Language use in education and human capital formation: Evidence from the Ethiopian educational reform

doi:10.1016/j.worlddev.2017.04.029

World Development

Volume 98, October 2017, Pages 195-213

https://doi.org/10.1016/j.worlddev.2017.04.029 Get rights and content

Highlights

•
Explores the role of language as a key input into the process of student learning.
•
Provides causal estimates of mother tongue schooling on indicators of human capital.
•
Mother tongue instruction increases the probability of being able to read an entire sentence by 40%.
•
It increases completed years of schooling by 0.65 years and the probability of completing primary schooling by 5
•
It increases newspaper readership by 25%.

Summary

In 1994, Ethiopia introduced mother tongue instruction in primary schooling for the largest ethnic group in the country. Combining data from before and after the policy change, from the regions that gain access to mother tongue schooling, along with the regions and ethnic groups that face no change in language policy, I estimate the effect of mother tongue instruction on human capital formation. The results show that access to mother tongue schooling increases the ability to read by 40%. Moreover, it increases the completed years of schooling by more than half a year, and the probability of completing primary schooling by five percentage points. The ability to read by permitting access to printed sources of information, increases newspaper readership by around 25% and makes it 17% points more likely that individuals report using pamphlets, posters or leaflets as a source of information about family planning. My preferred interpretation of the results is that provision of schooling in a language spoken at home, and used for social interactions, reduces the cost and increases the efficiency of learning. The result highlights the use of non-native languages, an almost ubiquitous practice in Sub-Saharan Africa, as a potentially important factor underlying the large educational failures in this part of the world.

Introduction

The question of language use in education is particularly relevant for the Sub-Saharan African continent. In Sub-Saharan Africa, with the exception of Eretria, Ethiopia, and Tanzania no country provides the entire span of primary schooling in a local language, let alone the mother tongue (Albaugh, 2014). Along with the wide-scale reliance on the use of non-native languages, extremely low levels of student learning characterize the schooling system in this part of the world. For instance, a recent evaluation of the state of education and learning in Sub-Saharan Africa estimates that out of 110 million children entering schooling, more than 40 million will not even acquire emergent reading and numeracy skills (VanFleet, Watkins, & Greubel, 2012). However, the role of language in affecting student learning and knowledge acquisition in contexts such as these remains an open question. In this paper I examine whether using as a medium of instruction in schools, a language that is different from the language used at home, can explain the failure to obtain basic skills such as literacy and also its impact on other facets of human capital.

I base my analysis on the Ethiopian educational reform of 1994. The reform introduced mother tongue instruction in primary schooling for the largest ethnic group—the Oromos—in four of the 11 regions of the country. Additionally, the reform also introduced a package of non-language educational reforms (curricula revision, teacher training, and decentralization) that were implemented uniformly across the country, that is, for all groups and regions. To isolate the effect of the language policy change, I exploit the fact that the exposure to mother tongue instruction is jointly determined by the interaction of year of birth, region of residence, and ethnicity dummies and implement a triple difference-in-differences (D-I-D) estimator. After controlling for year of birth, region of residence, and ethnicity dummies, as well as all the two-way interaction terms, the triple interaction term—year of birth, region of residence, and ethnicity—plausibly isolates the causal effect of the provision of mother tongue schooling on indicators of human capital.¹ I provide evidence in support of the identifying assumption underlying the estimation strategy, that is, the groups being compared demonstrate parallel trends on the indicators of interest before the implementation of the educational reform in 1994.

The results show that access to mother tongue instruction increases the probability of being able to read a complete sentence, a key capability, by 18 percentage points. As only 45% of the individuals prior to the reform were able to read a complete sentence, introduction of mother tongue education leads to a gain of more than a 40% increase in the ability to read. Examining the impact on quantitative indicators of human capital, the data show that provision of mother tongue instruction increases the completed years of schooling by 0.65 years and the probability of completing primary schooling by 5 percentage points. I hypothesize that the increase in the ability to read permits access to printed sources of information, and consistent with this reasoning find that individuals affected by the language policy change are 8 percentage points more likely to report reading a newspaper, and 17 percentage points more likely to report using pamphlets, posters, or leaflets as a source of information about family planning. The increased readership of newspapers is an important outcome, as newspaper readership has been shown to be a proxy for social capital (Putnam, 2001), is associated with improved voter knowledge (Snyder, 2010) and electoral participation (Gentzkow, Shapiro, & Sinkinson, 2011), as well as having an important role in improving government responsiveness to its citizens (Besley & Burgess, 2002). I also explore whether the extension of mother tongue instruction has a positive effect on the state of knowledge about the existence of major diseases and the availability of simple cures. I find small positive effects on the probability that a person who gains access to mother tongue instruction, and hence increases their ability to read, has heard of all the five diseases or cures considered, that is, the existence of modern contraceptives, tuberculosis (TB), human immunodeficiency virus (HIV), sexually transmitted infections (STIs), and oral rehydration solution (ORS).

