The above results are insufficient for drawing economic conclusions regarding the need for educated workers. True, education levels are rising in almost every subsector of the economy. But this could simply reflect a shortage of educated workers in the first place. To understand whether educated workers are over- or undersupplied, the discussion now turns to an analysis of the relative prices (wages) of workers of different education levels.
To do so, Box 3.2.1 introduces two important terms-the Mincerian and full rate of return-and explains how each of them could be interpreted to assess whether the labor force is "overeducated" relative to the job opportunities available.
Table 3.2.7 presents Mincerian returns to years of different levels of schooling for workers in two experience brackets-those who left school 10–15 years before the sample was drawn ("midcareer workers"), and those who did so 20–25 years before ("senior workers"). These returns are calculated from median wages paid to employees in each education category. Returns to education for all employees are presented first, followed by the returns among agricultural, industrial, and services sector workers. Returns for each experience bracket are provided from the initial survey year (10–13 years before) and the final survey year (2004 or 2005). There are several gaps in the table corresponding to returns that could not be estimated because there were too few sector workers with the necessary education and experience levels in the sample.
Several trends are apparent in the Mincerian returns. They are lower at lower levels of the school system, probably because private costs of basic education are lower as well. Moreover, returns to basic education (primary, middle, and sometimes LS school) have fallen in most sectors. Nowhere is this clearer than in the industry sector, where the returns to basic education are now moderately low. The immediate implication of falling returns is that if the supply of educated workers has been a constraint on industrial growth, this bottleneck is easing. One interpretation of this result is that the expansion of supply of suitably educated workers for industrial jobs has reduced scarcity. A possible complementary explanation is that, consistent with the writings of Harry Braverman (1974), industrialization and the splitting of the industrial supply chain across ever-more workers have led to tasks being fragmented to an extent where mental acuity and flexibility no longer matter as much.
In contrast with basic education, tertiary returns have risen. This is consistent with many studies documenting a polarizing income distribution in Asia.
3.2.1 Mincerian returns vs. full returns to education
The Mincerian return (Mincer 1958) to a particular education qualification is the percentage difference in wages earned by two workers who differ only in that one has the qualification and the other does not. The full (private) rate of return captures this wage premium, aggregated over the workers' careers, but also factors in the direct costs (fees, uniforms, books, etc.) of obtaining the qualification. Given that wages rise slowly over a worker's career, and that schooling is relatively expensive, the full rate of return is usually lower than the Mincerian return (for illustration, see Psacharopoulos and Ng 1994; for emphasis, see McEwan 1999).
Thus, if Mincerian returns are zero, full rates of return are usually negative. Or, if the Mincerian return is smaller than the return to other forgone investments in the economy, the full rate of return is likely to be below the opportunity cost of investment as well. In such a circumstance, the workforce may be said to be "overeducated" (Harberger 1965). Whereas a growing body of work on overeducation provides several other, generally weaker definitions of the phenomenon (McGuinness 2006), given the variety of other roles served by education in development, this chapter uses the definition least likely to take issue with investments in schooling economically.
This brief overview of the relationship between Mincerian and full returns was necessary, because while the results in the previous section imply that it will be useful to ask whether investments in education are paying off, good estimates of the direct cost of schooling are lacking. In the following section, any situation wherein the Mincerian return is less than the real cost of funds is taken to be indicative of "overeducation. "The return on a 10-year bond, corrected for expected inflation (box table) is taken as an estimate of the economywide real cost of funds. The economic cost of funds for education investment would certainly be higher than these rates, as sovereign loans carry lower rates of interest than private ones, especially private loans to finance education. Using these rates therefore deliberately biases the study against finding evidence of overeducation.
A cautionary word: overeducation may be short term, while investments in schooling are not, so the distance from such results to education policy should not be traversed in a single bound. Further, several benefits of education are not captured in wages, implying that private returns underestimate the social returns. Conversely, if talent is not observable, academic credentials may be used to dispense better jobs or promotions, in which case the private returns overstate the true social returns of education.
