unemployment rate was 13 percent, with 

youth unemployment being 30 percent (Guyana Labour Force Survey, 2019).5  

 

In 1998, the total working-age population was 259,138 people with 62 percent having lower than 

or equal to primary education (National Employment Report Guyana, 2006). However, in 2019, 

91.2 percent had higher than or equal to Primary education (Guyana Labour Force Survey, 2019). 

Moreover, in 2019, Guyana’s percentage share of the working-age population by ethnicity stood 

at 45 percent East Indians, 26 percent Africans/Blacks, 20 percent Mixed, 9 percent Amerindians 

and 0 percent Other.  

 

3.3 Data sources and descriptive statistics 
 

3.3.1 The Guyana Labour Force Survey (GLFS)  

 

This paper uses quarterly data from the Guyana Labour Force Survey (GLFS) and statistical 

tabulated data from the Digest of Education Statistics (DES 1999-2019). The GLFS is a continuous 

survey that uses a rotation scheme (or rotating panel) for 2017-2021. The GLFS samples 

approximately 4,000 households every quarter or a total of 11,000 individuals aged 15 years and 

above. The GLFS is based on a stratified two-stage probability design, each quarterly sample is 

spread over 12 consecutive weeks across 15 strata corresponding to urban, rural, coastal, and 

 
4At the time of this research, the conversion of Gyd to USD is $207.98 to 1 $USD (Bank of Guyana, September 2022).  
5A person is deemed unemployed if they are actively looking for employment and are available to work within a short 

time span.  



75 

interior areas in the 10 administrative regions in Guyana. In the first sampling stage, there are 360 

Primary Sampling Units (PSUs), which are groups of Census Enumeration Districts (EDs).  

 

The GLFS used a rotating panel design to measure the change over time more accurately. All PSUs 

are kept the same from one quarter to the next. However, some households within each PSUs are 

kept in the sample while new households replace the rest. As a result, one part of the panel sample 

overlaps in both quarters (Guyana Bureau of statistics, IDB-GLFS, Methodological Report, 2018). 

The GLFS provides information on a wide range of outcomes related to education, employment, 

migration, wages and work-hours outcomes among others.  

 

Given that the GLFS uses a rotating panel, we used one wave per block of respondents in this 

paper.  To analyse a complete sample and use an individual-level dataset, we used a total of three 

blocks from survey quarters 3, 1 and 4 of 2018 and 2019 respectively. Furthermore, the sample 

consists of a total of 37,907 individuals that were selected from the three surveys. These surveys 

were used because there is no overlapping. In addition, as these samples are pre-2020, they are 

also unaffected by the COVID-19 pandemic.  

 

In this research, Chapter 2 and 3 use different datasets. Chapter 2 uses MICS (2000-2014), and 

Chapter 3 uses GLFS (2018 and 2019). The samples in the Chapter 2 are female samples and 

Chapter 3 consist of both male and female samples. In Guyana, men are generally less educated 

than women (Digest of Education Statistics, 2018-2020; Guyana Chronicle, 2021).  Moreover, 

based on the Human Development Report (2010), the percentage of female completing at least 

some secondary education in 2008 was 30 percent. Unfortunately, the figure for 2009 is 

unavailable. These results are close to the DHS 2009 Guyana data which recorded 43 percent 

(based on the new regression results) with completed secondary or tertiary education in 2009. 

Furthermore, the Human Development Report (2016) reported that female completing at least 

secondary education for the period 2005-2015 was 68.1 percent. Consequently, it is evident that 

the MICS (2000-2014) data over-represented the female completion of secondary or higher 

education, as it stood at 75 percent.  

 



76 

3.3.2 Digest of Education Statistics (DES)  

 

The Digest of Education Statistics (DES) of Guyana is a collection of statistical tables compiled 

by the Ministry of Education (MOE), Guyana. It presents nationally aggregated yearly information 

on all the public educational institutions in the country, ranging from nursery to university. The 

DES also provides the education expenditure as a percentage of the national budget and gross 

domestic product (GDP). The DES (1999-2019) provides educational census data for Guyana for 

all the secondary institutions, i.e., Primary Tops, CHSs and GSSs. In addition, the DES contains 

variables such as: the number of schools at each level (i.e., nursery, primary, secondary, technical 

and university) by education district/region, grade and type of schools, enrolment by education 

district/region and type of schools, average enrolment, average attendance and percentage average 

attendance by education district/region and school type, admissions by education district/region 

and grade, class size by education district/region, number of classes by education district/region 

and the dropout rate by education district/region, grade and sex. Furthermore, the DES also 

provides information on the student/teacher ratio by education district/region by levels and type of 

schools, the number of teachers by education district/region, qualification and sex and teacher 

attrition by type of schools, education district/region, and sex.  

 

Given that the DES provides detailed information on the school types. We have detailed 

information on all the secondary institutions (i.e., Primary Tops, CHSs and GSSs), which we have 

used for our analysis.  

 

For this paper, we used the DES for the period 1999-2019 to effectively assess the impact of the 

USE (2003-2007) policy reform. Consequently, the DES (1999-2019) data was merged with the 

GLFS (2018 & 2019) data using the region in which the respondents were surveyed.         

                                            

3.4 Empirical Strategy 
 

To examine the relationship between the USE (2003-2007) policy reform and schooling, we 

employ a difference-in-differences estimation framework. The treatment is defined as those who 

were of secondary school-going age and thus exposed to the USE (2003-2007) policy reform while 



77 

the control group were those who would have completed secondary school before the start of the 

policy reform period. Furthermore, the USE policy reform had a higher intensity in regions where 

there was a lower share of General secondary schools before the reform. Moreover, our empirical 

strategy is similar to that of Ajayi and Ross (2020) and Brudevold-Newman (2019). The estimating 

equation is as follows:  

𝑆𝑖𝑗𝑟𝑞 = 0 + 𝛽1 (𝑈𝑆𝐸𝑗 ∗  𝐼𝑟) + 𝛽2 𝑋𝑖 +  𝛿𝑟 +  𝛾𝑗 + 
𝑞
+ 𝑖𝑗𝑟𝑞   (1) 

where 𝑆𝑖𝑗𝑟𝑞  are education outcomes, represented by schooling dummies  (i.e., completion of 

secondary or tertiary education, the completion of tertiary education and the attendance of 

secondary or tertiary education) and labour market outcomes (i.e., dummies for employment, wage 

employed, formal wage worker, work hours and log wage) of individual i in cohort j in region r 

and survey quarter q, 𝑈𝑆𝐸𝑗  is the treatment, i.e., a dummy variable equal to one for individuals 

born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the 

control) if not (aged 18-24 in 2003 at the time of the reform). The DES is used to construct the 

intensity variable of the USE program in the respondents’ region of residence. Consequently,  𝐼𝑟 

is the intensity of the USE program in region (r) of residence (defined as the share of non-GSS 

schools out of all secondary schools in the region in 2002, based on the DES).  𝑋𝑖 represents a 

vector of individual characteristics, such as gender and ethnicity.  𝛿𝑟 , 𝛾𝑗  and 
𝑞

 represent the 

region fixed effects, cohort fixed effects and survey quarter fixed effects, respectively. The 

interaction coefficient, 𝛽1 is the estimate of the effect of the USE on educational attainment and 

labour market outcomes. Given that we are using a fixed effects model to analyse our difference-

in-difference estimation, the region fixed effects absorb the intensity, and the cohort fixed effects 

are absorbed by the treated. Therefore, these two variables are collinear and are not included 

independently in our regression analysis. Moreover, 𝑖𝑗𝑟𝑞  is the stochastic error term clustered at 

treatment levels. We used the reghdfe package by Sergio Correia’s which allowed us to control for 

fixed effects and to cluster our standard errors at the treatment level.6 

 
6Although we are using binary dependent variables, the argument that infers or supports the use of reghdfe being 

appropriate over logit and probit models is posited by an article by Cruz-Gonzalez et al, 2017. See Bias Corrections 

for Probit and Logit Models with Two-way Fixed Effects (sagepub.com).  

https://journals.sagepub.com/doi/epdf/10.1177/1536867X1701700301
https://journals.sagepub.com/doi/epdf/10.1177/1536867X1701700301


78 

In our baseline regressions, we assumed that the survey region was the region in which the 

respondents were educated. However, in our first robustness check we redefined the intensity 

variable intensity as the average share of non-GSS institutions in the region for the period 1999-

2002 based on the DES.  

 

3.5 Result 
 

3.5.1 Summary Statistics 
 

 

Table 3.1 below presents summary statistics of the key variables used in our analysis. Within the 

GLFS sample, we created a dummy variable equal to 1 if the highest education completed is 

secondary and 0 otherwise and find that 28 percent of the respondents completed secondary 

education. We also defined the completion of secondary or tertiary (i.e., including tertiary 

education) as a dummy equal 1 and 0 otherwise and find that 43 percent of the respondents 

completed secondary or tertiary education. In addition, we defined completed tertiary education as 

a dummy equal 1 if the highest education (i.e., technical/vocational certificate or Diploma, 

University Certificate or Diploma, Bachelor’s degree, Postgraduate Certificate or Diploma, 

Masters, and Doctorate) and 0 otherwise. Moreover, we find that 15 percent of the respondents 

completed tertiary education.  

 

With regards to the attendance of these institutions, we created a dummy equal to 1 of the highest 

education attended is secondary or tertiary (i.e., including tertiary education) and 0 otherwise and 

find that 85 percent of the respondents attended secondary or tertiary education. Moreover, when 

looking at the labour market outcomes, 66 percent of the respondents were employed with 69 

percent earning an income. Furthermore, amongst the employed, 44 percent of the wage workers 

worked in the formal sector and the average work hours per week was 47 hours. Moreover, the 

mean of the log wage here is 11.15 (i.e., equivalent to a monthly income of $70,000 Guyanese 

dollars). This is due to missing wage data.  

 

 



79 

Table 3.1: Summary statistics 

Variable Mean 

Standard  

deviation    Min   Max N 

Outcome variables      
Secondary education (Completed) 0.278 0.448 0 1 6326 

Secondary or tertiary (Completed) 0.428 0.495 0 1 6326 

Tertiary education (Completed) 0.148 0.355 0 1 6326 

Secondary or tertiary (Attended) 0.879 0.326 0 1 6326 

Employed 0.656 0.475 0 1 6326 

Wage employed 0.689 0.463 0 1      3999 

Formal wage worker 0.437 0.496 0 1 2759 

Work hours 46.908 18.117 3 99 3999 

Log wage 11.15 0.634 3.47 14.22 2312 

USE & Intensity      
USE  0.526 0.499 0 1 6326 

Intensity 0.727 0.109 0.60 0.95 6326 

Individual characteristics      
Female 0.539 0.498 0 1 6326 

East Indian 0.424 0.494 0 1 6326 

African 0.287 0.453 0 1 6326 

Mixed 0.196 0.397 0 1 6326 

Amerindian 0.089 0.285 0 1 6326 

Other 0.004 0.062 0 1 6751 

Region 1 0.018 0.132 0 1 6326 

Region 2 0.061 0.239 0 1 6326 

Region 3 0.145 0.352 0 1 6326 

Region 4 0.285 0.451 0 1 6326 

Region 5 0.162 0.369 0 1 6326 

Region 6 0.195 0.396 0 1 6326 

Region 7 0.017 0.131 0 1 6326 

Region 8 0.026 0.157 0 1 6326 

Region 9 0.013 0.113 0 1 6326 

Region 10  0.078 0.268 0 1 6326 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., 

those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy 

variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to 

zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of 

non-GSS schools out of all secondary schools in the region in 2002, based on the DES).   

 

 

In terms of individual characteristics, we find that 54 percent of the respondents are females, 33 

percent of the respondents ranges between the ages of 26 and 40,  42 percent of the respondents 

were of East Indian descent, 29 percent were African, 20 percent were Mixed, 9 percent were 



80 

Amerindian.7 When looking at the regions, we find that 29 percent of the respondents were from 

region 4, 20 percent from region 6, 16 percent from region 5 and 35 percent were from the 

remaining regions.  

 

3.5.2 The effect of the USE (2003-2007) policy reform on schooling in 

Guyana 

 

Table 3.2 below presents the main results where we regress our main outcome variables for 

schooling, (i.e., completed secondary education, completion of secondary or tertiary education, 

tertiary education and secondary or tertiary education attended). In columns 1 and 2, the coefficient 

of interest, 𝛽1 is statistically significant at 1%, which means that greater exposure to the USE 

policy reform increases the completion of secondary education and secondary or tertiary education 

by 37.4 and 42.6 percentage points respectively. This finding is similar to that of Ajayi and Ross 

(2020), who find that greater exposure to Free Primary Education (FPE) program in Kenya 

increased the highest education level achieved. Consequently, the USE (2003-2007) resulted in an 

increased number of students completing and subsequently graduating from secondary or tertiary 

institutions in Guyana. 

 

Furthermore, in column 3, we find that greater exposure to the USE policy reform does not have a 

statistically significant effect on the completion of tertiary education. In addition, when examining 

column 4, we find that greater exposure to the USE policy reform increases the attendance at 

secondary or tertiary education level by 16.1 percentage points. We also find that more females 

attended secondary or tertiary education as compared to their male counterparts.  

 

 

 

 

 

 

 
7The East Indians are the most dominant ethnic group in Guyana, followed by the Africans, Mixed, Amerindians, 

Portuguese, Chinese, Whites and Other (i.e., those who did not identified with any of the listed race/ethnic groups. 

See: https://statisticsguyana.gov.gy/wp-content/uploads/2019/11/Final_2012_Census_Compendium2.pdf 

https://statisticsguyana.gov.gy/wp-content/uploads/2019/11/Final_2012_Census_Compendium2.pdf


81 

Table 3.2: Effect of the USE on Educational Attainment 

 (1) (2) (3) (4) 

 Secondary 

(Completed) 

Secondary/ 

tertiary education 

(Completed) 

Tertiary 

(Completed) 

Secondary/ 

tertiary 

(Attended) 

USE x Intensity 0.374***  0.426*** 0.032 0.161** 

 (0.097) (0.107) (0.065) (0.068) 

Female 0.067*** 0.070*** 0.002 0.032*** 

 (0.011) (0.011) (0.008) (0.007) 

African 0.051***  0.189*** 0.136*** 0.121*** 

 (0.016) (0.017) (0.012) (0.011) 

Mixed 0.031*  0.101*** 0.068*** 0.077*** 

 (0.018) (0.021) (0.015) (0.013) 

Amerindian -0.056** -0.079*** -0.020 -0.045* 

 (0.025) (0.030) (0.019) (0.026) 

Other 0.008 0.225** 0.219** 0.127*** 

 (0.089) (0.101) (0.109) (0.040) 

Constant 0.083** 0.160*** 0.083*** 0.754*** 

 (0.038) (0.043) (0.026) (0.029) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.278 

✓ 

✓ 

✓ 

6326 

0.428 

✓ 

✓ 

✓ 

6326 

0.148 

✓ 

✓ 

✓ 

6326 

0.879 

✓ 

✓ 

✓ 

R2 0.039 0.078 0.068 0.069 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 

a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES.  Standard errors 

in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

 

 

In Table 3.3, when restricting the regression to across the various ethnic groups in terms of 

educational outcomes, we find there was greater completion and attendance of secondary and 

tertiary education among the African ethnic group (see Columns 1-4). Although, the results for 

Amerindian students were not statistically significant, we find that the coefficients for the 

completion of secondary and attendance of secondary or higher were negative.  There are many 

systematic, socioeconomic, and potential discriminatory factors can have a negative effect on the 

Amerindians. However, in Guyana, Amerindian communities were said to be at a disadvantage, 



82 

as GSSs were mainly located on the outskirts of Amerindian villages which are situated in 

Guyana’s hinterland, thus making access to secondary schools difficult, as many students either 

have to travel by boat or walk long distances to get to their schools (UNICEF, 2017).88 In Table 

3.4, column 5 and 6, we find that 65 percentage points of  East Indian students were more likely 

to complete secondary education and 68 percentage point were more likely to complete secondary  

or higher education in Guyana.  

 

Overall, we find that the USE (2003-2007) policy reform did not close the achievement gaps in 

education across ethnicities. However, the Africans and East Indians benefited more from the USE 

policy reform. 

 

 

 

 
88 The study of indigenous women and children in Guyana, 2017, pp 70 for factors influencing attendance and dropouts 

in Amerindian communities in Guyana,   https://www.unicef.org/guyanasuriname/media/451/file/SitAn-

Indigenous-women-children-in-Guyana.pdf)  

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf


83 

Table 3.3: Educational Attainment - Restricted regressions (Ethnic groups: Africans, Mixed and Amerindians) 
 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) 

 Secondary 

(Completed) 

(African) 

Secondary/ 

higher 

education 

(Completed 

- African) 

Tertiary 

(Completed- 

African) 

Secondary/ 

higher 

(Attended- 

African) 

Secondary 

(Completed) 

(Mixed) 

Secondary/ 

higher 

education 

(Completed 

- Mixed) 

Tertiary 

(Completed- 

Mixed) 

Secondary/ 

higher 

(Attended- 

Mixed) 

Secondary 

(Completed - 

Amerindian) 

Secondary/ 

higher 

education 

(Completed - 

Amerindian) 

Tertiary 

(Completed- 

Amerindian) 

Secondary/ 

higher 

(Attended- 

Amerindian) 

USE x Intensity 0.807*** 1.173*** 0.339** 0.286** 0.226 0.152 -0.136 0.177 -0.387 -0.073 0.314 -0.285 

 (0.154) (0.174) (0.170) (0.077) (0.192) (0.178) (0.143) (0.127) (0.335) (0.374) (0.192) (0.317) 

Female 0.766*** 0.106*** 0.028 0.022** 0.061**  0.090*** 0.029 0.028** -0.014 -0.033 -0.019 -0.023 

 (0.026) (0.022) (0.022) (0.009) (0.027) (0.028) (0.021) (0.014) (0.035) (0.039) (0.015) (0.034) 

Constant -0.008 0.099 0.114* 0.851*** 0.166** 0.358*** 0.211*** 0.832*** 0.364** 0.279 -0.006 0.906*** 

 (0.056) (0.062) (0.060) (0.030) (0.074) (0.070) (0.056) (0.052) (0.155) (0.169) (0.059) (0.145) 

Observations 1922 1922 1922 1922 1319 1319 1319 1319 596 596 596 596 

R2 0.055 0.054 0.036 0.045 0.042 0.073 0.058 0.047 0.071 0.083 0.091 0.108 

Note: The omitted category for ethnicity is East Indian and Other consists of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any if the listed race/ethnic groups in 
Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not 
(aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors in 
parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

Table 3.4: Educational Attainment - Restricted regressions (Ethnic groups: Other and East Indian – reference group) 

 (1) (2) (3) (4) (5) (6) (7) 

 Secondary 

(Completed - 

Other) 

Secondary/ 

higher education 

(Completed - 

Other) 

Tertiary 

(Completed- 

Other) 

Secondary 

(Completed – 

East Indians – ref 

group) 

Secondary/ 

higher education 

(Completed – 

East Indians) 

Tertiary 

(Completed- 

East Indian) 

Secondary/ higher 

(Attended- East 

Indian) 

USE x Intensity 1.542 2.312 0.771 0.650*** 0.684*** 0.009 -0.043 

 (2.861) (2.229) (2.044) (0.182) (0.195) (0.115) (0.147) 

Female -0.179 -0.768* -0.589 0.072*** 0.054*** -0.020* 0.045*** 

 (0.611) (0.315) (0.556) (0.016) (0.018) (0.011) (0.013) 

Constant -0.271 0.316 0.587 0.001 0.082 0.089** 0.810*** 

 (1.195) (0.815) (0.984) (0.064) (0.069) (0.042) (0.052) 

Observations 18 18 18 2888 2888 2888 2888 

R2 0.662 0.899 0.842 0.048 0.057 0.034 0.034 

Note: Results for Other (secondary/higher attended) were dropped due to zero observations for secondary or higher atteneded. The USE is the treatment, i.e., a dummy variable 
equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the 

reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors in parentheses are 

clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 



84 

3.5.3 The effect of the USE (2003-2007) on labour market outcomes in 

Guyana 

 

Based on the analysis of the impact of the USE policy reform on education attainment in Guyana, 

it is observed that the USE policy reform resulted in an increase in both the completion and 

attendance of secondary or tertiary education in Guyana. In this section, we examine the impact of 

the USE policy reform on labour market outcomes.  

 

In Table 3.5, column 1, we find that the USE policy reform led to an increase in the employment, 

but the result is not statistically significant. In column 2, we find that the USE policy reform had 

a negative impact on wage employment, however the result is not statistically significant. 

Estimates in column 3 indicate a that the USE policy led to an increase in weekly work hours, 

however it is not statistically significant. Moreover, women tend to work 10.04 hours less when 

compared to men. We also find that there was an increase in work hours by 3.38, 2.51 and 10.18 

among Africans, Mixed and the other ethnic groups respectively when compared to East Indians. 

However, we find a reduction in the work hours for the Amerindian workers by 2.87 hours. These 

results are consistent with the literature that shows that higher education affects work hours (Card, 

1999; Heckman et al., 2006; Ye Chen, 2021). 

 

When looking at wages in column 4, we find that an additional year of exposure to the USE policy 

is associated with an increase in wage by 52.5 percentage points. This finding supports the 

argument that better-educated individuals earn higher wages (Kampelmann et al., 2018).  

We also find that females earned lower than their male counterparts. Consequently, we find that 

the USE policy led to an increase in the wage of people who were exposed to the policy. 

Furthermore, we find that log wage increases for Africans and Mixed ethnic groups but reduced 

for the Amerindians when compared to the East Indians. 

 

Columns 5, Table 3.5 presents the effects of the USE policy reform on formal wage workers. We 

find that the policy has no significant effect on the probability of being a formal wage worker.



85 

 

 

Table 3.5: Effect of the USE on the Labour Market in Guyana 

 (1) (2) (3) (4) (5) 

 Employed Wage 

employed 

Work hours Log wage Formal wage 

worker 

USE x Intensity 0.039 -0.112 3.155 0.525** 0.210 

 (0.076) (0.118) (3.803) (0.257) (0.135) 

Female -0.358*** 0.034** -10.041*** -0.214*** 0.128*** 

 (0.015) (0.015) (0.592) (0.029) (0.020) 

African 0.150*** 0.142*** 3.379*** 0.115*** 0.222*** 

 (0.014) (0.021) (0.732) (0.030) (0.025) 

Mixed 0.062*** 0.070*** 2.509*** 0.090** 0.074*** 

 (0.015) (0.021) (0.755) (0.036) (0.024) 

Amerindian -0.016 0.086* -2.870* -0.197** -0.001 

 (0.031) (0.038) (1.590) (0.078) (0.043) 

Other 0.152** -0.061 10.182** -0.094 0.052 

 (0.069) (0.112) (4.205) (0.281) (0.183) 

Constant 0.780*** 0.651*** 48.353*** 10.990*** 0.201*** 

 (0.030) (0.045) (1.436) (0.102) (0.060) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.656 

✓ 

✓ 

✓ 

3999 

0.689 

✓ 

✓ 

✓ 

3999 

46.908 

✓ 

✓ 

✓ 

2312 

11.15 

✓ 

✓ 

✓ 

2759 

0.437 

✓ 

✓ 

✓ 

R2 0.176 0.055 0.100 0.081 0.171 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any if the listed 

race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools 

in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

 

  



86 

Table 3.6: Labour Market Outcomes - Restricted regressions (Ethnic groups: Africans and Mixed) 

 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) 

 Employed 

(African) 

Wage 

employed 

(African) 

Work 

hours 

(African) 

Log wage 

(African) 

Formal 

wage 

worker 

(African) 

Employed 

(Mixed) 

Wage 

employed 

(Mixed) 

Work 

hours 

(Mixed) 

Log wage 

(Mixed) 

Formal 

wage 

worker 

(Mixed) 

USE x Intensity -0.049 0.257 8.616 0.299 0.595*** 0.006 -0.388* 3.004 0.151 -0.308 

 (0.166) (0.220) (8.305) (0.379) (0.223) (0.182) (0.231) (8.677) (0.428) (0.280) 

Female -0.184*** 0.091*** -10.954*** -0.214*** 0.108*** -0.311*** 0.004 -10.268*** -0.198*** 0.205*** 

 (0.020) (0.021) (0.903) (0.035) (0.030) (0.026) (0.030) (1.313) (0.072) (0.047) 

Constant 0.870*** 0.634*** 50.741*** 11.225*** 0.310*** 0.838*** 0.825*** 52.306*** 11.244*** 0.460*** 

 (0.058) (0.077) (2.866) (0.135) (0.084) (0.071) (0.085) (3.200) (0.170) (0.119) 

Observations 1922 1419 1419 887 1082 1319 864 864 496 591 

R2 0.066 0.053 0.119 0.061 0.144 0.126 0.079 0.129 0.073 0.154 
Note: Results for Other was dropped due to zero observations for secondary or higher atteneded. The USE is the treatment, i.e., a dummy variable equal to one for individuals born 

in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is 

defined as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, 

there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

Table 3.7: Labour Market Outcomes - Restricted regressions (Ethnic groups: Africans and Mixed) 

 (1) (2) (3) (4) (5) (6) 

 Employed 

(Amerindian) 

Wage 

employed 

(Amerindian) 

Work hours 

(Amerindian) 

Log wage 

(Amerindian) 

Formal wage 

worker 

(Amerindian) 

Employed 

(Other) 

USE x Intensity 0.357 0.071 51.368*** 0.678 0.806** 3.775** 

 (0.345) (0.396) (19.395) (0.830) (0.383) (1.182) 

Female -0.260*** -0.010 -11.295*** -0.260** 0.214*** -0.307 

 (0.041) (0.063) (2. 660) (0.121) (0.071) (0.549) 

Constant 0.623*** 0.603*** 27.009*** 10.749*** -0.146 -0.374 

 (0.159) (0.179) (8.533) (0.376) (0.183) (0.593) 

Observations 596 302 302 116 190 18 

R2 0.217 0.138 0.229 0.254 0.173 0.716 
Note: Results for Other (wage employed, work hours, log wage and formal wage worker) were dropped due to small number of observations for secondary or higher atteneded. 

