







































 
 

 

9 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 
 

Economy 
Vol. 9, No. 1, 9-17, 2022 

ISSN(E) 2313-8181/ ISSN(P) 2518-0118 
DOI: 10.20448/economy.v9i1.4104 

© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 
 

 
 
 
Education and Livelihood in Ghana: Evidence from Instrumental Variable Analysis 

 
Samuel OSEI-GYEBI   

 

 
 

 
Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. 
Email: Kwakuoseigyebi@gmail.com Tel: +234-80106629848 

 
Abstract 

The research verified if schooling makes any difference in the consumption expenditure and 
poverty status of individuals in Ghana. Using the recent round of the Ghana Living Standards 
Survey data, we quantified the difference that schooling makes in the livelihood of individuals 
through the method of instrumental variables (IV). We found that individuals who have 
completed some level of schooling experience a significant increase in their consumption 
expenditure relative to those who have not been to school at all. Again, educated individuals who 
are poor had a significantly lower consumption expenditure compared to educated individuals who 
are non-poor. Also, individuals with some level of schooling had a lower probability to be poor 
compared to those without any schooling. We suggest massive investment in educational 
infrastructure and a revision of the schooling curriculum to equip individuals with workable skills 
that can enable them to undertake livelihood activities to improve their welfare. 

 
Keywords: Ghana, Consumption expenditure, Schooling, Livelihood, Poverty, Instrumental variables. 

JEL Classification: C12; E21; I26. 
 

Citation | Samuel OSEI-GYEBI (2022). Education and Livelihood 
in Ghana: Evidence from Instrumental Variable Analysis. Economy, 
9(1): 9-17. 
History:  
Received: 4 April 2022 
Revised: 5 July 2022 
Accepted: 21 June 2022 
Published: 11 August 2022 
Licensed: This work is licensed under a Creative Commons 

Attribution 4.0 License  
Publisher:  Asian Online Journal Publishing Group 
 

Funding: This study received no specific financial support. 
Competing Interests: The author declares that there are no conflicts of 
interests regarding the publication of this paper. 
Transparency: The author confirms that the manuscript is an honest, 
accurate, and transparent account of the study; that no vital features of the 
study have been omitted; and that any discrepancies from the study as planned 
have been explained. 
Ethical: This study followed all ethical practices during writing. 

 

 

Contents 
1. Introduction ...................................................................................................................................................................................... 10 
2. Literature Review ............................................................................................................................................................................ 10 
3. Methodology ..................................................................................................................................................................................... 12 
4. Results ................................................................................................................................................................................................ 14 
5. Conclusion ......................................................................................................................................................................................... 16 
References .............................................................................................................................................................................................. 16 
 

 
 

 

 

 

 

 

 

 

mailto:Kwakuoseigyebi@gmail.com
https://creativecommons.org/licenses/by/4.0/
https://creativecommons.org/licenses/by/4.0/
https://www.doi.org/10.20448/economy.v9i1.4104
https://orcid.org/0000-0003-1435-6334


Economy, 2022, 9(1): 9-17 

10 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Contribution of this paper to the literature 
We established the difference that schooling makes in the consumption expenditure of 
individuals in Ghana and how this influences their livelihoods. We also analyzed the interplay 
of education and poverty on the consumption expenditure of individuals in Ghana to provide a 
new perspective on the nexus of education and livelihood. 

 
1. Introduction 

Every individual requires some resources to sustain life (Adjei, 2010) and these livelihood assets are what 
people dwell on to make a living (Oduro, Mohren, Pena-Claros, Kyereh, & Arts, 2015). Different individuals go 
through their livelihood activities in different places; some undertake their activities in their homes while others go 
outside of their homes (Mahama & Maharjan, 2018). Livelihood issues are of great importance in combating 
poverty; hence the government of Ghana has initiated the Livelihood Empowerment Against Poverty (LEAP) 
program as one of the key policy directives to address the vulnerability in the country (Mahama & Maharjan, 
2018). Education contributes to the enhancement of assets and a good foundation in education for that matter 
enhances one’s efforts of building human capital to earn a decent living (Tangney, Baumeister, & Boone, 2004). 
Again, education is potent in enabling peoples’ capabilities to be enhanced to achieve sustainable livelihoods. Also, 
Sen (1990) emphasizes that education serves both as a means to a better life and also an end in itself which makes it 
important for all individuals.  

However, despite the important role education plays in livelihood, there have been concerns about how 
education is being accessed and delivered (Tangney et al., 2004). According to United Nation International 
Children’s Emergency Fund (United Nations Children’s Fund (UNICEF), 2016) approximately 124 million 
adolescents and children lack the opportunity to attend and complete school. Approximately 59 million primary 
students and 65 million adolescents in lower secondary levels are unable to complete school with more than 50% of 
this number living in Sub-Saharan Africa.  

According to Corcoran et al. (2020) no matter the efforts people put into learning, if the teaching and learning 
environment is not favorable, then the quality of education is a problem. Even though the education system has 
experienced reforms and reviews over the years, their cumulative effects have not yielded the desired goals as 
significant issues still exist (Pajibo & Tamanja, 2017). Corcoran et al. (2020) noted that Ghana has inadequate 
schools, a low teacher-to-pupil ratio, and high illiteracy rates, especially in rural parts of the country. For example, 
the centers of Ghana’s Early Childhood Education (ECE) are lacking in terms of major learning and teaching 
materials per the same report. Another challenge has been the huge opportunity cost for some parents to allow 
their children to attend school because their children serve as a significant source of labor (Debrah, 2013). This 
makes it difficult to convince most parents to enroll their children in school. 