The estimation strategy (the triple D-I-D estimator) by design isolates the effect of the introduction of mother tongue education from the non-language educational reforms, however, I undertake several additional analyses to verify that the increase in human capital should be attributed to the change in the language of education and not the other aspects of the educational and political reform. I analyze factors such as increase in political power for the Oromos due to the creation of ethnic homelands, programs undertaken outside of formal schooling, changes in teacher composition, and the increase in school enrollment. I find little support to conclude that those other variables can explain the measured increase in the ability to read.

I discuss why findings from earlier studies such as Angrist et al. (1997) might differ from the ones found in this paper. I also elaborate on how my paper adds to the existing literature analyzing the role of language use in education in the context of Sub-Saharan Africa and also highlights the limited relevance of the findings from bilingual programs, primarily based on evidence from the United States, in informing us about the role of language in student learning and knowledge accumulation in Sub-Saharan Africa.

The paper is structured as follows: the next section provides an overview of the related literature and highlights the distinctions and commonalities. Section 3 provides information on the linguistic groups and the educational and language policy in Ethiopia. Section 4 presents the data and the empirical strategy. In Section 5, I show that introducing primary schooling in the native language increases the individual’s ability to read, completed years of schooling and newspaper readership. In Section 6, I show that the language policy change and not the other aspects of the reform explain the increase in human capital and Section 7 concludes.

Section snippets

Related literature

The impact of the choice of medium of instruction policies on educational outcomes in developing countries has been a vibrant area of research in the disciplines of sociolinguistics and education. The positive effects of mother tongue education on academic achievement and student performance has been demonstrated in a variety of contexts (Alidou, Bamgbose, 1984, Desai and Qorro, 2006). It also has been shown that use of a proximate language has positive impacts on attendance, participation in

The linguistic groups

Ethiopia is a landlocked country located in the northeast of Africa. It has a population of around 80 million, making it the second most populous nation on the continent. Ethiopia is an ethnically diverse country, containing more than 80 ethnic groups and 90 different languages (Lewis, Simon, & Fennig, 2014). Most languages belong to the Afro-Asiatic language phylum of which three branches are represented: Semitic, Cushtic, and Omotic.

The “Oromo” are the largest ethnic group in the country

Data

The Ethiopian Demographic and Health Surveys (DHS) are nationally representative data on health, education, and demographic trends in developing countries and the principal data source for the empirical analysis. The data are from 2011 and include information from the nine regions and the two city administrative units in the country.

The schooling system in Ethiopia involves eight years of primary schooling followed by four years of secondary schooling. The children in Ethiopia typically go to

The impact of mother tongue instruction on the ability to read

To estimate the triple D-I-D, I run the following reduced form equation: $H_{ijkn} = δ_{1} D_{j} C_{k} R_{n} + δ_{2} D_{j} C_{k} + δ_{3} D_{j} R_{n} + δ_{4} C_{k} R_{n} + δ_{5} D_{j} + B_{k} + T_{n} + ε_{ijkn},$ where $H$ is a dummy taking the value 1 for individuals able to read a complete sentence. The subscript $ijkn$ indexes individual $i$ , from language group $j$ , of cohort $k$ and resident in region $n$ . $D_{j}$ is a dummy variable taking the value 1 if the individual belongs to the Oromo language group and zero otherwise. $C_{k}$ is a dummy variable which takes the value 1 if the individual

Creation of ethnic homelands

The new Ethiopian constitution of 1994 created a federal republic. The country was divided into nine regions and two city administrative units based on ethnolinguistic lines. The creation of ethnic homelands could have led to greater political power and representation for the Oromos and this could be an alternative explanation for the observed increase in human capital. To test the strength of this competing hypothesis, in column (4) of Table 3, I exclude the regional homelands of the Oromo