The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the 

control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools in the region in 2002, 

based on the DES. Standard errors in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01.



87 

When examining the labour market results in Tables 3.6 and 3.7, we find that the results were 

mixed, with Africans (in column 5) being more likely to be employed in the formal sector. While 

there is a negative impact on the wage employed for the mixed ethnic group (see column 7). 

Furthermore, in Table 3.7, column 3, we find that Amerindians work more hours.  

 

  3.5.4 Robustness Checks  

 

We perform several robustness checks to examine whether the results are robust. Consequently, 

we check the robustness of our treatment intensity measure. Firstly, we check whether our results 

change if we redefine our intensity variable as: (𝐼𝑟) - the intensity of the USE program in region 

(r) of residence is defined as the average share of non-GSS institutions in the region for the period 

1999-2002 based on the DES rather than the share of non-GSS institutions in the immediate pre-

policy year of 2002.  

Table 3.8: Robustness Check - Effect of the USE on Educational Attainment in Guyana 

 (1) (1) (2) (3) 

 Secondary 

(Completed) 

Secondary/tertiary 

education 

(Completed) 

Tertiary 

(Completed) 

Secondary/ tertiary 

(Attended) 

USE x Intensity 0.411*** 0.464*** 0.032 0.155** 

 (0.098) (0.107) (0.066) (0.068) 

Female 0.067*** 0.070*** 0.002 0.032*** 

 (0.011) (0.011) (0.008) (0.007) 

African 0.052*** 0.189*** 0.136*** 0.121*** 

 (0.015) (0.017) (0.012) (0.011) 

Mixed 0.031** 0.101*** 0.068*** 0.077*** 

 (0.018) (0.021) (0.015) (0.013) 

Amerindian -0.056** -0.079** -0.020 -0.045** 

 (0.025) (0.030) (0.019) (0.026) 

Other 0.009 0.226*** 0.219** 0.127*** 

 (0.088) (0.101) (0.109) (0.040) 

Constant 0.064 0.139*** 0.083*** 0.754*** 

 (0.039) (0.044) (0.027) (0.030) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.278 

✓ 

✓ 

✓ 

6326 

0.428 

✓ 

✓ 

✓ 

6326 

0.148 

✓ 

✓ 

✓ 

6326 

0.879 

✓ 

✓ 

✓ 

R2 0.039 0.079 0.064 0.069 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 



88 

a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors 

in parentheses are clustered at the treatment level; there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

 

Table 3.8 presents our robustness check where we regress our main outcome variables for 

schooling, (i.e., completion of secondary education, completion of secondary or tertiary education, 

tertiary education and secondary or tertiary education attended) on our explanatory variables. In 

columns 1 and 2, we find that the results are robust when compared to our main results. 

Furthermore, the result in column 1 shows that the coefficient for  𝛽1 (𝑈𝑆𝐸𝑗 ∗  𝐼𝑟) is significant at 

the 1% level, which means that greater exposure the USE policy reform increased the completion 

of secondary and secondary or tertiary education by 41.10 and 46.40 percentage points 

respectively. These results are similar to our baseline result. This finding is similar to that of Ajayi 

and Ross (2020), who find that greater exposure to Free Primary Education (FPE) program in 

Kenya increased the highest education level achieved.  

In Table 3.8, column 4, we find that the results were robust when looking at the effect of the USE 

policy reform on the attendance of secondary or tertiary education, i.e., exposure to the USE policy 

reform increased the attendance of secondary or tertiary education by 15.5 percentage points.  

 

Table 3.9 below presents the results for the robustness check when looking at the effect of the USE 

on the labour market in Guyana using the average share of non-GSS institutions in the pre-policy 

period (1999-2002). Like our baseline results, we find that exposure to the USE policy reform 

increases log wage. Furthermore, we did not find any effect when examining the USE policy 

reform on the other outcome variables such as employment, wage employed work hours and 

formal wage workers.  

Secondly, we redefined the intensity as the share of non-GSS students in secondary institutions in 

the region in 2002 who would have never completed general secondary schools, i.e., those who 

would have dropped out of non-GSSs in grade 9, based on the DES. This was done in order to 

determine whether the USE policy would have had a positive effect on students who would have 

dropped out of secondary education at grade 9 which is grade in which children in CHS would 



89 

have normally dropped out at (Williams, 1993; Jennings, 2000; Gill-Marshall, 2000). 

  

In column 1, Table 3.10, we find that the results are robust when compared to our main results. 

Furthermore, the result in column 1 shows that the coefficient for  𝛽1 (𝑈𝑆𝐸𝑗 ∗  𝐼𝑟) is significant at 

the 1% level, which means that greater exposure the USE policy reform increased the completion 

of secondary education and secondary or tertiary education by 22 and 15.40 percentage points 

respectively. However, we find that tertiary education decline by 6.5 percentage points and 

increased secondary or tertiary education attended by 5.3 percentage points. These results are 

similar to our baseline results. In Table 3.11, columns 3 and 5, we also find that exposure to the 

USE policy reform increases work hours and formal wage worker.  



90 

Table 3.9: Robustness Check – Effect of the USE on the Labour Market in Guyana 

 (1) (2) (3) (4) (5) 

 Employed Wage employed Work hours Log wage Formal wage 

worker 

USE x Intensity 0.035 -0.095 3.613 0.451* 0.204 

 (0.076) (0.117) (3.747) (0.255) (0.136) 

Female -0.358*** 0.034** -10.040*** -0.214***     0.128*** 

 (0.015) (0.015) (0.592) (0.029) (0.020) 

African 0.150*** 0.142*** 3.380*** 0.116***     0.222*** 

 (0.014) (0.021) (0.733) (0.030)    (0.025) 

Mixed 0.062*** 0.070*** 2.510*** 0.090**     0.074*** 

 (0.015) (0.021) (0.755) (0.036)  (0.024) 

Amerindian -0.016 0.086* -2.871* -0.196** -0.001 

 (0.031) (0.038) (1.590) (0.078) (0.043) 

Other 0.152** -0.061 10.186** -0.094 0.052 

 (0.069) (0.112) (4.206) (0.281)   (0.183) 

Constant 0.781*** 0.645*** 48.136*** 11.013***    0.201*** 

 (0.031) (0.047) (1.465) (0.104)  (0.062) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.656 

✓ 

✓ 

✓ 

3999 

0.689 

✓ 

✓ 

✓ 

3999 

46.908 

✓ 

✓ 

✓ 

2312 

11.15 

✓ 

✓ 

✓ 

2759 

0.437 

✓ 

✓ 

✓ 

R2 0.176 0.055 0.100 0.080 0.171 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any 

if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy 

(i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share 

of non-GSS schools out of all secondary schools in the region in 2002, based on the DES.  Standard errors in parentheses are clustered at the treatment level; there 

are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 



91 

Table 3.10: Robustness Check- the effect of the USE on Educational Attainment 

 (1) (1) (2) (3) 

 Secondary 

(Completed) 

Secondary/tertiary 

education 

(Completed) 

Tertiary 

(Completed) 

Secondary/ 

tertiary (Attended) 

USE x Intensity 0.220*** 0.154*** -0.065** 0.053** 

 (0.028) (0.038) (0.027) (0.023) 

Female 0.067*** 0.071*** 0.002 0.032*** 

 (0.011) (0.011) (0.008) (0.007) 

African 0.053*** 0.190*** 0.136*** 0.124*** 

 (0.016) (0.017) (0.012) (0.011) 

Mixed 0.032* 0.101*** 0.067*** 0.077*** 

 (0.016) (0.021) (0.015) (0.012) 

Amerindian -0.050* -0.075** -0.021 -0.043* 

 (0.026) (0.030) (0.019) (0.025) 

Other 0.009 0.225** 0.219** 0.127*** 

 (0.089) (0.101) (0.109) (0.040) 

Constant 0.155*** 0.273*** 0.116*** 0.798*** 

 (0.016) (0.018) (0.013) (0.012) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.278 

✓ 

✓ 

✓ 

6326 

0.428 

✓ 

✓ 

✓ 

6326 

0.148 

✓ 

✓ 

✓ 

6326 

0.879 

✓ 

✓ 

✓ 

R2 0.042 0.079 0.065 0.069 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 

a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors 

in parentheses are clustered at the treatment level; there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 
  

 

 

 

 

 

 

 

 

 

 



92 

Table 3.11: Robustness Check – Effect of the USE on the Labour Market in Guyana 

 (1) (2) (3) (4) (5) 

 Employed Wage employed Work hours Log wage Formal wage 

worker 

USE x Intensity          -0.014 0.006 2.643** -0.100 0.085* 

 (0.029) (0.043) (1.274) (0.085) (0.047) 

Female -0.358*** 0.034** -10.041*** -0.215*** 0.128*** 

 (0.015) (0.015) (0.593) (0.029) (0.020) 

African 0.150*** 0.142*** 3.409*** 0.116*** 0.224*** 

 (0.014) (0.021) (0.736) (0.029) (0.025) 

Mixed 0.062*** 0.070** 2.522*** 0.088** 0.075** 

 (0.015) (0.021) (0.755) (0.036) (0.024) 

Amerindian -0.016 0.085** -2.840** -0.192** 0.000 

 (0.031) (0.038) (1.593) (0.078) (0.043) 

Other 0.151** -0.060 10.172** -0.103 0.052 

 (0.069) (0.112) (4.215) (0.284) (0.183) 

Constant 0.799*** 0.607*** 48.724*** 11.231*** 0.257*** 

 (0.013) (0.020) (0.585) (0.033) (0.026) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

6326 

0.658 

✓ 

✓ 

✓ 

3999 

0.653 

✓ 

✓ 

✓ 

3999 

46.559 

✓ 

✓ 

✓ 

2312 

11.12 

✓ 

✓ 

✓ 

2759 

0.422 

✓ 

✓ 

✓ 

R2 0.176 0.055 0.100 0.080 0.171 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any 

if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy 

(i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share 

of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level; there 

are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01.  

 

 

 



93 

Table 3.12: Summary statistics (respondents who have not migrated vs those who migrated) 

Variable Mean SD 

 Have not   

Migrated  

 N 

  

Mean SD 

 

Migrated 

N 

Outcome variables       

Secondary education (Completed) 0.276 0.447 5183 0.277 0.448 1030 

Secondary or tertiary (Completed) 0.423 0.494 5183 0.442 0.497 1030 

Tertiary education (Completed) 0.145 0.352 5183 0.163 0.369 1030 

Secondary or tertiary (Attended) 0.878 0.327 5183 0.883 0.321 1030 

Employed 0.666 0.472 5183 0.616 0.487 1030 

Wage employed 0.688 0.463 3315 0.698 0.459   626 

Formal wage worker 0.491 0.500 2282 0.556 0.497   437 

Work hours 46.550 17.847 3315 48.562 19.134   626 

Ln wage 11.17 0.631 2129 11.255 0.679   410 

USE & Intensity       

USE  0.525 0.499 5183 0.513 0.500 1030 

Intensity 0.725 0.109 5183 0.752 0.126 1030 

Individual characteristics       

Female 0.524 0.499 5183 0.618 0.486 1030 

East Indian 0.444 0.497 5183 0.318 0.466 1030 

African 0.291 0.454 5183 0.290 0.454 1030 

Mixed 0.178 0.383 5183 0.278 0.448 1030 

Amerindian 0.085 0.279 5183 0.114 0.317 1030 

Others 0.002 0.048 5183 0 0 1030 

Region 1 0.020 0.140 5183 0.006 0.076 1030 

Region 2 0.064 0.244 5183 0.046 0.209 1030 

Region 3 0.119 0.324 5183 0.262 0.440 1030 

Region 4 0.285 0.451 5183 0.284 0.451 1030 

Region 5 0.165 0.371 5183 0.158 0.365 1030 

Region 6 0.213 0.410 5183 0.106 0.309 1030 

Region 7 0.017 0.123 5183 0.019 0.138 1030 

Region 8 0.026 0.160 5183 0.021 0.145 1030 

Region 9 0.015 0.121 5183 0.002 0.044 1030 

Region 10  0.076 0.265 5183 0.094 0.292 1030 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who 
did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for 

individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-

24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools in 

the region in 2002, based on the DES.  

  

 

 

 

 
 

 

 



94 

 

Table 3.13: Robustness Check: Effect of the USE on Educational Attainment in Guyana 

 (1) (1) (2) (3) 

 Secondary 

(Completed) 

Secondary/tertiary 

education 

(Completed) 

Tertiary 

(Completed) 

Secondary/ 

tertiary (Attended) 

USE x Intensity 0.450*** 0.524*** 0.056 0.163** 

 (0.103) (0.117) (0.068) (0.075) 

Female 0.064*** 0.077*** 0.012 0.034*** 

 (0.012) (0.012) (0.008) (0.008) 

African 0.055** 0.191*** 0.133*** 0.120*** 

 (0.017) (0.018) (0.013) (0.012) 

Mixed 0.046** 0.118*** 0.069*** 0.096*** 

 (0.019) (0.023) (0.019) (0.014) 

Amerindian -0.051* -0.080** -0.024 -0.010 

 (0.030) (0.036) (0.024) (0.031) 

Other 0.099 0.306** 0.210 0.081 

 (0.132) (0.153) (0.163) (0.073) 

Constant 0.051 0.113** 0.068** 0.746*** 

 (0.041) (0.046) (0.027) (0.032) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

5183 

0.276 

✓ 

✓ 

✓ 

5183 

0.423 

✓ 

✓ 

✓ 

5183 

0.145 

✓ 

✓ 

✓ 

5183 

0.878 

✓ 

✓ 

✓ 

R2 0.045 0.086 0.069 0.072 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who 
did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for 

individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-

24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools in 

the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level; there are 200 clusters * p < 

0.10, ** p < 0.05, *** p < 0.01. 

 

Thirdly, to reduce the influence of migration, we restricted the results to the persons who were still 

residing in the same regions of birth at the time of the GLFS survey. This was done to determine whether 

the results would still hold when considering the impact of migration. The assumption made in the 

baseline regression is that where the individual is currently residing is where they would have completed 

their education.  

 

 Based on the survey, a total of 5183 of the survey sample were still residing in their region of birth, 

which means that they did migrate to other regions; this accounts for 82 percent of the observation (see 

Table 3.12). Furthermore, in Tables 3.13-3.14, we restricted the results of the regressions to persons who 

were still residing in their same region of birth at the time of the survey and find that the results are 

robust. In Table 3.13, based on our baseline equation, we find that greater exposure to the USE policy 



95 

reform led to an increase in the completion of secondary and secondary or tertiary education by 45.0 and 

52.4 percentage points. In addition, when examining the effects of the USE policy reform on the 

attendance of secondary or tertiary education, we find that greater exposure to the USE policy reform 

led to an increase in the attendance of secondary or tertiary education by 16.3 percentage points. 

Furthermore, we find that the results were robust when looking at the labour market impacts in Table 

3.14.  

 



96 

 

 

Table 3.14: Effect of the USE on the Labour Market in Guyana 

 (1) (2) (3) (4) (5) 

 Employed Wage employed Work hours Log wage Formal wage worker 

USE x Intensity 0.024 -0.049 4.789 0.493* 0.289** 

 (0.083) (0.135) (4.525) (0.259) (0.145) 

Female   -0.340*** 0.046** -9.386*** -0.167***   0.187*** 

 (0.016) (0.016) (0.650) (0.030) (0.024) 

African    0.138***   0.125*** 2.874*** 0.107*** 0.211*** 

 (0.016) (0.022) (0.784) (0.029) (0.023) 

Mixed 0.061*** 0.072** 1.449* 0.066* 0.091** 

 (0.017) (0.023) (0.842) (0.035) (0.034) 

Amerindian -0.005 0.058 -4.694** -0.284** -0.086 

 (0.039) (0.046) (2.139) (0.110) (0.065) 

Other 0.227** 0.058 1.916 -0.124 0.187 

 (0.102) (0.136) (5.150) (0.357) (0.187) 

Constant 0.784*** 0.629*** 47.619*** 11.011** 0.206** 

 (0.034) (0.053) (1.740) (0.104) (0.063) 

Observations 

Mean 

Region fixed effects 

Age fixed effects 

Survey quarter FE 

5183 

0.664 

✓ 

✓ 

✓ 

3315 

0.649 

✓ 

✓ 

✓ 

3315 

46.109 

✓ 

✓ 

✓ 

2129 

11.15 

✓ 

✓ 

✓ 

2282 

0.483 

✓ 

✓ 

✓ 

R2 0.162 0.050 0.090 0.072 0.141 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any if the listed 

race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools 

in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01.  

   

 



97 

Fourthly, we performed additional robustness checks were done to ascertain whether the results 

still hold when dropping the regions one at a time (i.e., regions 1 to 10) from the survey data. These 

results are presented in Tables 3.15-3.16. When using the survey region as the region in which the 

respondents were educated in Table 3.15, we find that the results are robust for the completion of 

secondary, secondary or tertiary when dropping regions 1,2,3,5,6,7,8,9 and 10. Consequently, the 

results hold across nine of the ten administrative regions in Guyana. The results were also robust 

for the attendance of secondary or tertiary for regions 2,3,5,6,7,8,9 and 10. In addition, the results 

are also robust for wage when looking at the labour market outcomes in Table 3.16. 

 

For our final robustness check, we applied multiple hypothesis testing using the Westfall-Young 

and use the Bonferroni Holm and Sidak-Holm adjusted p-value were used for comparison 

purposes. Here we compare the three-hypothesis tests by examining their adjusted p-value to a cut-

off of 0.05 and 0.01 respectively; given that the nominal cut-off in our baseline specification is 

significant at the 0.01 and 0.05 levels respectively. When looking at the Westfall-Young adjusted 

p-values, we find that completion of secondary and secondary or tertiary education is significant 

at the 1% level, while secondary or tertiary education attended is significant at the 5% level after 

adjusting for Type 1 error (Jones and Reif, 2019).91Similarly, we find that log wage is also 

significant at the 5% level after adjusting for Type 1 error (see Table 3.17, below for the multiple 

hypothesis tests). Furthermore, when looking at the Bonferroni-Holm and Sidak-Holm adjusted p-

values, we also find that the results were also significant for the completion of secondary or tertiary 

education and secondary or tertiary education attended. Consequently, we conclude that the real 

effect was observed for four outcome variables, i.e., the completion of secondary, secondary or 

tertiary education, secondary or tertiary education attended and log wage. Otherwise, the effects 

observed for work hours might be due to Type 1 error.  

 

 
9 Wyoung is a Stata command which applied free step-down resampling methodology of Westfall and Young (1993). 

It is used to calculate the adjusted p-values. The Stata command programmed by Julian Reif, is a precursor to the 

Romano-Wolf procedure, thus using bootstrap resampling to allow for dependence across outcomes. Furthermore, it 

also controls the Family Wise Error Rate (FWER) and allow for dependence amongst p-values. In addition, it also 

computes the Bonferroni-Holm and Sidak-Holm adjusted p-values (Jones and Reif, 2019). 



98 

Table 3.15: Robustness Check: Effect of the USE on Educational Attainment in Guyana (Share of non-GSS institutions) 

Dropped 

Regions Diff-in-Diff 

Secondary 

(Completed) 

Secondary/ tertiary 

education 

(Completed) 

Tertiary 

(Completed) 

Secondary/ 

tertiary 

(Attended) 

Total 

Observations 

Region 1 USE x Intensity 

  0.389*** 

(0.103) 

   0.460*** 

(0.114) 

0.049 

(0.071) 

0.083 

(0.067) 6214 

Region 2 USE x Intensity 

  0.367*** 

(0.098) 

   0.406*** 

 (0.108) 

0.021 

 (0.068) 

  0.171** 

 (0.068) 5941 

Region 3 USE x Intensity 

  0.321*** 

(0.100) 

 0.369*** 

(0.112) 

0.030 

(0.070) 

0.144** 

(0.072) 5408 

Region 4 USE x Intensity 

-0.096 

(0.099) 

0.029 

(0.133) 

0.122 

(0.100) 

0.168 

(0.107) 4521 

Region 5 USE x Intensity 

  0.285*** 

(0.098) 

  0.362*** 

(0.107) 

0.046 

(0.067) 

  0.171** 

(0.072) 5299 

Region 6 USE x Intensity 

  0.472*** 

(0.092) 

   0.476*** 

(0.105) 

-0.019 

(0.061) 

  0.175** 

(0.069) 5092 

Region 7 USE x Intensity 

  0.414*** 

(0.095) 

  0.470*** 

(0.106) 

0.036 

(0.068) 

 0.172** 

(0.070) 6216 

Region 8 USE x Intensity 

 0.434*** 

(0.103) 

  0.525*** 

(0.110) 

0.069 

(0.072) 

0.146** 

(0.072) 6166 

Region 9 USE x Intensity 

 0.380*** 

(0.101) 

  0.449*** 

(0.111) 

0.043 

(0.066) 

  0.186*** 

(0.071) 6244 

Region 10 USE x Intensity 

0.465*** 

(0.098) 

  0.454*** 

(0.111) 

-0.025 

(0.057) 

  0.195*** 

(0.072) 5833 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any 

if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy 

(i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share 

of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, there 

are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01.  



99 

Table 3.16: Robustness Check: Effect of the USE on the Labour Market in Guyana (Share of 

non-GSS institutions) 

Dropped 

Regions Diff-in-Diff Employed 

Wage 

Employed Work hours Log wage 

Formal wage 

worker 

Region 1 USE x Intensity 

0.013 

(0.079) 

-0.127 

(0.121) 

-0.614 

(3.824) 

0.446* 

(0.265) 

0.226 

(0.141)  

Region 2 USE x Intensity 

0.031 

(0.078) 

-0.126 

(0.122) 

6.685* 

(3.804) 

0.460* 

(0.242) 

0.237* 

(0.138) 

Region 3 USE x Intensity 

0.074 

(0.081) 

-0.083 

(0.127) 

1.541 

(4.080) 

0.438 

(0.280) 

0.200 

(0.140) 

Region 4 USE x Intensity 

0.118 

(0.115) 

-0.293 

(0.191) 

-1.575 

(5.484) 

     1.130*** 

(0.370) 

 0.368* 

(0.214) 

Region 5 USE x Intensity 

0.032 

(0.080) 

-0.162 

(0.123) 

2.450 

(4.107) 

0.755*** 

(0.268) 

0.110 

(0.140) 

Region 6 USE x Intensity 

0.020 

(0.076) 

-0.080 

(0.119) 

4.237 

(3.660) 

0.378 

(0.253) 

0.163 

(0.143) 

Region 7 USE x Intensity 

0.057 

(0.078) 

-0.088 

(0.120) 

4.132 

(3.894) 

0.592** 

(0.269) 

0.273** 

(0.137) 

Region 8 USE x Intensity 

0.049 

(0.081) 

-0.100 

(0.124) 

2.647 

(4.183) 

0.438 

(0.273) 

0.154 

(0.138) 

Region 9 USE x Intensity 

0.037 

(0.077) 

-0.046 

(0.120) 

5.471 

(3.875) 

0.529** 

(0.267) 

0.205 

(0.141) 

Region 10 USE x Intensity 

0.011 

(0.075) 

-0.141 

(0.131) 

3.061 

(3.966) 

0.494* 

(0.268) 

0.250* 

(0.137) 
Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who 

did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one 

for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) and equal to zero (i.e., the control) if not (aged 

18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools out of all secondary schools 

in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, there are 200 clusters * p 

< 0.10, ** p < 0.05, *** p < 0.01.  



100 

Table 3.17: Robustness Check (Baseline specification): Multiple Hypothesis Testing (Westfall-Young, Bonferroni and Sidak-Holm 

Corrections) 

 

Outcome variable 

 

 

Coeff (se) 

(Diff-in-diff) 

Unadjusted 

p-value 

(Diff-in-Diff) 

Westfall-Young 

(Adjusted p- value) 

(Diff-in-Diff) 

Bonferroni-Holm 

(Adjusted p- value) 

(Diff-in-Diff) 

Sidak-Holm 

(Adjusted p-value) 

Westfall 

Young 

(sig) 

Bonf 

(sig) 

Sidak 

(sig) 

Secondary education 

(Completed) 

  0.374*** 

(0.097) 

0.000 0.00 0.000 0.000 yes yes yes 

Secondary/tertiary 

education (Completed) 

   0.426*** 

(0.107) 

0.000 0.00 0.000 0.000 yes yes yes 

Tertiary  

(Completed) 

0.032 

(0.065) 

0.620 0.46 0.620 0.620 no no no 

Secondary/ tertiary 

(Attended) 

0.161** 

(0.068) 

0.020 0.00 0.041 0.040 yes yes yes 

Employed 0.039 

(0.076) 

0.605 0.56 1.000 0.716 no no no 

Wage Employed -0.112 

(0.118) 

0.343 0.56 1.000 0.716 no no no 

Work hours 3.155 

(3.803) 

0.408 0.56 1.000 0.716 no no no 

Log wage 0.525** 

(0.257) 

0.043 0.05 0.215 0.198 yes no no 

Formal wage worker 0.209 

(0.135) 

0.123 0.28 0.492 0.408 no no no 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other (i.e., those who did not identified with any if the 

listed race/ethnic groups in Guyana). The USE is the treatment, i.e., a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 

11-17 in 2003) and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined as the share of non-GSS schools 

out of all secondary schools in the region in 2002, based on the DES. Standard errors in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** 

p < 0.05, *** p < 0.01.  

 

 

 

 



101 

In addition, we also did a DID-IV results on the impact of education on labour market outcomes 

with USE * intensity as the instrumental variable. We use USE*intensity as an instrumental 

variable for education to conduct 2SLS estimates. We chose to focus mainly on log wage, as it 

was the only statistically significant variable among the other labour market outcome variables. 