Estimated cost of funds, %
Period bond rate observed
Annual inflation in year of observation
10-year bond rate
Minimum real cost of funds
India
June 2004
6.4
5.9
-0.5
Indonesia
August 2004
6.1
6.1
0.0
Philippines
October 2004
6.0
13.5
7.0
Thailand
October 2005
4.5
6.1
1.5
Sources:Asian Development Outlook database; CEIC Data Company Ltd.; Money Market Association of the Philippines, available: http://www.mart.com.ph; Bloomberg; all downloaded 26 February 2007.
While the evidence from a very limited set of developing countries suggests that this type of "credentialism" is not empirically important for determining wages (Glewwe 2002), the issue could well be important in extreme environments where unemployment is high and education levels are rising fast. Thus, the following analysis of wage returns should only be taken as a crude indicator of where overeducation is occurring.
Rather than drawing education policy conclusions from this, it is hoped that the results will draw attention to other distortions that may be preventing the full benefits of schooling from being realized.
Returns in services are higher than in industry and agriculture. This suggests that the concentration of more-educated workers in services is not entirely residual. In contrast, comparing the returns between agriculture and industry does not yield obvious trends.
Finally, returns to higher secondary and tertiary education, and in the Philippines, LS education, have mostly held up much better for senior workers than their midcareer counterparts. These effects are most pronounced in industry and services (except in Thailand). Combined with falling returns to basic education and rising returns to tertiary, this yields a picture highly evocative of Nelson and Phelps' view of the links between education and structural change. Highly skilled workers in senior positions who can introduce new ideas and technologies are rewarded well for their education, while more junior workers are not.
3.2.7 Returns to education by sector, cohort, and country
All sectors
Agriculture
Industry
Services
Initial
Subsequent
Initial
Subsequent
Initial
Subsequent
Initial
Subsequent
MC
S
MC
S
MC
S
MC
S
MC
S
MC
S
MC
S
MC
S
India
1993/94
2004
1993/94
2004
1993/94
2004
1993/94
2004
Middle
9.1
10.1
6.3
12.3
0.0
2.3
3.4
0.0
11.1
11.9
7.3
3.6
5.2
13.1
8.7
14.5
Secondary
14.5
26.4
10.1
23.3
0.0
10.1
3.6
0.0
11.2
20.3
2.8
14.5
16.7
12.8
10.3
18.6
Higher secondary
22.5
7.3
14.6
16.0
0.0
12.8
3.3
8.0
6.9
16.3
5.9
18.4
13.1
Postsecondary
15.9
15.4
29.6
20.1
18.6
17.4
28.9
23.3
13.9
15.2
20.5
20.5
Indonesia
1994
2004
1994
2004
1994
2004
1994
2004
Primary school
3.1
12.2
5.8
6.0
6.6
5.3
17.4
Junior high school
11.6
14.0
11.2
18.6
12.0
26.6
10.1
21.6
7.4
11.9
7.7
6.8
11.6
9.1
14.1
21.1
Senior high school
11.9
14.9
10.9
19.7
10.1
14.5
10.1
10.8
18.6
10.1
12.6
11.7
13.5
13.0
14.5
Tertiary
9.3
10.3
12.5
5.7
24.8
25.7
17.0
18.9
8.4
9.0
12.5
5.7
Philippines
1991
2004
1991
2004
1991
2004
1991
2004
Elementary graduate
3.7
7.0
12.2
High school graduate
16.1
10.4
7.5
9.8
5.7
0.0
4.7
2.0
9.3
2.6
3.6
5.7
24.3
17.2
10.1
10.9
College graduate
12.6
11.6
18.9
22.2
11.0
17.0
11.1
18.9
16.0
10.4
19.6
21.3
Thailand
1995
2005
1995
2005
1995
2005
1995
2005
Elementary
7.4
10.6
11.0
6.2
7.0
3.1
10.0
3.8
4.2
7.6
11.4
11.3
5.8
Lower secondary
7.7
26.8
5.3
13.4
5.3
-1.4
12.6
5.4
21.3
3.7
7.1
11.1
25.0
8.4
15.5
Upper secondary
13.8
8.7
9.5
17.0
8.1
9.0
15.9
36.5
5.3
14.5
9.6
8.7
8.5
16.5
Postsecondary
16.1
15.6
19.4
28.7
13.6
7.4
23.3
36.1
16.7
14.7
17.2
24.6
- = insufficient sample size for calculation of returns.
MC = midcareer workers; S = senior workers.
