The Role of Cognitive Skills in Economic Development
Total Page:16
File Type:pdf, Size:1020Kb
Journal of Economic Literature 2008, 46:3, 607–668 http:www.aeaweb.org/articles.php?doi=10.1257/jel.46.3.607 The Role of Cognitive Skills in Economic Development Eric A. Hanushek and Ludger Woessmann* The role of improved schooling, a central part of most development strategies, has become controversial because expansion of school attainment has not guaranteed improved economic conditions. This paper reviews the role of cognitive skills in pro- moting economic well-being, with a particular focus on the role of school quality and quantity. It concludes that there is strong evidence that the cognitive skills of the population—rather than mere school attainment—are powerfully related to indi- vidual earnings, to the distribution of income, and to economic growth. New empiri- cal results show the importance of both minimal and high level skills, the comple- mentarity of skills and the quality of economic institutions, and the robustness of the relationship between skills and growth. International comparisons incorporating expanded data on cognitive skills reveal much larger skill deficits in developing coun- tries than generally derived from just school enrollment and attainment. The mag- nitude of change needed makes clear that closing the economic gap with developed countries will require major structural changes in schooling institutions. 1. Introduction raise the schooling levels of the population. And, indeed, this is exactly the approach of t takes little analysis to see that schooling the Education for All initiative and a central Ilevels differ dramatically between devel- element of the Millennium Development oping and developed countries. Building Goals (see, e.g., UNESCO 2005 and David upon several decades of thought about E. Bloom 2006). human capital—and centuries of general But there are also some nagging uncer- attention to education in the more advanced tainties that exist with this strategy. First, countries—it is natural to believe that a pro- developed and developing countries differ ductive development strategy would be to in a myriad of ways other than schooling * Hanushek: Stanford University, CESifo, and NBER. Paul Gertler, Manny Jimenez, Ruth Kagia, Beth King, Woessmann: Ifo Institute, University of Munich, and Lant Pritchett, Emiliana Vegas, the editor, and three CESifo. This project developed through conversa- anonymous referees. Jason Grissom provided important tions with Harry Patrinos, who provided useful com- research assistance. Support has come from the World ments and suggestions along the way. We have also Bank, CESifo, the Program on Education Policy and Gov- benefited from comments by Martha Ainsworth, Luis ernance of Harvard University, and the Packard Humani- Benveniste, François Bourguignon, Deon Filmer, ties Institute. 607 608 Journal of Economic Literature, Vol. XLVI (September 2008) levels. Second, a number of countries—both We will show in graphic terms the dif- on their own and with the assistance of oth- ferences in cognitive skills that exist, even ers—have expanded schooling opportunities after allowing for differences in school without seeing any dramatic catch-up with attainment. Most people would, in casual developed countries in terms of economic conversation, acknowledge that a year of well-being. Third, countries that do not schooling in a school in a Brazilian Amazon function well in general might not be more village was not the same as a year of school- able to mount effective education programs ing in a school in Belgium. They would also than they are to pursue other societal goals. agree that families, peers, and others con- Fourth, even when schooling policy is made tribute to education. Yet, the vast major- a focal point, many of the approaches under- ity of research on the economic impact of taken do not seem very effective and do not schools—largely due to expedience—ig- lead to the anticipated student outcomes. In nores both of these issues. The data suggest sum, is it obvious that education is the driv- that the casual conversation based on dis- ing force or merely one of several factors that parities in school attainment may actually are correlated with more fundamental devel- understate the magnitude of differences in opment forces? true education and skills across countries. The objective of this study is to review We think of education as the broad mix of what research says about the role of educa- inputs and processes that lead to individual tion in promoting economic well-being. We knowledge. Yet, because schooling and edu- pay particular attention to the robustness cation are conflated in common usage, we of the relationship between education and will generally refer specifically to cognitive economic outcomes across tests of alterna- skills, the part of education for which we tive specifications and hypotheses about the have good measures. underlying determinants of outcomes. We provide strong evidence that ignoring The discussion also has one distinctive ele- differences in cognitive skills significantly ment. We have come to conclude