 

Table 3.18: Robustness Check (DID-IV: 2SLS Regressions) 

 
Log wage Log wage Log wage Log wage 

2SLS 2SLS 2SLS 2SLS 

1 2 3 4 

Secondary education (Completed) 1.330    

  (0.869)    

 [0.128]    

Secondary/Tertiary education (Completed)   0.997*   

  (0.530)   

  [0.062]   

Tertiary education (Completed)   4.244  

   (4.546)  

   [0.352]  

Secondary/Tertiary (Attended)    2.324 

    (1.455) 

    [0.113] 

Observations 2312 2312 2312 2312 

Controls ✓ ✓ ✓ ✓ 

Region fixed effects ✓ ✓ ✓ ✓ 

Age fixed effects  ✓ ✓ ✓ ✓ 

Survey quarter fixed effects  ✓ ✓ ✓ ✓ 

Kleibergen-Paap     6.626 8.812 0.775 6.269 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 
a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES. Where the 

brackets ( ) contains the standard errors and [ ] contains the p-values. Standard errors in parentheses are clustered at 

the treatment level; there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01.  

 

In Table 3.18 above, our coefficient for the completion of secondary/tertiary education when used 

as an instrumented variable with respect to the log wage is positive and statistically significant at 

the 10 percent level, suggesting that the USE policy increases wage income. Furthermore, although 

the other three coefficients are not statistically significant, the coefficients for the completion of 



102 

secondary education and secondary/tertiary education attended are close to significant with p-

values less than 0.15.  

However, when synthesizing various pieces of evidence, the conclusion that USE policy resulted 

in higher wage is still valid. 

 

In addition, we have also present placebo analysis to test that our main results are not driven by 

simultaneous impacts affecting both the region and the (treated group) 11-17 age groups, we 

conduct placebo tests. 

 

 

Table 3.19: Placebo test - Effect of the USE on Educational Attainment 

 (1) (2) (3) (4) 

 Secondary 

education  

(Completed) 

Secondary/ 

Tertiary 

education 

(Completed) 

Tertiary 

education 

(Completed) 

Secondary/ 

Tertiary 

education 

(Attended) 

USE x Intensity   0.374***  0.426*** 0.032   0.161** 

 (0.097) (0.106) (0.065) (0.68) 

 [0.000] [0.000] [0.004] [0.000] 

     

Observations  6326 6326 6326 6326 

Controls ✓ ✓ ✓ ✓ 

Region fixed effects ✓ ✓ ✓ ✓ 

Age fixed effects ✓ ✓ ✓ ✓ 

Survey quarter fixed effects ✓ ✓ ✓ ✓ 

R2 0.039 0.078 0.064 0.069 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 

a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES.  Where the 

brackets ( ) contains the standard errors and [ ] contains the p-values. Standard errors in parentheses are clustered at 

the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

 

 

 

 

 

 

 



103 

Table 3.20: Placebo test - Effect of the USE on the Labour Market in Guyana 

 (1) (2) (3) (4) (5) 

 Employed Wage 

employed 

Work hours Log wage Formal wage 

worker 

USE x Intensity 0.039  -0.112 3.155   0.525** 0.210 

 (0.076) (0.118) (3.803) (0.257) (0.135) 

 [0.011] [0.000] [0.000] [0.000] [ 0.000] 

      

Observations  6326 3999 3999 2312 2759 

Controls ✓ ✓ ✓ ✓ ✓ 

Region fixed effects ✓ ✓ ✓ ✓ ✓ 

Age fixed effects ✓ ✓ ✓ ✓ ✓ 

Survey quarter fixed effects ✓ ✓ ✓ ✓ ✓ 

R2 0.176 0.055 0.100 0.081 0.171 

Note: The omitted category for ethnicity is East Indian and Other consist of Chinese, Portuguese, White and the Other 

(i.e., those who did not identified with any if the listed race/ethnic groups in Guyana). The USE is the treatment, i.e., 

a dummy variable equal to one for individuals born in cohort j if exposed to the policy (i.e., those aged 11-17 in 2003) 

and equal to zero (i.e., the control) if not (aged 18-24 in 2003 at the time of the reform). Intensity of the USE is defined 

as the share of non-GSS schools out of all secondary schools in the region in 2002, based on the DES.  Standard errors 

in parentheses are clustered at the treatment level, there are 200 clusters * p < 0.10, ** p < 0.05, *** p < 0.01. 

 

 

When performing randomization inference on the outcome variables. The results in Table 3.19 

show that the coefficients for the completion of secondary education and secondary/tertiary 

education and the attendance of secondary/tertiary education are significant, with p=c/n being 

0.000 in all cases. 

Furthermore, when looking at the coefficient for log wage in Table 3.20, we find that the result 

was also statistically significant, with p=c/n being 0.000. However, when examining the other 

labour market variables, the results was not significant.  

 



104 

3.6 Conclusion  
 

This study uses the DES (1999-2018) and the GLFS (2018-2019) to analyse the effects of the USE 

(2003 -2007) school conversion policy on educational attainment and labour market outcomes in 

Guyana. The study contributes to the body of literature on supply-side policies aimed at increasing 

the supply of academic schooling based on the conversion of vocational or technical education 

into an academic track. Studies have shown that educational reform policies that convert 

vocational education to general education are likely to result in higher educational achievements 

and better-paying jobs (e.g., Malamud and Pop-Eleches, 2010; Brunello and Rocco, 2017).  

We use a difference-in-difference estimation strategy exploiting variation across birth cohorts and 

Guyana’s ten administrative regions where the conversion of CHSs and Primary Tops into GSSs 

occurred. We combine this with nationally representative data from the GLFS (2018-2019) that 

contains information on a wide range of outcomes related to education, migration, wages and 

work-hours outcomes. We find that greater exposure to the USE policy increases the completion 

of secondary and secondary or tertiary education by 37.4 and 42.6 percentage points respectively. 

Therefore, the USE policy reform resulted in an increase in the completion and graduation rates in 

secondary or tertiary education in Guyana.  

We find that greater exposure to the USE policy increases the attendance of secondary or tertiary 

education. When examining the labour market outcomes, we find that an additional year of 

exposure to the USE policy reform is associated with an increase in wage by 52.5 percentage 

points.  This finding supports the argument that better-educated individuals earn higher wages. We 

also find that the USE (2003-2007) policy reform did not close the achievement gaps in education 

across ethnicities. Moreover, our results are robust when redefining the intensity of the USE 

program as the average share of non-GSS institutions in the period 1999-2002 based on the DES; 

when restricting the results to the persons who were still residing in the same regions of birth at 

the time of the survey, i.e., examining the results without the impact of migration and when 

dropping the regions individually (i.e., regions 1 to 10) from the survey data. 

Effectively, our paper provides a comprehensive analysis of the effects of school conversion policy 

on educational attainment and labour market outcomes in Guyana. To date, this is the first study 



105 

to analyse the effects of this policy on Guyana’s educational system and labour markets. 

Consequently, this study provides evidence to support the ongoing conversion of Primary Tops to 

GSSs. Moreover, this study can be used by the Ministry of Education to speed up the process of 

their ongoing conversion, since the implementation of the USE (2003-2007) has thus far achieved 

some positive results in increasing the supply and improving access to secondary education in 

Guyana. Furthermore, this has reduced the dropout rates which was a phenomenon (in CHSs and 

Primary Tops) under the old system.  

 

Although the USE (2003-2007) policy reform did not show much impact on the labour market 

outcomes in Guyana, we noticed that it has led to an increase in the wages that were reported in 

the GLFS. Consequently, this supports the literature on better-educated individuals who earn 

higher wages, experience less unemployment and work in more prestigious occupations than their 

less-educated counterparts. 

 

It is evident that Guyana’s Ministry of Education has made strides in increasing and improving the 

access to secondary education, however, more can be done to increase the speed of the conversion 

of the remaining Primary Tops to GSSs. In addition, the MoE should evaluate the USE policy 

impact across ethnicities and employ measures to help close the current achievement gaps. 

Furthermore, we find that the policy also led to the completion of secondary or tertiary education 

when dropping eight out of the ten administrative regions; this highlights the realised benefits of 

Guyana’s USE (2003-2007) policy reform. However, the broader societal impacts and 

intergenerational effects are topics that warrant further exploration in future studies.  

 



106 

References 

 

Ajayi, K.F. and Ross, P.H., (2020). The effects of education on financial outcomes: Evidence 

from Kenya, Economic Development and Cultural Change, 69(1), pp.253-289. 
 

Ashenfelter, O. and Krueger, A., (1994). Estimates of the economic return to schooling from a 

new sample of twins, American Economic Review, 84(5), pp.1157-1173. 

 

Akresh, R., Halim, D. and Kleemans, M., (2018). Long-term and intergenerational effects of 

education: Evidence from school construction in Indonesia (No. w25265). National Bureau of 

Economic Research. 

 

Alderman, H., Orazem, P.F. and Paterno, E.M., (2001). ‘School quality, school cost, and the 

public/private school choices of low-income households in Pakistan’, Journal of Human 

resources, pp.304-326. 

 

Barrera-Osorio, F., Linden, L.L. and Urquiola, M., (2013). The effects of user fee reductions 

on enrolment: Evidence from a quasi-experiment. 

 

Bhalotra, S., (2004). Early childhood investments in human capital: Parental resources and 

preferences. University of Bristol Discussion Paper, 4, p.561. 

 

Bravo, D., Mukhopadhyay, S. and Todd, P.E., (2010). Effects of school reform on education 

and labour market performance: Evidence from Chile's universal voucher system. Quantitative 

economics, 1(1), pp.47-95. 

 

Brudevold-Newman, A., (2021). Expanding access to secondary education: Evidence from a 

fee reduction and capacity expansion policy in Kenya. Economics of Education Review, 83, 

p.102127. 

 

Brunello, G. and Rocco, L., (2017). The labour market effects of academic and vocational 

education over the life cycle: Evidence based on a British cohort. Journal of Human 

Capital, 11(1), pp.106-166. 

 

Card, D., (1999). The causal effect of education on earnings. Handbook of labour economics, 3, 

pp.1801-1863. 

 

Cyril Potter College of Education., (2022). Available at: https://cpce.edu.gy/ (Accessed: 10 

March 2022) 

 

Dearden, L., McIntosh, S., Myck, M. and Vignoles, A., (2002). The returns to academic and 

vocational qualifications in Britain. Bulletin of economic research, 54(3), pp.249-274. 

 

Deininger, K., (2003). Does cost of schooling affect enrolment by the poor? Universal primary 

education in Uganda. Economics of Education review, 22(3), pp.291-305. 

 

Driessen, G., (2000). The limits of educational policy and practice? The case of ethnic 

minorities in The Netherlands. Comparative education, 36(1), pp.55-72. 

 



107 

Duflo, E., (2001). Schooling and labour market consequences of school construction in 

Indonesia: Evidence from an unusual policy experiment. American economic review, 91(4), 

pp.795-813. 

 

Duflo, E., (2004). The medium run effects of educational expansion: Evidence from a large 

school construction program in Indonesia. Journal of Development Economics, 74(1), pp.163-

197. 

 

Duflo, E., Dupas, P. and Kremer, M., (2021). The impact of free secondary education: 

Experimental evidence from Ghana (No. w28937). National Bureau of Economic Research. 

 

Enu, P., Hagan, E., Ahouandjinou, E. and Attah-Obeng, P., (2014). Relationship between 

education and wage differentials in Ghana: a case study of Accra-a Suburb of greater Accra 

Region. International Journal of Academic Research in Business and Social Sciences, 4(1), 

p.277. 

 

Gafar, J., (2003). Guyana: From state control to free markets. Nova Publishers. 

 

Garlick, R., (2013). How price sensitive is primary and secondary school enrolment? Evidence 

from nationwide tuition fee reforms in South Africa. Unpublished working paper. Available 

at: http://pubdocs.worldbank.org/en/3241466186067633/garlickschoolfees.pdf   

 

Gill-Marshall, B., (2000). Child abuse in Guyana: A study of teacher abuse of children 

(Unpublished thesis). University of Guyana, Guyana. 

  

Guyana Labour Force Survey, Methodological Report., (2018). Available at:  

https://sistemasintegrales.cl/project/guyana-labor-force-survey/ (Accessed: 20 August 2021) 

Guyana Bureau of Statistics. (2022), Available at: https://statisticsguyana.gov.gy/ (Accessed: 

30 August 2022) 

Harmon, C., Oosterbeek, H. and Walker, I., (2003). The returns to education: 

Microeconomics. Journal of economic surveys, 17(2), pp.115-156. 

 

Heckman, J.J., Lochner, L.J. and Todd, P.E., (2006). Earnings functions, rates of return and 

treatment effects: The Mincer equation and beyond. Handbook of the Economics of 

Education, 1, pp.307-458. 

 

Human Development Report., (2010), Available at: 

https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-

english.human-development-report-2010-complete-english[Accessed: 5 Aug, 2023] 

Human Development Report., (2016), Available at: 

https://hdr.undp.org/system/files/documents/2016humandevelopmentreportpdf1pdf.pdf 

[Accessed: 5 Aug, 2023] 

İşcan, T.B., Rosenblum, D. and Tinker, K., (2015). School fees and access to primary 

education: Assessing four decades of policy in sub-Saharan Africa. Journal of African 

Economies, 24(4), pp.559-592. 

 

http://pubdocs.worldbank.org/en/3241466186067633/garlickschoolfees.pdf
https://sistemasintegrales.cl/project/guyana-labor-force-survey/
https://statisticsguyana.gov.gy/
https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-english.human-development-report-2010-complete-english
https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-english.human-development-report-2010-complete-english
https://hdr.undp.org/system/files/documents/2016humandevelopmentreportpdf1pdf.pdf


108 

International Labour Organisation., (2004). National Labour Law Profile: Guyana, Available 

at: https://www.ilo.org/ifpdial/information-resources/national-labour-law-

profiles/WCMS_158900/lang--en/index.htm (Accessed: 5 Sept 2022) 

 

Jennings, Z. and Miller, E., (1999). Educational reform in Guyana in the post-war 

period. Educational reform in the Commonwealth Caribbean, pp.149-198. 

 

Jones, D., Molitor, D. and Reif, J., (2019). What do workplace wellness programs do? Evidence 

from the Illinois workplace wellness study. The Quarterly Journal of Economics, 134(4), 

pp.1747-1791. 

 

Kan, S. and Klasen, S., (2021). Evaluating universal primary education in Uganda: School fee 

abolition and educational outcomes. Review of Development Economics, 25(1), pp.116-147. 

 

Kampelmann, S., Rycx, F., Saks, Y. and Tojerow, I., (2018). Does education raise productivity 

and wages equally? The moderating role of age and gender. IZA Journal of Labour 

Economics, 7(1), pp.1-37. 

 

London, T., (2018). Assessing Competing Perspectives: A Critical Analysis of Guyana's 

National Grade Six Assessment. Journal of Pan African Studies, 11(4). 

 

Malamud, O. and Pop-Eleches, C., (2010). General education versus vocational training: 

Evidence from an economy in transition. The review of economics and statistics, 92(1), pp.43-

60. 

 

Massing, N. and Schneider, S.L., (2017). Degrees of competency: the relationship between 

educational qualifications and adult skills across countries. Large-Scale Assessments in 

Education, 5(1), pp.1-34. 

 

Ministry of Education, Guyana., (2022). Available at: https://education.gov.gy/en/ (Accessed: 

20 November 2021).  

 

Murnane, R.J., Waldman, M.R., Willett, J.B., Bos, M.S. and Vegas, E., (2017). The 

consequences of educational voucher reform in Chile (No. w23550). National Bureau of 

Economic Research. 

 

Organisation for Economic Co-operation and Development., (2000). Literacy in the 

information age: final report of the International Adult Literacy Survey. OECD Publishing. 

Ozier, O., (2018). The Impact of Secondary Schooling in Kenya A Regression Discontinuity 

Analysis. Journal of Human Resources, 53(1), pp.157-188. 

 

Peet, E.D., Fink, G. and Fawzi, W., (2015). Returns to education in developing countries: 

Evidence from the living standards and measurement study surveys. Economics of Education 

Review, 49, pp.69-90. 

 

Psacharopoulos, G., (1994). Returns to investment in education: A global update. World 

development, 22(9), pp.1325-1343. 

 

https://www.ilo.org/ifpdial/information-resources/national-labour-law-profiles/WCMS_158900/lang--en/index.htm
https://www.ilo.org/ifpdial/information-resources/national-labour-law-profiles/WCMS_158900/lang--en/index.htm
https://education.gov.gy/en/


109 

Psacharopoulos, G. and Patrinos, H.A., (2018). Returns to investment in education: a decennial 

review of the global literature. Education Economics, 26(5), pp.445-458. 

 

Riphahn, R.T., (2012). Effect of secondary school fees on educational attainment. The 

Scandinavian Journal of Economics, 114(1), pp.148-176. 

 

Schultz, T.P., (2004). School subsidies for the poor: evaluating the Mexican Progresa poverty 

program. Journal of development Economics, 74(1), pp.199-250. 

 

Shavit, Y. and Muller, W., (2000). Vocational secondary education. European societies, 2(1), 

pp.29-50. 

Sianesi, B. and Van Reenen, J., (2002). The returns to education: a review of the empirical 

macro-economic literature. 

 

Study on Indigenous Women & Children in Guyana,  

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-

in-Guyana.pdf   

 

UNECIF., (2017). The study of indigenous women and children in Guyana, Available at: ,   

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-

in-Guyana.pdf) (Accessed: 20 October 2021). 

 

UNESCO, I., (2006/07). World data on education: sixth edition. Principles and General 

Objectives of Education. 

 

UNESCO, I., (2010/11). World data on education: seventh edition. Principles and General 

Objectives of Education. 

 

University of Guyana., (2022) Available at: https://www.uog.edu.gy/ (Accessed: 9 April 2022). 

 

World Bank., (2008). Universal Secondary Education (USE) in Guyana, Available at:  

https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941epo

rt0by0Geoff0Howse.pdf?sequence=1&isAllowed=y (Accessed: 20 August 2021).  

 

Ye, C., (2021) March. How Does Education Affect the Time Spending on Work and Leisure? 

In 6th International Conference on Financial Innovation and Economic Development (ICFIED 

2021) (pp. 274-280). Atlantis Press. 

 

 
 

 

  

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.uog.edu.gy/
https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941eport0by0Geoff0Howse.pdf?sequence=1&isAllowed=y
https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941eport0by0Geoff0Howse.pdf?sequence=1&isAllowed=y


 
 

110 
 

CHAPTER 4  

 

Birth Order Effects and Early Childhood Health in Guyana 

 

4.1 Introduction 
 

Literature has shown that first-born children have lower birth weight and other gestational 

problems, however ,first-born children also tend to have higher cognitive and non-cognitive 

skills, acquire more education, and receive higher earnings than their later-born siblings.1 

Consequently, the negative health impacts of first-born children might be mitigated by the 

differences in parental behaviour (such as postnatal investments), in order to reverse the 

negative health outcomes at birth and in early childhood (Saldarriaga, 2015).  

 

In this study, we analyse the effects of birth order on early childhood health in Guyana. This is 

the first study to examine this question in the context of Guyana. We use a rich set of 

anthropometric data for children aged 0-59 months from the 2009 Guyana Demographic and 

Health Survey (GDHS) to examine whether first born children are at a disadvantage at birth 

when compared to their later-born siblings. Previous literature on child health and birth order 

effects reveals that the health for first born children is considerably disadvantaged and they 

also tend to have lower birth weight when compared to later born children (Björkegren and 

Svaleryd 2016; Brenøe and Molitor, 2017; Pruckner et al., 2019). This study is important for 

Guyana because the weight at birth, in particular, ‘low birth weight’ often leads to neonatal 

deaths and poor adult health which can negatively impact the human capital formation of the 

country. Consequently, according to UNICEF-WHO (2019), low birthweight new-borns have 

a higher risk of dying in the first 28 days of life. Moreover, studies have also shown that there 

is a negative correlation between birth weight and academic achievements measured using 

standardised test scores (Kohara and Ohtake, 2009).  

Consequently, we focus on several measures of early childhood health, i.e., birth weight, low 

birth weight, height for age z score (HAZ), stunting and child height indicators.2 Low-birth 

 
1 See Björkegren and Svaleryd (2016), Brenøe and Molitor (2017), Pruckner, Schneeweis, Schober and 

Zweimüller (2019) for evidence from developed countries. In contrast, De Haan, Plug and Rosero (2014), 

Saldarriaga (2015) provide evidence for birth order and birth weight in developing countries.  
2 According to literature, child height is predictive of adult height (Healy et al., 1956) and taller adults have greater 

cognitive skills (Glewwe and Miguel 2008; Guven and Lee 2015), fewer functional impairments (Barker and 

Osmond 1986; Barker et al. 1993; Gould 1989), and higher earnings (Strauss and Thomas 1998; Case and Paxson 

2008; Hoddinott et al. 2013). 



 
 

111 
 

weight is the leading indicator of poor health among infants, which may lead to negative health 

impacts, such as delay brain and somatic development that affects a wide range of subsequent 

health outcomes later in life (Nakamuro et al., 2013). A child’s height for age z score is an 

important measure since it captures lingering effects of growth disturbance that may have 

occurred early in life (Leroy et al., 2015). Stunted is defined as height for age z scores below -

2 standard deviation. We also examine the height of the child (in cm) at the time of the 2009 

GDHS. Moreover, we also examined parental postnatal investment to determine which group 

of children it favours more. 

 

Overall, we find that birth outcomes improve with birth order. Consequently, birth order is 

positively correlated with birth weight and negatively correlated with low birth weight for 

children born between 2005-2009 from the 2009 GDHS. These findings contribute to the 

broader body of literature such as Björkegren and Svaleryd 2017; Brenøe and Molitor, 2017; 

Pruckner et al., 2019 whom all found that health at birth improves with increasing birth orders. 

In addition, we find that compared to first born children, second and third born children tend 

to grow taller in Guyana. These results are analogous to Sahn and Stifel (2002) who found that 

compared to first born children, second and third-borns had higher-for-age z scores in fourteen 

African countries. Moreover, we find similar results when we examined the height of the 

children at the time of the survey, i.e., we find that when compared to first born children, height 

increases for the second and third-borns.  

 

We also examine postnatal intervention such as indicators for exclusive breastfeeding and 

completed vaccination for children between the ages of 0-59 months and found that mother’s 

both breastfed extensively and invested more in vaccinations. It is argued that extensive 

breastfeeding (i.e., 6 months or more) and postnatal investments can both offset the negative 

health impacts at birth for first-born children. Moreover, we see that given that parents tend to 

invest more in pentavalent and polio vaccinations, such actions may offset the negative health 

outcome at birth for first-borns and possibly leads to a reversal effect (Elliot, 1992; Saldarriaga, 

2015). When examining the heterogeneous effects, we find that the height for age z score 

increases more for second and third born male children when compared to first born children. 

We also find that compared to firstborn, third born and fourth born or higher Muslim children 



 
 

112 
 

tend to have higher birthweight.3  Moreover, the results are robust when clustering the standard 

errors at the 2009 GDHS Primary Sampling Units (PSUs) level. Consequently, we find that 

birth outcomes improve with birth order.  

 

The rest of the paper unfolds as follows. Section 4.2 present the literature review. Section 4.3 

describes the data sources and Section 4.4 introduces the empirical strategy. Section 4.5 

presents the results and Section 4.6 concludes.  

 

4.2 Literature Review 
 

 

There are several studies that focused on birth order effects on child health since health status 

during childhood is important for outcomes in later life. Childhood health is an important 

determinant for later life outcomes such as educational attainment, labour market outcomes 

and adult health (Case et al., 2005; Currie et al., 2010).  

Birth order refers to the order in which children are born into a family setting. Children are 

defined as firstborn, second born, third born and so on. Moreover, birth order categories have 

also been used to describe the order of birth, i.e., firstborn children, only or singleton children, 

middle children, early and later-born children, and last-born children (Elliot, 1992).  

There are several studies that posited that firstborn children generally do better academically 

than later-born children. Consequently, studies have shown that first-born children have higher 

IQs and higher educational attainment, and earnings and are more likely to live a healthier life 

than later-born children (Kantarevic and Mechoulan, 2006; Kristensen and Bjerkedal, 2007; 

Booth and Kee, 2008; Bu, 2014; Black et al.,2005,2011,2016).   Furthermore, a study has 

shown that firstborn children are more likely to have more cognitive ability (due to greater 

parental human capital investments) and acquire more prestigious jobs, such as managers, 

while later-born children are more likely to be self-employed (Black et al., 2017). Esposito et 

al., 2020 find that firstborns’ advantage in educational outcomes is amplified for males when 

their other siblings are females in Mexico. Isungest et al., 2022, studied birth order differences 

in education and found that it originates in postnatal environments because although early-born 

 
3 In Guyana during the month of Ramadan, women who are pregnant or nursing, the sick, or elderly people and 

children, are all exempted from fasting. See Guyana Times Newspaper, April 2020:  

https://guyanatimesgy.com/islams-holy-month-of-ramadan-commences-on-saturday-ciog/  

https://guyanatimesgy.com/islams-holy-month-of-ramadan-commences-on-saturday-ciog/


 
 

113 
 

children had lower birth weights, they had higher educational outcomes, which made them 

conclude that birth order differences arise postnatally.  

However, as it relates to health outcomes at birth, several studies have shown that first-born 

children have worst health outcomes than later-born children. Previous literature on child 

health and birth order effects reveals that the health of firstborn children is considerably 

disadvantaged, and they also tend to have lower birth weight when compared to later-born 

children (Elliot, 1992; Björkegren and Svaleryd 2017; Brenøe and Molitor, 2017; Pruckner et 

al., 2019). Brenøe and Molitor (2018) examine birth order differences in the health of newborns 

using data from Denmark (1981-2010) and find that firstborn children are less healthy at birth 

compared to later-born children. They posited that later-born children had better health due to 

a biological mechanism which improves nutrient supply to the foetus with each pregnancy 

(Khong et al., 2003; Gluckman and Hanson, 2004). Moreover, according to Black et al., (2016), 

there is higher prevalence of obesity in first- born adult. However, Wells et al., (2017) 

concluded that this might be due to lower nutrient supply to first-borns in-utero. 

Furthermore, Pruckner et al., 2019 found that later-born siblings exhibit greater health than 

firstborns. Consequently, childhood health increases with birth order. They also find 

differences in parental health investment by birth order. First-born children are more likely to 

receive greater postnatal care and immunisation against measles, mumps, and rubella. 

