Education and Wages

1
Education and Wages

• Pedro Telhado Pereira
• May 2004
• Part 4

2
The paradigm of the two islands Pereira and
Martins (2004)

• two islands
• (I1) with a productivity per capita of P
• (I2) with a productivity per capita of Q
• PltQ.
• The inhabitants of the islands live an eternal
  life (all are born at the same moment) and
  maximize the present value of their production.

3

• The inhabitants of I1 can move to I2 if they
  spend one period learning how to swim.
• The only cost is the product they forgo during
  that period.
• The discount rate is uniformly distributed
  between (r1, r2) f(r)1/(r2-r1).

4
The decision of learning to swim is made by
comparing (E1)
and (E2)

5

• There is a value of r, rc, such that E1E2.
• Let us assume that r1 lt rc lt r2.
• If r lt rc then E1 lt E2 and the individual decides
  to learn to swim and thereafter swims from I1 to
  I2.
• If r gt rc then E1 gt E2 and the individual does
  not learn to swim and stays on I1.

6
Therefore, there will be S1 proportion of
swimmers, where

• We end up with three groups of individuals
• living on island one,
• living on island two and born there, and
• living on island two and born on island one.

7

• The labour economists want to explain the
  differences in productivity and therefore they
  can do so by examining the place where the
  person is living (the wage equation approach).
• But suppose that we are interested in studying
  the returns to education (learning to swim). We
  must then look at the persons born on island one
  and see the differentials in their productivities
  as the return we are looking for. This is what
  education economists look for.
• We could never find this return if we considered
  the place of residence as one of the explanatory
  variables of the productivity differential, the
  reason being that this covariate is a result of
  learning to swim for people born on island one.

8
This example illustrates

• the difference between wage equations - where
  education is one of the explanatory variables -
  and returns to education - where all the indirect
  effects should be accounted for.

9
When an individual (or his family) decides the
amount of education he wants to have

• he knows that education is going to bring him a
  better paid job,
• but at the same time gives him more opportunities
  in choosing other characteristics as, for
  instance, the sector and firm where he ends up
  working in.
• Part of the private return from education to the
  individual is going to be through the return he
  will get from the posterior choices he can make.

10
In econometric terms
Let us assume that wages depend on education (S)
and other covariates, represented by an indicator
(C). To make it as simple as possible and
following the Mincer specification we have
where b1gt0 and b2gt0 or b2lt0
11

• If b2gt0, people with more education will choose
  the largest C and therefore there is a positive
  relation between S and C.
• If b2lt0, people with more education will choose
  the smallest C and therefore there is a negative
  relation between S and C.

12
If we assume this relation to be linear, we have
and
13
The rate of return of Education
14
Main conclusions

• To obtain the full effect of the education on
  wages we should only regress the ln(wage) in
  education and/or not consider the covariates that
  depend on schooling
• If we include covariates that depend on education
  on the regression the coefficient of schooling
  decreases (at least in the expected value of the
  estimators).

15
What has been considered in the wage equation

• The efficiency wages hypothesis (see, inter alia,
  Krueger and Summers, 1988) warrants the inclusion
  of sectors of activity, firm size and firm age.
• Hartog, Pereira, Vieira (2000)

16

• Following the seminal work of Krueger and Summers
  (1988), several studies have been carried out in
  a large number of countries and shown the
  existence of inter industry wage differentials
  for apparently equally skilled workers. Although
  part of these differentials can be explained by
  unobserved heterogeneity, this does not explain
  all the variation
• These findings pose a challenge to textbook
  competitive models of the labour market and
  alternative explanations based on efficiency wage
  mechanisms or rent sharing have been put forward
• Nevertheless, the existence of such differentials
  has not been clearly understood and remains an
  intricate and unresolved puzzle.

17

• The data were drawn from Quadros de Pessoal for
  the years of 1982, 1986 and 1992.
• Conclusion
• The study shows Portugal has a high
  inter-industry wage inequality. The size of the
  inter-industry wage dispersion in Portugal seems
  similar to that of countries rated as having a
  decentralised wage setting.
• Nevertheless, the dispersion decreased during
  the1980s.
• We argue that shifts towards a more centralised
  and co-ordinated wage setting may have played a
  role here.

18

• The existence of rents and trade unions or
  agency models (see Freeman and Medoff (1986),
  Pencavel (1991), Hart and Holmstrom (1987), among
  others) justify the inclusion of the bargaining
  regimes.
• Hartog, Pereira and Vieira (2002)

19

• labour economists have increasingly realised that
  wage structures and labour contracts may also be
  instruments to the firm.
• Moreover, the influence of the institutional
  environment of the labour market has led to
  greater interest in the role of bargaining
  between trade unions and management.

20

• The data used - Quadros de Pessoal for 1986 and
  1992
• Conclusions
• The results show that after controlling for a
  large set of individual and job related
  characteristics, the firm bargaining regime
  coverage apparently is important in explaining
  the variability of wages.
• The main results reveal that wage differentials
  between the bargaining regimes are in some cases
  sizeable.

21

• that multi-firm contracts tend to generate higher
  wages
• sectoral contracts tend to generate the lowest
  wages.
• Compulsory regimes and single-firm contracts
  align at an intermediate level in the ranking.

22

• Internal wage structures (see Lazear (1998), for
  instance) justify the inclusion of seniority
  (tenure).
• Lima, F. and P. T. Pereira (2003)

23

• The relationship between employers and employees
  is rather complex and subject to different
  dimensions of analysis.
• One such dimension is the wage growth associated
  with the employees career progression inside the
  firm.
• The real world is replete with examples of job
  ladders that individuals climb during their
  working lives and of the major pay changes that
  accompany those steps.