Conclusion

The paper analyzes the effects that the choice of medium of instruction had on human capital formation in Ethiopia. The empirical strategy employs a triple D-I-D estimator to uncover the causal effects of the provision of schooling in the mother tongue. I provide evidence in support of the identifying assumptions underlying the empirical strategy and also analyze the role of other confounding factors. The estimates show that provision of mother tongue schooling increases the individual’s

Acknowledgements

I would like to thank Tommaso Agasisti, Ana Rute Cardoso, Francesco Fasani, Lidia Farre, Florian Hett, David Laitin, Emmy Lindstam, Jan Luskic, Hannes Mueller, Urska Petrovcic, Christopher Rauh, Devesh Rustagi, Emilia Soldani, Alessandro Tarozzi and Jack Willis, and the critical audiences at the University of Barcelona, Universitat Autonoma de Barcelona, Universite Libre de Bruxelles, Goethe University, Stockholm School of Economics, the summer school in development economics, Italy, the

References (60)

H.A. Patrinos et al.
Costs and benefits of bilingual education in Guatemala: A partial analysis
International Journal of Educational Development
(2009)
Y. Seid
Does learning in mother tongue matter? Evidence from a natural experiment in ethiopia
Economics of Education Review
(2016)
S. Taylor et al.
Estimating the impact of language of instruction in South African primary schools: A fixed effects approach
Economics of Education Review
(2016)
E.A. Albaugh
State-building and multilingual education in Africa
(2014)
H. Alidou et al.
Optimizing learning and education in Africa – The language factor
(2006)
J. Angrist et al.
Is Spanish-only schooling responsible for the Puerto Rican language gap?
Journal of Development Economics 85.
(2008)
J.D. Angrist et al.
The effect of a change in language of instruction on the returns to schooling in Morocco
Journal of Labor Economics
(1997)
D. August et al.
Developing reading and writing in second-language learners: Lessons from the report of the national literacy panel on language-minority children and youth
(2008)
A. Bamgbose
Mother-tongue medium and scholastic attainment in Nigeria
Prospects: Quarterly Review of Education
(1984)
P. Bender et al.
In their own language: Education for all
(2005)

Benson, C. (2005). The importance of mother tongue-based schooling for educational quality. EFA Global Monitoring...

T.J. Besley et al.

The political economy of government responsiveness: Theory and evidence from India

The Quarterly Journal of Economics

(2002)

A.C. Cameron et al.

Bootstrap-based improvements for inference with clustered errors

The Review of Economics and Statistics

(2008)

A.C. Cameron et al.

Robust inference with multiway clustering

Journal of Business & Economic Statistics

(2012)

A.C. Cameron et al.

Microeconometrics using Stata

(2009)

Central Statistical Agency [Ethiopia] and ICF International

Ethiopia demographic and health survey 2011

(2012)

Chin, A. (2015). Impact of bilingual education on student achievement. IZA World of...

S. Courtois et al.

The black book of communism: Crimes, terror, repression

(1999)

J. Cummins

Linguistic interdependence and the educational development of bilingual children

Review of Educational Research

(1979)

J. Cummins

Interdependence of first-and second-language proficiency in bilingual children

Language Processing in Bilingual Children

(1991)

T.S. Dee

A teacher like me: Does race, ethnicity, or gender matter?

The American Economic Review

(2005)

Z. Desai et al.

Focus on fresh data on the language of instruction debate in tanzania and south africa

African Minds

(2006)

De Waal, A. (1991). Evil days: Thirty years of war and famine in Ethiopia (Vol; 3169) (No. 69). Human Rights...

C. Dustmann et al.

Educational achievement of second-generation immigrants: An international comparison

Economic Policy

(2012)

C. Dustmann et al.

Ethnicity and educational achievement in compulsory schooling

The Economic Journal

(2010)

N. Dutcher et al.

The use of first and second languages in education – A review of international experience

(1996)

R.G. Ehrenberg et al.

Do teachers’ race, gender, and ethnicity matter? Evidence from the National Educational Longitudinal Study of 1988

Industrial & Labor Relations Review 483.

(1995)

K. Eriksson

Does the language of instruction in primary school affect later labour market outcomes? Evidence from South Africa

Economic History of Developing Regions

(2014)

Federal Democratic Republic, Government of Ethiopia (1994). Education and training...

