1. 1 Education, Skills, and Labour Market Outcomes: Exploring the Linkages in Canada W. Craig Riddell Department of Economics University of British Columbia and Team for Advanced Research on Globalization, Education and Technology (TARGET) TARGET website: www.econ.ubc.ca/ine Presentation to Association of Professional Economists of B.C. October 2004

2. 2 Education, Skills, and Labour Market Outcomes Objective is to assess recent evidence on the relationship between the resources devoted to education and skill formation, and the labour market consequences of those expenditures. The widespread tendency to associate advances in technology and other sources of economic change with the need for greater emphasis on student achievement presumes that there are clear linkages between what students learn in school and their subsequent economic success. This presentation will examine recent research on the nature of these linkages, and will assess the implications of this research.

3. 3 The presentation is organized as follows: 1.Canadian educational expenditures and educational outcomes are compared to those of other countries. Several educational outcomes are discussed: educational attainment, student achievement, and the literacy skills of the adult population. This comparative examination of educational “inputs” and “outcomes” provides a basis for assessing whether Canada obtains good value from its public and private investments in education. 2.I then examine the labour market consequences of education. The question of how best to interpret the strong positive correlation between education and economic success has long been a subject of debate and controversy. Substantial recent progress has been made on this issue. 3.The third part discusses a new area of research opened up by advances in data collection: the linkages between education and literacy skills, and the labour market consequences of such skills.

4. 4 Motivation and background Education, training and skill formation have become prominent public policy issues. Several factors account for the increased attention being paid to the knowledge, skills and competencies of the population: 1.Technological change -- especially advances in information and computer technologies -- has resulted in growing demand for highly skilled workers and changes in the nature of skills needed in the workplace.

5. 5 Motivation and background cont. 2.Growing concern about future skills shortages, in part due to the fact that the leading cohorts of the well-educated "baby boom" generation are now approaching retirement age and are being replaced by much smaller (though even better-educated) cohorts. 3.Resurgence of interest in the determinants of long-term growth in living standards. "New growth theory" emphasizes the importance of human capital in the creation of new knowledge and in the growth of living standards.

6. 6 As economic activity becomes more knowledge-based, human capital is also increasingly being viewed as a central component of social policy: 1.Major reassessment of the “welfare state” programs that emerged in the post war period. Governments are moving away from "passive" income maintenance programs toward "active" labour market and social policies that facilitate adjustment to change, assist the jobless to find work, and encourage labour force participation.

7. 7 2.Greater emphasis on individual responsibility and on providing those in need of assistance with the opportunity to improve their economic situation -- providing a "hand up" rather than a "handout". As stated by Paul Martin, “Providing security and opportunity for Canadians in the future means investing in their skills, in their knowledge and capacity to learn....good skills are an essential part of the social safety net of the future."

8. 8 3.Education and training may ameliorate pressures for widening inequality in economic and social outcomes. 4.Education is also often regarded as a mechanism for promoting equality of opportunity and social mobility. From the perspective of social policy, equality of opportunity may contribute to social cohesion and a belief in common interests among citizens.

9. 9 Differing perspectives Commenting on this emergence of a common emphasis on human capital formation in both economic and social policy, Courchene (2001) states that we are presented "... with an historically unprecedented window of opportunity...[in which]... a societal commitment to a human capital future is emerging as the principal avenue by which to promote both economic competitiveness and social cohesion.” However, some observers disagree with the emphasis being placed on education and skill formation. Examples include: David Livingstone, The Education Jobs Gap 1999 Gordon Lafer, The Job Training Charade 2002

10. 10 Canadian Investment in Education Education systems vary substantially across countries, making international comparisons of educational inputs and outcomes difficult Relative to other developed countries, Canada invests a substantial amount on education. Most of this expenditure is publicly financed. Canadian expenditure per student on elementary and secondary is near the top of the OECD and second highest (after the U.S.) among the G-7

11. 11 Canadian Investment in Education cont. Expenditure on post-secondary is also among the highest in the OECD and G-7, but substantially below the U.S. Relative to other countries, Canadian expenditure is especially high at non- university post-secondary level