The abundance of highly educated people undoubtedly is a prerequisite for socio-economic development. The 
effect of education on certain indicators has therefore received enormous attention across the globe. Mention can be 
made of Coelli and Green (2012) which analyzed the effect of high school education on the welfare of young 
Canadians while Little, Aboud, and Lenachuru (2009) studied how education influences the livelihood of farmers in 
Kenya. In Ghana, Sackey (2005) looked at how education affect the labor force participation and the fertility of 
women whereas Peters, Baker, Dieckmann, Leon, and Collins (2010) examined the effect of education on the 
health-protecting behaviors of people. Based on these studies and discussions, it is obvious that the direct link 
between education and livelihood in Ghana has not been sufficiently highlighted in the literature. Precisely, the 
effect of education has been examined on certain aspects of livelihood but not on a livelihood measure like the 
consumption expenditure of individuals in Ghana. Consumption forms a critical aspect of the individual’s livelihood 
and underscores the need to examine the direct effect of education on it. This will provide more insight into how 
the lives of Ghanaians can be enhanced by improving the access and quality of education.  

The contribution of this study to literature is three-fold: First, we find out if Ghanaians with some level of 
schooling significantly differ from those who have never been to school in terms of their consumption expenditure. 
To do this, we estimate the magnitude of the difference in consumption expenditure between individuals who have 
had some level of schooling and those who have not been to school at all. Second, we estimate the effect of 
education on the consumption expenditure of poor individuals in Ghana. This in particular allowed us to analyze 
the interplay of education and poverty on the consumption expenditure of individuals in Ghana. Third, we estimate 
the effect of education on the poverty status of individuals in Ghana. This objective reveals further insight into the 
role of education in influencing the poverty status of Ghanaians. 

The remainder of this paper is organized as follows. Section two reviews the literature in addition to an 
overview of the educational system in Ghana. The next section outlines the methodology adopted by the study 
whiles section four presents empirical results and its discussions. The paper is concluded in section five by 
presenting a summary of findings, conclusions, and some policy recommendations. 
 

2. Literature Review 
2.1. Theoretical Review 

Endogenous growth models with emphasis on human capital are the theoretical basis for this study. 
Endogenous growth models that stressed knowledge accumulation became necessary after classical growth 
theories failed to explain what determines the vast increase in income over time and the wide disparities between 
income levels of poor and rich countries. Essentially, endogenous models of growth focused on explaining what 
determines technical progress (Romer, 2012). For instance, Aghion and Howitt (2008) mentioned formal education 
as the first and foremost channel of knowledge accumulation. These models emphasized education because it is 
important for the development of human capital. The reason is that education has a progressive effect as individuals 
with some level of education can obtain an extra level of education which improves their human capital and makes 
them more productive. The Lucas model for instance noted that an individual with human capital h(t) is as 
productive as two individuals with half each of h(t) (Crockett, Duffy, & Izhakian, 2019). Also, education has an 
externality effect as individuals from highly educated families are able to attain more education than those from 



Economy, 2022, 9(1): 9-17 

11 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

families that are less educated. This argument is extended to include the fact that individuals from highly educated 
neighborhoods are more likely to reach higher levels of education compared to those from other neighborhoods 
(Heckman & Karapakula, 2019). 

The increase in human capital and productivity of the individual enhances their capacity to secure a meaningful 
source of livelihood. It follows that the more years of schooling attained by an individual, the higher will be their 
consumption which enhances their livelihood (Tran, Tran, & Nguyen, 2020). Based on this, education is considered 
a gateway for individuals all over the world to obtain a sustainable source of income either through formal 
employment or self-employment. Although the ascending rate of unemployment especially in developing countries 
may cause individuals not to secure a meaningful source of livelihood after several years of education, it is difficult 
to deny the wholistic improvement education brings to a person. Education remains a catalyst for economic growth 
which further enhances the livelihood of individuals. Education has been described as a defining factor for the 
difference between rich and poor countries as evidence shows that developed countries have more educated 
individuals than developing ones (Lim, 2018).  
 

2.2. Empirical Review 
The review of literature is done along two lines. We first consider studies that examined the effect of education 

on other indicators in jurisdictions other than Ghana. The second part focused on studies that investigated 
education and livelihood in Ghana. Coelli and Green (2012) found Canadians who graduate from a high school 
reduce their receipt of welfare by 50% and this effect is much more noticeable in more deprived neighborhoods. 
Tran et al. (2020) confirmed the positive effect of education on the consumption and negative effect on poverty in 
Vietnam. A crucial aspect of their study showed that the positive effect of education is higher for relatively better-
off households which suggests a form of inequality. Little et al. (2009) also noted the positive effect of education on 
the livelihood of pastoralists in Kenya as it helps them reduce the risk of famine and increase their employability. 

     Sackey (2005) found positive effects of education on labor force participation and a negative effect on the 
fertility of women. Peters et al. (2010) also found evidence that education enhances the health-protecting behaviors 
of people in Ghana. Another study is Porter et al. (2011) which noted that inadequate transport infrastructure 
affects the livelihood of young Ghanaians particularly by reducing their attendance and performance in school. 
Essentially, the paper suggests that improvement in the transport infrastructure will enhance the education of 
young people in Ghana. Cameron and Ananga (2015) also showed that saving groups improve the livelihood of 
rural households by helping them to pay for the education of their children. It further noted that households 
generate income from the loans that they source from the saving program. More recently, Mabe, Konja, Addo, and 
Awuni (2022) examined how the location and gender aspects of the LEAP program affect the education of children 
in Ghana. Their results indicate that social protection projects like the LEAP improve education for children of 
beneficiary households but this impact is higher for urban areas compared to rural areas. 

The general conclusion from the review of empirical studies is that education improves the livelihood of 
individuals through channels such as increased labor force participation, increased productivity, and incomes. It is 
also clear that an improvement in the livelihood of individuals improves the quality of education for their children. 
However, the direct effect of education on the consumption expenditure of individuals in Ghana is yet to be studied. 
Since livelihood improvement ends up with increased consumption, it is important to examine the direct effect of 
education on consumption expenditure in Ghana. 
 