24
The main questions

• What is the effect of promotion on the wage
  growth?
• How does the wage growth depend on the
  hierarchical position held by the employee?
• To what extent are human capital accumulation and
  learning about individual abilities reflected in
  wage paths?
• What is the effect of demotions?

25
Data used

• The data set used is a sample of firms drawn from
  the survey Quadros de Pessoal collected annually
  by the Ministry of Employment.
• A random sample of firms was drawn, stratified
  according to economic activity, location, firms
  legal form, and number of employees.
• The sample is a longitudinal matched
  employer-employee
• panel of 74 large firms from the manufacturing
  sector, with more than 500 workers each year and
  followed between 1991 and 1995. T
• he employment history of all workers in the firm
  and several firms characteristics are available.
  
• The original sample has 391,618 observations.

26
Conclusions

• Promotions and/or transitions to upper job levels
  imply a positive and important wage premium.
• Demotions are associated with negative wage
  premiums.
• The negative wage premiums are more important at
  the bottom job levels, probably as a result of
  the (negative) learning effect.
• The wage premiums for promotions are higher at
  the tails of the hierarchy, generating a U-shaped
  relationship between the workers career path and
  wage growth.
• In the context of the model discussed, it
  suggests a stronger learning and/or human capital
  accumulation effect at the bottom of the
  hierarchy and a stronger job assignment effect at
  the top.

27
The influence of covariates on returns to
education

• Pereira, P. T. and P. S. Martins (2004), Returns
  to Education and Wage Equations, Applied
  Economics, forthcoming.
• We a meta-analysis - is a regression that takes
  as dependent variable the outcomes from different
  studies that focus on the same topic and employ
  the same general methodology. The regressors
  describe the characteristics (in terms of
  equation specification, in sample size, in year
  of estimation, and so on) underlying those
  different results and/or studies. A meta-analysis
  is therefore a useful tool for summarizing
  several results on a given topic, allowing a
  researcher to have a global and quantifiable view
  on the link between the structure of a research
  project and its results. In the present study,
  we examine the influence of covariates in the
  return to education.

28
Data used

• A meta-analysis uses two kinds of data, which we
  label here as foreground and background data.
• The former is the directly-used information,
  which includes the coefficients of education that
  were obtained in different studies, and the
  presumably relevant characteristics of those
  studies. By such characteristics, we mean the
  regressors used, sample size, and so on.
• Background data, on the other hand, is simply the
  primary sources (data sets) used for computing
  the returns to education (or better, a
  coefficient to education).

29
Wage equations in Portugal

• Have used between two ( experience and experience
  squared) to a maximum was 37 explanatory
  variables.
• The number of explanatory variables - in the
  background regressions - corresponds to the
  number of regressors used besides education and a
  constant.

30
Conclusions

• The use of the Mincer equation in its simpler
  form (education, experience and experience
  squared) seems to give an approximate value for
  the total return to education.
• If more covariates are used in this equation and
  these covariates are choice variables that depend
  on education, then the coefficient of the
  education should fall.
• This result is supported by the meta-analysis we
  performed using data for Portugal.
• The coefficient decreases with all combinations
  of variables used and can drop to half of its
  size, especially when the sector of activity is
  one of the covariates used.
• The education-related choice of sector is an
  aspect that should reflect itself in
  over-education in the better paying sectors.

31

• Sample size, the use of monthly wages instead of
  hourly wages, the interaction between education
  and experience and tenure do not seem to
  influence the coefficient, which shows its
  robustness to sample size, specification of the
  simple Mincer equation and variables that are
  independent of education.
• The increase of the return to education when
  regions is used as one of the covariates needs
  further research, as it seems to show that in the
  Portuguese case the mobility due to job
  opportunities is rare.

32
Final Remark

• Returns to education and changes in productivity
  can be very distinct realities.
• Both are worth studying but one should
  distinguish between them.

33

• References
• Freeman, R. B. and J. L. Medoff, (1986). The Two
  Faces of Unionism. In L. G. Reynolds, S.H.
  Masters and C. H. Moser, eds,. Readings in labor
  economics and labor relations, Fourth edition,
  Englewood Cliffs, N.J. Prentice Hall.
• Hart, O. and B. Holmstrom (1987). A theory of
  contracts. In T. F. Bewley, Advances in Economic
  Theory, Fifth World Congress, Cambridge
  Cambridge University Press.
• Hartog, H., P.T. Pereira and J. C. Vieira, (2000)
  'Inter-Industry Wage Dispersion in Portugal High
  But Falling', Empirica, 27/4, 353-364.
• Hartog, H., P.T. Pereira and J. C. Vieira, (2002)
  "Bargaining Regimes and Wages in Portugal",
  Portuguese Economic Journal, vol.1, n. 3,
  237-268.
• Krueger, A. and L. Summers (1988). Efficient
  wages and the inter-industry wage structure.
  Econometrica 56(2), 259-293.
• Lazear, E. P (1998). Personnel Economics for
  Managers, New York John Wiley and Sons.
• Lima, F. And P. T. Pereira (2003), Careers and
  Wage Growth within Large Firms ", International
  Journal of Manpower No 7, Volume 24, 2003.
• Pencavel, J. (1991). Higher Education,
  Productivity, and Earnings A Review. Journal of
  Economic Education 224, 331-359.
• Pereira, P. T. and P. S. Martins (2004), Returns
  to Education and Wage Equations, Applied
  Economics, forthcoming.