12. 12 Educational Expenditures in Canada and G-7 Countries, 1995 Expenditure per student from public and private sources by level of education, in U.S. dollars converted using purchasing power parity (PPP) exchange rates CanadaFranceGermanyItalyJapanU.K.U.SOECD Average Elementary & Secondary 5,4015,0414,6905,0994,2823,8106,2814,162 Post- secondary 11,4716,5698,8975,0138,7687,22516,2628,134 All levels of Education 6,3965,0016,0575,1574,9914,2227,9054,717 Source: Council of Ministers of Education, Canada and Statistics Canada (2000)

13. 13 Expenditure relative to GDP Expenditure on education as a percentage of GDP is highest in the G-7 countries and among the highest in the OECD Canada's relatively high percentage of GDP spent on education reflects both the substantial per-student expenditures on education at all levels, and Canadian's comparatively high participation rates in education, especially at the non-university post-secondary level. At post-secondary level, Canada and US have similar expenditures, both substantially above other G-7 and OECD countries However, dramatic difference in composition of PSE expenditure between Canada and US. Canada spends much less on university sector and much more on non-university PSE sector

14. 14 Educational expenditure from public and private sources for educational institutions as a percentage of GDP, by level of education, Canada and G-7 countries, 1995 CanadaFranceGermanyItalyJapanU.K.U.SOECD Average Elementary & Secondary 4.34.43.83.23.1---3.93.7 Post- secondary 2.51.1 0.81.0 2.41.3 All levels of Education 7.06.35.84.7 ---6.75.6 Source: Council of Ministers of Education, Canada and Statistics Canada (2000)

15. 15 Differences across provinces There are also moderately large differences in expenditure per student across Canadian provinces, especially at the elementary and secondary level. For example, Ontario -- the province with the highest per student expenditure -- spends about 50% more than low spending provinces such as Nova Scotia, New Brunswick and Prince Edward Island.

16. 16 Expenditure per student by province in Canadian dollars, 1995 NfldPEINSNBQueOnt Elementary & Secondary 5,5164,7615,3945,1207,0637.617 Post- secondary 12,72413,77013,46312,53212,68713,122 All levels of Education 7,0796,3276,9756,6238,4067,963

17. 17 Expenditure per student by province in Canadian dollars, 1995 cont. ManSaskAltaBCCanada Elementary & Secondary6,6605,5075,5206,3996,677 Post-secondary15,61515,05112,99414,73814,182 All levels of Education7,6897,0326,9318,1627,907 Source: Council of Ministers of Education, Canada and Statistics Canada (2000)

18. 18 Outcomes of Educational Expenditures By various measures, Canadian educational attainment is high by international standards 80% of Canada's adult population has completed high school, versus OECD average of 64% –Similar to that of Germany, Japan, and the U.K. –Substantially below the U.S. (87%)

19. 19 Outcomes of Educational Expenditures cont. Canada stands out in terms of the fraction of the adult population with completed post-secondary education –Canada's proportion (52%) is more than double the OECD average of 25%, and highest in the OECD countries –Also substantially higher than the U.S., the country ranked second (where 35% have completed post-secondary education).

20. 20 Outcomes of Educational Expenditures cont. Canada's extremely high ranking on this dimension arises principally because of the very substantial fraction of the population with non- university post-secondary education—at 33%, triple the OECD average and more than double any other G-7 country. At the university level, Canada (19%) is above the OECD average (14%) and similar to Japan and the U.K., but substantially below the U.S. where 27% have graduated from university

21. 21 Educational Attainment in Canada and G-7 Countries Proportion of the population aged 25-64 years by highest level of educational attainment, 1999 CanadaFranceGermanyItalyJapanU.K.U.SOECD Average Less than upper secondary 2038195719181336 Upper secondary graduate 2841533049575140 Non- university post- secondary 3310154138811 University graduate 191113918172741 Source: OECD (2001)

22. 22 Average completed years of schooling of the population aged 25- 64 years, 1995 CanadaFranceGermanyItalyJapanU.K.U.S.OECD Average 13.211.213.410.0---12.113.511.9

23. 23 Average completed years of schooling of the population aged 25-64 years, 1995 CanadaFranceGermanyItalyJapanU.K.U.S.OECD Average 13.211.213.410.0---12.113.511.9 Source: OECD (1998)

24. 24 Canadian distribution of educational attainment The distribution of the educational attainment of Canadians has unique features. At the bottom and top of the distribution—specifically, those with less than completed high school and those with a university degree—Canadian educational attainment is similar to that of several other OECD countries and significantly lower than that of the U.S. However, in the middle of the distribution the proportion of Canadians with a community college diploma or trade school certificate is unusually high and the proportion of high school graduates relatively low.