G. Garcia

Bilingual children’s reading

Handbook of Reading Research

(2000)

Cited by (31)

English medium instruction in multilingual contexts: Empirical evidence from Ethiopia
2024, International Journal of Educational Development
Language acquisition and learning literatures favour mother-tongue education policies, particularly in the early years of schooling. However, English Medium Instruction (EMI) remains a popular language policy position in multilingual contexts. This paper studies EMI transition policies using a longitudinal dataset from Ethiopia, by leveraging regional variation in education language policy reforms for causal identification. Employing a dynamic value-added model, this paper shows that students in schools using English as a medium of instruction have lower mathematics test scores (−0.25 standard deviations) compared to students in mother tongue education schools. Furthermore, English medium learners do not perform any better in their English test scores compared to mother tongue learners. These findings are in line with the international education literature on skill development, learning and second language acquisition. The main results support the prolonged utilisation of mother tongue instruction in primary education. These results are particularly relevant for policy makers in linguistically diverse contexts.
Mother tongue reading materials as a bridge to literacy
2022, Economics of Education Review
Citation Excerpt :
In 1994 Ethiopia passed a law which increased the offering of mother tongue primary education. Ramachandran (2017) finds the change substantially increased literacy in one of the five major languages, defined as being able to read a simple sentence. He also shows that this increase in literacy is accompanied by a greater use of written materials and better health knowledge.
Children whose mother tongue is different from the language of instruction at school face additional challenges developing literacy skills. One approach to this favours immersion in the language of instruction, while another favours a transitional period of bilingual education. This paper evaluates the impact of a primary school literacy intervention. In addition to multi-faceted literacy support, the programme included a component of transitional bilingual education for students whose mother tongue is not the usual language of instruction. Over two years, the intervention raised literacy scores in the language of instruction by ＋0.41 sd, and literacy in mother tongue for minority language speakers by ＋0.75 sd. Our findings suggests that a light-touch transitional bilingual component can support minority language literacy without crowding out acquisition of the language of instruction.
The Value of Foreign Language Skills in the German Labor Market
2022, Labour Economics
This article explores the relationship between foreign language skills and individuals’ labor income in Germany, focusing on the English language. Using the 2012 and 2016 waves of the German Socio-Economic Panel’s Innovation Sample (SOEP-IS), we find that native speakers of German with English language skills earn a wage premium of 13 percent, on average. Incremental improvements in the level of skills, e.g., from basic to independent user, increase wages by 11 percent, on average. We address endogeneity issues by using novel data that combine comprehensive information about individuals’ characteristics with fine-grained self-assessments of language skills based on descriptors derived from the Common European Framework of Reference for Languages (CEFR). Any remaining sources of endogeneity in the level of language proficiency are addressed by an instrumental variable approach that exploits exogenous variation in individuals’ exposure to foreign language acquisition in school. We also show that wage differentials cannot be explained by the value of foreign language skills as a general ability signal, but they are driven by the productive value of such skills. Finally, by examining language skills instead of the use of such skills in the workplace, we identify individual returns to foreign languages for the general population. As education policy is the main determinant of English language acquisition (not only in Germany), this information is highly relevant for policy-makers.
How natural disasters affect energy innovation? The perspective of environmental sustainability
2022, Energy Economics
Citation Excerpt :
Human capital helps to transform resources into the innovative process. In particular, skill labor in any given country is central to firms' competitiveness and output innovation (Chang et al., 2016; Ramachandran, 2017). Therefore, we control human capital (proxied by Education) using gross secondary school enrollment rate.
The nexus between natural disasters and energy technological innovation is of great importance in the energy technology innovation theory and related literature. This research thus investigates the influence of natural disasters on energy technology innovation by using panel data technology from 1975 to 2018 for the samples of 29 OECD countries and first finds that natural disasters may have a significantly negative impact on energy technology innovation, which not only is impactful in the year of the natural disasters, but also lasts over the next 4 years. In addition, the effects of natural shocks on energy innovation appears to be weaken by greater economic development, trade openness, financial openness, interpersonal globalization and government stability. Furthermore, epidemic has the biggest negative impact on energy innovation. Finally, natural disasters are significantly negatively related to low-carbon energy innovation, but not related to non-low carbon energy innovation. Overall, the findings provide a research reference for how the policy makers can improve the level of energy innovation in the face of natural disasters and environmental sustainability.
Linguistic diversity, official language choice and human capital
2022, Journal of Development Economics
The mechanisms linking high ethnic fractionalization (ELF) to poor economic outcomes in sub-Saharan Africa remain obscure. This paper proposes that high ELF raises the costs of implementing indigenous language education policies, and that the concomitant reliance on colonial languages as media of instruction lowers human capital development, our measure for economic failure. Exploiting the fact that 101 linguistic groups in sub-Saharan Africa cross country borders that have different language-use-in-education systems, we are able to include ethnicity fixed effects in our regression estimates. This allows us to control for a large set of cultural, geographical, and environmental factors that could confound the estimates. Our within-ethnicity estimates show that even the partial use of indigenous languages in schools is associated with an economically meaningful increase in the probability of reading for individuals with seven or fewer years of schooling. Once including language-use-in education as a regressor, the coefficient for ELF in explaining lower human capital turns insignificant. Our results suggest that retention of colonial languages is an important factor in explaining economic failures in postcolonial Africa.
The spillover effect of direct competition between marketing cooperatives and private intermediaries: Evidence from the Thai rice value chain
2021, Food Policy
Citation Excerpt :
Could the language spoken at home affect farmers’ ability? Education economists have investigated the impact of language used in education on human capital formation (e.g., Ramachandran, 2017). Using a language that is different from the language spoken at home as a medium of instruction in school can increase the cost and reduce the efficiency of learning.
Despite the widespread belief that marketing cooperatives’ benefits may extend beyond participating farmers, little is known about the cooperative’s effect on nonparticipating farmers. This paper exploits exogenous variation in language spoken at home in Thailand to obtain instrumental variable estimates of the spillover effect of marketing cooperatives. We hypothesize that farmers who sell rice to private intermediaries in the area where there is direct competition between marketing cooperatives and private intermediaries (treated areas) are likely to receive a higher price than those who sell rice in other areas. Using household-level data of rice farmers in Thailand in the marketing year 2018/19, we find strong evidence that farmers in treated areas receive 10.9% higher prices from private intermediaries than those in comparison areas. Our results provide crucial implications for food policy debates regarding the role of marketing cooperatives in agri-food value chains. In particular, evaluating the inclusiveness of marketing cooperatives toward poor farmers should not be limited to sampling and analyzing its members only. Failure to consider the spillover effect could lead to substantial underestimation of the impact of marketing cooperatives on societal welfare.