2.3 Overview of the Ghanaian Educational System 
Education is widely accepted as one of the important means by which any country can improve the lives of its 

people. Like most countries in Sub-Saharan Africa (SSA), Ghana is making a lot of efforts to improve education for 
its people. Some of these initiatives include working towards the attainment of universal primary education for 
every child. Ghana commenced Free Compulsory Basic Education (FCBE) in the middle of the 1990s to give every 
Ghanaian child a free basic education. This initiative did not yield the expected results as there were still several 
Ghanaians who did not have the privilege of basic education and most of them were from poor rural households 
(Akyeampong, 2009). Akyeampong (2009) further noted that one main problem that militated against the success 
of FCUBE was that the program did not provide much incentive for some children to give up whatever it is they 
were doing for school. Apart from the fact that there were still costs incurred by parents, poor households were 
unable to allow their children to attend school and forgo their child labor. Essentially, the program did little to 
eliminate the cost associated with primary education, especially for poor households in rural areas. 

Before this, Ghana had already experienced a mammoth number of its teachers who had left for Nigeria due to 
the poor economic conditions between the late 1970s and the middle of the 1980s. This situation, coupled with 
insufficient and deteriorating infrastructure negatively affected the quality of education leading to a decline in 
primary enrollment in the late 1980s (World Bank, 2004). The educational sector received a major boost with the 
introduction of the Structural Adjustment Program (SAP) and a huge investment worth about US$ 260 million 
between 1986 and to early 2000s (Akyeampong, 2009). To further boost education by improving free compulsory 
education, the government of Ghana in collaboration with its partners introduced the school feeding program 
which was to provide one healthy meal for children in public basic schools every day of school (Akyeampong, 
2009). . The program which encouraged the production and consumption of local foods encouraged more children 
to attend school because they would not have to worry about food anymore. According to Adu-Gyamfi, Donkoh, 
and Addo (2016), one phenomenon that has characterized the educational system in Ghana is the pendulum of four 
years and three years duration of senior secondary education that accompany various changes in political power. 
This inconsistency in secondary education does not lead to the effective transfer of knowledge to the students and 
may lead to poor academic performance. 

The most recent initiative by the government of Ghana to improve education in the country is the Free Senior 
High School program. This policy was introduced in fulfillment of article 25, section 1b of the 1992 constitution of 
Ghana which states that; “secondary education in its different forms including technical and vocational education, 
shall be made generally available and accessible to all by every appropriate means, and in particular, by the 
progressive introduction of free education”. It is also in line with one of the goals of Sustainable Development 



Economy, 2022, 9(1): 9-17 

12 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Goals of the United Nations (SDG-UN) to provide free, equitable, and quality primary and secondary education to 
all boys and girls by the year 2030 (United Nations Development Group (UNDP), 2016). This program has 
increased the number of students in second-cycle institutions and it’s widely regarded as a step in the right 
direction as far as providing accessible and quality education for all is concerned. 

There are still problems despite these initiatives and progress made as a country concerning education. 
Problems include the huge infrastructural deficit, high pupil-to-teacher ratio, and the inability of most students to 
attain tertiary education. The student-to-teacher ratio of primary schools in Ghana is 27.30 which is relatively high 
compared to countries like Mauritius (17.94) and Cuba (8.83) (Majgaard & Mingat, 2012). The inadequate 
infrastructure within the educational system is evident as most school children especially in the rural areas have to 
learn in deplorable structures which inhibit quality teaching and learning. In some rural communities, teachers 
have no option but to teach their students under trees and no one needs further explanations as to how that will 
affect students. The phenomenon of low tertiary completion is not surprising because of the many challenges that 
affect basic education in the country. The issue is uncommon in several countries in SSA as only 5% of students 
who begin primary school can complete tertiary education in the region (Majgaard & Mingat, 2012). The 
government of Ghana needs to do more especially in rural areas and at the basic level if the country has any hopes 
of using education to transform its economy and improve the lives of its people. 
 

3. Methodology 
3.1. Data and Description of Variables 

The study employed cross-sectional data from the most recent round of the Ghana Living Standard Survey 
(GLSS). This round of GLSS was conducted in 2016/2017 and provides a wide range of information on 59,864 
Ghanaians.  GLSS is the most widely used dataset as far as cross-sectional information on Ghanaian households is 
concerned. The study, therefore, found it prudent to employ it since the aim was to estimate how education 
influences the livelihood of Ghanaians. Specifically, GLSS provides data on education and several other factors that 
influence the livelihood of households such as income, gender, location, size of household, and poverty status. 

The study used consumption expenditure as a measure of livelihood. Chambers and Conway (1992) defined 
livelihood as comprising the assets, capabilities, and activities that one requires for a living. Essentially, individuals 
employ the resources at their disposal to earn a living and it is, therefore, prudent to measure their livelihood by 
their consumption expenditure. The main aim of livelihood strategies is to provide the necessities of life such as 
food, clothing, and shelter (Gecho, Ayele, Lemma, & Alemu, 2014). The majority of individuals in developing 
countries like Ghana spend almost all their income in fulfillment of these basic needs and justify the use of total real 
consumption expenditure as a measure for their livelihood.   

Education is measured as a binary variable of 0 and 1 respectively for individuals who have not attended school 
at all and those who have completed a given level of schooling. This allowed the study to find out how individuals 
with some level of schooling significantly differ from those who have not been to school at all. The years of 
schooling of the individual have been used by studies such as Pons and Gonzalo (2002) to estimate the effect of 
schooling but this study is unable to do the same due to limited information on the years of schooling as well on the 
course years completed. Apart from education which is the main explanatory variable, the study controls for 
individual characteristics such as household size, poverty status, location, gender, and total gross income. Location 
tells whether the individual resides in an urban or rural area whereas poverty status indicates whether an 
individual is non-poor or poor (GSS, 2017). 