25. 25 Canadian distribution of educational attainment cont. However, this “non-university post-secondary” group is heterogeneous. There are various routes to a community college diploma or trade school certificate. More than one-quarter of the “non-university post- secondary” group have not graduated from secondary school, and their average years of completed schooling is not much different from those whose highest educational attainment is high school completion. Quebec’s CEGEP system also results in some overstatement of importance of non-university PSE in Canada relative to other countries

26. 26 High School Completion Although the overall educational attainment of Canadians is impressive, high school completion has been a weak spot for many years. For example, the Canadian secondary school graduation rate is near the bottom of the G-7 countries and only marginally above that of the U.S., the bottom dweller on this dimension As of the mid-1990s, approximately 25% of 18 years olds had not graduated from high school. This non-completion rate is much higher among males (30%) than females (20%).

27. 27 High School Completion cont. Some of these dropouts graduate after the “normal age” of 18; e.g. the high school graduation rate is 81% by age 19-20 and 87% by age 25-29 In addition a significant number of high school dropouts obtain a college diploma or trade certificate. Nonetheless, Canada's high school dropout rate is a concern.

28. 28 Ratio of Upper Secondary Graduates to Population at a Typical Age of Graduation, 1996 CanadaFranceGermanyItalyJapanU.K.U.S Both sexes 7585867999---72 Males 7085867696---69 Females 8186 82102---76 Source: Council of Ministers of Education, Canada and Statistics Canada (2000)

29. 29 Outcomes of Educational Expenditures Student achievement We know a good deal more about student achievement than we did even a decade ago. Canada did not participate in the early rounds of international mathematics and science tests carried out in the 1960s and 1970s. However, some provinces took part in the Second International Mathematics and Science Studies carried out in the 1980s, and all Canadian jurisdictions except P.E.I. participated in the third round—the Third International Mathematics and Science Study or TIMSS—carried out in the 1990s.

30. 30 Outcomes of Educational Expenditures cont. In addition, the decade of the 1990s saw the introduction of the Canadian School Achievement Indicators Program, which has now completed several rounds of testing. Canadian secondary school students participated in the recent Program of International Student Assessment (PISA) tests carried out in several OECD countries

31. 31 Results from the 1990s TIMSS tests These data have the advantage of providing information on student performance on a common set of curriculum- based tests. One disadvantage is that the set of countries is very diverse. The set of countries participating in each test also varies, so the international average needs to be interpreted cautiously. According to these recent international tests, Canadian student achievement in mathematics is average or somewhat above average among a diverse set of countries. Within the G-7, Canada generally ranks in the middle of the participating countries.

32. 32 Results from the 1990s TIMSS tests cont. In science, Canadian student performance is above average among the full set of countries that took the tests but about average among the G-7 participants—above France and Italy but below Japan and similar to England, Germany and the U.S. Overall, these results indicate that Canadian student achievement is satisfactory but not as good as one might expect given Canada’s relatively high expenditure on elementary and secondary schooling.

33. 33

34. 34 Differences across provinces Substantial provincial variation is evident. In mathematics, Quebec student achievement is substantially above the Canadian average and high by international standards, albeit still significantly below the top-ranked countries (Korea, Singapore, Hong Kong, and Japan). Ontario's student performance tends to be below the national average, although the differences are not always statistically significant.

35. 35 Differences across provinces cont. Alberta student performance is impressive in science. At both Grade 4 and Grade 8 Alberta's students rank among the best in the world, exceeded only by Korea and Japan. Although overall national levels of Canadian student performance in mathematics and science are not outstanding, some provinces—such as Quebec in mathematics and Alberta in science—are able to obtain very high levels of achievement within the existing Canadian social, cultural, and fiscal framework. In other provinces, especially the Atlantic provinces, student achievement generally falls below the Canadian average and is relatively low by international standards.