View all citing articles on Scopus

View full text

Language use in education and human capital formation: Evidence from the Ethiopian educational reform

Highlights

Summary

Introduction

Section snippets

Related literature

The linguistic groups

Data

The impact of mother tongue instruction on the ability to read

Creation of ethnic homelands

Conclusion

Acknowledgements

International Journal of Educational Development

Economics of Education Review

Economics of Education Review

State-building and multilingual education in Africa

Optimizing learning and education in Africa – The language factor

Is Spanish-only schooling responsible for the Puerto Rican language gap?

Journal of Development Economics 85.

The effect of a change in language of instruction on the returns to schooling in Morocco

Journal of Labor Economics

Developing reading and writing in second-language learners: Lessons from the report of the national literacy panel on language-minority children and youth

Mother-tongue medium and scholastic attainment in Nigeria

Prospects: Quarterly Review of Education

In their own language: Education for all

The political economy of government responsiveness: Theory and evidence from India

The Quarterly Journal of Economics

Bootstrap-based improvements for inference with clustered errors

The Review of Economics and Statistics

Robust inference with multiway clustering

Journal of Business & Economic Statistics

Microeconometrics using Stata

Ethiopia demographic and health survey 2011

The black book of communism: Crimes, terror, repression

Linguistic interdependence and the educational development of bilingual children

Review of Educational Research

Interdependence of first-and second-language proficiency in bilingual children

Language Processing in Bilingual Children

A teacher like me: Does race, ethnicity, or gender matter?

The American Economic Review

Focus on fresh data on the language of instruction debate in tanzania and south africa

African Minds

Educational achievement of second-generation immigrants: An international comparison

Economic Policy

Ethnicity and educational achievement in compulsory schooling

The Economic Journal

The use of first and second languages in education – A review of international experience

Do teachers’ race, gender, and ethnicity matter? Evidence from the National Educational Longitudinal Study of 1988

Industrial & Labor Relations Review 483.

Does the language of instruction in primary school affect later labour market outcomes? Evidence from South Africa

Economic History of Developing Regions

Bilingual children’s reading

Handbook of Reading Research