Moreover, when comparing first-born with later-born siblings, it was found that later-born 

children had higher birth weights and were less likely to be born with low birth weights (i.e., 

below 2.5 kg) and less likely to be born premature. It is argued that mothers are more likely to 

attend prenatal care in their first pregnancy and to breastfeed the first-born child (Buckles and 

Kolka, 2014; Black et al., 2016; Brenøe and Molitor, 2018; Lehmann et al., 2018). Mmopelwa 

(2019) studied the household size, birth order and child health in Botswana and found that 

later-born children were more likely to be nutritionally deprived compared to firstborns. 

Several studies have shown that birth order is correlated with height-for-age z scores, stunting 

and child height. As such, Sahn and Stifel (2002) use the Demographic and Health Survey from 

fourteen African countries and find that birth order was positively correlated with the children’s 

height-for-age z score, i.e., the height-for-age z scores were higher for second and third born 

children when compared to first-born children. In a study, De Keyser and Van Rossem (2017), 

also find that compared to first-borns, the height-for-age z score for second and third-born 

children increases. However, it declines systematically with higher birth order.  



 
 

114 
 

Black et al., (2011) argue that endowments at birth increase the birth order effects on outcomes 

later in adulthood. Myrskyla et al., (2013) use military enlistment data from Sweden to examine 

the relationship between birth order and height. They find the negative effects of birth order on 

height and argue that parental postnatal investments are important in helping young children 

(especially those under the age of 2) achieve their childhood health milestones, as height is 

often used as a proxy for health in early life. Rahman (2016), in his study on birth order and 

chronic malnutrition of children in Bangladesh, finds that stunted growth increases with higher 

birth order. Björkegren and Svaleryd (2017), examine birth order effects on child health in 

Sweden and find that firstborns have worsened health at birth until age six. However, their 

study reveals that such disadvantaged at birth are reversed in later life since higher-born 

siblings are more likely to be hospitalised for injuries, avoidable conditions, alcohol-related 

problems, or mental health issues. Jayachandran and Pande (2017) in a comparative study on 

Indian and Sub-Saharan African children, find that Indian children tend to be taller at birth. 

However, the differences decrease for children with higher birth order.  

In the literature, there is a common trend that is observed for developing countries, i.e., later-

born children tend to receive less maternal and health care in developing nations when 

compared with developed countries. On the contrary, De Haan et al. (2014), in an Ecuadorian 

study found mothers spent more time with their male children than with female children. 

Moreover, they find that firstborns are breastfed for less time when compared to later borns. 

Furthermore, De Hann et al. (2014) argue that disadvantages between firstborns and later-borns 

are likely to reverse in adulthood in children from rich and highly educated families due to 

parental investments.  

 

4.3 Data sources and descriptive statistics 
 

4.3.1 The 2009 Guyana Demographic and Health Survey (GDHS) 

 

This study analysed data from the 2009 Guyana Demographic and Health Survey (GDHS) 

provided by the DHS Programs. The 2009 GDHS is a cross-sectional, nationally- 

representative dataset which samples women of childbearing age that uses the 2002 Population 

and Housing Census as the master sample for the survey. The survey used a stratified two-stage 

cluster sample of 330 primary sampling units/clusters (equivalent to very small regions); that 



 
 

115 
 

sampled 5,632 households.  

Guyana is divided administratively into ten regions. As such, the regions were then divided 

into enumeration districts (EDs) and in the first stage units (primary sampling units of PSUs) 

are the enumeration districts utilised in the 2002 Population and Housing Census. In the second 

stage, 25 households were selected by systematic random sampling from the updated household 

listing of the selected EDs (2009 GDHS). A total of 4,996 face-to-face interviews were 

successfully completed for women aged 15-49. The survey gathers information on 

socioeconomic and demographic characteristics of households and their members. 

Consequently, mothers provided data on their entire birth history and outcomes (such as birth 

weight), anthropometric z scores (such as height -for-age z scores, weight-for-age z scores, 

etc), health service utilisation (for prenatal care and immunisation), and nutritional status of 

mothers and young children under 5 years old.  

We use the children’s data from the Kid Record (KR) file in the 2009 GDHS data files, since 

it relates the child’s outcomes to characteristic of the mother. Furthermore, the KR file includes 

all the children aged 0-59 months whose mothers were samples. Information for 2,178 children 

were recorded, however when excluding the only/singleton children, we have a total of 1,730 

observations. As such, our regressions were restricted to 1,328 children aged 0-59 months year 

old whose birth weight was recorded in the KR file. Moreover, to measure the child health 

indicators, we use the haz06 Stata package to compute the standardised child nutritional 

indicators through the World Health Organisation (WHO) software for growth and 

development of the world’s children (WHO, 2011). The software uses information on child 

sex, height, weight, and age in months and transforms it to height-for-age, weight-for-age and 

weight-for-height into z scores, as standardised measures of their deviation from the median of 

the reference population.4 This method takes the difference between a child’s observation and 

median reference value and divides it by the standard deviation of the reference population 

(i.e., children of the same age).  

Following standard practice, we use anthropometric measures for height and weight of children 

aged 0-59 months collected from the DHS Guyana 2009 dataset. Moreover, we compute z-

scores for height-for-age z-scores using the haz06 Stata package. Binary outcome variables are 

 
4The composition of the reference population changed over time, from the US National Centre for Health Statistics 

(NCHS) data to the current reference population, which is an outcome of a WHO Multi Growth Reference study 

(MGRS) of a pooled sample of six participating countries: Brazil, Ghana Indian, Norway, Oman, and the USA, 

which consist of around 8,500 children (WHO, 2006).  



 
 

116 
 

constructed for stunted based on the z-score for HAZ. As such children who are more than two 

standard deviations (SDs) below the HAZ median for their age group are classified as stunted. 

Low birth weight is computed based on the birth weight data, i.e., birth weight below 2.5 kg 

based on the WHO’s standard is deemed as “low birthweight.” 

The measure above is widely used to assess children’s health and nutrition status, as they 

capture different aspects of childhood undernutrition. For instance, stunted reflects the 

cumulative effect of undernutrition and possible infections from a child’s birth and even in 

utero. Moreover, it often indicates poor environmental conditions or other long-term 

restrictions or lingering health problem to a child’s physical development.  

In addition, we control for mother’s characteristics such as mother’s age at current birth and 

ethnicity. These data are all taken from the Kid Record (KR) file in the DHS Guyana 2009 

dataset. 

 

4.3.2 Summary Statistics  

Below, we present the descriptive statistics of the key variables used in our analysis in Table 

4.1. The average birth weight for children born between 2005-2009 is 3.3 kg with a minimum 

and maximum of 0.8 kg and 8 kg respectively. Of the 1328 children in the sample, 

approximately 12 percent had a birth weight below 2.5 kg and are thus underweight. Birth order 

1 accounts for 33 percent, followed by Birth order 4 or higher, Birth Order second, third with 

shares of 28, 22 and 17 percent respectively. The height-for-age z score (HAZ) for children 

ranges from -5.99 and 5.53 with a mean of -1.077, thus suggesting that the children in the 2009 

GDHS sample are 1 standard deviation below the median of the reference population. On 

average 25 percent of the children had stunted growth.5  

 

 

 

 

 
5 Stunted is defined as z scores below -2 standard deviation. 



 
 

117 
 

Table 4.1: Summary statistics 

Variables      Mean 

Standard      

deviation       Min        Max        N 

Demographic Characteristics      
Birth order 1 0.325 0.469 0 1 1730 

Birth order 2 0.239 0.427 0 1 1730 

Birth order 3 0.158 0.366 0 1 1730 

Birth order 4 or higher 0.277 0.447 0 1 1730 

Child is Male 0.499 0.500 0 1 1730 

Urban 0.169 0.375 0 1 1730 

Health at birth      
Birth weight (in kilograms) 3.275 0.842 0.8 8 1328 

Low birth weight (<2.5 kg) 0.117 0.321 0 1 1328 

Childhood Health      
Height for age z score* -1.077 1.544 -5.99 5.53 1328 

Stunted 0.244 0.430 0 1 1328 

Child height (cm)* 86.468 13.864 42 123.3 1328 

Mother's Characteristics      
Mother's age at first birth 19.009 3.722 11 38 1730 

Nursery or no education 0.039 0.194 0 1 1730 

Primary education 0.278 0.448 0 1 1730 

Secondary education 0.631 0.483 0 1 1730 

Higher education  0.051 0.221 0 1 1730 

Ethnicity      

African 0.172 0.378 0 1 1730 

Indian 0.225 0.418 0 1 1730 

Mixed 0.228 0.420 0 1 1730 

Amerindian 0.373 0.484 0 1 1730 

Portuguese 0.012 0.034 0 1 1730 

Religion      

Christian  0.827 0.378 0 1 1730 

Hindu  0.127 0.333 0 1 1730 

Muslim 0.033 0.179 0 1 1730 

Rastafarian 0.001 0.034 0 1 1730 

Not Religious 0.009 0.096 0 1 1730 

Other 0.001 0.339 0 1 1730 

Postnatal Investments      
Breastfeed (Months) 0.944 0.230 0 1 1211 

All vaccinations 0.506 0.500 0 1 1290 

No vaccination 0.054 0.227 0 1 1290 

BCG 0.932 0.252 0 1 1290 

Polio 1 0.739 0.439 0 1 1290 

Pentavalent 3 0.771 0.421 0 1 1290 

Measles 0.678 0.468 0 1 1290 
Note: * Height-for-age z score and child health reflect the height and weight for the child at the time of the survey 

interview. The postnatal investments are restricted to children whose birth weight were recorded.  

Literature has shown that maternal age and birth order of an infant are highly correlated. 

Consequently, as maternal age increases, the influence of birth order gets increasingly stronger 

and is associated with increased birth weight in infants (Selvin and Janerich, 1971; Jarry et al., 

2013; Brenøe and Molitor, 2018).  The average age of a first-time mother is 19 years old, 



 
 

118 
 

however the minimum and maximum age at first time mothers between 11 and 38 years old 

respectively. Approximately 50 percent of children are males. On average, 2.3 percent of the 

mothers had no education, 0.4 percent had primary education, 64 percent had secondary 

education and 6 percent had higher education. About 19 percent of the children resided in urban 

locations. Approximately 51 percent of the children in the survey received all their 

vaccinations. 93 percent of the children received their BCG vaccinations, 73 percent received 

Polio, 77 percent received Pentavalent, and 68 percent received Measles vaccination.6 

4.4 Empirical Strategy 
 

To measure the relationship between birth order and child health at birth, we employ methods 

used by Pruckner et al. (2019) and Brenøe and Molitor (2018) by estimating an ordinary least 

square (OLS) regression of the following type: 

 

𝑌𝑖𝑓𝑦𝑚 =   + 
𝑗

∑ 1(Birth order𝑖

4

𝑗=2

= 𝑗) + Boy𝑖 +  𝑡 +  
𝑚

+ 𝛿𝑓 + 𝑋𝑖+𝑖𝑓𝑦𝑚          (1) 

Where 𝑌𝑖𝑓𝑦𝑚  is health outcome of child i, born in family f, born in year y and month m. The 

health outcome variables are birth weight, dummy for low birth weight (birth weight < 2.5 kg), 

height for age z score, a dummy for being stunted (i.e., HAZ < -2 SD) and child health (in cm). 

The sum represents a set of birth order dummies, 1(Birth order𝑖 = 𝑗) for j = 2, 3, and 4 where 

1(.) is the indicator function. As such children of birth order, one represent the omitted category 

therefore 
𝑗
, the vector of the coefficient of interest, captures differences with respect to birth 

order one. Furthermore, we include  Boy𝑖 is a dummy=1 if the child is male and 0 otherwise, 

year of birth fixed effects (𝑡), month of birth fixed effects (
𝑚

), cluster fixed effects  (𝛿𝑓) and 

(𝑋𝑖) a vector of mother characteristics (i.e., mother’s education and mother’s age at first birth).  

Boy𝑖 is a dummy that denotes the child being male, and 𝑖𝑓𝑦𝑚  is the stochastic error term. We 

cluster standard errors at the mother-level.  

 

The estimation approach in equation (1) exploits only within-family variations in birth order 

and accounts for time-constant family-level confounding factors, such as maternal age at first 

birth. Conditionally on the mother fixed effects, controlling for child’s birth cohort implies 

 
6 Based on the 2009 GDHS, Pentavalent is a multi-dose/combination vaccination which comprises of DPT, it is 

given in weeks 8, 16 and 24 respectively.  



 
 

119 
 

controlling for mother’s age at birth, since the two variables are perfectly collinear. Therefore, 

our estimation model accounts for the maternal age at birth increasing with the child’s birth 

order.  

The estimation approach is based on comparing the health outcomes of children born within 

the same family, in the birth order categories of second, third and fourth born or higher using 

the firstborns as the reference category, across different years and months. We also include 

cluster fixed effects to control for common unobserved influences at the cluster level that affect 

all units within the cluster.   

Moreover, we performed robustness check by clustering the standard errors at the Primary 

Sampling Units (PSUs) or enumeration district (clusters level). Therefore, in these regressions, 

we did not include cluster fixed effects. These regressions also used svyset or the command for 

survey setting. 

 

4.5 Results and Discussions  
 

Our OLS regression results in Table 4.2 below shows that compared to firstborns (reference 

category) birth order is positively correlated with birth weight and negatively correlated with 

low birth weight. These results are also comparable to Björkegren and Svaleryd 2017; Brenøe 

and Molitor, 2018; Pruckner et al., 2019 who all found that health at birth improves with 

increasing birth orders. Column 1 demonstrates that health at birth improves with increasing 

birth order. Consequently, we see that birth weight increases by 0.14kg and 0.31kg for birth 

order two, three compared to firstborns for mothers within the same cluster, however it 

decreases to 0.25kg for fourth born or higher when compared to first born. Based on column 

2, we see that third born and fourth born children are less likely to born with low birth weight. 

Furthermore, the magnitudes of the increase in the coefficient for second-borns are comparable 

to Pruckner et al., (2019), who found that second-born siblings are 158 grams (5 percent) 

heavier than their firstborn counterparts in Austria.  



 
 

120 
 

Table 4.2: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) 
 (1) (2) (3) (4) (5) 

 Birth weight 
(kg) 

Low birth weight 
(<2.5 kg) 

Height-for-age 
z score 

Stunted 
 

Child height 
(cm) 

2nd born         0.146* -0.024 0.299** -0.002 1.043* 

 (0.078) (0.034) (0.150) (0.041) (0.558) 

3rd born 0.313*** -0.075** 0.318* -0.065 1.287** 

 (0.090) (0.038) (0.165) (0.045) (0.607) 

4th born or higher 0.247** -0.073** -0.060 0.052 0.053 

 (0.099) (0.037) (0.148) (0.045) (0.555) 

Male 0.135** -0.025 -0.092 0.003 0.881** 

 (0.053) (0.019) (0.097) (0.027) (0.333) 

Urban -0.347 0.022 0.294 0.050 -1.090 

 (0.307) (0.049) (0.562) (0.253) (1.540) 

Mother’s age at 1st birth 0.006 -0.001 0.002 0.002 0.034 

 (0.008) (0.003) (0.015) (0.097) (0.054) 

Primary education 0.166 -0.039 0.369 -0.074 1.626 

 (0.218) (0.083) (0.341) (0.093) (1.174) 

Secondary education 0.161 -0.084 0.552* -0.142 2.436** 

 (0.214) (0.079) (0.325) (0.108) (1.130) 

Higher education 0.246 -0.083 0.965** -0.000 3.646** 

 (0.247) (0.093) (0.398) (0.004) (1.402) 

Constant 2.983*** 0.119 -1.896** 0.162 106.246*** 

 (0.413) (0.115) (0.578) (0.218) (1.731) 

Observations 1328 1328 1328 1328 1328 

R2 0.350 0.294 0.347 0.329 0.904 
                                 Note: All regressions include month and year (2005-2009) of birth fixed effects, cluster fixed effects and a male dummy variable. Standard errors are clustered at the Mother 
                                         level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01. 



 
 

121 
 

Moreover, in column 3, we examine the height-for-age z scores which were taken at the time 

of the survey. Consequently, we see that second and third born children tend to grow taller 

when compared to firstborn children. As such, height-for-age z score of second and third borns 

are 0.299 SD and 0.318 SD higher than that of firstborns. These results are comparable to Sahn 

and Stifel (2002) and De Keyser and Van Rossem (2017), who find that compared to first born 

children, second and third-borns have higher height-for-age z scores. Moreover, we also find 

that in general the height-for-age z scores increases for children whose mother had completed 

their secondary or higher education. Column 4, in Table 4.2, we see that stunting declines from 

second and third born children compared to first born children, although the results are not 

significant.  

When examining the height of the children at the time of the survey, we find that when 

compared to first born children, height increases for the second and third-borns by 1.04 cm and 

1.29 cm respectively. We also find that child height increases for children whose mother had 

completed their secondary and higher education.  

It is important to note here that that the proportion of births that are multiple birth increases 

with age, up to 30-34, and then declines thereafter. 

 

4.5.1 Postnatal Investments 
 

A large body of literature reveals that there is a positive effect of breastfeeding on child health 

in developing countries (Betran et al., 2001, Sanauer and Kassouf 2000; Victora et al., 1987). 

Moreover, postnatal investments, such as breastfeeding, vaccination, and early stimulation, are 

important in determining children’s physical development and, subsequently, their 

socioeconomic outcomes later in life. Postnatal health investments are often viewed as a 

method of correcting the growth profiles of disadvantaged children, such as those born with 

low birth weights (Saldarriaga, 2015). Consequently, we use indicators of exclusive 

breastfeeding and completed vaccination for children between the ages of 0-59 months year 

old. According to Saldarriaga (2015), postnatal investments in child health can mitigate 

stunting caused by low birth weight in newborns. As such, we examine the effects of postnatal 

investments on childhood health given their birth order.   



 
 

122 
 

Table 4.3 reveals postnatal investments by the parents in the form of vaccinations and months 

of breastfeeding (see Table C.1 in the Appendix C for the recommended vaccination schedule 

in Guyana). Studies have shown that mothers are less likely to breastfeed later born children 

(Lehmann et al., 2018, Black et al., 2016; Buckles and Kolka, 2014). Contrary to the literature, 

in Table 4.3, column 1 we find that when compared to their firstborns’ mothers tend to spend 

more months breastfeeding their second, third and fourth born or higher. Consequently, we 

find that birth order gaps are reinforced, as later-borns do better at birth and are breastfed more.  

These finding are comparable to De Haan et al., (2014), who found in their Ecuadorian study 

that firstborns are breastfed for less time when compared with later borns.  



 
 

123 
 

 

Table 4.3: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) 
 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 

 Breastfed 

(Months) 

All 

vaccination 

(either) 

No 

vaccination 

(either) 

Polio 1st 

dose 

(either) 

Polio 2nd 

dose 

(either) 

Polio 3rd  

dose (either) 

Measles  

(either) 

Pentavalent 

1st dose 

(either) 

Pentavalent 

2nd dose 

(either) 

Pentavalent 

3rd dose 

(either) 

BCG 

(either) 

2nd born 0.097** 0.052 -0.015 0.017 0.008 0.024 0.055 0.030 0.029 0.064* 0.031 

 (0.034) (0.047) (0.021) (0.040) (0.040) (0.043) (0.038) (0.025) (0.028) (0.033) (0.025) 

3rd born 0.099** 0.033 0.005 0.034 0.022 0.018 0.046 0.027 0.032 0.042 0.005 

 (0.035) (0.053) (0.026) (0.045) (0.046) (0.050) (0.040) (0.030) (0.033) (0.039) (0.029) 

4th born or higher 0.102** 0.047 -0.030 0.066 0.045 0.054 0.067 0.060** 0.062* 0.063* 0.027 

 (0.034) (0.052) (0.026) (0.047) (0.048) (0.050) (0.041) (0.030) (0.033) (0.038) (0.030) 

Male -0.016 0.009 -0.003 -0.006 -0.001 -0.014 0.000 -0.001 0.008 -0.027 0.011 

 (0.015) (0.025) (0.014) (0.023) (0.023) (0.025) (0.021) (0.016) (0.017) (0.021) (0.016) 

Urban 0.014 -0.123 -0.339* 0.413* 0.392 0.548** -0.096 0.360** 0.344* 0.510** 0.334* 

 (0.037) (0.183) (0.177) (0.231) (0.244) (0.239) (0.184) (0.172) (0.186) (0.201) (0.178) 

Mother’s age 1st first 0.003 0.004 0.003 0.001 0.006 0.007 -0.002 -0.005 -0.002 0.002 -0.003 

 (0.003) (0.004) (0.003) (0.004) (0.004) (0.005) (0.003) (0.003) (0.003) (0.004) (0.003) 

Primary education -0.083** 0.071 -0.043 0.064 -0.001 0.088 0.030 0.010 -0.046 0.012 0.072 

 (0.042) (0.112) (0.043) (0.091) (0.099) (0.103) (0.085) (0.055) (0.070) (0.081) (0.053) 

Secondary education -0.064 0.054 -0.022 0.003 -0.043 0.035 0.043 -0.017 -0.037 -0.009 0.054 

 (0.043) (0.111) (0.040) (0.090) (0.098) (0.102) (0.083) (0.052) (0.067) (0.077) (0.050) 

Higher education -0.113* 0.035 -0.016 -0.004 -0.055 0.031 0.020 0.008 -0.047 -0.028 0.028 

 (0.058) (0.125) (0.046) (0.108) (0.116) (0.121) (0.095) (0.060) (0.076) (0.088) (0.057) 

Constant 0.943*** 0.543** 0.265 0.643** 0.590** 0.503* 0.828*** 0.839*** 0.788*** 0.751** 0.709*** 

 (0.087) (0.234) (0.198) (0.265) (0.278) (0.275) (0.219) (0.197) (0.213) (0.231) (0.203) 

Observations 1211 1290 1290 1290 1290 1290 1290 1290 1290 1290 1290 

R2 0.311 0.542 0.350 0.494 0.532 0.520 0.647 0.474 0.573 0.575 0.340 
Note: All regressions include month and year (2005-2009) of birth fixed effects, cluster fixed effects and a male dummy variable. The results show the three doses of vaccination for Polio and Pentavalent. 

Pentavalent is a multi-dose/combination vaccination which comprises of DPT given in weeks 8, 16 and 24 respectively. Polio vaccination is given in weeks 8, 16 and 24 respectively. Standard errors are 
clustered at the PSUs level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01. 



 
 

124 
 

According to the 2009 GDHS, the coverage for Polio, Pentavalent and DPT declined with 

subsequent doses, thereby leading to higher dropout rates for the full course of the vaccines 

(GDHS 2009). The vaccinations were reported based on both the mother’s records, as well as 

what was recorded on the children’s immunisation cards. Columns 10 in Table 4.3 above shows 

that compared to their firstborn children, mothers tend to vaccinate their second and fourth-

born or higher more against pentavalent.  

The results also provide evidence on the relationship between health investments during 

pregnancy and early childhood. We see that postnatal investments may have mitigated the 

incident of stunting caused by low birthweight in new-borns, since we do not find any impact 

of stunting (see Table 4.2, column 4) on childhood health as parents tend to employ postnatal 

interventions such as breastfeeding and immunisation. 

 

4.5.2 Robustness Check  
 

For robustness check we used an alternative specification of clustering using the primary 

sampling units (PSUs) levels. Moreover, these results do not include cluster fixed effects.  In 

2000, the Guyana Bureau of Statistics (GBOS), in collaboration with the U.S. Census Bureau, 

designed a sampling frame from the census master sample. In the same year, GBOS updated 

the geographical location and household listing each primary sampling units or (PSUs) 

included in the master sample; this work was supported in part by USAID. Consequently, we 

present Table 4.4 by clustering at the PSUs (level 3) or using 330 cluster ids.  Column 1, in 

Table 4.4 show that (based on the point estimates) birthweight increases for second, third and 

fourth born or higher by 0.18kg, 0.24kg and 0.26kg respectively, when compared to firstborn 

children. These results are similar to our main results in Table 4.2. However, based on column 

2, third born are less likely to born with low birthweight when compared to first-borns.  

In column 3, we find that when compared to firstborn children, second and third born children 

tend to grow taller. This is reflected in the height for age z score being 0.48 SD and 0.58 SD 

higher than that of firstborns. Analogous to our main results, when examining the height of 

children at the time of the survey, we find that when compared to first born children height (in 

cm) increases for second and third-borns by 1.773 cm and 2.223 cm respectively when 

compared with first borns.  



 
 

125 
 

Table 4.4: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at PSUs level (OLS Regressions) 
 (1) (2) (3) (4) (5) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height-for-age 

z score 

Stunted Child height 

(cm) 

2nd born 0.184** 0.001 0.478** -0.066 1.773** 

 (0.074) (0.035) (0.187) (0.061) (0.789) 
3rd born 0.241** -0.062* 0.576** -0.126* 2.223** 

 (0.079) (0.038) (0.228) (0.073) (0.958) 

4th born or higher 0.260** -0.035 0.176 0.007 1.257 

 (0.093) (0.036) (0.209) (0.067) (0.905) 
Male 0.103** -0.014 -0.109 0.008 0.663* 

 (0.051) (0.020) (0.100) (0.027) (0.369) 

Urban -0.074 0.023 0.229* -0.042 0.116 
 (0.068) (0.025) (0.120) (0.029) (0.447) 

Mother’s age at 1st birth 0.013 -0.002 0.034** -0.008** 0.138** 

 (0.008) (0.003) (0.013) (0.004) (0.048) 
Primary education 0.120 -0.009 -0.008 0.049 0.398 

 (0.231) (0.078) (0.283) (0.084) (0.958) 

Secondary education 0.107 -0.039 0.372 -0.079 1.540* 

 (0.219) (0.078) (0.251) (0.073) (0.856) 
Higher education 0.099 -0.022 0.487 -0.123 2.318** 

 (0.229) (0.086) (0.325) (0.088) (1.136) 

Constant   2.878*** 0.222** -2.181*** 0.429** 104.144*** 
 (0.316) (0.110) (0.454) (0.133) (1.704) 

Observations 1328 1328 1328 1328 1328 

R2 0.028 0.021 0.091 0.090 0.880 
                       Note: All regressions include month and year (2005-2009) of birth fixed effects, and a male dummy variable. Standard errors are clustered at the PSUs level in the      
                       Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01. 