Table 1 depicts significant differences in individual characteristics based on their poverty status. The income 
and consumption of poor individuals were significantly lower than those who are better off. For instance, much 
better-off individuals consume about 3 times more than poor individuals highlighting the vulnerability of poor 
individuals in Ghana. Poor individuals also tend to have a relatively larger household size and are much younger. 
This implies that poverty is prevalent among young adults in Ghana, especially in the face of rising unemployment. 
The difference in course years completed by poor and non-poor individuals is small and it further shows the 
average course years completed by Ghanaians is very small. Essentially, this gives a snapshot of the state of 
education in Ghana and the need for more efforts to improve it. 
 

Table 1. Individual characteristics based on their poverty status. 

Variables Non-poor Poor Whole Sample 

Mean SD Mean SD Mean SD 

Consumption (GHS) 12,866.77 11,186.96 4,652.05 3,047.79 10,766.32 10,409.83 
Income (GHS) 33,186.5 174,681.8 8,197.1 26,606.1 26,799.1 151,693.4 
Household size 3.6 2.5 5.9 3.2 4.2 2.8 
Course years completed 2.4 1.2 2.2 1.3 2.3 1.3 
Age in years 45.2 15.7 20.7 18.5 24.9 20.3 
Gender (Female) 1.33 0.47 1.55 0.49 1.52 0.49 
Location (Rural) 1.46 0.49 1.67 0.47 1.63 0.48 
Observations 10,427  49,437  59,864  
Note: SD represents the Standard Deviation. 

 

Table 2 describes the total real consumption expenditure of individuals based on their educational status. It 
also presents the two-sample t-test for the relationship between the consumption expenditure of individuals and 
their educational status. Educated individuals were shown to consume more than those who are uneducated. 
Essentially, individuals who have completed some level of schooling consume GHS 2,165.14 more than those who 
have never been to school. Results of the two-sample t-test confirm that this difference in their consumption 
expenditure is significant at the 5% level of significance. This difference does not establish any causal effect 
between educated and uneducated individuals but points to the fact that education or schooling is an avenue for 
individuals to improve their livelihood through an increase in their consumption expenditure. 
 



Economy, 2022, 9(1): 9-17 

13 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Table 2. Cross-tabulation of individual’s consumption by their educational status. 

Educational Status Obs. Mean (GHS) Std. Err. Std. Dev. [95% conf. Interval] 

Uneducated 3,357 9120.074 124.57 7217.514 8875.834 9364.314 
Educated 10,652 11285.14 108.32 11179.53 11072.81 11497.47 
Combined 14,009 10766.32 87.95 10409.83 10593.93 10938.72 
Diff.  -2165.066 205.2358  -2567.356 -1762.777 
Two-sample t test H0: diff.=0 

H1: diff.! =0 
t=-10.549 

Df. = 14007 
Pr. (|T|>|t|) =0.000 

 
3.2. Instrumental Variable Estimation of the Effect of Education on Livelihood 

Following Pons and Gonzalo (2002), the IV model for the effect of education on the livelihood of individuals is 
given as; 

𝐶𝑖 = 𝑎 + 𝑋𝑖𝛼 + 𝑆𝑖𝛽 + 휀𝑖                                                                           (1) 
  𝑆𝑖 = 𝑞 + 𝑍𝑖𝜌 + 𝑣𝑖                                                                                     (2) 

Equation 1 shows the total real consumption expenditure of individuals (𝐶𝑖) is explained by a vector of 

exogenous variables (𝑋𝑖) and (𝑆𝑖) which indicates if the individual has completed any form of schooling or not.  (𝑆𝑖) 
is 1 if the individual has completed some level of education and 0 if he/she has not been to school at all. Equation 2 

presents the schooling equation in a reduced form in which 𝑍𝑖 is a vector of exogenous variables that influences the 

schooling decision so that 𝑋𝑖  is included in 𝑍𝑖 . a and q are intercepts of the two equations. Indeed, the decision to 
attend school and complete a given course of study is not exogenous and therefore requires the use of relevant 

instruments. The reason is that estimating Equation 1 by OLS in the presence of correlation between ε and v leads 

to a biased estimate of β.  
The second objective was to find the effect of education on the consumption expenditure of poor individuals in 

Ghana considering the relevance of poverty as far as livelihood issues are concerned. This requires the interaction 
of the schooling variable and the variable for poverty. The interacted variable will then be included in Equation 1 
as; 

𝐶𝑖 = 𝑎 + 𝑋𝑖𝛼 + 𝑆𝑖𝛽 + (𝑆𝑖 ∗ 𝑃𝑆𝑖)𝜃 + 𝑒𝑖                                                                         (3) 
Equation 3 expresses the consumption expenditure of poor individuals as a function of a vector of exogenous 

variables (𝑋𝑖) and (𝑆𝑖) which indicates if the individual has completed any form of schooling or not plus the 

interaction of the schooling variable and the variable for poverty. 𝜃 represents the effect of the interacted variable 

for schooling (𝑆𝑖) and poverty status (𝑃𝑆𝑖) of individuals in Ghana. The endogeneity of schooling implies the 
interacted variable is also endogenous and the instrumental variable method was employed to find out how 

education impacts the consumption of poor Ghanaians. 𝑒𝑖 is the error term, and the definition of other variables 
remained the same. 