that cogni- distorts the picture about the relationship tive skills—particularly in assessing policies between education and economic outcomes. related to developing countries—are THE This distortion occurs at three levels. It misses key issue. It is both conventional and conve- important differences between education and nient in policy discussions to concentrate on skills on the one hand and individual earnings such things as years of school attainment or on the other. It misses an important underly- enrollment rates in schools. These things are ing factor determining the interpersonal dis- readily observed and measured. They appear tribution of incomes across societies. And, it in administrative data and they are published very significantly misses the important ele- on a consistent basis in virtually all countries ment of education in economic growth. of the world. And, they are very misleading The plan of this study is straightforward. in the policy debates. Based on a simple conceptual framework, Cognitive skills are related, among other we document the importance of cognitive things, to both the quantity and quality skills in determining individual earnings of schooling. But schooling that does not and, by implication, important aspects of improve cognitive skills, measured here by the income distribution. We then turn to comparable international tests of mathemat- the relationship of education and economic ics, science, and reading, has limited impact growth. Research into the economics of on aggregate economic outcomes and on growth has itself been a growth area, but economic development. much of the research focuses just on school Hanushek and Woessmann: The Role of Cognitive Skills in Economic Development 609 attainment with no consideration of cogni- Yet, while recognizing the impact of these tive skill differences that might arise from overall institutions, we find that cognitive variations in school quality or other sources skills play an important role. Furthermore, of learning. We show, in part with new evi- there is ample evidence that a high-quality dence, that the previous estimation is highly school system can lead to improved cognitive biased by concentration on just quantity of skills. And from a public policy perspective, schooling. interventions in the schools are generally The simple answer in the discussion of viewed as both more acceptable and more economic implications of education is that likely to succeed than, say, direct interven- cognitive skills have a strong impact on indi- tions in the family. vidual earnings. More than that, however, Given the evidence on the importance cognitive skills have a strong and robust of cognitive skills for economic outcomes, influence on economic growth. Models that we turn to what can be said about their include direct measures of cognitive skills production in developing countries. Although can account for about three times the varia- information on enrollment and attainment tion in economic growth than models that has been fairly widely available, cognitive include only years of schooling; including the skills information has not. We use newly cognitive skills measures makes the coeffi- developed data on international assessments cient on years of schooling go to zero; and of cognitive skills (also employed in the anal- the estimates of such more inclusive models ysis of growth) to show that the education are far more robust to variations in the over- deficits in developing countries are larger all model specification. In both the earnings than previously appreciated. and the growth areas, we can reject the most Finally, in terms of framing the analysis frequently mentioned alternative explana- here, we motivate much of the discussion by tions of the relationships. a consideration of world development goals To be sure, while there is a policy link to and the impacts of educational quality on schools, none of this says that schools per se economic outcomes in the developing world. are the answer. Even though it is common to Yet the bulk of the evidence on outcomes treat education and schooling synonymously, has clear and direct implications for devel- it is important to distinguish between oped countries. In simplest terms, available knowledge and skills on the one hand and evidence suggests that the economic situa- schooling. This distinction has important tion in, for example, the United States and substantive underpinnings. Cognitive skills Germany is governed by the same basic edu- may be developed in formal schooling, cational forces as that in developing coun- but—as extensively documented—they may tries of South America. Thus, this should also come from the family, the peers, the not be thought of so much as a treatise on culture, and so forth. Moreover, other fac- policies toward Sub-Saharan Africa but as tors obviously have an important impact on an analysis of the fundamental role of cogni- earnings and growth. For example, overall tive skills on the operations of international economic institutions—a well-defined sys- economies. tem of property rights, the openness of the economy, the security of the nation—can be 2.