Sources: India National Sample Survey Organisation, Socio-economic Survey, Schedule 10, 1993/94, 2004; Indonesia SAKERNAS 1994, 2004; Philippine Labor Force Survey, 1991, 2004, October rounds; Thailand Labor Force Survey, 1995, 2005, October rounds.
Turning next to country specifics, consider India, where returns to basic education in agriculture, which employs a large majority of the primary educated, are very low. Indeed, given the low Mincerian returns, the full returns to education in India's agricultural sector are probably negative (the direct costs of schooling in India are reported to be high relative to the incomes of the poor)-a certain indicator of over-education. Overeducation in the context of rampant illiteracy implies only that the problem originates outside the education system. While India's public schools are in a sorry state, especially in rural areas, these startling results cannot be ascribed to school quality. Economywide returns to basic education in India were much higher in 1993/94, but have fallen since then. Problems detailed in PROBE (1999) indicate that quality has been a problem for a lot longer. It is therefore more likely that the low returns in agriculture reflect the well-documented stagnation of agricultural productivity. Reasons for this stagnation, and what to do about it, are presented in the chapter on India in Part 2 of ADO 2007.
As in other countries, India's returns to basic education in industry have fallen (as well as in agriculture), further dragging down the overall returns. Economywide, only returns to college and, for older workers, higher secondary education, have grown. However, returns at some levels of education and experience are rising in the services sector.
The Indonesian story is complicated by reverse migration. Less-educated workers in Indonesia were disproportionately pushed out of industry into lower-productivity agricultural jobs, driving up the returns to schooling. This was probably especially relevant for senior workers, among whom job stability would be positively correlated with education. This exemplifies the idea that if job security is tied to education, high returns in a turbulent environment do not necessarily reflect a real scarcity of educated workers. Conversely, education expansion may have exerted pressure in the other direction, driving returns down. The inability to track changes in Indonesian employment structure at a more disaggregate level precludes a more detailed interpretation of these results.
Returns to LS schooling fell sharply in the Philippines, albeit from a high base. In contrast, tertiary education has become more valued over time. Certainly, falling returns to LS schooling are consistent with the increased supply of educated workers, deindustrialization, and structural stagnation in services described in the previous section. Moreover, the next section shows that in some of the faster growing services occupations, returns to education are approximately zero.
It is also useful to note that in the Philippines there has been a substantial increase in education requirements for access to jobs. A survey compiled by the Bureau of Labor and Employment Statistics shows that LS or college degrees are now required for entry to every formal sector job in the country. An obvious interpretation is that as education levels rose, education criteria for the same old jobs became tighter, and the returns (probably associated with rank in the company hierarchy) simply migrated up the education ladder. This view sits well with the findings of the Philippine Presidential Commission on Educational Reform 2000 Report, which lamented that: "The country has too long suffered the imbalance of an overly credential-conscious society, which puts a premium more on diplomas than knowledge or skills, and values prestige institutions granting degrees more than the competence that the degree itself embodies. "
It is important to note that according to the Commission on Higher Education, 66% of college education in the Philippines in 2003/04 was privately provided, and therefore unsubsidized. This may help to explain why wage returns to college have not been driven down by supply.
The real cost of funds in the Philippines in 2004 was roughly equal to the LS Mincerian returns in 2004, so overeducation is likely, especially in agriculture and industry. Having said this, the high cost of funds at that time was due to the Philippine Government's poor external debt position, which is now on the mend. Thus overeducation could be temporary. However, unless long-term stagnation within and outside agriculture is reversed, returns could fall well below the cost of funds.
Finally, in Thailand, where secondary education was made free of charge in the period considered, the supply of LS graduates has boomed and returns have fallen. While midcareer workers are too old to have taken advantage of these public subsidies, they are probably more readily replaceable by younger workers than the older cohorts. Thus it seems reasonable that the entry of large numbers of new graduates would have driven down their returns on LS schooling. However, most college schooling is still not subsidized in Thailand, and returns to college are therefore still rising, particularly as the services sector takes off. There is also no apparent bottleneck in the supply of industrial workers with basic education. The somewhat low returns to basic education are compensated for by a low cost of funds, so it is not clear whether Thai workers are overeducated.
Top of page