 

  



 
 

126 
 

 

Table 4.5: DHS Guyana 2009 Birth Order Effects on Child Health (Postnatal Investments) - Clustered at PSUs level (OLS Regressions) 
 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 

 Breastfed 

(Months) 

All 

vaccination 

(either) 

No 

vaccination 

(either) 

Polio 1st 

dose 

(either) 

Polio 2nd  

dose 

(either) 

Polio 3rd 

dose 

(either) 

Measles  

(either) 

Pentavalent 

1st dose 

(either) 

Pentavalent 

2nd dose 

(either) 

Pentavalent 

3rd dose 

(either) 

BCG  

(either) 

2nd born 0.076* 0.077 -0.019 0.099** 0.089* 0.088* 0.066 0.067** 0.073** 0.100** 0.026 

 (0.040) (0.051) (0.025) (0.045) (0.046) (0.051) (0.045) (0.033) (0.034) (0.037) (0.026) 

3rd born 0.106** 0.085 -0.016 0.086* 0.087* 0.063 0.096** 0.061* 0.074** 0.077* 0.024 

 (0.043) (0.053) (0.033) (0.049) (0.051) (0.053) (0.041) (0.032) (0.035) (0.041) (0.035) 

4th born or higher 0.099** 0.077 -0.023 0.121** 0.091* 0.064 0.091** 0.066** 0.047 0.029 0.020 

 (0.047) (0.050) (0.028) (0.050) (0.050) (0.050) (0.039) (0.030) (0.031) (0.037) (0.029) 

Male -0.013 -0.007 0.002 0.010 0.018 -0.045* -0.004 -0.007 0.004 -0.052** -0.005 

 (0.017) (0.026) (0.015) (0.026) (0.027) (0.025) (0.020) (0.018) (0.020) (0.020) (0.017) 

Urban 0.020 -0.093* -0.034* -0.118* -0.132* -0.120* -0.007 0.030 0.004 -0.009 0.034* 

 (0.018) (0.056) (0.018) (0.068) (0.068) (0.064) (0.029) (0.020) (0.023) (0.032) (0.020) 

Mother’s age 1st birth 0.002 0.009** 0.000 0.006* 0.009** 0.011** 0.003 0.000 0.002 0.005* -0.000 

 (0.003) (0.003) (0.002) (0.004) (0.004) (0.004) (0.002) (0.002) (0.002) (0.003) (0.002) 

Primary education -0.055** 0.217* -0.114 0.166 0.125 0.237* 0.091 0.122 0.074 0.152 0.169 

 (0.025) (0.123) (0.105) (0.162) (0.173) (0.128) (0.116) (0.121) (0.135) (0.107) (0.112) 

Secondary education -0.063*** 0.188 -0.096 0.086 0.057 0.194 0.128 0.082 0.074 0.171 0.146 

 (0.016) (0.120) (0.113) (0.154) (0.157) (0.128) (0.121) (0.117) (0.120) (0.115) (0.117) 

Higher education -0.093* 0.118 -0.088 0.033 0.038 0.144 0.077 0.089 0.062 0.177 0.144 

 (0.048) (0.135) (0.117) (0.172) (0.174) (0.146) (0.129) (0.122) (0.128) (0.119) (0.121) 

Constant 0.845*** 0.513*** 0.120 0.595** 0.555** 0.461** 0.883*** 0.894*** 0.824*** 0.726*** 0.821*** 

 (0.088) (0.153) (0.109) (0.189) (0.185) (0.168) (0.155) (0.148) (0.145) (0.151) (0.120) 

Observations 1211 1290 1290 1290 1290 1290 1290 1290 1290 1290 1290 

R2 0.038 0.307 0.110 0.142 0.184 0.217 0.523 0.353 0.415 0.407 0.095 
Note: All regressions include month and year (2005-2009) of birth fixed effects, and a male dummy variable. The results show the three doses of vaccination for Polio and Pentavalent. Pentavalent is a 
multi-dose/combination vaccination which comprises of DPT given in weeks 8, 16 and 24 respectively. Polio vaccination is given in weeks 8, 16 and 24 respectively.  Standard errors are clustered at 

the PSUs level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01. 

 
 

 



 
 

127 
 

When looking at postnatal investments, we find that overall parents tend to invest more in 

vaccinations when clustering the standard errors at the Primary Sampling Units (PSUs) or 

enumeration district (clusters level).  compared to our main equation (1). In column 1 of Table 

4.5 above, we find that when compared to firstborn children, mothers tend to spend more 

months breastfeeding their second, third and fourth or higher. As it relates to investments in 

inoculations, we find that on average mothers tend to invest in Polio vaccinations for the second 

born, third born and fourth born or higher children when compared to their firstborns. 

Compared to first-born children mothers tend to invest more in Pentavalent vaccines for their 

second, third and fourth born or higher. In a similar manner, mothers tend to invest more in 

Measles vaccines for their third and fourth born or higher.  

 

 4.5.3 Heterogeneous effects  
 

So far, we have shown results that suggest that first-born children are at a disadvantage both at 

birth and in early childhood than their younger siblings. However, there are other factors that 

can influence the relationship between birth order and childhood health. Differences in birth 

order effects also can vary by ethnicity and gender etc. Consequently, we examine the 

heterogenous effects with respect to ethnicity, religion, maternal education, and gender on the 

effect of birth order and childhood health. In addition, we have also examined the heterogenous 

effect by the mother’s age at each birth (or current age of the mother) and the mother’s age at 

first birth.  

We have included heterogeneous effects by religion and ethnicity because Guyana is an 

ethnically diverse country with a strong religious background. According to the 2002 Guyana 

census, 57 percent of the Guyanese population are Christians, 28 percent are Hindus, 7 percent 

are Muslims, and the remainder are non-religious (Guyana Bureau of Statistics, 2023). These 

variables are relevant to birth order effects because religion and ethnicity can influence birth 

order. For instance, Buber-Ennser and Berghammer (2021), in their study revealed that 

Christians who regularly attend church services generally have more children than non-

affiliated persons in Europe. Religious belief can also influence how parents raise their 

firstborns. For instance, there are specific rituals or expectations for the firstborn child that may 

affect investment decisions. Furthermore, there can be a son-bias in some religious culture 

which can cause investment decisions to be skewed towards boys and impact birth order 



 
 

128 
 

(Barcellos, 2014; Jayachandran and Pande, 2017). Moreover, ethnicity can have influences on 

birth order and vary across different ethnic groups (Su et al., 2014; Marjoribanks, 1999).  As 

such, based on ethnicity and religion children are treated differently based on birth order.  

Guyana is an ethnically diverse nation with persons who are descendants from their African, 

Indian, Amerindian/Indigenous, Chinese, Portuguese, and other European ancestors (mainly 

French, Dutch, and English). There is also a subgroup of persons who are of Mixed ethnic 

backgrounds via a combination of any of the above-mentioned groups. 

According to the 2002 Guyana Population and Housing Census, the Indo-Guyanese or Indians 

were the largest ethnic group and accounted for 43.4 percent of the total population. The Afro-

Guyanese or African were the second largest group and accounted for 30.2 percent of the 

population while the Mixed race accounted for 16.7 percentage of the population.  For this 

research, we focused on these three main ethnic subgroups, i.e., Africans, Indian and Mixed, 

as the observations for these groups. Therefore, our omitted categories are the Amerindian and 

Portuguese subgroups. 



 
 

129 
 

Table 4.6: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) – Heterogeneous Effects (ETHNICITY) 
 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for age 

z score 

Child height 

(cm) 

2nd born 0.267** -0.062 0.546** 2.497** 

 (0.126) (0.059) (0.275) (0.940) 

3rd born 0.301* -0.059 0.455* 2.567** 

 (0.163) (0.071) (0.264) (0.972) 

4th born or higher 0.284* -0.103* 0.067 1.081 

 (0.154) (0.059) (0.235) (0.883) 

Male 0.135** -0.026 -0.111 0.808** 

 (0.053) (0.019) (0.097) (0.332) 

Urban -0.413 0.001 0.462 -0.447 

 (0.341) (0.069) (0.550) (1.571) 

African 0.085 0.067 0.514 3.107** 

 (0.238) (0.098) (0.406) (1.467) 

Indian -0.452** 0.122 0.578 3.131** 

 (0.183) (0.089) (0.361) (1.356) 

Mixed 0.247 -0.067 0.073 1.163 

 (0.201) (0.066) (0.323) (1.198) 

2nd born x African -0.140 0.048 -0.532 -3.048** 

 (0.241) (0.103) (0.433) (1.543) 

2nd born x Indian -0.150 0.042 -0.585 -2.855* 

 (0.170) (0.093) (0.400) (1.470) 

2nd born x Mixed -0.232 0.075 -0.038 -0.961 

 (0.220) (0.075) (0.404) (1.375) 

3rd born x African 0.034 -0.042 -0.504 -3.125* 

 (0.274) (0.116) (0.468) (1.706) 

3rd born x Indian 0.254 -0.114 -0.188 -2.275 

 (0.212) (0.104) (0.395) (1.488) 

3rd born x Mixed -0.305 0.112 0.026 -0.702 

 (0.253) (0.091) (0.422) (1.535) 

4th born or higher x African -0.197 0.026 -0.513 -2.778* 

 (0.300) (0.113) (0.429) (1.553) 

4th born or higher x Indian 0.248 -0.059 -0.553 -3.171** 

 (0.218) (0.097) (0.384) (1.450) 

4th born or higher x Mixed -0.267 0.172** 0.329 0.486 

 (0.233) (0.080) (0.372) (1.327) 

Mother’s age at 1st first 0.011 -0.003 0.001 0.031 

 (0.009) (0.003) (0.015) (0.056) 

Constant 2.809*** 0.164 -2.267*** 104.163*** 

 (0.432) (0.124) (0.588) (1.951) 

Observations 1328 1328 1328 1328 

R2 0.369 0.307 0.353 0.905 

 Note: All regressions include month and year (2005-2009) of birth fixed effects and clustered fixed effects, and a male dummy variable. We also control for the mother’s education (primary, secondary, and higher 

education) with the omitted category being no and nursery education. Standard errors are clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 
0.05, *** p < 0.01. 



 
 

130 
 

Table 4.7: DHS Guyana 2009 Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) – Heterogeneous Effects (RELIGION) 
 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for age  

z score 

Child height 

(cm) 

2nd born                 0.197 -0.115 1.049**   7.004*** 

 (0.366) (0.143) (0.459) (1.784) 

3rd born -0.502 -0.049 -0.793 -1.643 

 (0.332) (0.174) (1.427) (5.548) 

4th born or higher -0.568* 0.093 0.471 3.575 

 (0.326) (0.138) (0.744) (2.385) 

Male 0.125** -0.023 -0.097 0.877** 

 (0.054) (0.019) (0.099) (0.338) 

Urban -0.308 0.008 0.263 -1.198 

 (0.314) (0.050) (0.569) (1.544) 

Christian -0.253 0.080 0.803 4.063** 

 (0.218) (0.067) (0.667) (1.968) 

Hindu -0.645** 0.214** 0.922 5.182** 

 (0.281) (0.106) (0.699) (2.099) 

Muslim -0.633** 0.202 1.559 6.107 

 (0.285) (0.159) (1.005) (3.919) 

2nd born x Christian -0.062 0.114 -0.696 -5.806** 

 (0.373) (0.144) (0.497) (1.912) 

2nd born x Hindu 0.021 -0.012 -0.984* -6.865*** 

 (0.387) (0.166) (0.518) (1.974) 

2nd born x Muslim 0.016 0.103 -1.360 -7.434* 

 (0.437) (0.203) (1.006) (4.252) 

3rd born x Christian 0.738** 0.018 1.063 2.933 

 (0.341) (0.176) (1.441) (5.583) 

3rd born x Hindu 1.105** -0.195 1.381 2.915 

 (0.374) (0.197) (1.468) (5.701) 

3rd born x Muslim 1.388** -0.260 0.913 2.600 

 (0.427) (0.238) (1.621) (6.519) 

4th born or higher x Christian 0.759** -0.132 -0.540 -3.478 

 (0.349) (0.142) (0.774) (2.499) 

4th born or higher x Hindu 1.067** -0.298* -0.429 -3.767 

 (0.395) (0.171) (0.812) (2.632) 

4th born or higher x Muslim  1.733*** -0.501** -1.052 -5.126 

 (0.495) (0.223) (1.275) (4.752) 

Mother’s age at 1st birth 0.005 -0.001 0.004 0.041 

 (0.009) (0.003) (0.015) (0.054) 

Constant 3.277*** 0.004 -2.717** 102.046*** 

 (0.466) (0.134) (0.885) (2.639) 

Observations  1328 1328 1328 1328 

R2 0.363 0.305 0.352 0.905 

Note: All regressions include month and year (2005-2009) of birth fixed effects and clustered fixed effects, and a male dummy variable. We also control for the mother’s education (primary, secondary, and higher 

education) with the omitted category being no and nursery education. Standard errors are clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 
0.05, *** p < 0.01.



 
 

131 
 

 

Table 4.6 above presents the heterogenous effect of birth order and ethnicity. We did not find 

any heterogeneity as it relates to birth weight, although we do find 3rd and 4th born or higher, 

however we find that in general Indian children tends to have lower birth weights when 

compared to other subgroups (see column 1). Moreover, we find that the low birthweight 

increases for fourth born or higher Mixed-race children. In general, we find that African and 

Indian children tend to grow taller when compared to Mixed race. However, when looking at 

the heterogenous effect of child height (in Column 4), we find that second and fourth born or 

higher African and Indian children tend to grow taller. 

Table 4.7 above displays the heterogeneous effects with respect to religion. In Guyana, 

Christianity is the dominant religion. According to the 2002 Guyana Population and Housing 

Census, approximately 57 percent of the population were Christians, 28 percent were Hindus, 

7 percent Muslims, 4 percent Rastafarian and other and 4 percent of the population did not 

conform to any religious belief. In Table 4.7, we restrict our regressions primarily to the three 

main religions (i.e., Christian, Hindu, and Muslim). We find that when compared to firstborns 

(in column 1), third and fourth born or higher children from all religious groups had higher 

birthweight. Consequently, fourth born or higher children born to Christians, Hindu and 

Muslims weighed 0.76kg, 1.06kg and 1.73kg more respectively when compared to firstborns. 

We also find that for fourth born or higher low birth weight reduces (see column 2) for Hindu 

and Muslim families when compared to firstborns. When examining the height-for-age z score 

(column 3), we find that second born children of the Hindu religion tend to be shorter. When 

examining the child height at the time of the survey, we see that 2nd born Muslim children tends 

to be shorter across all the religious groups when compared to first-borns. 

Table 4.8, (column 1) below shows that mothers who had secondary and higher education had 

children with higher birth weight when compared to those mothers with only nursery and 

primary education. However, when looking at the heterogenous effects, we see that birthweight 

for third and fourth born or higher children reduces for mothers with secondary and higher 

education. Table 4.9 displays the gender heterogeneity, and we find that the height-for-age z 

score increases more for second and third born male children when compared to first born 

children. 



 
 

132 
 

Table 4.8: Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) – Heterogeneous Effects (EDUCATION) 
 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for age  

z score 

Child height 

(cm) 

2nd born              0.946 0.261 1.364*  5.396** 
 (0.665) (0.240) (0.818) (2.203) 

3rd born 2.690*** -0.002 1.208 4.502 

 (0.798) (0.128) (1.133) (4.990) 

4th born or higher 2.579*** -0.061 0.348 1.801 

 (0.744) (0.148) (0.612) (1.849) 

Male 0.149** -0.027 -0.090 0.871** 

 (0.053) (0.019) (0.099) (0.339) 

Urban -0.347 0.026 0.317 -1.014 

 (0.312) (0.048) (0.524) (1.463) 

Nursery or Primary education 2.183** 0.013 0.413 2.006 

 (0.722) (0.132) (0.627) (1.964) 

Secondary or Higher education 2.227** -0.033 1.053* 4.586** 

 (0.714) (0.120) (0.582) (1.714) 

2nd born x Nursery or Primary educ -0.736 -0.243 -0.642 -2.626 

 (0.683) (0.251) (0.894) (2.572) 

2nd born x Secondary or Higher educ -0.822 -0.299 -1.246 -5.077** 

 (0.674) (0.243) (0.824) (2.245) 

3rd born x Nursery or Primary educ -2.323** -0.074 -0.426 -1.815 

 (0.816) (0.153) (1.177) (5.143) 
3rd born x Secondary or Higher educ -2.419** -0.075 -1.081 -3.796 

 (0.802) (0.136) (1.156) (5.052) 

4th born or higher x Nursery or Primary educ -2.349** -0.033 -0.185 -0.724 

 (0.761) (0.160) (0.682) (2.209) 

4th born or higher x Secondary or Higher educ -2.356** 0.005 -0.484 -2.141 

 (0.747) (0.154) (0.632) (1.937) 

Mother’s age at 1st birth 0.006 -0.001 0.005 0.043 

 (0.008) (0.003) (0.015) (0.055) 

Constant 0.954 0.071 -2.230** 104.725*** 

 (0.795) (0.138) (0.782) (2.215) 

Observations 1328 1328 1328 1328 

R2 0.360 0.297 0.348 0.904 
Note: All regressions include month and year (2005-2009) of birth fixed effects and clustered fixed effects, and a male dummy variable. The omitted category is no education. Standard errors are 
clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01.



 
 

133 
 

Table 4.9: Birth Order Effects on Child Health - Clustered at Mother’s level (OLS Regressions) – Heterogeneous Effects (GENDER) 
 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for age 

z score 

Child height 

(cm) 

2nd born 0.112 -0.053 0.017 -0.148 

 (0.115) (0.050) (0.199) (0.752) 

3rd born 0.272** -0.127** 0.037 0.081 

 (0.124) (0.053) (0.217) (0.814) 

4th born or higher 0.182 -0.056 -0.268 -0.686 

 (0.135) (0.051) (0.197) (0.714) 

Male 0.046 -0.051 -0.575** -1.044 

 (0.136) (0.059) (0.258) (1.001) 

Urban -0.337 -0.014 0.283 -1.166 

 (0.313) (0.053) (0.580) (1.668) 
2nd born x Male 0.072 0.059 0.619* 2.602** 

 (0.159) (0.077) (0.318) (1.218) 

3rd born x Male 0.089 0.097 0.612* 2.596** 

 (0.171) (0.073) (0.317) (1.230) 

4th born or higher x Male 0.133 -0.031 0.478 1.735 

 (0.161) (0.066) (0.310) (1.136) 

Primary education 0.158 -0.026 0.382 1.698 

 (0.219) (0.082) (0.340) (1.168) 

Secondary education  0.153 -0.073 0.565* 2.505** 

 (0.216) (0.079) (0.326) (1.126) 

Higher education 0.240 -0.075 0.972** 3.679** 
 (0.250) (0.093) (0.400) (1.407) 

Mother’s age 1st first 0.006 -0.001 0.002 0.035 

 (0.009) (0.003) (0.015) (0.054) 

Constant 3.029*** 0.112 -1.685** 107.047*** 

 (0.421) (0.117) (0.582) (1.732) 

Observations 1328 1328 1328 1328 

R2 0.351 0.298 0.350 0.904 
Note: All regressions include month and year (2005-2009) of birth fixed effects and clustered fixed effects, and a male dummy variable. The omitted category is no education. Standard errors are 
clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p < 0.05, *** p < 0.01.



 
 

134 
 

Furthermore, we have done heterogenous effect by the mother’s age at each birth (see Table 

4.10 below), i.e.., the current age of the mother at each birth. However, we did not find any 

birth order effects when looking at these sub-groups.  

Table 4.10 DHS Guyana 2009 Birth Order Effects on Child Health – Heterogeneous effects 

(Mother’s age at each birth subgroups). 

 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for 

age z score 

Child height 

(cm) 

2nd born -0.676 0.676 -5.976 -30.762 

 (2.779) (1.667) (6.733) (25.110) 

3rd born -0.126 1.028 -4.197 -23.196 

 (2.535) (1.340) (6.069) (21.226) 

4th born or higher -0.763 0.473 -3.975 -22.570 

 (2.404) (1.159) (6.021) (20.576) 

Male 0.169** -0.022 0.007 0.838 

 (0.078) (0.035) (0.178) (0.581) 

Urban 0.490 -0.019 0.857 -3.157 

 (0.447) (0.094) (1.675) (2.686) 

Mother’s age at 2nd birth 0.037 -0.021 -0.038 -0.113 

 (0.033) (0.014) (0.066) (0.215) 

Mother’s age at 3rd birth -0.006 0.010 -0.080 -0.330 

 (0.042) (0.019) (0.081) (0.269) 

Mother’s age at 4th birth -0.001 0.013 -0.032 -0.324 

 (0.110) (0.053) (0.271) (0.914) 

2nd born x Mother’s age at 2nd birth -0.028 -0.014 0.220 1.150 

 (0.139) (0.080) (0.345) (1.258) 

3rd born x Mother’s age at 3rd birth -0.055 -0.025 0.118 0.691 

 (0.122) (0.063) (0.297) (1.011) 

4th born or higher x Mother’s age at 4th birth -0.021 -0.006 0.098 0.602 

 (0.108) (0.052) (0.273) (0.916) 

Primary education 0.072 -0.074 0.515 2.108 

 (0.176) (0.110) (0.454) (1.622) 

Secondary education 0.080 -0.090 0.955** 3.664** 

 (0.193) (0.109) (0.457) (1.625) 

Higher education 0.053 -0.070 2.223** 7.474** 

 (0.281) (0.189) (0.696) (2.671) 

Constant 3.595 -0.225 3.236 131.897*** 

 (2.489) (1.202) (6.196) (21.237) 

Observations 542 542 602 602 

R2 0.603 0.490 0.435 0.918 

Note: All regressions include month and year (2005-2009) of birth fixed effects and a male dummy variable. Standard errors 

are clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p 

< 0.05, *** p < 0.01. 

 

 

 



 
 

135 
 

Furthermore, when looking at the heterogenous effect by mother age at first birth (see Table 

4.11, column 1), we find that birthweight increases for the interaction for 2 and 3rd born x age 

at first birth, however the results were not statistically significant. However, we find that the 

coefficient for 4th born x age at first birth was statistically significant at the 10 percent level, 

i.e., we find that birth weight increases. Furthermore, we find that height and height-for-age z-

score decreases and are significant at the 5 percent and 10 percent level respectively. 

Table 4.11 DHS Guyana 2009 Birth Order Effects on Child Health – Heterogeneous effects 

(Mother’s age at first birth). 

 (1) (2) (3) (4) 

 Birth weight 

(kg) 

Low birth weight 

(<2.5 kg) 

Height for age  

z score 

Child height 

(cm) 

2nd born -0.027 -0.226* 1.576** 5.997** 

 (0.365) (0.136) (0.770) (2.992) 

3rd born -0.151 -0.056 1.253 5.913* 

 (0.416) (0.159) (0.878) (3.462) 

4th born or higher -0.645 -0.189 1.582* 7.055** 

 (0.483) (0.159) (0.887) (3.444) 

Male 0.129** -0.023 -0.087 0.904** 

 (0.053) (0.020) (0.097) (0.331) 

Urban -0.335 0.020 0.275 -1.175 

 (0.305) (0.049) (0.567) (1.545) 

Mother’s age at 1st birth -0.014 -0.007 0.062* 0.287* 

 (0.016) (0.006) (0.036) (0.150) 

2nd born x Mother’s age at 1st birth 0.009 0.010 -0.065* -0.253* 

 (0.017) (0.007) (0.038) (0.150) 

3rd born x Mother’s age at 1st birth 0.024 -0.001 -0.047 -0.236 

 (0.020) (0.008) (0.044) (0.176) 

4th born x Mother’s age at 1st birth 0.047* 0.006 -0.086* -0.367** 

 (0.025) (0.008) (0.046) (0.178) 

Primary education 0.130 -0.045 0.415 1.819 

 (0.217) (0.085) (0.346) (1.194) 

Secondary education 0.114 -0.092 0.610* 2.688** 

 (0.213) (0.082) (0.333) (1.156) 

Higher education 0.212 -0.089 0.992** 3.774** 

 (0.243) (0.095) (0.401) (1.411) 

Constant 3.440*** 0.232 -3.114*** 101.081*** 

 (0.486) (0.162) (0.864) (3.177) 

Observations 1328 1328 1328 1328 

R2 0.354 0.296 0.350 0.904 

Note: All regressions include month and year (2005-2009) of birth fixed effects and a male dummy variable. Standard errors 

are clustered at the Mother level in the Guyana Demographic and Health Survey (GDHS), noted in parentheses* p < 0.10, ** p 

< 0.05, *** p < 0.01. 

 
 



 
 

136 
 

4.5.4 Policy Implications 
 

A child's birth order cannot be predetermined, but knowledge of its impact can be used to 

formulate policies to alleviate the pressures that parents feel with their newborns and to 

promote the development of later-born children. 

Policy implications of birth order are not very straightforward because it is not in the hands of 

policymakers, thus making it difficult to alter. At conception, each sibling inherits the same 

genes from both parents, but the genes expressed by each individual after birth vary depending 

on prenatal and postnatal factors (Björkegren and Svaleryd, 2016). Essentially, this means that 

differences by birth order are likely to be dependent on pre-and postnatal influences rather than 

pre-determined conditions, which opens discussions for policy inventions. According to 

Björkegren and Svaleryd (2016), policy implications depend on findings. 

Our findings reveal that mothers are more likely to invest more months in breastfeeding their 

later-born children, and these children are more likely to receive important vaccinations, such 

as Polio, Measles and Pentavalent. Consequently, alone time with parents is crucial to newborn 

infants, thus indicating that day care for older children would be important so that parents can 

better tend to their newborns while on parental leave. One possible policy implication would 

be free childcare for working parents with 2-3 children aged five and under. This would help 

parents to boost the development of later-born children. Currently, Guyana’s Childcare and 

Development Service Act Number 12 of 2011 (Part 1b), states that pre-school, day care, home 

care, playgroup etc, must be provided for a fee or reward (Part 1b (i)).71 As such, there is no 

free provision to assist working parents with childcare services. Furthermore, given the 

statistics taken from the 2009 GDHS vaccine coverage rate for Polio, Pentavalent and DPT, we 

see that the dropout rates for the full vaccine course are extremely high. Another policy 

implication would be the creation of programmes by the Ministry of Health Guyana to better 

educate first-time mothers on the impact of low birth weight on firstborn children and the 

importance of postnatal inventions such as exclusive breastfeeding and being fully vaccinated 

to mitigate such negative impacts.    