36. 36

37. 37 Program of International Student Assessment (PISA) tests These tests assessed the skills of 15-year-old students in reading, mathematics and science in over 30 (mainly OECD) countries. Whereas TIMSS is curriculum-based and assesses students' knowledge of the subject matter, PISA assesses the ability of 15-year-ols to apply their knowledge to real world issues and challenges. In contrast to the TIMSS results, the performance of Canadian secondary school students in the PISA tests was excellent. Across the three assessed subjects, Canada, Japan and the U.K. are the top three countries, followed by France and the U.S. in the middle and Germany and Italy at the bottom.

38. 38 Program of International Student Assessment (PISA) tests cont. Consistent with the TIMSS findings, Canadian student achievement was below that of Japan in math and science. However, the PISA results paint a more favourable picture than do earlier international assessments. The gap between Canada and the bottom four G-7 countries (France, U.S., Germany and Italy) is large in all three of the subject areas, and in reading Canadian students were the top performers among the G-7 countries. Canadian student achievement was equally impressive among the larger group of OECD countries.

39. 39

40. 40 Provincial results in PISA Variation in student achievement across provinces is similar to that observed in TIMSS, with Alberta, Quebec and B.C. students being in the top group and students in the four Atlantic provinces being at the bottom. Quebec students continue to achieve the highest scores in math and Alberta students perform best in science (as well as in reading). Ontario, Manitoba and Saskatchewan tend to have scores close to the national average.

41. 41

42. 42 Summary of evidence on student achievement Caution is appropriate in interpreting these summary statistics on student performance in reading, mathematics and science. Many factors, in addition to the resources devoted to the school system, influence student achievement. For example, Canada has a high proportion of immigrant children (for whom English or French is often a second language) in its schools. Furthermore, countries may differ in the extent to which they aim to raise average performance or to principally improve achievement among those who would otherwise perform poorly.

43. 43 Summary of evidence on student achievement cont. Nonetheless, these measures of student achievement -- especially the PISA results -- suggest that Canada appears to obtain reasonably good "value for money" from the elementary and secondary school system. Canada ranks at or near the top of the G-7 countries in terms of expenditure per student on elementary and secondary schooling and --according to the PISA findings -- places in the top three of the G-7 countries in terms of student performance. However the TIMSS results lead to a somewhat less favourable assessment. According to these tests, Canada is in the middle or upper middle of the pack in student achievement in math and science.

44. 44 Provincial variation in student achievement Although overall national levels of Canadian student performance in mathematics and science are not outstanding, some provinces— such as Quebec in mathematics and Alberta in science—are able to obtain very high levels of achievement. In other provinces, especially the Atlantic provinces, student achievement generally falls below the Canadian average and is relatively low by international standards.

45. 45 Provincial variation in student achievement cont. There is a positive relationship between expenditure per student and student achievement as measured by TIMSS and PISA. However, the relationship between expenditure and test scores is relatively weak, with Ontario having the greatest expenditure per student but student achievement that is typically about average. Similarly, Alberta combines frugality in its expenditure on elementary and secondary education with very high levels of student achievement. The source of these provincial variations is an important subject for future research.

46. 46 Outcomes of Educational Expenditures Literacy skills of the adult population Student achievement tests provide information on the skills of those who will be entering the labour force in the future -- that is, the flow of new entrants. Until recently, however, no nationally representative measures of the skills and knowledge of the existing stock—the adult population—were available.

47. 47 Outcomes of Educational Expenditures cont. The International Adult Literacy Survey (IALS), which was carried out in over 20 countries during the 1994–98 period, represents a breakthrough in international data collection, providing for the first time measures of the literacy and numeracy skills of the adult population that are comparable across countries and language groups. The survey provided three measures of literacy: prose, document, and quantitative literacy (or numeracy). These measures correspond to information-processing skills needed to perform everyday tasks at home, at work, and in the community.

48. 48 Key findings from the IALS Results are reported for Canada and other participating G-7 countries (Germany, U.K., and U.S.) The top panel shows the mean score and the score at the 25th and 75th percentiles of the literacy distribution. The average scores rank Germany at the top (with the exception of the prose scale, on which Canada ranks first), followed by Canada, the U.S., and the U.K. at the bottom.

49. 49 Key findings from the IALS cont. Although the differences in mean scores may not appear large, they are non-trivial. For example, on the document scale, the mean score in Germany, the top-ranked country, is 285, while that in the U.K., the bottom-ranked country, is 268. An individual with a score of 268 is in the middle of the distribution in the U.K. but would be at approximately the 33rd percentile of the distribution in Germany—that is, about two- thirds of the adult population would have superior document literacy skills.

