The Economics of the Public Sector Second Half

1
The Economics of the Public Sector Second Half

• Topic 5 Analysis of Human Capital Policies
• The Returns to Schooling

2

• In the rest of the course we will analyze some
  specific policies to foster skill formation over
  different stages of the life cycle. Assuming that
  there are reasons for government intervention in
  education then we should be investigate what are
  the best policies that can be implemented to
  achieve a certain goal.
• As much as possible, we will document what is
  known about the costs and benefits of each
  policy. Once we know these we can compute and
  compare the returns to different policies.

3

• THE RETURNS TO SCHOOLING
• The economic return to schooling is one of the
  most estimated parameters in empirical economics.
  It measures the percent increase in wages of an
  individual due to additional schooling and is
  usually estimated from the following equation
  (Mincer, 1974)
• ln(W)aßSe
• where ln(W) are log wages and S are years of
  schooling. ß is the return to schooling.

4

• Why do we call ß a return?
• The most important costs of schooling are
  foregone earnings. For example, let W(S) be the
  potential earnings of an individual with S years
  of schooling. If he drops out of school and
  starts working he can earn W(S) in the labour
  market. If he decides to stay one more year in
  school we can earn W(S1) next year, but he
  cannot work immediately and loses W(S) for a
  year. Therefore W(S) is the cost of staying in
  school an extra year.

5

• In summary, the cost of the investment is W(S).
  However, once I incur this cost, I will get a
  benefit of W(S1)-W(S) starting next year (the
  increase in the wage due to the extra year in
  school). Therefore the return is
• W(S1)-W(S)/W(S).
• However
• W(S1)-W(S)/W(S)ln(W(S1))-ln(W(S))ß.

6

• How large is the return to schooling? Is
  schooling a good or a bad investment?
• Psacharopoulos and Patrinos (2002) survey
  estimates of returns around the world.

7

• Using data from 1978-86, Harmon and Walker
  (1995) estimate that the return to schooling in
  the UK is around 15 per year of schooling.

8

• Using 1991 data for the UK, Dearden (1999) finds
  different numbers. She measures the returns of
  having a given qualification vs. having no
  qualification at all.

9

• Card (1999) also presents a survey of estimates
  for the US and other countries. He finds
  estimates in the range of 6-15 per year of
  schooling.
• Krueger (2002) documents the increase in the
  returns to schooling over time.

10

• Heckman, Lochner and Todd (2003) recalculate the
  returns to different levels of schooling taking
  into account tuition costs and taxes, which are
  usually ignored in standard estimates. They also
  allow different years of schooling to have
  different returns. They find that there is a
  large return to graduating from high school in
  the US, that has dramatically increased over time.

11
(No Transcript)
12
(No Transcript)
13

• Estimates of the returns from many different
  countries show high rates of return to schooling.
  Schooling appears to be a good investment and its
  payoffs have increased over time. However, the
  estimates I presented so far hide an enormous
  amount of HETEROGENEITY the returns to schooling
  vary across individuals. In many cases the
  average returns I presented are not useful guides
  to policy.

14
(No Transcript)
15

• Different individuals have different returns. If
  are going to implement policies that increase the
  level of schooling in the population we need to
  know what section of the distribution of returns
  is being affected by the policy.
• Different policies affect different people.
  Therefore we need different returns to evaluate
  different policies. This is a general problem in
  any type of policy evaluation where heterogeneity
  is judged to be important. The average individual
  is often a meaningless concept in policy
  evaluation.

16

• Carneiro, Heckman and Vytlacil (2003) examine
  how the return to college in the US varies across
  individuals with different levels of cognitive
  skills and unobserved ability.

17

• Carneiro, Hansen and Heckman (2003) estimate
  distributions of returns to college. A
  substantial fraction of the population has
  negative returns to college.

18

• THE EFFECT OF SCHOOLING ON TEST SCORES
• Schooling has an important effect on test
  scores, especially at low levels. Test scores are
  important determinants of future schooling and
  wages.

19

• On top of affecting earnings, schooling affects
  several other outcomes. Some examples are
• Crime Lochner and Moretti
• Health Lleras-Muney, Banks et. al
• Outcomes of Offspring Carneiro, Meghir and Parey

20
(No Transcript)
21

• Banks et al. compare socioeconomic health
  gradients in the US and the UK

22

• Carneiro, Meghir and Parey examine the effect of
  maternal education on a variety of child outcomes.

23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)