 

 
7 See: GUY99448.pdf (ilo.org)  

 

https://www.ilo.org/dyn/natlex/docs/ELECTRONIC/99448/118729/F-1718165102/GUY99448.pdf


 
 

137 
 

4.6 Conclusion 
 

This study uses the 2009 Guyana Demographic and Health Survey to analyse the effects of 

birth order on early childhood health in Guyana. This study contributes to the body of literature 

that postulates that birth order positively affects health at birth and in early childhood. 

Consequently, we provide new evidence for Guyana that supports the hypothesis that compared 

to first-born children, birth order is positively correlated with birth weight and negatively 

correlated with low birth weight. Our finding also shows that the height for age z score taken 

for the children at the time of the survey increases for second and third born children when 

compared to firstborns. Moreover, our findings reveal that the height (in cm) of the child at the 

time of the survey, increases for more for second and third born children when compared to 

first-borns.  

This study also examines the use of postnatal intervention which may offset the disadvantages 

health that first-born faces at birth (Saldarriaga, 2015; Brenøe and Molitor, 2017; Pruckner, 

2019). Differential parental behaviour as it relates to postnatal investments can be capture 

through our findings from postnatal investments. Our findings show that compared to first-

borns mothers tend to extensively breastfeed their later-born babies and they also tend to invest 

in postnatal inoculations (viz Pentavalent and Polio vaccinations) in Guyana. 

Effectively, our paper provides a comprehensive analysis of the relationship between birth 

order and childhood health using the 2009 DHS from Guyana. Consequently, we also examined 

the heterogeneity by ethnicity, religion, maternal education, and child gender. Overall, we find 

that in general Indian children tends to have lower birth weights when compared to other 

subgroups. In addition, we find that second born children of Hindu religion tend to be shorter 

when compared to firstborns. We also find that third and fourth born or higher Muslim children 

across the various religious groups had higher birth weight compared to firstborns. While birth 

order is not quite a policy intervention debatable subject; given the statistics taken from the 

2009 GDHS vaccine coverage rate for Polio, Pentavalent and DPT, we see that the dropout 

rates for the full vaccine course are extremely high. As such, the Ministry of Health Guyana 

should create programmes to better educate first time mothers on the impact of low birth weight 

on first born children and the importance of postnatal invention such as exclusive breastfeeding 

and full vaccination doses in mitigating such negative impacts.  



 
 

138 
 

References 

Barcellos, S.H., Carvalho, L.S. and Lleras-Muney, A., 2014. Child gender and parental 

investments in India: Are boys and girls treated differently?. American Economic Journal: 

Applied Economics, 6(1), pp.157-189. 

Betran, A.P., de Onis, M., Lauer, J.A. and Villar, J., (2001). Ecological study of effect of breast 

feeding on infant mortality in Latin America. Bmj, 323(7308), pp.303. 

Björkegren, E. and Svaleryd, H., (2017). Birth order and child health (No. 2017:3). Working 

Paper. 

Black, S.E., Devereux, P.J. and Salvanes, K.G., (2016). Healthy (?), wealthy, and wise: Birth 

order and adult health. Economics and Human Biology, 23, pp.27-45. 

Brenøe, A.A. and Molitor, R., (2018). Birth order and health of newborns. Journal of 

Population Economics, 31(2), pp.363-395. 

Buber‐Ennser, I. and Berghammer, C., (2021). Religiosity and the realisation of fertility 

intentions: A comparative study of eight European countries. Population, Space and 

Place, 27(6), p.e2433. 

Buckles, K. and Kolka, S., (2014). Prenatal investments, breastfeeding, and birth order. Social 

Science and Medicine, 118, pp.66-70. 

Case, A., Fertig, A. and Paxson, C., (2005). The lasting impact of childhood health and 

circumstance. Journal of health economics, 24(2), pp.365-389. 

Currie, J., Stabile, M., Manivong, P. and Roos, L.L., (2010). Child health and young adult 

outcomes. Journal of Human resources, 45(3), pp.517-548. 

De Haan, M., Plug, E. and Rosero, J., (2014). Birth order and human capital development 

evidence from Ecuador. Journal of Human Resources, 49(2), pp.359-392. 

De Keyser, L. and Van Rossem, R., (2017). Birth order and height development in early 

childhood: empirical evidence for twenty-first-century Flanders. The History of the 

Family, 22(2-3), pp.236-257. 

Elliott, B. A. (1992). Birth order and health: Major issues. Social Science and Medicine, 35(4), 

443–452. 

Esposito, L., Kumar, S.M. and Villaseñor, A., (2020). The importance of being earliest: birth 

order and educational outcomes along the socioeconomic ladder in Mexico. Journal of 

Population Economics, 33(3), pp.1069-1099. 

Gluckman, P.D. and Hanson, M.A., (2004). October. Maternal constraint of fetal growth and 

its consequences. In Seminars in fetal and neonatal medicine (Vol. 9, No. 5, pp. 419-425). WB 

Saunders. 

Guyana Demographic and Health Survey., (2009). Available at: 

https://dhsprogram.com/publications/publication-fr232-dhs-final-reports.cfm [Accessed: 

January 2019] 

Guyana Ministry of Health., (2021). Available at: https://www.health.gov.gy/ [Accessed: Jan 

2021] 

https://dhsprogram.com/publications/publication-fr232-dhs-final-reports.cfm
https://www.health.gov.gy/


 
 

139 
 

Guyana., (2002) Population and Housing Census, Guyana Bureau of Statistics, 2021. Available 

at: https://statisticsguyana.gov.gy/publications/#cen2002  [Accessed: January 2021] 

Isungset, M.A., Freese, J., Andreassen, O.A. and Lyngstad, T.H., (2022). Birth Order 

Differences in Education Originate in Post-Natal Environments. bioRxiv, pp.2021-06. 

Jarry, V., Gagnon, A. and Bourbeau, R., (2013). Maternal age, birth order and other early-life 

factors: a family-level approach to exploring exceptional survival. Vienna Yearbook of 

Population Research, pp.267-294. 

Jayachandran, S. and Pande, R., (2017). Why are Indian children so short? The role of birth 

order and son preference. American Economic Review, 107(9), pp.2600-2629. 

Khong, T.Y., Adema, E.D. and Erwich, J.J.H., (2003). On an anatomical basis for the increase 

in birth weight in second and subsequent born children. Placenta, 24(4), pp.348-353. 

Lehmann, J.Y.K., Nuevo-Chiquero, A. and Vidal-Fernandez, M., (2018). The early origins of 

birth order differences in children’s outcomes and parental behaviour. Journal of Human 

Resources, 53(1), pp.123-156. 

Leroy, J.L., Ruel, M., Habicht, J.P. and Frongillo, E.A., (2015). Using height-for-age 

differences (HAD) instead of height-for-age z-scores (HAZ) for the meaningful measurement 

of population-level catch-up in linear growth in children less than 5 years of age. BMC 

pediatrics, 15(1), pp.1-11. 

Mmopelwa, D., (2019). Household size, birth order and child health in Botswana (No. 19/10). 

CREDIT Research Paper. 

Myrskylä, M., Silventoinen, K., Jelenkovic, A., Tynelius, P. and Rasmussen, F., (2013). The 

association between height and birth order: evidence from 652 518 Swedish men. J Epidemiol 

Community Health, 67(7), pp.571-577. 

Nakamuro, M., Uzuki, Y. and Inui, T., (2013). The effects of birth weight: Does fetal origin 

really matter for long-run outcomes? Economics Letters, 121(1), pp.53-58. 

Pruckner, G.J., Schneeweis, N., Schober, T. and Zweimüller, M., (2021). Birth order, parental 

health investment, and health in childhood health care utilization. Journal of Health 

Economics, 76, pp.102426. 

Rahman, M., (2016). Association between order of birth and chronic malnutrition of children: 

a study of nationally representative Bangladeshi sample. Cadernos de saude publica, 32, 

p.e00011215. 

Sahn, D.E. and Stifel, D.C., (2002). Robust comparisons of malnutrition in developing 

countries. American Journal of Agricultural Economics, 84(3), pp.716-735. 

Saldarriaga, V., (2015). Birth weight and early childhood physical health: evidence from a 

sample of Latin American twins. Economía, 15(2), pp.161-197. 

Selvin, S. and Janerich, D.T., (1971). Four factors influencing birth weight. British journal of 

preventive and social medicine, 25(1), p.12. 

Victora, C., Vaughan, J.P., Lombardi, C., Fuchs, S.C., Gigante, L., Smith, P., Nobre, L., 

Teixeira, A., Moreira, L. and Barros, F., (1987). Evidence for protection by breast-feeding 

against infant deaths from infectious diseases in Brazil. The Lancet, 330(8554), pp.319-322. 

WHO Multicentre Growth Reference Study Group and de Onis, M., (2006). Assessment of 

https://statisticsguyana.gov.gy/publications/#cen2002


 
 

140 
 

differences in linear growth among populations in the WHO Multicentre Growth Reference 

Study. Acta Paediatrica, 95, pp.56-65. 

World Health Organization., (2019). UNICEF-WHO low birthweight estimates: levels and 

trends 2000-2015 (No. WHO/NMH/NHD/19.21). World Health Organization. 

 

 



 
 

141 
 

Chapter 5: - Conclusion 

 

5.1 Introduction 
 

This final chapter summarises the overall findings and contribution of the thesis, which focuses 

on the research questions that were answered concerning the effects of weather shocks, 

educational reform and birth order on health and education-related outcomes in Guyana. The 

chapter presents significant findings on how weather shocks and birth order can impact early 

childhood health in Guyana, as well as the impact of educational policy reforms on the 

completion of secondary and higher education in Guyana. In addition, the chapter highlights 

the negative impacts of in-utero exposure to weather shocks (i.e., positive and negative rainfall 

shocks) and early gestation, as well as the benefits of being later-born siblings and the positive 

impacts of secondary school reform policy on the completion of secondary and the completion 

of secondary or tertiary education. The chapter then presents some of the policy implications 

for local mitigation of weather shocks on early childhood health and possible policy 

implications for birth order effects. Furthermore, this chapter also provides empirical evidence 

of the impact of the USE (2003-2007) policy reform on educational and labour market 

outcomes in Guyana and supports the need for the swift conversion of secondary schools with 

vocational curriculums into academic curriculums (i.e., Primary Tops into General Secondary 

Schools). There is a need to contextualise early childhood vulnerabilities to climate change and 

its impact on childhood health, understand how birth order effects can influence policy changes 

and how secondary educational policy reform can influence secondary or tertiary completion 

and, ultimately, the human capital of Guyana.  

 

5.2 Research Highlights 
 

Climate change has led to an increase in the phenomenon of natural disasters, which has 

catastrophic impacts on economies at large. Chapter Two, based on weather shocks' effects on 

Guyana's childhood health, highlights how vulnerability to floods and climate variability in 

flood-prone regions can lead to negative childhood health impacts on outcomes such as height-

for-age z-scores, and stunted and severely stunted growth in Guyana.  Le and Nguyen (2021) 

reveal that exposure to excess rainfall in the first trimester in 55 low and middle-income 



 
 

142 
 

countries resulted in reduced children’s HAZ and increased stunting. My research 

demonstrated similar findings in the context of middle-income Guyana, which is susceptible to 

flooding and sea levels rises. Chapter Two reveals that exposure to rainfall shocks, primarily 

in the first trimester, results in lower height-for-age z-scores and an increase in the incidence 

of stunted and severely stunted growth. My research also found that positive rainfall shocks in 

the first trimester increases the likelihood of stunted and severely stunted growth. Similarly, 

positive rainfall shock in the second trimester increases the likely of stunted growth. Therefore, 

the research provides many insights into the effects of in-utero exposure to weather shocks and 

its impact on early childhood health in Guyana. The research also provides some plausible 

policy implications which can be used by the Government of Guyana to combat the incidence 

of positive and negative rainfall shocks due to climate change.  

Chapter Three of this thesis examined the impact of a secondary school policy reform on the 

completion of secondary and the completion of secondary or tertiary education in Guyana. 

Furthermore, the research is the first-ever study to analyse a USE (2003-2007) policy reform 

that converts secondary schools with a vocational curriculum into academic curriculums. It is 

also the first study to provide an empirical evaluation of the impact of the USE (2003-2007) 

policy reform on Guyana’s educational sector.  

The findings of this research show that greater exposure to USE (2003-2007) had positive 

impacts on the completion of secondary and the completion of secondary or tertiary education 

and the potential to earn higher wages in Guyana. Research from Germany, Denmark, Austria 

and Switzerland shows that some developed countries operate under a dual-educational system 

where schools with vocational education and training curriculums are encouraged alongside 

schools with academic curriculums to suit the various students’ abilities (Eichhorst, 2015; 

Baethge and Wolter, 2015). However, my research findings revealed that the conversion of 

pre-vocational secondary schools with vocational curriculums into academic curriculums led 

to the completion of secondary or tertiary education and an increase in wages of the people 

exposed to the policy. On the other hand, the finding also revealed that the USE (2003-2007) 

policy reform did not close the achievement gaps in education across all ethnic groups. As 

such, the USE policy only benefitted the two larger ethnic groups, i.e., East Indians and 

Africans.  

Chapter Four is based on birth order effects and early childhood health in Guyana. My research 

revealed that health at birth improves with increasing birth orders. Furthermore, we find that 



 
 

143 
 

third- and fourth-born children are less likely to be born with low birth weight. It also revealed 

that second and third-born children had higher height-for-age z-scores when compared to 

firstborn children. Furthermore, when looking at postnatal investments in the form of 

breastfeeding, we also find that when compared to their firstborn, mothers spend more months 

breastfeeding their later borns. Overall, when looking at postnatal investments in inoculations, 

we also find that mothers tend to vaccinate their later borns against Pentavalent and Polio 

compared to their firstborns.   

This thesis presents several policy implications that can be used by the Government of Guyana. 

For Chapter Two, one potential policy recommendation is for the Ministry of Health, Guyana, 

to collaborate with a multidisciplinary research team to specifically assess the medium-and 

long-term effects of adverse weather shocks on child and maternal health. Furthermore, the 

Ministry of Health should consider the nutrition, and medical outreach programmes for 

pregnant mothers, in regions that are more susceptible to flooding. It is also proposed that a 

climatologist should be added to this team to effectively monitor and predict the onset of these 

shocks, which would better put provisions in place to improve maternal and early childhood 

health in weather shocks aftermath.  

 

As it relates to Chapter Three, it was recognised that the USE (2003-2007) reform policy 

introduced by Guyana’s Ministry of Education had made strides in increasing and improving 

access to secondary education. Although it was found that the policy led to the completion of 

secondary and the completion of secondary or higher education, thus realising the benefits of 

Guyana’s USE (2003-2007) policy reform. Overall, I find that the USE (2003-2007) policy 

reform did not close the achievement gaps in education across ethnicities. It is also 

recommended that the MoE evaluate the USE (2003-2007) policy reform and assess whether 

it is fulfilling its equity goal and employ measures to ensure that the achievement gaps are 

closed across all ethnic groups. It is also recommended that more should be done to increase 

the speed of the conversion of the remaining Primary Tops to GSSs. With the greater 

conversion of Primary Tops to GSSs, it is likely that this will lead to greater completion of 

secondary and the completion of secondary or tertiary education in Guyana. As such, it is 

imperative that Guyana’s Ministry of Education works towards making this a reality.  

Chapter Four examined the birth order effects and early childhood health in Guyana. My 

finding revealed that mothers are more likely to invest in postnatal investments such as 

breastfeeding practices and inoculations for their later-born children. Consequently, one 



 
 

144 
 

possible policy implication is the provision of free childcare for working parents with 2-3 

children aged five and under by the Government of Guyana to help parents boost the 

development of later-born children.  

 

In addition, the statistics taken from the 2009 GDHS revealed that the dropout rates for the full 

vaccine coverage for Polio, Pentavalent and DPT are extremely high. Therefore, another policy 

implication would be the creation of programmes by Guyana’s Ministry of Health to better 

educate first-time mothers on the impact of low birth weight on firstborn children and the 

importance of postnatal investments such as exclusive breastfeeding and being fully vaccinated 

to mitigate such negative impacts.    

 

The policy implications proposed by this research can be used by the Government of Guyana 

to better deal with the negative effects of climate change on early childhood health. Moreover, 

the provision of childcare services for parents with 2-3 children five years old and under should 

be a top priority for the Government of Guyana. This research also analysed the effectiveness 

of the USE (2003-2007) policy reform and found that the policy reform was beneficial for the 

students who were exposed to the USE. It is also recommended that the conversion of the 

remaining Primary Tops into GSSs by Guyana’s Ministry of Education should be done in a 

swift manner to encourage greater completion of secondary or tertiary education.  

 

5.3 Needs for Further Research 
 

This thesis presents findings from research that took place in several different periods, i.e., 

during in-utero periods of weather shocks for children born in 1995-2014, using MICS 2000, 

2006-07 and 2014; using the GLFS for the period 2018 and 2019 to analyse the USE (2003-

2007) policy reform and using Guyana 2009 DHS to analyse birth order effects. Consequently, 

there is a need for similar studies in the same areas with extended data, for instance, including 

the MICS 2018 dataset for Guyana, in order to find out if there will be significant changes in 

terms of childhood health outcomes and also to ascertain how mothers are coping with weather 

shocks given the current policy changes in response to flooding made by the Government of 

Guyana. In addition, with the conversion of the remaining Primary Tops to GSSs, there will be 

a need for an extension of Chapter Three, which can provide empirical evidence for the benefits 

of the USE (2003-2007) reform policy as it relates to educational attainment and labour market 



 
 

145 
 

outcomes. Finally, I recommend more research examining, understanding, and exploring birth 

order effects and childhood health in Guyana. It is evident that ‘good’ childhood health is very 

important in strengthening the human capital of Guyana and, ultimately, the economic wealth 

of Guyana. As health and education are two important human capital investments essential to 

any society, the Government of Guyana must continue to invest in these sectors at the national 

and community level to improve its human capital.  

 

5.4 Summary 
 

Chapter Two of this study provides evidence that weather shocks can lead to negative 

childhood outcomes, potentially impacting the labour market and the economy as a whole. The 

study provides evidence that children who were in-utero during weather shocks are likely to 

have lower height-for-age z-scores and an increase in the incidence of stunted and severely 

stunted growth in Guyana. Furthermore, it was found that mothers exposed to positive rainfall 

shocks were more likely to breastfeed for longer periods of time and invest in some postnatal 

vaccinations when looking at the trimester-specific effect. Furthermore, the study recommends 

that in an effort to reduce vulnerability to climate variability, Guyana’s Ministry of Health, in 

collaboration with climatologists, should better monitor the onsets of floods and other climatic 

disasters to provide the necessary health for mothers affected by these natural disasters.  

The empirical evidence provided in Chapter Three is beneficial to both the Ministry of 

Education and the wider society, as there are no empirical studies, nor no known studies done 

by the MoE to effectively analyse the USE (2003-2007) policy reform in Guyana. 

Consequently, this study provides knowledge to both the Government of Guyana and the 

economy at large. The education sector accounts for approximately 5 percent of Guyana’s 

yearly GDP, and it is one of the country’s most important sectors as it provides the country’s 

human capital (World Bank Data, 2022).  This study recommends that the MoE speed up the 

conversion of Primary Tops into GSSs to encourage greater completion of secondary or tertiary 

education in Guyana. It is also recommended that the MoE assess the impact of the policy 

across the ethnic groups and use the appropriate measures to address achievement gaps. Based 

on the findings of this study, it is evident that the students exposed to the USE (2003-2007) 

policy reform were more encouraged to complete their secondary education in their new 

General Secondary Schools.  



 
 

146 
 

Chapter Four provided evidence that later-born children have better birthweight and mothers 

tend to invest in postnatal vaccinations for later-born siblings. As such, it was recommended 

that the Government of Guyana provide free childcare services for parents with 2-3 children 

five years old and under.   

Overall, this dissertation provides empirical evidence supporting the need for policy changes 

to improve Guyana’s Health and Education sectors to strengthen the country’s human capital 

and thereby encourage greater economic growth.    

 

 



 
 

147 
 

Appendix A 
 

CHAPTER 2 - The effects of weather shock on childhood health  

 

Appendix 1 Further details on data and variable description 

 

Individual Characteristics 

 

Height-for-age z scores (HAZ) – Defined as height-for-age z-score between -9.62 SD and 

6.58 SD. NB:// The variable was defined using the variables (HAZ2 - MICS 2000, 2006/07 

and 2014). 

 

Stunted – Defined as a dummy variable equal to 1 if the HAZ is less than -2 SD and 0 

otherwise.  

 

Severely stunted – Defined as a dummy variable equal to 1 if the HAZ less than -3 SD and 0 

otherwise.  

 

Male – Defined as a dummy variable equal to 1 if the male and 0 otherwise.  

 

Postnatal investments  

 

Child ever breastfed – Defined as a dummy, equal to 1 if the child has ever been breastfed 

and 0 otherwise. The variable was defined using the variables (bf1 (MICS 2000), BF1 (MICS 

2006/07) and BD2 (MICS 2014) in the MICS survey– the month of BCG immunization. 

 

Measles vaccination – Defined as a dummy, equal to 1 if the child has received their measles 

vaccination within their first year of birth and 0 otherwise. In Guyana, measles vaccinations 

are distributed at 9-11 months after birth. NB:// The variable was defined using the variables 

(im5m (MICS 2000), im6m (MICS 2006/07) and im3mm (MICS 2014) in the MICS survey – 

the month of BCG immunization.  

 

 

DPT Vaccination – Defined as a dummy equal to 1 if the child had received their DPT 

vaccinations within their first year of birth and 0 otherwise. NB:// The variable was defined 

using the variables (im4cm (MICS 2000), im5cm (MICS 2006/07) and im3d3m (MICS 2014) 

in the MICS survey – the month of BCG immunization.  

 

 

BCG Vaccination - Defined as a dummy equal to 1 if the child had received their BCG 

vaccinations within their first year of birth and 0 otherwise. NB:// The variable was defined 

using the variables (im2m (MICS 2000), im2m (MICS 2006/07) and im3bm (MICS 2014) in 

the MICS survey – the month of BCG immunization.  



 
 

148 
 

Mother Characteristics 

 

No education – Defined as a dummy equal to 1 if the mother has no education (i.e., variable 

melevel=1) and 0 otherwise.  

 

Primary education – Defined as a dummy equal to 1 if the mother has primary education 

(melevel =3) and 0 otherwise.  

 

Higher education – Defined as a dummy equal to 1 if the mother had secondary or higher 

education (i.e., variable melevel=4 and 5) and 0 otherwise.  

 

Urban – Defined as a dummy equal to 1 if the mother lived in an urban area and 0 otherwise.  

 

Ethnicity 

 

African – Defined as a dummy equal to 1 if the child is of African descent and 0 otherwise.  

 

East Indian - Defined as a dummy equal to 1 if the child is of East Indian descent and 0 

otherwise.  

 

Mixed - Defined as a dummy equal to 1 if the child is of Mixed descent and 0 otherwise.  

 

Amerindian - Defined as a dummy equal to 1 if the child is of Amerindian descent and 0 

otherwise.  

 

Portuguese - Defined as a dummy equal to 1 if the child is of Portuguese descent and 0 

otherwise.  

 

Chinese – Defined as a dummy equal to 1 if the child is of Chinese descent and 0 otherwise.  

 

 

 

 

 

 

 

 

 



 
 

149 
 

Appendix B 
 

CHAPTER 3 - Effects of a School Conversion Policy on Educational Attainment and 

Labour Market Outcomes in Guyana 

 

 

Appendix 1 Further details on data and variable description 

 

 

Educational Attainment outcomes 

 

Secondary or higher education (Completed) – Defined as a dummy equal to 1 if the highest 

education completed is Secondary or higher (including tertiary education) and 0 otherwise. 

NB//: This variable was constructed by combining the response to questions q1_14 and q1_15 

in the GLFS questionnaires.  

Q1_14 = What class did you Complete?  

Q1_15 = What is the highest post-secondary level/degree that you have completed?  

 

 

Tertiary (Completed) – Defined as a dummy equal to 1 if the highest education (i.e., 

technical/vocational certificate or Diploma, University Certificate or Diploma, Bachelor’s 

degree, Postgraduate Certificate or Diploma, Master’s and Doctorate) and 0 otherwise. NB:// 

This variable was constructed based on the combined variable above (using q1_14 and q1_15).  

 

Secondary or higher education (Attended) – Defined as a dummy equal to 1 if the highest 

education attended is Secondary & higher and 0 otherwise. NB:// This variable was constructed 

using question Q1_13, i.e., What is the highest level of education that you have attended? 

 

 

Labour Market Outcomes 

 

Employed – Defined as a dummy equal to 1 if in the last seven days the respondent worked for 

a wage salary or tips (q2_04), or if in the last seven days the respondent did any kind of business 

or farming (q2_05), or if in the last seven days the respondent did even one hour of cooking 

for sale or other food processing activities for sale, making baskets, mats or other handicrafts 

for sale (q2_06), or if in the last seven days, the respondent helped in an unpaid business or 

farm owned by a household (q2_07), or if in the last seven days, the respondent was absent 

from a paid job or business including a household business, cook for sale or other food 

processing activities for sale or making handicrafts for sale (q2_10), given that the age of the 

respondent, i.e., age 15 and above (q1_04) and 0 otherwise. NB:// This variable used to 

construct this variable are q1_04>=15, q2_04, q2_05, q2_06, q2_07, and q2_10.  

Wage employers – Defined as a dummy equal to 1 if the individual reported working as an 

employee for a company and 0 if the self-employed or unpaid family worker, conditional on 



 
 

150 
 

employed. NB:// This variable was constructed using question Q3_16, i.e., What is your 

employment status in your main job? 