50. 50 Literacy Skills in Canada and G-7 Countries, 1994-98 Mean scores and scores at the 25th and 75th percentiles of the prose, document and quantitative literacy scales Literacy scale CanadaGermany 25 th mean75 th 25 th mean75th Prose243279322245276308 Document243279326256285318 Quantitative247281323265293324 Sources: OECD (1998) and OECD and Statistics Canada (2000).

51. 51 Literacy Skills in Canada and G-7 Countries, 1994-98 cont. Mean scores and scores at the 25th and 75th percentiles of the prose, document and quantitative literacy scales Literacy scale U.K.U.S. 25 th mean75 th 25 th mean75th Prose 233267311237274320 Document 230268314230268316 Quantitative 231268314237275322 Sources: OECD (1998) and OECD and Statistics Canada (2000).

52. 52 Distribution of literacy skills Compared with many countries participating in the IALS survey, especially continental European countries, Canada, the U.K., and the U.S. display substantial variation in the literacy skills of the adult population. This phenomenon is evident from a comparison of the lower and upper tails of the literacy distributions for Canada and Germany. At the 25th percentile, the German score exceeds that of Canada on all three literacy scales, with the differential being especially large for document and quantitative literacy.

53. 53 Distribution of literacy skills cont. However, at the 75th percentile the Canadian score exceeds that of Germany in both prose and document literacy and is approximately equal to that of Germany in quantitative literacy. In general, individuals in the top 25 percent of the Canadian literacy distribution have higher literacy skills than their German counterparts, while individuals in the bottom one-quarter of the Canadian literacy distribution have lower skills than their German counterparts.

54. 54 Incidence of low literacy Next table shows the percent of adults with low literacy skills (level 1 or level 2) by broad age groups. For the adult population as a whole, the ranking is the same as before: Germany has the lowest percent of adults with low literacy (the exception being the prose scale, on which Canada ranks at the top), followed by Canada, the U.S., and the U.K.

55. 55 Incidence of low literacy cont. However, important differences in the country rankings are evident among age groups. Among young adults (16-25 years of age), Canadians perform best, followed closely by Germans. In the U.K., and especially in the U.S., the incidence of low literacy skills among young adults is much higher. In contrast, Canada has the highest incidence of low literacy skills among those 46-55 years of age, exceeding even the U.K. on this dimension and substantially above Germany.

56. 56 Percent of adults with low literacy skills Age group Literacy scaleCanadaGermanyU.K.U.S. 16-65Prose42495247 16-65Document434250 16-65Quantitative43335146 16-25Document33344456 46-55Document54425350 16-65Document434250 Sources: OECD (1998) and OECD and Statistics Canada (2000).

57. 57 Education and literacy Among those with less than a completed secondary school education, literacy skills of Canadians are very poor, substantially below the U.K. and Germany but above the U.S. However, average literacy scores improve substantially with educational attainment, and this gradient is steepest in Canada. Canadian high school graduates rank second (after Germany) among this group of countries, and post- secondary graduates rank at the top, despite the very large fraction of the Canadian population with completed post-secondary education.

58. 58 Mean document literacy score and educational attainment EducationCanadaGermanyU.K.U.S. Less than high school 227276247200 High school graduate 288295286266 Post- secondary graduate 3183153123 03 All adults279285268 Sources: OECD (1998) and OECD and Statistics Canada (2000).

59. 59 Summary comments on adult literacy Canadian literacy skills are reasonably good by international standards, especially among younger cohorts and post-secondary graduates. However, the literacy skills of older Canadians and those with less than a high school education are relatively poor. In all of these countries, a disturbingly large fraction of the population has low levels of prose, document, and quantitative literacy.

60. 60 Summary comments on adult literacy cont. Nonetheless, to the extent that these information-processing skills used in daily activities are an outcome of the education system, the IALS data suggests that Canada's education system may be doing a reasonably good job of enhancing literacy skills. This favourable assessment is similar to that resulting from the recent PISA tests of student achievement, which also assessed the ability to apply knowledge to challenges that arise in daily activities.