To further delve into the effect of education in enhancing the livelihood of Ghanaians, the third objective was 
to estimate the effect of schooling on the poverty status of individuals. This was done in an instrumental variable 
probit (IV probit) framework as the dependent variable was binary. Following Haji and Legesse (2017) the study 
specifies the poverty status of the individuals as a function of their schooling and other significant variables as; 

Pr(𝑃𝑆𝑖 ≠ 0|𝑋𝑖) = 𝛷(𝑋𝑖𝛽)                                                                                               (4) 
Equation 4 expresses the poverty likelihood of individuals as a function of their schooling and other significant 

variables. 𝑃𝑆𝑖 is poverty status of the individual which is the dependent variable, Φ is the cumulative standard 

normal and 𝑋𝑖 is a vector of explanatory variables which includes the instrumented variable for schooling (𝑆𝑖). The 
poverty status of the individual is a binary variable with a value of 1 representing poor and 0 for non-poor. The 
probit model, therefore, assumes how different the effect of education on the poor is from the non-poor. In this case, 
how likely is the poverty status of educated individuals to differ from those who are uneducated (Wooldridge, 
2019). 

Information on the family background has been used as an instrument for schooling by several studies. Pons 
and Gonzalo (2002) for instance, used the education of parents and their occupation as part of instruments to 
estimate the effect of schooling on wages. This study follows in that direction and employs the educational level of 
the father and mother as instruments for the education of the individual. Pons and Gonzalo (2002) are one of many 
studies that have asserted the significance of parents’ education to the educational achievement of the individual. 
Individuals whose parents have a higher level of education tend to have significantly longer years of schooling. The 
reverse is true for individuals whose parents have a low level of education significantly and have shorter years of 
schooling. 
 

3.3. Hausman’s Test for Endogeneity 
Self-reported values associated with survey data make variables endogenous (Wooldridge, 2019). Again, the 

decision to attend school and complete a given level of study is also influenced by several factors. Suspecting the 
possibility of endogeneity of the schooling variable is not enough. There is the need to ascertain if the variable is 
endogenous. Verification of the schooling variable will inform the use of appropriate instruments through the 
method of instrumental variable estimation to obtain consistent and valid estimates (Wooldridge, 2019).   

According to Davidson and MacKinnon (1993), an augmented regression needs to be run after performing a 
first auxiliary regression which specifies the schooling variable as a function of the instruments. This specification 
is given as; 

𝑆𝑖 = 𝑘 + 𝛾𝑆𝑓 + 𝛿𝑆𝑚 + 𝑢𝑖                                                                                              (5) 

Equation 5 represents the represents the auxiliary regression and  𝑆𝑓 and 𝑆𝑚 represents the educational level of 

the individual’s father and mother respectively. k is the intercept,  𝑢𝑖 is the stochastic error term, and the definition 

of 𝑆𝑖 remained unchanged.  Equation 4 is estimated and its residuals are predicted and stored as 𝑆𝑖
∗. 𝑆𝑖

∗ is hence 
included as a regressor in the consumption function in Equation 1 as; 

𝐶𝑖 = 𝑎 + 𝑋𝑖𝛼 + 𝑆𝑖𝛽 + 𝜙𝑆𝑖
∗+ 𝜔𝑖                                                                                    (6) 



Economy, 2022, 9(1): 9-17 

14 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Equation 6 represents the augmented regression that contains the predicted residuals from Equation 5.  𝜙 
which is the co-efficient of the predicted residuals for Equation 4 is therefore tested. The definition of other 

variables remains unchanged and 𝜔𝑖 is the stochastic error term for Equation 5. Rejecting the null hypothesis of 

exogeneity means the coefficient of 𝑆𝑖
∗  is significantly different from zero. This means the variable correlates with 

the error term and hence employing the ordinary least square method will yield inconsistent results. This test is 
also done for the interacted variable for the third objective. 
 

4. Results 
4.1. What Difference Does Schooling Make in the Consumption Expenditure of Individuals in Ghana? 

Table 3 presents how the consumption of individuals with education differs from those without any formal 
education. It also contains results for various tests, particularly for the exogeneity of education. Hausman’s test 
results imply the rejection of the null hypothesis which suggests that the education of the individual is exogenous. 
This means education is endogenous and justifies the use of instrumental variable estimation techniques to find out 
how the consumption of educated individuals significantly differs from those who have not had any education. 
 

Table 3. Effect of education on consumption of individuals in Ghana. 

Consumption expenditure OLS Reduced Form IV 

Schooling 1.37*** 
(0.171) 

1.22*** 
(0.176) 

33.32*** 
(5.954) 

Income 0.009*** 
(0.004) 

0.011*** 
(0.0005) 

0.0102*** 
(0.001) 

Household size 1.46*** 
(0.028) 

1.52*** 
(0.029) 

1.47*** 
(0.055) 

Age in years 0.004 
(0.0046) 

0.014*** 
(0.0048) 

-0.067*** 
(0.015) 

Poverty status (Poor) -9.781*** 
(0.19) 

-9.349*** 
(0.199) 

-10.21*** 
(0.387) 

Gender (Female) -0.972*** 
(0.162) 

-1.012*** 
(0.165) 

2.164*** 
(0.659) 

Location (Rural) -4.107*** 
(0.159) 

-3.627*** 
(0.165) 

-1.44** 
(0.564) 

Intercept 8.326*** 
(0.276) 

6.726*** 
(0.306) 

-15.146*** 
(4.371) 

Observations 14,009 13,142 13,142 

𝑅2 0.325 0.332 . 

Test results 

Hausman t-test on exogeneity n/a n/a (87.39) *** 
F-test statistic F (2, 28286) 59.62*** F (9, 13132) 727.12*** Wald chi (7) 1825.29*** 
Sargan’s identification test n/a n/a (20.65) *** 

          Notes: Robust std. errors in parenthesis. *** represents 5 percentage significance levels. 
 

Results from the OLS and Reduced form equations show that the effect of education on the consumption of 
individuals is undercast due to the presence of endogeneity. The IV results which correct for the problem of 
endogeneity show educated individuals consume 33.32 cedis more than those with no education. This confirms the 
theory of Becker (1993) and empirical studies like (Tran et al., 2020) which asserted the significance of education in 
enhancing the livelihood of individuals through an expansion in their consumption. Indeed, estimating the effect of 
education in the presence of endogeneity leads to bias as the effect was undercast by about 32 cedis.  