Work hours – Hours worked in the past week conditional on employed, i.e., missing for non-

working individuals. NB:// This variable was constructed using question Q3_03, i.e., How 

many hours do you usually work per week in your main job (the one that takes more hours)? 

Log wage – The logarithm of the monthly wage variable.  NB:// This variable was constructed 

using question Q6_01, i.e., What was your income in the last month in terms of net salary (in 

numbers)? 

Formal wage worker – Defined as a dummy equal to 1 if the individual was provided with a 

written contract, as opposed to a verbal agreement (informal workers). NB:// This variable was 

constructed using question Q3_18, i.e., Are you working on the basis of a written contract or 

is it on the basis of a verbal agreement? 

 

 



 
 

151 
 

Appendix B.1: The USE policy reform and teacher’s qualification 

 

The section below provides some descriptive statistics analysed using the DES (1999-2018) 

which explains changes as it relates to trained, untrained and unqualified teachers.  

 

 
NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 

Figure B1: The share of students in secondary educational institutions by type during (1999-2019). 

Source: Digest of Education Statistics (1999-2019). 

 

The DES (1999-2019) provides us with data on the number of trained and untrained teachers 

in the secondary institutions. Trained teachers are defined as teachers with the necessary 

professional qualifications, i.e., a minimum of five subjects inclusive of Mathematics and 

English at CSEC level coupled with an educational certification from the Cyril Potter College 

of Education (CPCE). Untrained teachers are teachers who do not have the professional 

qualification to teach, for instance, those without any educational certification from the CPCE. 

On the other hand, unqualified teachers do not have the requisite qualification to enter the 

CPCE to become certified to teach (Chief Planning Officer, MOE, 2022).  

 

In addition, during the post-and-pre policy periods, we find that GSSs had a greater share of 

trained teachers when compared to CHSs and Primary Tops who both had lower share of 

trained teachers before the policy change. During the pre-policy period, we can see a steady 

increase in trained teachers in the GSSs, but on average a slight decline in trained teachers in 



 
 

152 
 

Primary Tops and CHSs (see Figure B2 below for the share of trained teachers in the secondary 

institutions for the period (1999-2019)).  

 
NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 

Figure B2: The share of trained teachers for types of secondary educational institutions during (1999-2019). 

Source: Digest of Education Statistics (1999-2019) 

 

 

Furthermore, GSSs saw successive increase in the number of trained teachers over the years. 

When looking at the share of untrained teachers for the period, we find that there were higher 

shares of untrained teachers for GSSs but lower shares in Primary Tops and CHSs overtime 

(see Figure B3 below for the share of untrained teachers for the period (1999-2019). 

 

 
NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 
Figure B3: The share of untrained teachers for the period (1999-2019). 

Source: Digest of Education Statistics (1999-2019). 

 



 
 

153 
 

However, when looking at the share of unqualified teachers, we find that in the years (1999-

2009) the Primary Tops had a greater proportion of unqualified teachers when compared to the 

CHSs and GSSs (see Figure B4 below for the share of unqualified teachers for the period (1999-

2019)).  

 

 
NB:/Missing data for the year (2012-2013) - Ministry of Education (MOE). 

Figure B4: The share of unqualified teachers in the secondary institutions for the period (1999-2019). 

Source: Digest of Education Statistics (1999-2019). 

 

 

 

 

 

 

 

 

 

 

 

  



 
 

154 
 

Appendix C 
 

CHAPTER 4 - Birth Order Effects on Early Childhood Health in Guyana  

 

Appendix 1 Further details on data and variable description 

 

 

Demographic Characteristics 

 

Birth order 1 – Defined as a dummy equal to 1 if the child was the first born and 0 otherwise. 

 

Birth order 2- Defined as a dummy equal to 1 if the child was the second born and 0 otherwise. 

 

Birth order 3 – Defined as a dummy equal to 1 if the child was the third born and 0 otherwise.  

 

Birth order 4 or higher – Defined as a dummy equal to 1 if the child was a fourth born or 

higher and 0 otherwise. 

 

Child is male – Defined as a dummy equal to 1 if the child is male and 0 otherwise.  

 

Urban – Defined as a dummy equal to 1 if the child resided in Urban areas and 0 otherwise.  

 

Health at birth  

 

Birth weight (kg) – Defined as the birth weight in kilogram ranging between 0.8 kg – 8kg. 

 

Low birth weight (<2.5 kg) – Defined as a dummy if the childbirth weight is 2.5 kg or lower.  

 

Childhood Health  

 

Height-for-age z-score - Defined as height-for-age z-score that lies between -5.99 SD and 5.53 

SD. 

 

Stunted - Defined as a dummy variable equal to 1 if the HAZ is less than -2 SD and 0 otherwise.  

 

Child height (cm) – Defined as the height of children ranging between 42 cm to 123.3 cm. 

 

Mother’s Characteristics 

 

Mother’s age at first birth – Defined as the mother’s age (continuous variable) ranging from 

11 to 38 years old.  

 

Nursery or no education – Defined as a dummy equal to 1 if the mother has a nursery or no 

education and 0 otherwise. 

 

Primary education – Defined as a dummy equal to 1 if the mother has primary education and 

0 otherwise. 

 



 
 

155 
 

Secondary education – Defined as a dummy equal to 1 if the mother has secondary education 

and 0 otherwise. 

 

Higher education – Defined as a dummy equal to 1 if the mother has higher education and 0 

otherwise. 

 

 

Ethnicity 

 

African – Defined as a dummy equal to 1 if the child is of African descent and 0 otherwise.  

 

East Indian - Defined as a dummy equal to 1 if the child is of East Indian descent and 0 

otherwise.  

 

Mixed - Defined as a dummy equal to 1 if the child is of Mixed descent and 0 otherwise.  

 

Amerindian - Defined as a dummy equal to 1 if the child is of Amerindian descent and 0 

otherwise.  

 

Portuguese - Defined as a dummy equal to 1 if the child is of Portuguese descent and 0 

otherwise.  

 

Religion 

 

Christian – Defined as a dummy equal to 1 if the child is from a Christian family and 0 

otherwise.  

 

Hindu – Defined as a dummy equal to 1 if the child is from a Hindu family and 0 otherwise.  

 

Muslim – Defined as a dummy equal to 1 if the child is from a Muslim family and 0 otherwise.  

 

Rastafarian – Defined as a dummy equal to 1 if the child is from a Rastafarian family and 0 

otherwise.  

 

Not Religious – Defined as a dummy equal to 1 if the child is from a family without a religious 

background and 0 otherwise.  

 

Other - Defined as a dummy equal to 1 if the child is from a family who has “other” religious 

belief that was not listed and 0 otherwise. 

 

Postnatal Investments 

 

Breastfed – Defined as a dummy equal to 1 if the child has ever been breastfed and 0 otherwise.  

 

All vaccination (either) – Defined as a dummy equal to 1 if the child has received all their 

vaccinations ‘either’ declared by the mother or on their vaccination card and 0 otherwise.  

 

No vaccination (either) - Defined as a dummy equal to 1 if the child has not received any 

vaccination ‘either’ declared by the mother or on their vaccination card and 0 otherwise. 

 



 
 

156 
 

Polio 1st dose (either) - Defined as a dummy equal to 1 if the child has received their first dose 

of polio vaccine ‘either’ declared by the mother or on their vaccination card and 0 otherwise. 

 

Polio 2nd dose (either) - Defined as a dummy equal to 1 if the child has received their second 

dose of polio vaccine ‘either’ declared by the mother or on their vaccination card and 0 

otherwise.    

 

Polio 3rd dose (either) - Defined as a dummy equal to 1 if the child has received their third 

dose of polio vaccine ‘either’ declared by the mother or on their vaccination card and 0 

otherwise.    

 

Measles (either) - Defined as a dummy equal to 1 if the child has received their measles 

vaccine ‘either’ declared by the mother or on their vaccination card and 0 otherwise.  

 

Pentavalent 1st dose (either) - Defined as a dummy equal to 1 if the child has received their 

first dose of pentavalent vaccine ‘either’ declared by the mother or on their vaccination card 

and 0 otherwise. 

 

Pentavalent 2nd dose (either) - Defined as a dummy equal to 1 if the child has received their 

second dose of pentavalent vaccine ‘either’ declared by the mother or on their vaccination card 

and 0 otherwise. 

 

Pentavalent 3rd dose (either) - Defined as a dummy equal to 1 if the child has received their 

third dose of pentavalent vaccine ‘either’ declared by the mother or on their vaccination card 

and 0 otherwise.  

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 

157 
 

Appendix 2: Recommended vaccination schedule in Guyana 

 

Table C.1: Recommended Vaccination Schedule in Guyana 

VACCINE NO. of DOSES AGE at administration 

BCG 1 Birth 

OPV 

3 8, 16, 24 weeks 

1st booster 18 months 

2nd booster 3-5 years 

3rd booster 11+ years 

  every 10 years thereafter 

IPV 3 8, 16, 24 weeks 

Pentavalent 
3 8, 16, 24 weeks 

(DPT, Hib, Hepatitis B) 

PCV 13 3 8, 16, 24 weeks 

DPT 
1st Dose after Pentavalent series 18 months 

2nd Dose 3 years 9 months 

DT   Given to children in whom DPT is  
(Paediatric)   contraindicated 

DT (adult) 

1st Dose  Pregnant women at contact  

2nd Dose 4 weeks after 

1st Booster 6 months after 2nd dose 
    Repeat every 10 years if necessary 

MMR 2 doses 
12 months 

3 years 9 months and over 

Yellow fever 
1 12 months 

  Repeat every 10 years if necessary 

Hepatitis B 

  Health workers at risk 

1st dose At any age 

2nd dose 4 weeks later 

3rd dose 5 months later 

Source: Ministry of Health, Guyana, Jan 20th, 2020 

  



 
 

158 
 

Bibliography 

 

Adhvaryu, A. and Nyshadham, A., (2016). Endowments at birth and parents' investments in 

children. The Economic Journal, 126(593), pp.781-820. 

 

Agüero, J., (2014). Long-term effect of climate change on health: Evidence from heat waves 

in Mexico, IDB-WP-481. 

 

Aguilar, A. and Vicarelli, M., (2022). El Nino and children: medium-term effects of early-life 

weather shocks on cognitive and health outcomes. World Development, 150, 105690. 

 

Aizer, A. and Cunha, F., (2014). Child endowments, parental investments, and the production 

of human capital. National Bureau for Economic Research Working paper, 18429. 

 

Ajayi, K.F. and Ross, P.H., (2020). The effects of education on financial outcomes: Evidence 

from Kenya, Economic Development and Cultural Change, 69(1), pp.253-289. 

 

Akresh, R., Halim, D. and Kleemans, M., (2018). Long-term and intergenerational effects of 

education: Evidence from school construction in Indonesia (No. w25265). National Bureau of 

Economic Research. 

Alderman, H., Hoddinott, J., and Kinsey, B., (2006). Long Term Consequences of Early 

Childhood Malnutrition. Oxford Economic Papers, 53(4), pp. 450-474.  

Alderman, H., Orazem, P.F. and Paterno, E.M. (2001). ‘School quality, school cost, and the 

public/private school choices of low-income households in Pakistan’, Journal of Human 

resources, pp.304-326. 

Almond, D., Currie, J. and Duque, V., (2018). Childhood circumstances and adult outcomes: 

Act II. Journal of Economic Literature, 56(4), pp.1360-1446. 

Almond, D., Edlund, L. and Palme, M., (2009). Chernobyl's subclinical legacy: prenatal 

exposure to radioactive fallout and school outcomes in Sweden. The Quarterly Journal of 

Economics, 124(4), pp. 1729-1772.  

Almond, D. and Mazumder, B., (2013). Foetal origins and parental responses. Annual. Rev. 

Econ., 5(1), pp.37-56. 

Altshuler, K., M. Berg, L.M. Frazier, J. Laurenson, J. Longstreth, W.Mendez, and C. A 

Molgaard., (2003). Critical periods in development. OCHP Paper Series on Childrens Health 

and the Environment, 3:1–31 

Ashenfelter, O. and Krueger, A. (1994). Estimates of the economic return to schooling from a 

new sample of twins, American Economic Review, 84(5), pp.1157-1173. 

 

Baethge, M. and Wolter, A., 2015. The German skill formation model in transition: from dual 

system of VET to higher education?. Journal for labour market research, 48(2), pp.97-112. 



 
 

159 
 

Banerjee, Abhijit, Duflo, Esther, Postel‐Vinay, Gilles, and Watts, Tim., (2010). “Long run 

health impacts of income shocks: Wine and phylloxera in 19th century France,” The Review 

of Economics and Statistics 92(4): 714– 728. 

Barcellos, S.H., Carvalho, L.S. and Lleras-Muney, A., 2014. Child gender and parental 

investments in India: Are boys and girls treated differently?. American Economic Journal: 

Applied Economics, 6(1), pp.157-189. 

Barker, D.J., (1995). Fetal origins of coronary heart disease. British Medical 

Journal, 311(6998), pp.171-174. 

 

Barrera-Osorio, F., Linden, L.L. and Urquiola, M., (2013). The effects of user fee reductions 

on enrolment: Evidence from a quasi-experiment. 

BBC., (2019). Available at: Will Guyana soon be the richest country in the world? - BBC News 

[Accessed: 25th June 2022] 

Behrman, J.R., Pollak, R.A. and Taubman, P., (1982). Parental preferences and provision for 

progeny. Journal of political economy, 90(1), pp.52-73. 

Betran, A.P., de Onis, M., Lauer, J.A. and Villar, J., (2001). Ecological study of effect of breast 

feeding on infant mortality in Latin America. Bmj, 323(7308), pp.303. 

Beuermann, D. and Pecha, C., (2020). The effects of weather shocks on early childhood 

development: Evidence from 25 years of tropical storms in Jamaica. Economics & Human 

Biology, 37, p. 100851. 

Bhalotra, S., (2004). Early childhood investments in human capital: parental resources and 

preferences. University of Bristol Discussion Paper, 4, p.561. 

Björkegren, E. and Svaleryd, H., (2017). Birth order and child health (No. 2017: 3). Working 

Paper. 

Black, S.E., Devereux, P.J. and Salvanes, K.G., (2016). Healthy (?), wealthy, and wise: Birth 

order and adult health. Economics and Human Biology, 23, pp.27-45. 

Bonjean, C., Brunelin, S. and Simonet, C., (2012). Impact of climate related shocks on child’s 

health in Burkina Faso (No. 201232). CERDI. 

Botzen, W.W., Deschenes, O. and Sanders, M., (2019). The economic impacts of natural 

disasters: A review of models and empirical studies. Review of Environmental Economics and 

Policy, 13(2), pp.167-188. 

 

Bovolo, C.I., Pereira, R., Parkin, G., Kilsby, C. and Wagner, T., (2012). Fine‐scale regional 

climate patterns in the Guianas, tropical South America, based on observations and reanalysis 

data. International Journal of climatology, 32(11), pp.1665-1689. 

 

Bozzoli, C. and Quintana-Domeque, C., (2014). The weight of the crisis: Evidence from 

newborns in Argentina. Review of Economics and Statistics, 96(3), pp.550-562. 

 

https://www.bbc.co.uk/news/world-latin-america-48185246


 
 

160 
 

Bravo, D., Mukhopadhyay, S. and Todd, P.E., (2010). Effects of school reform on education 

and labour market performance: Evidence from Chile's universal voucher system. Quantitative 

economics, 1(1), pp.47-95. 

 

Breining, S., Daysal, N.M., Simonsen, M. and Trandafir, M., (2015). Spillover effects of early-

life medical interventions. 

 

Brenøe, A.A. and Molitor, R., (2018). Birth order and health of newborns. Journal of 

Population Economics, 31(2), pp.363-395. 

 

Brudevold-Newman, A., (2021). Expanding access to secondary education: Evidence from a 

fee reduction and capacity expansion policy in Kenya. Economics of Education Review, 83, 

p.102127. 

 

Brunello, G. and Rocco, L., (2017). The labour market effects of academic and vocational 

education over the life cycle: Evidence based on a British cohort. Journal of Human 

Capital, 11(1), pp.106-166. 

 

Buber‐Ennser, I. and Berghammer, C., (2021). Religiosity and the realisation of fertility 

intentions: A comparative study of eight European countries. Population, Space and 

Place, 27(6), p.e2433. 

 

Buckles, K. and Kolka, S., (2014). Prenatal investments, breastfeeding, and birth order. Social 

Science and Medicine, 118, pp.66-70. 

Burgess, R., Deschenes, O., Donaldson, D. and Greenstone, M., (2014). The unequal effects of 

weather and climate change: Evidence from mortality in India. Cambridge, United States: 

Massachusetts Institute of Technology, Department of Economics. Manuscript. 

 

Camacho, A., (2008). Stress and birth weight: evidence from terrorist attacks. American 

Economic Review, 98(2), pp. 511-15.  

 

Card, D. (1999) The causal effect of education on earnings. Handbook of labour economics, 3, 

pp.1801-1863. 

CARICOM., (2022). Available at: CARICOM-Caribbean Community - CARICOM 

[Accessed: 3rd July 2022] 

Caribbean Disaster Emergency Management Agency., (2012), Available at: 

https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-

guyana-situation-report--3&catid=125:guyana [Accessed: 30 June 2023] 

Carrillo, B., 2020. Early Rainfall Shocks and Later-Life Outcomes: Evidence from 

Colombia. The World Bank Economic Review, 34(1), pp.179-209. 

 

Case, A., Fertig, A. and Paxson, C., (2005). The lasting impact of childhood health and 

circumstance. Journal of health economics, 24(2), pp.365-389. 

CEIC., (2022). Available at: (https://www.ceicdata.com/en/guyana/health-statistics/gy-

prevalence-of-stunting-height-for-age--of-children-under-5 [Accessed: 25th June 2022] 

https://caricom.org/
https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-guyana-situation-report--3&catid=125:guyana
https://www.cdema.org/index.php?option=com_content&view=article&id=1040:flooding-in-guyana-situation-report--3&catid=125:guyana
https://www.ceicdata.com/en/guyana/health-statistics/gy-prevalence-of-stunting-height-for-age--of-children-under-5
https://www.ceicdata.com/en/guyana/health-statistics/gy-prevalence-of-stunting-height-for-age--of-children-under-5


 
 

161 
 

Chandisingh, R., (1983). The State, the Economy, and Type of Rule in Guyana: An Assessment 

of Guyana's" Socialist Revolution". Latin American Perspectives, 10(4), pp.59-74. 

Climate Change (2007): Impacts, Adaptation and Vulnerability. Ipcc.ch. (2007). Available at: 

https://www.ipcc.ch/site/assets/uploads/2018/03/ar4_wg2_full_report.pdf [Accessed 12th Apr. 

2020].  

Climate Change Legislation in Guyana, (2015). Available at 

http://www.lse.ac.uk/GranthamInstitute/wp-content/uploads/2015/05/GUYANA.pdf 

[Accessed: 5th Apr. 2020] 

Currie, J. and Rossin-Slater, M., (2013). Weathering the storm: Hurricanes and birth 

outcomes. Journal of Health Economics, 32(3), pp.487-503. 

 

Currie, J., Stabile, M., Manivong, P. and Roos, L.L., (2010). Child health and young adult 

outcomes. Journal of Human resources, 45(3), pp.517-548. 

Currie, J. and Vogl, T., (2013). Early-Life Health and Adult Circumstance in Developing 

Countries. Annual Review of Economics, 5(1), pp. 1-36.  

Cyril Potter College of Education., (2022). Available at: https://cpce.edu.gy/ (Accessed: 10 

March 2022) 

 

Dearden, L., McIntosh, S., Myck, M. and Vignoles, A., (2002). The returns to academic and 

vocational qualifications in Britain. Bulletin of economic research, 54(3), pp.249-274. 

 

De Haan, M., Plug, E. and Rosero, J., (2014). Birth order and human capital development 

evidence from Ecuador. Journal of Human Resources, 49(2), pp.359-392. 

De Keyser, L. and Van Rossem, R., (2017). Birth order and height development in early 

childhood: empirical evidence for twenty-first-century Flanders. The History of the 

Family, 22(2-3), pp.236-257. 

Deininger, K., (2003).  Does cost of schooling affect enrolment by the poor? Universal primary 

education in Uganda. Economics of Education review, 22(3), pp.291-305. 

Del Bono, E., Ermisch, J. and Francesconi, M., (2012). Intrafamily resource allocations: a 

dynamic structural model of birth weight. Journal of Labour Economics, 30(3), pp.657-706. 

Demographic and Health Survey., (2023), Available at: https://dhsprogram.com/data/Guide-

to-DHS-Statistics/Nutritional_Status.htm; 

https://microdata.worldbank.org/index.php/catalog/688/data-dictionary/F1?file_name=Child 

). [Accessed: 5 July 2023] 

Dercon, S. and Porter, C., (2014). Live Aid Revisted: Long-term impacts of the 1984 Ethiopian 

Famine on Children. Journal of the European Economic Association, 12(4), pp. 927-948.  

Dimitrova, A. and Bora, J.K., (2020). Monsoon weather and early childhood health in 

India. PloS one, 15(4), p.e0231479. 

http://www.lse.ac.uk/GranthamInstitute/wp-content/uploads/2015/05/GUYANA.pdf
https://dhsprogram.com/data/Guide-to-DHS-Statistics/Nutritional_Status.htm
https://dhsprogram.com/data/Guide-to-DHS-Statistics/Nutritional_Status.htm
https://microdata.worldbank.org/index.php/catalog/688/data-dictionary/F1?file_name=Child


 
 

162 
 

Douglas, B., (1995). 'Global Sea level change: Determination and interpretation', Revs. 

Geophys. (supp.), pp. 1425-1432. 

Driessen, G., (2000). The limits of educational policy and practice? The case of ethnic 

minorities in The Netherlands. Comparative education, 36(1), pp.55-72. 

Duflo, E., (2001). Schooling and labour market consequences of school construction in 

Indonesia: Evidence from an unusual policy experiment. American economic review, 91(4), 

pp.795-813. 

 

Duflo, E., (2004). The medium run effects of educational expansion: Evidence from a large 

school construction program in Indonesia. Journal of Development Economics, 74(1), pp.163-

197. 

Duflo, E., Dupas, P. and Kremer, M., (2021). The impact of free secondary education: 

Experimental evidence from Ghana (No. w28937). National Bureau of Economic Research. 

Duque, V., Rosales-Rueda, M. and Sanchez, F., (2018) May. How do early-life shocks interact 

with subsequent human-capital investments? Evidence from administrative data. In IZA World 

of Labor Conference. 

Edoka, I.O., (2013). Weather shocks and nutritional status of disadvantaged children in 

Vietnam. Health Econometrics and Data Group, The University of York, WP, 13(10). 

Eichhorst, W., Rodríguez-Planas, N., Schmidl, R. and Zimmermann, K.F., (2015). A road map 

to vocational education and training in industrialized countries. Ilr Review, 68(2), pp.314-337. 

Elliott, B. A., (1992). Birth order and health: Major issues. Social Science and Medicine, 35(4), 

443–452. 

Enu, P., Hagan, E., Ahouandjinou, E. and Attah-Obeng, P., (2014). Relationship between 

education and wage differentials in Ghana: a case study of Accra-a Suburb of greater Accra 

Region. International Journal of Academic Research in Business and Social Sciences, 4(1), 

p.277. 

 

Esposito, L., Kumar, S.M. and Villaseñor, A., (2020). The importance of being earliest: birth 

order and educational outcomes along the socioeconomic ladder in Mexico. Journal of 

Population Economics, 33(3), pp.1069-1099. 

ExxonMoble., (2015). Available at: ExxonMobil announces significant oil discovery offshore 

Guyana  [Accessed: 5th June 2022] 

 

ExxonMoble., (2022). Available at: Guyana project overview | ExxonMobil[Accessed: 5th June 

2022] 

Gafar, J., (1996). Guyana: From Cooperative Socialism to Economic Liberalization and 

Growth: 1976-1994. The Journal of Developing Areas, 31(1), pp.41-74. 

Garlick, R., (2013). How price sensitive is primary and secondary school enrolment? Evidence 

from nationwide tuition fee reforms in South Africa. Unpublished working paper. Available 

at: http://pubdocs.worldbank.org/en/3241466186067633/garlickschoolfees.pdf   

https://corporate.exxonmobil.com/Locations/Guyana/News-releases/ExxonMobil-announces-significant-oil-discovery-offshore-Guyana
https://corporate.exxonmobil.com/Locations/Guyana/News-releases/ExxonMobil-announces-significant-oil-discovery-offshore-Guyana
https://corporate.exxonmobil.com/Locations/Guyana/Guyana-project-overview
http://pubdocs.worldbank.org/en/3241466186067633/garlickschoolfees.pdf


 
 

163 
 

 

Gill-Marshall, B., (2000). Child abuse in Guyana: A study of teacher abuse of children 

(Unpublished thesis). University of Guyana, Guyana. 

Glewwe, P. and King, E., (2001). The Impact of Early Childhood Nutritional Status on 

Cognitive Development: Does the Timing of Malnutrition Matter? The World Bank Economic 

Review, 15(1), pp. 81-113.  

Gluckman, P.D. and Hanson, M.A., (2004), October. Maternal constraint of fetal growth and 

its consequences. In Seminars in fetal and neonatal medicine (Vol. 9, No. 5, pp. 419-425). WB 

Saunders. 

Greenwood, B., (2014). The contribution of vaccination to global health: past, present and 

future. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1645), 

p.20130433. 

 

Gribble, K.D., McGrath, M., MacLaine, A. and Lhotska, L., (2011). Supporting breastfeeding 

in emergencies: protecting women's reproductive rights and maternal and infant 

health. Disasters, 35(4), pp.720-738. 

 

Guyana (2000, 2006-07 and 2014) Multiple Indicator Cluster Survey. Available at: 

https://www.unicef.org/statistics/index_24302.html [Accessed 9th May 2020] 

 

Gafar, J. (2003) Guyana: From state control to free markets. Nova Publishers. 

 

Guyana Bureau of Statistics., (2023), Available at: Road map15 (rev7June2016).cdr 

(statisticsguyana.gov.gy)  (Accessed: 30 May, 2023). 