61. 61 Education and labour market outcomes Education is one of the best predictors of “who gets ahead” Better-educated workers earn higher wages, have greater earnings progression over their lifetimes, experience less unemployment, and work longer. Higher education is also associated with longer life expectancy, better health, and reduced participation in crime The strong positive association between education and earnings is one of the most well established relationships in social science

62. 62 Education and labour market outcomes cont. Many social scientists have, however, been reluctant to interpret this correlation as evidence that education exerts a causal effect on earnings The positive relationship between earnings and schooling may arise because both education and earnings are correlated with unobserved factors such as ability, perseverance, and ambition In these circumstances, standard estimates of the return to schooling are likely to be biased upwards because they do not take into account unobserved “ability”

63. 63 Education and labour market outcomes cont. Assessing the question of whether education exerts a causal influence on earnings has been the subject of much recent research in labour economics The answer to this question is crucial not only for how to interpret the relationship between education and earnings, but also for economic and social policy relating to education To the extent that estimates of the return to schooling are biased upwards because of unobserved factors, estimated average rates of return to education may substantially over-predict the economic benefits that a less-educated person would receive if he/she acquired additional schooling.

64. 64 Education and labour market outcomes cont. The estimated average rates of return in the population reflect both the causal effect of schooling on productivity and earnings and the average return to the unobserved ability of the well-educated. If those with low levels of education are also, on average, those with low ability or ambition, they can only expect to receive from any additional schooling the return associated with the causal effect of schooling on earnings. The marginal return—the impact of additional schooling for someone with low levels of education—may be substantially below the average return.

65. 65 Education and labour market outcomes cont. In these circumstances, education may not be very effective in improving the employment or earnings prospects of relatively disadvantaged groups. Estimates of the causal effect of education on earnings are thus important for economic and social policy. How can such estimates be obtained? The most reliable method would be to conduct an experiment. Individuals randomly assigned to the treatment group would receive a larger “dose” of education than those assigned to the control group.

66. 66 Education and labour market outcomes cont. By following the two groups through time we could observe their subsequent earnings and obtain an unbiased estimate of the impact of schooling on labour market success. Random assignment would ensure that, on average, treatment and control groups would be equally represented by “high ability” and “low ability” individuals. In the absence of experimental evidence, labour economists have studied quasi-experiments or “natural experiments” that isolate the influence of education from the possible effects of unobserved ability.

67. 67 Education and labour market outcomes cont. A large number of such studies have now been carried out, using data on identical twins or on sources of variation in education such as those implied by compulsory schooling laws or proximity to a college or university. A consistent result of these studies is that conventional OLS estimates of the return to schooling tend, if anything, to under-estimate rather than over-estimate the causal impact of education on earnings.

68. 68 OLS and IV Estimates of the Return to Education StudyCountry, Instrument Returns to Schooling OLSIV Angrist & Kreuger (1991) US compulsory schooling laws 0.070 0.063 0.052 0.101 0.060 0.078 Staiger & Stock (1977) US compulsory schooling laws 0.063 0.052 0.098 0.088 Harmon & Walker (1995) UK compulsory schooling laws 0.0610.153 Kane & Rouse (1993) Tuition, distance to college, US 0.080 0.063 0.091 0.094 Card (1995)US, distance to nearest college 0.0730.132 0.097

69. 69 StudyCountry, Instrument Returns to Schooling OLSIV Conneely & Uusitalo (1997) Finland, living in university town 0.085 0.083 0.110 0.098 Lemieux & Card (2001) Canada, WWII veterans rehab act, Ont/Que 0.070 0.062 0.164 0.076 Meghir & Palme (2000) Sweden education reforms 0.0280.036 Sweetman (2000) Canada, Nfld education reform females males 0.146 0.108 0.170 0.221 OLS and IV Estimates of the Return to Education cont.

70. 70 Why do conventional estimates generally understate the true return to schooling, when the presence of “omitted ability bias” should cause these estimates to be upward biased? The reason appears to be that there are two additional sources of bias that operate in the opposite direction. First, is the presence of measurement error in educational attainment (especially years of completed schooling). Measurement error in an explanatory variable causes the estimated coefficient to be biased toward zero. The downward bias due to measurement error thus acts in the opposite direction to any upward bias associated with unobserved ability. Second, is what is sometimes referred to as “discount rate bias.” The returns to schooling are not the same for all individuals in the population; rather, there is a distribution of such returns.