Results from Table 3 also show that income significantly influences the consumption of individuals in Ghana. A 
one cedi rise in the individual’s income increases the consumption of the individual by less than one cedi. This 
finding confirms the fundamental psychological law that states that a change in income increases consumption but 
not as much as the change in income. This implies consumption of individuals is less sensitive to changes in their 
income (Sloman & Wride, 2009). Again, an additional increase in the size of an individual’s household increases 
his/her consumption by 1.47 cedis. Age of the individual was found to significantly reduce the consumption of the 
individual whiles poor individuals experience a fall of 10.21 cedis in their consumption relative to those who are 
better-off. Females were found to consume 2.164 cedis more than their male counterparts while individuals that 
reside in rural areas consume 1.44 cedis less than their counterparts in the urban areas, holding all other factors 
constant. 
 

4.2. Effect of Schooling on the Consumption Expenditure of Poor Individuals in Ghana 
It is hard to talk about livelihood, especially in developing countries like Ghana without mentioning poverty. 

The poverty rate of Ghana as of 2016 was 56.30% which is significant, especially in influencing the consumption 
expenditure of individuals (GSS, 2017). The focus of this section is to find out how poverty influences the 
consumption expenditure of educated individuals in Ghana. Alternatively, how will education influence the 
consumption expenditure of poor individuals? Because education was endogenous, interacting it with the poverty 
status of individuals makes the interacted variable also endogenous. Hausman’s test for endogeneity of the 
interacted variable presented in Table 4 confirmed the presence of endogeneity and justifies the need to employ 
instrumental variable estimation to estimate the effect of education on poor individuals. 

Table 4 indicates that poverty hurts the consumption of individuals even if they are educated. Educated 
individuals who are poor consume 39.65 cedis less than educated individuals who are non-poor. This negative effect 
substantiates the severity of poverty in Ghana and the damage it does to the livelihood of individuals even if they 
have attained some level of schooling (World Bank, 2019).  

 
 



Economy, 2022, 9(1): 9-17 

15 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Table 4. Effect of education on the consumption expenditure of poor individuals in Ghana. 

Consumption expenditure OLS Reduced Form IV 

Schooling 2.132*** 
(0.198) 

1.89*** 
(0.202) 

38.92*** 
(6.16) 

Schooling*Poverty -2.938*** 
(0.388) 

-2.822*** 
(0.441) 

-39.65*** 
(6.155) 

Income 0.0091*** 
(0.004) 

0.011*** 
(0.0012) 

0.0098*** 
(0.001) 

Household size 1.46*** 
(0.028) 

1.52*** 
(0.029) 

1.57*** 
(0.056) 

Age in years 0.0063 
(0.0046) 

0.016*** 
(0.0048) 

-0.028*** 
(0.010) 

Poverty status (Poor) -7.596*** 
(0.346) 

-7.24*** 
(0.366) 

-19.54*** 
(4.537) 

Gender (Female) -0.921*** 
(0.162) 

-0.964*** 
(0.165) 

2.45*** 
(0.642) 

Location (Rural) -4.05** 
(0.159) 

-3.587*** 
(0.165) 

-1.095** 
(0.565) 

Intercept 7.575*** 
(0.293) 

6.081*** 
(0.320) 

-21.79*** 
(4.912) 

Observations 14,009 13,142 13,142 

𝑅2 0.328 0.335 . 

Test results 
Hausman t-test on exogeneity n/a n/a (119.86) *** 
F-test statistic/Wald chi2 F (8, 14000) 852.6*** F (10, 13131) 661.4*** Wald chi (8) 1820.29*** 
Sargan’s identification test n/a n/a (11.40) *** 

           Notes: Robust std. errors in parenthesis. *, ** & *** represents 10, 5 & 1 percentage significance levels. 

 
Individuals with some years of schooling were also found to consume more than those who have not been to 

school at all. The significant difference in their consumption even increased from 33.32 cedis to 38.92 cedis which 
goes to affirm the role schooling plays in improving the livelihood of individuals. In the counterfactual sense, 
individuals who have never been to school would have consumed 38.92 cedis more if they have had some level of 
schooling. Such an increase would have improved their livelihood. 

The effect of control variables remained the same. Consumption expenditure remained less sensitive to changes 
in income, household size had a positive effect on consumption expenditure of individuals and poor individuals 
consume 19.54 cedis less relative to non-poor ones. Females were shown to consume more than their male 
counterparts whiles individuals who dwell in rural areas were found to consume less than those in urban areas. 
 

4.3. Effect of Schooling on the Poverty Status of Ghanaians 
After establishing that poverty reduces consumption expenditure even for educated individuals, we went 

further to estimate the effect of schooling on the poverty status of individuals. The aim is to find out how education 
influences the poverty status of individuals in Ghana. 

Results of the Wald test presented in Table 5 show that the null hypothesis of exogeneity of the schooling is 
rejected at the 5% significance level. This implies that the schooling variable is endogenous and justifies the use of 
an instrumental variable to estimate the effect of schooling on poverty in Ghana. 
 

Table 5. Effect of schooling on the poverty status of individuals in Ghana. 

Poverty status Coef. Marginal Effects 

Schooling  -2.356*** 
(0.0439) 

0.0025 
(0.0143) 

Household size 0.204 
(0.2977) 

0.083*** 
(0.0026) 

Gender (Female) -0.259*** 
(0.0403) 

-0.0103*** 
(0.0028) 

Age in years 0.0061*** 
(0.00107) 

0.000243*** 
(0.000075) 

Location (Rural) -0.112*** 
(0.0408) 

0.0083*** 
(0.0033) 

Consumption expenditure -0.151 
(0.225) 

-0.0625*** 
(0.00036) 

Constant 1.594*** 
(0.1044) 

n/a 

Other Statistics 
Wald chi2 (7) 28746.58*** 
Observation 13,902 
Corr. (education, poverty) 0.995 
Wald test of exogeneity 4.38** 

Notes: Robust std. errors in parenthesis. ** & *** represents 10, 5 percentage significance levels. 