Guyana Bureau of Statistics., (2022), Available at: https://statisticsguyana.gov.gy/ [Accessed: 

30 August 2022] 

Guyana Bureau of Statistics., (2023), Available at: https://statisticsguyana.gov.gy/wp-

content/uploads/2019/10/2012_Preliminary_Report.pdf [Accessed: 5 June, 2023] 

Guyana Chroncile., (2011). Available at: Race relations as a factor in national development 

(Part III) - Guyana Chronicle [Accessed: 4th July 2022] 

 

Guyana Demographic and Health Survey., (2009). Available at: 

https://dhsprogram.com/publications/publication-fr232-dhs-final-reports.cfm [Accessed: 

January 2019] 

Guyana, Gov., (2012). Guyana Second National Communication to the United Nations 

Framework Convention on Climate Change, pp. 1-437 

 

Guyana Labour Force Survey, Methodological Report., (2018). Available at:  

https://sistemasintegrales.cl/project/guyana-labor-force-survey/ [Accessed: 20 August 2021] 

Guyana National Land Use Plan (2013), Available at: 

https://observatorioplanificacion.cepal.org/sites/default/files/plan/files/2013%20NLUP%20G

uyana.pdf [Accessed: 15 July, 2023]  

Guyana Ministry of Health., (2021). Available at: https://www.health.gov.gy/ [Accessed: Jan 

https://www.unicef.org/statistics/index_24302.html
https://statisticsguyana.gov.gy/wp-content/uploads/2019/11/Final_2012_Census_Compendium2.pdf
https://statisticsguyana.gov.gy/wp-content/uploads/2019/11/Final_2012_Census_Compendium2.pdf
https://statisticsguyana.gov.gy/
https://statisticsguyana.gov.gy/wp-content/uploads/2019/10/2012_Preliminary_Report.pdf
https://statisticsguyana.gov.gy/wp-content/uploads/2019/10/2012_Preliminary_Report.pdf
https://guyanachronicle.com/2011/02/27/race-relations-as-a-factor-in-national-development-part-iii/
https://guyanachronicle.com/2011/02/27/race-relations-as-a-factor-in-national-development-part-iii/
https://dhsprogram.com/publications/publication-fr232-dhs-final-reports.cfm
https://sistemasintegrales.cl/project/guyana-labor-force-survey/
https://observatorioplanificacion.cepal.org/sites/default/files/plan/files/2013%20NLUP%20Guyana.pdf
https://observatorioplanificacion.cepal.org/sites/default/files/plan/files/2013%20NLUP%20Guyana.pdf
https://www.health.gov.gy/


 
 

164 
 

2021] 

Guyana., (2002). Population and Housing Census, Guyana Bureau of Statistics, 2021. 

Available at: https://statisticsguyana.gov.gy/publications/#cen2002  [Accessed: January 2021] 

Guyana Times., (2016). Available at: https://guyanatimesgy.com/1-in-4-amerindian-children-

considered-stunted-unicef/ [Accessed: 20th June 2022] 

 

Hallegatte, S., Vogt-Schilb, A., Rozenberg, J., Bangalore, M. and Beaudet, C., 2020. From 

poverty to disaster and back: A review of the literature. Economics of Disasters and Climate 

Change, 4, pp.223-247. 

 

Harmon, C., Oosterbeek, H. and Walker, I., (2003) The returns to education: 

Microeconomics. Journal of economic surveys, 17(2), pp.115-156. 

 

Heckman, J.J., Lochner, L.J. and Todd, P.E., (2006) Earnings functions, rates of return and 

treatment effects: The Mincer equation and beyond. Handbook of the Economics of 

Education, 1, pp.307-458. 

Hoddinott, J. and Kinsey, B., (2001). Child Growth in the Time of Drought. Oxford Bulletin of 

Economics and Statistics, 63(4), pp. 409-436.  

Human Development Report., (2010), Available at: 

https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-

english.human-development-report-2010-complete-english[Accessed: 5 Aug, 2023] 

Human Development Report., (2016), Available at: 

https://hdr.undp.org/system/files/documents/2016humandevelopmentreportpdf1pdf.pdf 

[Accessed: 5 Aug, 2023] 

Hunt-McCool, J. and Bishop, D.M., (1998). Health economics and the economics of education: 

Specialization and division of labor. Economics of Education Review, 17(3), pp.237-244. 

International Labour Organisation., (2004). National Labour Law Profile: Guyana, Available 

at: https://www.ilo.org/ifpdial/information-resources/national-labour-law-

profiles/WCMS_158900/lang--en/index.htm (Accessed: 5 Sept 2022). 

 

İşcan, T.B., Rosenblum, D. and Tinker, K., (2015). School fees and access to primary 

education: Assessing four decades of policy in sub-Saharan Africa. Journal of African 

Economies, 24(4), pp.559-592. 

 

Isungset, M.A., Freese, J., Andreassen, O.A. and Lyngstad, T.H., (2022). Birth Order 

Differences in Education Originate in Post-Natal Environments. bioRxiv, pp.2021-06. 

Jarry, V., Gagnon, A. and Bourbeau, R., (2013). Maternal age, birth order and other early-life 

factors: a family-level approach to exploring exceptional survival. Vienna Yearbook of 

Population Research, pp.267-294. 

Jayachandran, S., (2006). Selling Labor Low: Wage Responses to Productivity Shocks in 

Developing Countries. Journal of Political Economy, 114(3), pp. 538-575. 

https://statisticsguyana.gov.gy/publications/#cen2002
https://guyanatimesgy.com/1-in-4-amerindian-children-considered-stunted-unicef/
https://guyanatimesgy.com/1-in-4-amerindian-children-considered-stunted-unicef/
https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-english.human-development-report-2010-complete-english
https://hdr.undp.org/system/files/documents/human-development-report-2010-complete-english.human-development-report-2010-complete-english
https://hdr.undp.org/system/files/documents/2016humandevelopmentreportpdf1pdf.pdf
https://www.ilo.org/ifpdial/information-resources/national-labour-law-profiles/WCMS_158900/lang--en/index.htm
https://www.ilo.org/ifpdial/information-resources/national-labour-law-profiles/WCMS_158900/lang--en/index.htm


 
 

165 
 

Jayachandran, S. and Pande, R., (2017). Why are Indian children so short? The role of birth 

order and son preference. American Economic Review, 107(9), pp.2600-2629. 

Jennings, Z. and Miller, E., (1999). Educational reform in Guyana in the post-war 

period. Educational reform in the Commonwealth Caribbean, pp.149-198. 

 

Jones, D., Molitor, D. and Reif, J., (2019). What do workplace wellness programs do? Evidence 

from the Illinois workplace wellness study. The Quarterly Journal of Economics, 134(4), 

pp.1747-1791. 

 

Kaieteur News., (2020). Available at: The Breadbasket of the Caribbean – Kaieteur News 

(kaieteurnewsonline.com) [Accessed: 5th July 2022] 

Kaieteur News., (2014). Available at: Poor sanitation and stunting in Guyana – Kaieteur News 

(kaieteurnewsonline.com) [Accessed: 10th July 2022] 

Kampelmann, S., Rycx, F., Saks, Y. and Tojerow, I. (2018). Does education raise productivity 

and wages equally? The moderating role of age and gender. IZA Journal of Labour 

Economics, 7(1), pp.1-37. 

Kan, S. and Klasen, S., (2021). Evaluating universal primary education in Uganda: School fee 

abolition and educational outcomes. Review of Development Economics, 25(1), pp.116-147 

Khemraj, T., (2016). The political economy of Guyana's underdevelopment. The Review of 

Black Political Economy, 43(3-4), pp.325-342. 

Khong, T.Y., Adema, E.D. and Erwich, J.J.H., (2003). On an anatomical basis for the increase 

in birth weight in second and subsequent born children. Placenta, 24(4), pp.348-353. 

Kramer, M., (1987). Determinants of low birth weight: methodological assessment and meta-

analysis. Bulletin of the world health organization, 65(5), pp. 663. 

Kramer, M., Aboud, F., Mironova, E., Vanilovich, I., Platt, R.W., Matush, L., Igumnov, S., 

Fombonne, E., Bogdanovich, N., Ducruet, T. and Collet, J., (2008). Breastfeeding and child 

cognitive development: new evidence from a large randomized trial. Archives of general 

psychiatry, 65(5), pp. 578-584. 

Kumar, S., Molitor, R. and Vollmer, S., (2016). Drought and early child health in rural 

India. Population and Development Review, pp.53-68. 

 

Lakhan, V.C., (2014). Use of geospatial techniques to relocate flood-prone coastal 

communities in Guyana. Contract Report No. 11. Caribbean Research Solutions, Guyana and 

Toronto 

 

Le, K. and Nguyen, M., (2021). In-utero exposure to rainfall variability and early childhood 

health. World Development, 144, p.105485. 

Lee, B. E., E. H. Ha, H. S. Park, Y. J. Kim, Y. C. Hong, H. Kim, and J. T. Lee., (2003). 

“Exposure to Air Pollution during Different Gestational Phases Contributes to Risks of Low 

Birth Weight.” Human Reproduction 18 (3): 638–43 

https://www.kaieteurnewsonline.com/2020/08/12/the-breadbasket-of-the-caribbean/
https://www.kaieteurnewsonline.com/2020/08/12/the-breadbasket-of-the-caribbean/
https://www.kaieteurnewsonline.com/2014/08/16/poor-sanitation-and-stunting-in-guyana/
https://www.kaieteurnewsonline.com/2014/08/16/poor-sanitation-and-stunting-in-guyana/


 
 

166 
 

Lehmann, J.Y.K., Nuevo-Chiquero, A. and Vidal-Fernandez, M., (2018). The early origins of 

birth order differences in children’s outcomes and parental behaviour. Journal of Human 

Resources, 53(1), pp.123-156. 

Leight, J., Glewwe, P. and Park, A., (2015). The impact of early childhood rainfall shocks on 

the evolution of cognitive and non-cognitive skills.  

Leroy, J.L., Ruel, M., Habicht, J.P. and Frongillo, E.A., (2015). Using height-for-age 

differences (HAD) instead of height-for-age z-scores (HAZ) for the meaningful measurement 

of population-level catch-up in linear growth in children less than 5 years of age. BMC 

pediatrics, 15(1), pp.1-11. 

 

López, R.E., Thomas, V. and Troncoso, P., (2015). Climate change and natural 

disasters. Santiago. 

 

London, T., (2018) Assessing Competing Perspectives: A Critical Analysis of Guyana's 

National Grade Six Assessment. Journal of Pan African Studies, 11(4). 

Maccini, S. and Yang, D., (2009). Under the Weather: Health, Schooling, and Economic 

Consequences of Early-Life Rainfall. American Economic Review, 99(3), pp. 1006-1026.  

Mach, O., Lu, L., Creek, T., Bowen, A., Arvelo, W., Smit, M., Masunge, J., Brennan, M. and 

Handzel, T., (2009). Population-based study of a widespread outbreak of diarrhea associated 

with increased mortality and malnutrition in Botswana, Jan–Mar 2006. The American journal 

of tropical medicine and hygiene, 80(5), pp.812-818. 

 

Marjoribanks, K., 1999. Ethnicity, birth order, and family environment. Psychological 

Reports, 84(3), pp.758-760. 

 

Malamud, O. and Pop-Eleches, C., (2010). General education versus vocational training: 

Evidence from an economy in transition. The review of economics and statistics, 92(1), pp.43-

60. 

 

Massing, N. and Schneider, S.L., (2017). Degrees of competency: the relationship between 

educational qualifications and adult skills across countries. Large-Scale Assessments in 

Education, 5(1), pp.1-34. 

Ministry of Education, Guyana., (2022). Available at: Ministry of Education, Guyana 

[Accessed: 5th Jan 2022] 

Ministry of Communities., (2016). Capital Towns and Economic Development. Guyana: 

communities. Available at: https://moc.gov.gy/capital-towns-and-economic-development 

[Accessed 10th April 2020]. 

Ministry of Health, Guyana., (2022). Available at: Ministry of Health - Home [Accessed: 25th 

June 2022] 

Mmopelwa, D., (2019). Household size, birth order and child health in Botswana (No. 19/10). 

CREDIT Research Paper. 

https://education.gov.gy/en/
https://moc.gov.gy/capital-towns-and-economic-development
https://health.gov.gy/


 
 

167 
 

Murnane, R.J., Waldman, M.R., Willett, J.B., Bos, M.S. and Vegas, E., (2017) The 

consequences of educational voucher reform in Chile (No. w23550). National Bureau of 

Economic Research. 

 

Myrskylä, M., Silventoinen, K., Jelenkovic, A., Tynelius, P. and Rasmussen, F., (2013). The 

association between height and birth order: evidence from 652 518 Swedish men. J Epidemiol 

Community Health, 67(7), pp.571-577. 

Nakamuro, M., Uzuki, Y. and Inui, T., (2013). The effects of birth weight: Does fetal origin 

really matter for long-run outcomes? Economics Letters, 121(1), pp.53-58. 

National Institute for Health Care Excellence., (2017). Available at: https://www.nice.org.uk/ 

[Accessed:1st May 2020] 

 

Ogasawara, K. and Yumitori, M., (2019). Early-life exposure to weather shocks and child 

height: Evidence from industrializing Japan. SSM-population health, 7, p.100317 

 

OilNow., (2022). Available at: Guyana is number 17 for largest oil reserves in the world | 

OilNOW[Accessed: 15th June 2022] 

 

Organisation for Economic Co-operation and Development., (2000).  Literacy in the 

information age: final report of the International Adult Literacy Survey. OECD Publishing. 

Ozier, O., (2018). The Impact of Secondary Schooling in Kenya A Regression Discontinuity 

Analysis. Journal of Human Resources, 53(1), pp.157-188. 

 

Peet, E.D., Fink, G. and Fawzi, W., (2015). Returns to education in developing countries: 

Evidence from the living standards and measurement study surveys. Economics of Education 

Review, 49, pp.69-90. 

 

Pörtner, H.O., Roberts, D.C., Adams, H., Adler, C., Aldunce, P., Ali, E., Begum, R.A., Betts, 

R., Kerr, R.B., Biesbroek, R. and Birkmann, J., (2022). Climate change 2022: Impacts, 

adaptation and vulnerability. IPCC Sixth Assessment Report. 

 

Psacharopoulos, G. (1994). Returns to investment in education: A global update. World 

development, 22(9), pp.1325-1343. 

 

Psacharopoulos, G. and Patrinos, H.A., (2018). Returns to investment in education: a decennial 

review of the global literature. Education Economics, 26(5), pp.445-458. 

 

Pruckner, G.J., Schneeweis, N., Schober, T. and Zweimüller, M., (2021). Birth order, parental 

health investment, and health in childhood health care utilization. Journal of Health 

Economics, 76, pp.102426. 

Quinn, D., Lavigne, S.V., Chambers, C., Wolfe, L., Chipman, H., Cragan, J.D. and Rasmussen, 

S.A., (2008). Addressing concerns of pregnant and lactating women after the 2005 hurricanes: 

the OTIS response. MCN: The American Journal of Maternal/Child Nursing, 33(4), pp.235-

241. 

 

Rahman, M., (2016). Association between order of birth and chronic malnutrition of children: 

a study of nationally representative Bangladeshi sample. Cadernos de saude publica, 32, 

p.e00011215. 

https://www.nice.org.uk/
https://oilnow.gy/featured/worlds-largest-oil-reserves-by-country/
https://oilnow.gy/featured/worlds-largest-oil-reserves-by-country/


 
 

168 
 

 

Randell, H., Gray, C. and Grace, K., (2020). Stunted from the start: Early life weather 

conditions and child undernutrition in Ethiopia. Social Science & Medicine, p.113234. 

 

Riphahn, R.T., (2012). Effect of secondary school fees on educational attainment. The 

Scandinavian Journal of Economics, 114(1), pp.148-176. 

Rocha, R. and Soares, R., (2015). Water scarcity and birth outcomes in the Brazilian 

semiarid. Journal of Development Economics, 112, pp. 72-91. 

Rocha, R. and Soares, R., (2011). September. Early life health and educational performance: 

evidence from rainfall fluctuations in the Brazilian northeast. In 33º Meeting of the Brazilian 

Econometric Society. 

Rosales-Rueda, M., (2018). The impact of early life shocks on human capital formation: 

evidence from El Niño floods in Ecuador. Journal of Health Economics, 62, pp. 13-44.  

Rose, S. and Corbin, H., (2017). Prospects and constraints of rehabilitating and expanding 

Guyana’s road network, ISSN 15169111, Papers Donaea, 371. 

Rözer, V., Mehryar, S. and Surminski, S., 2022. From managing risk to increasing resilience: 

a review on the development of urban flood resilience, its assessment and the implications for 

decision making. Environmental Research Letters. 

Ryan., (2020), Guyana elections 2020: a crude awakening. Available at: Guyana elections 

2020: a crude awakening? | LSE Latin America and Caribbean  [Accessed: 25th June 2022] 

Sahn, D.E. and Stifel, D.C., (2002). Robust comparisons of malnutrition in developing 

countries. American Journal of Agricultural Economics, 84(3), pp.716-735. 

Saldarriaga, V., (2015). Birth weight and early childhood physical health: evidence from a 

sample of Latin American twins. Economía, 15(2), pp.161-197. 

Salm, M. and Schunk, D., (2012). The relationship between child health and developmental 

gaps, and parental education: Evidence from administrative data. Journal of the European 

Economic Association, 10(6), pp. 1425-1449. 

Schultz, T.P., (2004) School subsidies for the poor: evaluating the Mexican Progresa poverty 

program. Journal of development Economics, 74(1), pp.199-250. 

 

Selvin, S. and Janerich, D.T., (1971). Four factors influencing birth weight. British journal of 

preventive and social medicine, 25(1), p.12. 

Shavit, Y. and Muller, W., (2000). Vocational secondary education. European societies, 2(1), 

pp.29-50. 

Sianesi, B. and Van Reenen, J., (2002). The returns to education: a review of the empirical 

macro-economic literature. 

 

Skoufias, E., Vinha, K. and Conroy, H.V., (2011). The impacts of climate variability on welfare 

in rural Mexico. The World Bank. 

https://blogs.lse.ac.uk/latamcaribbean/2020/02/27/guyana-elections-2020-a-crude-awakening/
https://blogs.lse.ac.uk/latamcaribbean/2020/02/27/guyana-elections-2020-a-crude-awakening/


 
 

169 
 

Smith, R.C., Ainley, D., Baker, K., Domack, E., Emslie, S., Fraser, B., Kennett, J., Leventer, 

A., Mosley-Thompson, E., Stammerjohn, S. and Vernet, M., (1999). Marine ecosystem 

sensitivity to climate change: Historical observations and paleoecological records reveal 

ecological transitions in the Antarctic Peninsula region. BioScience, 49(5), pp. 393-404. 

Stein, A.D. and Lumey, L.H., (2000). The relationship between maternal and offspring birth 

weights after maternal prenatal famine exposure: the Dutch Famine Birth Cohort 

Study. Human biology, pp.641-654. 

Study on Indigenous Women and Children in Guyana,  

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-

in-Guyana.pdf   

 

Su, C.T., McMahan, R.D., Williams, B.A., Sharma, R.K. and Sudore, R.L., 2014. Family 

matters: Effects of birth order, culture, and family dynamics on surrogate decision‐

making. Journal of the American Geriatrics Society, 62(1), pp.175-182. 

 

Thomson Reuters Foundation News., (2011). Available at 

https://news.trust.org/item/20110601141800-sh9ro  [Accessed: 15th May 2020] 

 

UNESCO, I., (2006/07). World data on education: sixth edition. Principles and General 

Objectives of Education. 

 

UNESCO, I., (2010/11). World data on education: seventh edition. Principles and General 

Objectives of Education. 

UNICEF., (2019). Available at: SOWC-2019.pdf (unicef.org) [Accessed: 1st July 2022] 

UNECIF., (2017). The study of indigenous women and children in Guyana, Available at: ,   

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-

in-Guyana.pdf) (Accessed: 20 October 2021). 

United Nation International Children’ Emergency Fund (UNICEF) (2007). “Child & 

Development 2005 Survey (Multiple Indicators Cluster Survey 3) Final Report”, Ulaanbaatar.  

United Nation International Children’ Emergency Fund (UNICEF) (2020). 

https://data.unicef.org/country/guy/ [Accessed: 1st Sept 2020] 

 

United Nations., (2019), Unprecedented Impacts of Climate Change Disproportionately 

Burdening Developing Countries, Delegate Stresses, as Second Committee Concludes General 

Debate, https://www.un.org/press/en/2019/gaef3516.doc.htm [ Accessed: 10th Jan 2021] 

 

University of Guyana., (2022). Available at: https://www.uog.edu.gy/ (Accessed: 9 April 

2022). 

Valle., (2022). Available at: Exxon makes three new oil discoveries in Guyana and boosts 

reserves | Reuters [Accessed: 15th June 2022] 

Victora C.G., Adair L., Fall C. Maternal and child undernutrition: Consequences for adult 

health and human capital. Lancet., (2008);371:340–357. [PMC free article] [PubMed] [Google 

Scholar] 

https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://news.trust.org/item/20110601141800-sh9ro
https://www.unicef.org/media/63016/file/SOWC-2019.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://www.unicef.org/guyanasuriname/media/451/file/SitAn-Indigenous-women-children-in-Guyana.pdf
https://data.unicef.org/country/guy/
https://www.un.org/press/en/2019/gaef3516.doc.htm
https://www.uog.edu.gy/
https://www.reuters.com/business/energy/exxon-makes-three-new-discoveries-guyana-increases-oil-reserves-2022-04-26/
https://www.reuters.com/business/energy/exxon-makes-three-new-discoveries-guyana-increases-oil-reserves-2022-04-26/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2258311/
https://pubmed.ncbi.nlm.nih.gov/18206223
https://scholar.google.com/scholar_lookup?journal=Lancet&title=Maternal+and+child+undernutrition:+Consequences+for+adult+health+and+human+capital&author=C.G.+Victora&author=L.+Adair&author=C.+Fall&volume=371&publication_year=2008&pages=340-357&pmid=18206223&
https://scholar.google.com/scholar_lookup?journal=Lancet&title=Maternal+and+child+undernutrition:+Consequences+for+adult+health+and+human+capital&author=C.G.+Victora&author=L.+Adair&author=C.+Fall&volume=371&publication_year=2008&pages=340-357&pmid=18206223&


 
 

170 
 

 

Victora, C., Vaughan, J.P., Lombardi, C., Fuchs, S.C., Gigante, L., Smith, P., Nobre, L., 

Teixeira, A., Moreira, L. and Barros, F., (1987). Evidence for protection by breast-feeding 

against infant deaths from infectious diseases in Brazil. The Lancet, 330(8554), pp.319-322. 

World Atlas., (2022). Available at: What Are The Major Natural Resources Of Guyana? - 

WorldAtlas [Accessed: 1st July 2022] 

World Bank., (2022). Available at: GDP per capita (current US$) - Guyana | Data 

(worldbank.org) [Accessed: 1st July 2022] 

World Bank Data., (2023), Available at:  

https://data.worldbank.org/indicator/SH.DYN.MORT?end=2014&locations=XO&start=2000 

[Accessed: 5 June 2023] 

World Bank Report., (2005), Available at: 

https://www.gfdrr.org/sites/default/files/publication/pda-2005-guyana.pdf [Accessed: 30 May, 

2023] 

World Bank, GFDRR., (2014), Available at 

https://www.gfdrr.org/en/floodpreventioninguyana [Accessed: 5 June 2023] 

World Bank., (2022), Supporting recovery and strengthening flood risk management in 

Guyana. Available at: World Bank Document [Accessed: 25th June 2022] 

 

World Bank., (2020), Available at: https://data.worldbank.org/country/GY [Accessed: 10th 

May 2020] 

 

World Bank Climate Change Knowledge Portal., (2020), Available at: 

https://climateknowledgeportal.worldbank.org/ [Accessed: 10th June 2020] 

World Health Organisation., (2020). Global Database on Child Growth and Malnutrition. 

Who.int. Available at: https://www.who.int/nutgrowthdb/about/introduction/en/index5.html. 

[Accessed date: 6th April. 2020]. 

World Bank., (2008) Universal Secondary Education (USE) in Guyana, Available at:  

https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941epo

rt0by0Geoff0Howse.pdf?sequence=1&isAllowed=y (Accessed: 20 August 2021).  

 

WHO Multicentre Growth Reference Study Group and de Onis, M., (2006). Assessment of 

differences in linear growth among populations in the WHO Multicentre Growth Reference 

Study. Acta Paediatrica, 95, pp.56-65. 

World Health Organization, (2019). UNICEF-WHO low birthweight estimates: levels and 

trends 2000-2015 (No. WHO/NMH/NHD/19.21). World Health Organization. 

 

Yakusheva, O. and Fletcher, J., (2015). Learning from teen childbearing experiences of close 

friends: Evidence using miscarriages as a natural experiment. Review of Economics and 

Statistics, 97(1), pp.29-43. 

https://www.worldatlas.com/articles/what-are-the-major-natural-resources-of-guyana.html
https://www.worldatlas.com/articles/what-are-the-major-natural-resources-of-guyana.html
https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=GY
https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=GY
https://data.worldbank.org/indicator/SH.DYN.MORT?end=2014&locations=XO&start=2000
https://www.gfdrr.org/sites/default/files/publication/pda-2005-guyana.pdf
https://www.gfdrr.org/en/floodpreventioninguyana
https://documents1.worldbank.org/curated/en/099307104052251719/pdf/IDU0468c78c10092804f1c082d004de6dfcd66cc.pdf
https://data.worldbank.org/country/GY
https://climateknowledgeportal.worldbank.org/
https://www.who.int/nutgrowthdb/about/introduction/en/index5.html
https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941eport0by0Geoff0Howse.pdf?sequence=1&isAllowed=y
https://openknowledge.worldbank.org/bitstream/handle/10986/12327/535160ESW0P0941eport0by0Geoff0Howse.pdf?sequence=1&isAllowed=y


 
 

171 
 

Ye, C., (2021) March. How Does Education Affect the Time Spending on Work and Leisure? 

In 6th International Conference on Financial Innovation and Economic Development (ICFIED 

2021) (pp. 274-280). Atlantis Press.