71. 71 Consider the case of individuals with high potential returns to education who do not pursue higher education—perhaps because of low family income, limited ability to borrow in order to finance human capital formation, or a family background in which the importance of education is not emphasized. For these “high potential return” individuals, a policy intervention that results in increased educational attainment would have a substantial payoff. Indeed, the marginal return to the investment may exceed the average return in the population.

72. 72 Two recent Canadian studies have pursued this “natural experiment” approach. Lemieux and Card (2001) study the impact of the Veterans Rehabilitation Act—the Canadian “G.I. Bill”. In order to ease the return of World War II veterans into the labour market, the federal government provided strong financial incentives for veterans to attend university or other sorts of educational programs. Because many more young men from Ontario than Quebec had served as soldiers, those from Ontario were significantly more likely to be eligible for these benefits.

73. 73 Lemieux and Card estimate that the VRA increased the education of the veteran cohort of Ontario men by 0.2 to 0.4 years. Further, they estimate the rate of return to schooling to be 14 to 16 percent, substantially higher than the OLS estimate with their data of 7 percent. Sweetman (2000) investigates the impact on education and earnings of the education policy change in Newfoundland that raised the number of years of schooling required for high school graduation from 11 to 12.

74. 74 He estimates that this intervention increased educational attainment of affected Newfoundland cohorts by 0.8 to 0.9 years. Estimated rates of return to the additional schooling are substantial: 17.0% for females (versus an OLS estimate of 14.6%) and 11.8% for males (compared to an OLS estimate of 10.8%). As with this growing body of research, these Canadian studies conclude that conventional OLS estimates of the return to schooling are likely, if anything, to be biased downwards, as opposed to being inflated by unobserved ability.

75. 75 Two principal conclusions follow from this body of research. First, rates of return to investments in education are high. Typical OLS estimates produce real rates of return in Canada in the order of 6-9% for males and 8-10% for females. Such estimates compare favourably with rates of return on physical capital investments. Second, the payoff to marginal investments in education may exceed the average return in the population. There is no evidence that investments in higher education are experiencing diminishing returns because they require society to “reach lower into the ability barrel”.

76. 76 Policy interventions that result in additional schooling being acquired by individuals from disadvantaged backgrounds or those who face other barriers to acquiring human capital, may yield a substantial return in the form of enhanced employability and earnings, in addition to contributing to equity objectives.

77. 77 Education, skills and labour market outcomes Most research on the determinants of labour market success uses only relatively crude indicators of human capital such as educational attainment and years of work experience. Individuals with the same educational attainment and years of experience may have substantially different skills, depending on numerous other factors. More generally, education and experience are "inputs" into the production of human capital, not direct measures of the "outcomes" -- a set of skills, competencies and knowledge.

78. 78 Education, skills and labour market outcomes cont. Although the relationships between inputs such as education and experience and outcomes such as employment and earnings have been extensively investigated, relatively little is known about the relationship between direct measures of skills and labour market outcomes. Green and Riddell (2003) use the Canadian component of the International Adult Literacy Survey (IALS) to investigate the relationship between education, skills and labour market earnings.

79. 79 Education, skills and labour market outcomes cont. Conventional estimates of the return to schooling and to experience confound two effects. The first is the impact of education and experience on skill production -- the relationship between human capital inputs such as education and experience and outputs such as literacy skills or problem-solving skills. The second is the value placed on various skills in the labour market -- the relationship between literacy or problem-solving skills and market earnings.

80. 80 Education, skills and labour market outcomes cont. When skills are not directly observed, the best researchers can do is to analyse the relationship between human capital inputs and labour market outcomes. However, the availability of directly observed skills in data sets that also contain information on labour force behaviour allows researchers to "unpack" these two effects to some extent -- to obtain estimates of both the skill production effect and the market valuation effect.

81. 81 Education, skills and labour market outcomes cont. Green and Riddell find that formal education exerts a substantial effect on the production of literacy skills in Canada. However, they conclude that labour market experience has essentially no net effect on literacy production. These results suggest that policies aimed at improving cognitive skills such as literacy should focus on formal schooling. Policies designed to increase work experience can lead to earnings growth but they appear unlikely to enhance the cognitive skills of the workforce.

82. 82 Education, skills and labour market outcomes cont. Another important result is that the Canadian labour market places a high value on literacy skills. A 20-point increase in the literacy score -- equivalent to 1/3rd of a standard deviation of the literacy score distribution -- produces an increase in earnings equal to that associated with an extra year of formal schooling.