 
According to Table 5, individuals with some years of schooling are less likely to be poor compared to those 

who have not been to school at all. Put differently, schooling reduces the likelihood of individuals being poor as it 
enhances their abilities to make a meaningful living. 

The marginal effect of household size on poverty status showed that individuals with a higher household size 
are 8.3% more likely to be poor. The simple explanation is that the income per head of a household reduces as its 
size increases and reduces the relative provision each member receives, holding all other factors constant. An 



Economy, 2022, 9(1): 9-17 

16 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

increase in household size also increases the dependency burden which increases the vulnerability of that household 
to poverty, assuming all other factors are fixed. Compared to their male counterpart, females were shown to be less 
likely to be poor. Their marginal effect confirms that females are 1.03% less likely to be poor compared to males in 
Ghana. Table 5 also revealed that Individuals in rural areas of Ghana were less likely to be poor relative to those in 
urban areas. Its marginal effect was however contradictory as rural dwellers are 0.83% more likely to be poor. The 
results further showed that individuals with a higher consumption expenditure are 6.25% less likely to be poor. 
This means individuals with a higher level of consumption expenditure have the means to do so and are hence able 
to have a sustainable living which makes them less likely to be poor. 

 
5. Conclusion 

The purpose of this study was to verify if schooling makes any difference as far as the consumption expenditure 
and poverty status of individuals in Ghana are concerned. In doing so, we quantified the magnitude of this 
difference through the method of instrumental variables (IV). The use of IV became important since schooling was 
found to be endogenous which means the use of OLS will yield invalid and inconsistent estimates. The study also 
evaluated the effect of schooling on the consumption of poor individuals since poverty is considered paramount in 
livelihood issues, especially in developing countries. The influence of education on the poverty status of individuals 
was also investigated. 

Using the education of individuals’ parents as instruments, we found a significant difference in the consumption 
expenditure of educated and uneducated individuals in Ghana. Essentially, individuals who have completed some 
level of schooling experience a significant increase in their consumption expenditure relative to those who have not 
been to school at all. As evidence of endogeneity, the results of OLS underestimated the difference in consumption 
expenditure between these groups of individuals. 

 Again, educated individuals who are poor had a significantly lower consumption expenditure compared to 
educated individuals who are non-poor. The fall in consumption for the educated but poor significantly exceeds 
that of uneducated individuals. This evidence suggests the gravity of poverty on the livelihood of individuals even 
if they have had some level of schooling. It also suggests that putting individuals in schools is not enough but 
efforts must be made to adequately resource them to undertake livelihood activities.  

Also, individuals with some level of schooling had a lower probability of poverty compared to those without 
any schooling. Essentially, schooling to some extent empowers the individual with some skills and knowledge set 
that makes them less likely to be poor. Finally, this paper recommends serious attention and investments in the 
Ghanaian educational system especially at the lower level to increase enrollment. The schooling curriculum must 
be revised to equip individuals with some workable set of skills that can enable them to undertake livelihood 
activities to improve their welfare. This will help individuals to undertake start-up businesses and reduce the high 
rate of unemployment in the country. 
 

References 
Adjei, R. S. (2010). Effects of action aid and land reclamation program on food production in Ghana: Online Library, University of Cape 

Coast. Retrieved from: 
https://ir.ucc.edu.gh/xmlui/bitstream/handle/123456789/2669/ADJEI%202010.pdf?sequence=1&isAllowed=y. 

Adu-Gyamfi, S., Donkoh, W. J., & Addo, A. A. (2016). Educational reforms in Ghana: Past and present. Journal of Education and Human 
Development, 5(3), 158-172. 

Aghion, P., & Howitt, P. W. (2008). The economics of growth. Cambridge, United States: MIT Press. 
Akyeampong, K. (2009). Revisiting free compulsory universal basic education (FCUBE) in Ghana. Comparative Education, 45(2), 175-

195.Available at: https://doi.org/10.1080/03050060902920534. 
Becker, G. S. (1993). Human Capital: A theoretical and empirical analysis with special reference to education. Chicago: University of Chicago Press. 
Cameron, S., & Ananga, E. D. (2015). Savings groups, livelihoods and education: Two case studies in Ghana. Journal of International 

Development, 27(7), 1027-1041.Available at: https://doi.org/10.1002/jid.3067. 
Chambers, R., & Conway, G. (1992). Sustainable rural livelihoods: Practical concepts for the 21st century. Discussion Paper No. 296. IDS, 

Sussex. 
Coelli, M., & Green, D. A. (2012). Leadership effects: School principals and student outcomes. Economics of Education Review, 31(1), 92-

109.Available at: https://doi.org/10.1016/j.econedurev.2011.09.001. 
Corcoran, J. L., Davies, S. L., Knight, C. C., Lanzi, R. G., Li, P., & Ladores, S. L. (2020). Adolescents’ perceptions of sexual health education 

programs: An integrative review. Journal of Adolescence, 84(1), 96-112.Available at: 
https://doi.org/10.1016/j.adolescence.2020.07.014. 

Crockett, S., Duffy, J., & Izhakian, Y. (2019). An experimental test of the Lucas asset pricing model. The Review of Economic Studies, 86(2), 
627-667.Available at: https://doi.org/10.1093/restud/rdy035. 

Davidson, R., & MacKinnon, J. G. (1993). Estimation and inference in econometrics. New York: Oxford University Press. 
Debrah, E. (2013). Alleviating poverty in Ghana: the case of livelihood empowerment against poverty (LEAP). Africa Today, 59(4), 41-67. 
Gecho, Y., Ayele, G., Lemma, T., & Alemu, D. (2014). Rural household livelihood strategies: Options and determinants in the case of Wolaita 

Zone, Southern Ethiopia. Social Sciences, 3(3), 92-104.Available at: https://doi.org/10.11648/j.ss.20140303.15. 
GSS. (2017). Ghana living standards survey: Study document for the seventh round. Accra: Ghana Statistical Service. 
Haji, J., & Legesse, B. (2017). Impact of sedentarization program on the livelihood and food security of Ethiopian pastoralists. Journal of Arid 

Environments, 136, 45-53.Available at: https://doi.org/10.1016/j.jaridenv.2016.10.007. 
Heckman, J. J., & Karapakula, G. (2019). Intergenerational and intragenerational externalities of the Perry preschool project. NBER 

Working Paper No. 25889 National Bureau of Economic Research. 
Lim, D. (2018). Quality assurance in higher education: A study of developing countries: Routledge. 
Little, P., Aboud, A., & Lenachuru, C. (2009). Can formal education reduce risks for drought-prone pastoralists?: A case study from Baringo 

District, Kenya. Human Organization, 68(2), 154-165.Available at: https://doi.org/10.17730/humo.68.2.n70t617197x4w778. 
Mabe, F. N., Konja, D. T., Addo, M. D. D., & Awuni, J. A. (2022). The locational and gendered impacts of livelihood empowerment against 

poverty (LEAP) on children education in Ghana. International Journal of Social Economics, 49(6), 850-866.Available at: 
https://doi.org/10.1108/ijse-10-2021-0596. 

Mahama, T. A., & Maharjan, K. L. (2018). Generalised ordered model for conceptualising and ascertaining the determinants of livelihood 
satisfaction in Ghana. Journal of Happiness Studies, 19(4), 1195-1216. 

Majgaard, K., & Mingat, A. (2012). Education in Sub-Saharan Africa: A comparative analysis. Paper presented at the © 2012 International Bank 
for Reconstruction and Development / the World Bank, 1818 H Street NW, Washington DC 20433.  

Oduro, K., Mohren, G., Pena-Claros, M., Kyereh, B., & Arts, B. (2015). Tracing forest resource development in Ghana through forest 
transition pathways. Land Use Policy, 48, 63-72.Available at: https://doi.org/10.1016/j.landusepol.2015.05.020. 



Economy, 2022, 9(1): 9-17 

17 
© 2022 by the authors; licensee Asian Online Journal Publishing Group 

 

 

Pajibo, E., & Tamanja, E. M. (2017). Influence and challenges of the capitation grant on education delivery in basic schools in Ghana. Asian 
Journal of Education and Training, 3(1), 53-63.Available at: https://doi.org/10.20448/journal.522.2017.31.53.63. 

Peters, E., Baker, D. P., Dieckmann, N. F., Leon, J., & Collins, J. (2010). Explaining the effect of education on health: A field study in Ghana. 
Psychological Science, 21(10), 1369-1376.Available at: https://doi.org/10.1177/0956797610381506. 

Pons, E., & Gonzalo, M. T. (2002). Returns to schooling in Spain: How reliable are instrumental variable estimates? Labour, 16(4), 747-
770.Available at: https://doi.org/10.1111/1467-9914.00215. 

Porter, G., Hampshire, K., Abane, A., Tanle, A., Esia-Donkoh, K., Obilie Amoako-Sakyi, R., & Asiedu Owusu, S. (2011). Mobility, education 
and livelihood trajectories for young people in rural Ghana: A gender perspective. Children's Geographies, 9(3-4), 395-410.Available 
at: https://doi.org/10.1080/14733285.2011.590705. 

Romer, D. (2012). Advanced macroeconomics (4th ed.). New York: McGraw-Hill. 
Sackey, H. A. (2005). Female labor force participation in Ghana: The effects of education. AERC Research Paper No. 150, African Economic 

Research Consortium, Nairobi. 
Sen, A. (1990). Gender and cooperative conflict. In: Tinker, I., Ed., Persistent inequality (pp. 123-148). Oxford: University Press Oxford. 
Sloman, J., & Wride, A. (2009). Economics. Edinburg Gate, England: Pearson Education Limited. 
Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good judgement, less pathology, better grades, and 

interpersonal success. Journal of Personality, 72(2), 271-324.Available at: https://doi.org/10.1111/j.0022-3506.2004.00263.x. 
Tran, T. A., Tran, T. Q., & Nguyen, H. T. (2020). The role of education in the livelihood of households in the Northwest region, Vietnam. 

Educational Research for Policy and Practice, 19(1), 63-88.Available at: https://doi.org/10.1007/s10671-018-9242-6. 
United Nations Children’s Fund (UNICEF). (2016). The State of the world’s children. Retrieved from 

https://www.unicef.org/media/50076/file/UNICEF_SOWC_2016-ENG.pdf. 
United Nations Development Group (UNDP). (2016). The sustainable development goals are coming to life: Stories of country 

implementation and UN support. Retrieved from https://undg.org/wp-content/uploads/2016/12/SDGs-are-Coming-to-Life-
UNDG-1.pdf. 

Wooldridge, J. M. (2019). Correlated random effects models with unbalanced panels. Journal of Econometrics, 211(1), 137-150.Available at: 
https://doi.org/10.1016/j.jeconom.2018.12.010. 

World Bank. (2004). World development report 2004: Making services work for poor people: The World Bank. 
World Bank. (2019). Report on agriculture and livelihood in Africa. Washington D.C: The World Bank. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
Asian Online Journal Publishing Group is not responsible or answerable for any loss, damage or liability, etc. caused in relation to/arising out of the use of the content. 
Any queries should be directed to the corresponding author of the article. 
 

http://www.unicef.org/media/50076/file/UNICEF_SOWC_2016-ENG.pdf