83. 83 Education, skills and labour market outcomes cont. Together these results imply that a significant amount of the "return to education" as conventionally measured represents the combined effects of the contribution of schooling to producing literacy skills and the value placed on literacy in the labour market. Indeed, Green and Riddell estimate that about one-quarter to one-third of the "return to education" is associated with these effects. The remainder reflects the impact of education on the production of other (unobserved) skills that are valued in the labour market.

84. 84 Summary and Conclusions Canada invests heavily in education. Relative to other G- 7 or OECD countries, Canada ranks near the top in terms of expenditure per student or the fraction of GDP devoted to elementary, secondary, and post-secondary education. Within Canada there are large differences across provinces in expenditure per student, especially at the elementary and secondary level. One consequence of this substantial expenditure is a population that is well educated by international standards. Canada compares favourably with other G-7 and OECD countries in terms of most measures of educational attainment.

85. 85 Summary and Conclusions cont. Compared to the U.S., Canada has lower educational attainment at both the bottom (less than completed high school) and top (university degree) of the education distribution. Where Canada stands out is in the middle of the distribution—those who have completed high school or a non-university secondary program. The proportion of Canada's population with a non-university post-secondary education is much higher than that of any other OECD country.

86. 86 Summary and Conclusions cont. Recent international data on Canadian student achievement paint a mixed picture. Results from TIMSS, curriculum-based tests of achievement in mathematics and science, indicate that Canadian student achievement is satisfactory but not as good as one might expect given Canada's relatively high expenditure on elementary and secondary schooling. In contrast, the performance of 15-year-old Canadian students in the recent PISA tests, which assessed the ability to apply knowledge in reading, mathematics and science, was excellent.

87. 87 Summary and Conclusions cont. The PISA results suggest that Canada appears to obtain reasonably good "value for money" from the elementary and secondary school system. However the TIMSS results lead to a somewhat less favourable assessment. There are substantial variations in student achievement across Canadian provinces. Some provinces—such as Quebec in mathematics and Alberta in science—are able to obtain very high levels of achievement within the existing Canadian social, cultural, and fiscal framework. In other provinces, especially the Atlantic provinces, student achievement generally falls below the Canadian average and is relatively low by international standards.

88. 88 Summary and Conclusions cont. The literacy skills of the adult population are above average among the G-7 countries that participated in the IALS survey. Canada, like the U.S. and U.K., has a high variance across the population in its literacy skills compared to European countries such as Germany. By international standards, older and less well-educated Canadians have relatively poor literacy skills, whereas younger and well-educated Canadians have relatively good literacy skills compared to their counterparts in other G-7 countries. The IALS data thus suggest that Canada's education system is doing a reasonably good job of enhancing literacy skills.

89. 89 Summary and Conclusions cont. Canadian studies using conventional methods to analyse the relationship between education and earnings obtain estimates of the “return to schooling” that are similar to those obtained in many studies carried out in other developed countries: real rates of return of approximately 8- 10%. Such estimates compare favourably with rates of return on physical capital investments.

90. 90 Summary and Conclusions cont. Important recent advances have taken place in our understanding of the relationship between education and labour market success. Conventional estimates of the return to schooling appear, if anything, to be biased downward—so the causal effect of education on earnings appears to be higher than previously believed. Further, the marginal return to incremental investments in education may exceed the average return from previous investments. There is no evidence that investments in schooling are running into diminishing returns.

91. 91 Summary and Conclusions cont. Conventional estimates of the return to schooling and to work experience confound two effects: the impact of education and experience on skill production and the value placed on skills in the labour market. The availability of data on directly observed skills allows researchers to "unpack" these two effects. Recent evidence indicates that formal education exerts a substantial effect on the production of literacy skills in Canada. However, work experience has no net effect on literacy production.

92. 92 Summary and Conclusions cont. Canada's labour market places a high value on literacy skills. A 20-point increase in the literacy score -- equivalent to 1/3rd of a standard deviation of the literacy score distribution -- produces an increase in earnings equal to that associated with an extra year of formal schooling. A substantial fraction -- one-quarter to one-third -- of the "return to education" as conventionally measured represents the combined effects of the contribution of schooling to producing literacy skills and the value placed on literacy in the labour market.