The
literature of endogenous growth theory has stimulated economists’ interest in
the empirical evidence available from cross-country comparisons, bearing on the
main-level relationship between human capital formation and the growth rate of
real output. The growth models that view human capital as a simple input to
production predict that growth rates will be positively associated with changes
in the stock of education, whereas models in which human capital has a role in
the development of innovations and its diffusion throughout the economy imply
that it is the stock (rather than the flow) of human capital that affects the
overall productivity growth rate of the country.
Early
studies of the effects of human capital on growth, such as Mankiw, Romer, and
Weil (1992) and Barro (1991), were based on data sets pertaining to a very
diverse array of (more than 100) countries during the post – 1960 era. They
used narrow flow measures of human capital such as the school enrolment rates
at the primary and secondary levels, which were found to be positively
associated with output growth rates. Barro reported that the process of
catching up was firmly linked to human capital formation: only those poor
countries with high levels of human capital formation relative to their GDP
tended to catch up with the richer countries.
Barro
and Sala- I –Martin (1995) among many others, have also included life
expectancy and infant mortality in the growth regressions as a proxy of
tangible human capital, complementing the intangible human capital measures
derived from school inputs or cognitive tests considered; their finding is that
life expectancy has a strong, positive relation with growth.
A
recent survey by Krueger and Lindahi (1998) from the econometric studies of
cross-country growth equations shows more robust results. First, changes in the
human capital stock do not seem to affect growth rates, as would be implied by
the model in Lucas (1988). This contrasts with the robust evidence from the
micro literature of education on income. When allowances are made for
measurement errors, the change in stock measures of education is positively
correlated with economic growth. Secondly, the evidence with respect to the
positive effect of the level of human capital stock on growth rates is much
stronger. But the size of this effect varies cross countries. Two other well-established
results that emerged from the cross-country studies examined by Krueger and
Lindahi are: (a) the greater effect of secondary and higher education on
growth, compared with primary education, and (a) the seemingly insignificant,
or even negative, effect of female education on the growth of output. With
respect to the latter, they follow. Barro (1999) in suggesting that the
insignificant effect of female education may be a result of gender
discrimination in some countries’ labour markets. The argument is that females
receive education in these countries but are discouraged from participating in
the labour market, and thus cannot contribute directly to the growth of output.
This may explain part of the problem, but it seems that other mechanisms also
are at work: in countries with high female labour market participation,
variations in the extent of female education have an insignificantly small
positive effect on output growth rates.
While
there is persuasive evidence about the positive relation between initial human
capital levels and output growth and (weaker) empirical support for the
relation between changes in human capital and growth, it is not all clear that
this implies a causal relationship running from human capital to growth.
Motivated by the fact that schooling has increased dramatically in the last 30
years at the same time that the “productivity slowdown” became manifest in many
of the hgiehr income economies, Bils and Klenow (2000) suggest that the causal
direction may run from growth to schooling. That relationship would be
predicted by a Mincerian model in which high anticipated growth leads to lower
discount rates in the population, and so to higher demands for schooling. Of
course, both variables might be driven by other factors. From the results of
different empirical tests, Bils and Klenow conclude that the channel from
schooling to growth is too weak to explain the strong positive association
found by Barro (1991), and Barro and Lee (1993), as described above. But, they
argue, the “growth to schooling” connection is capable of generating a
coefficient of the magnitude reported by Barro. Lucas (1988) includes human
capital as an additional input in the production of goods, while retaining the
other features of the neoclassical growth model. In the model, the labour force
can accumulate human capital, which is then used together with physical capital
to generate the output of the economy. In one version of the model, human
capital is acquired through time spent in an (non-productive) educational
process, introducing a trade-off for workers between employing time to produce
output and using it to gain further human capital that will increase their
marginal productivity when working in subsequent periods. In another version of
the model, human capital is gained by the workers through on-the-job-training,
and so the time employed working increases their productivity later on. The
accumulation of human capital involves a sacrifice of current utility in the
form of less current consumption in the case of education, or a less desirable
mix of current consumption goods when on-the-job training is considered.
In
the Solow-Swan and Ramsey models, the equation describing physical capital
accumulation is sufficient to determine the dynamic evolution of output. To
specify the growth path when human capital is included, it is necessary to
consider an additional sector where the growth of human capital takes place.
Given that physical capital still has diminishing returns, the required assumption
for the model to exhibit a positive growth rate of output per worker in the
steady state is that the “technology” for generating human capital has constant
returns. This means that the growth of human capital is assumed to be the same
for a given level of effort whatever the level of human capital attained. With
this assumption, the rate of output growth (per worker) is positive and
increasing in the productivity of education or on-the-job training in the
creation of human capital.
Azariads
and Drazen (1990) model the mechanism of human capital transmission across generations
in the more plausible framework of an overlapping generation’s model (Lucas
followed Ramsey in the simplifying assumption that households, as well as
firms, are infinitely lived). In these models agents inherit the human capital
accumulated by the previous generation; they then decide how much time to
devote to training a young graduate in acquiring further skill in technology
that increases labour quality, thereby affecting their marginal productivity
when older. Since a given generation deciding
its own human capital investment does not take into account the inter temporal
spill-over effect upon the human capital endowment of future generations, there
is a technological eternality that can result in constant or increasing returns
to human capital at the social level. This state of affairs could be ascribed
to the impossibility of contracting with the future generations, and sometimes
is described as allocation inefficiency due to “incompleteness of markets”. The
sources of this problem affecting human capital investment is therefore rather
different from the set of conditions previously seen to impair the allocative
efficiency of markets that do exist.
Acemoglu
(1998) has offered a formal demonstration of how positive spill-over effects
(pecuniary externalities) created by worker’s educational and training
investment decisions can give rise to macro-level increasing returns in human
capital. His model supposes that workers and firms make their investments in
human and physical capital, respectively, before being randomly matched with
one another. The direct consequence of random matching is that the expected
rate of return on human capital is increasing in the expected amount of
(Complementary) physical capital with which a worker will be provided;
similarly, the return on physical capital is increasing in the average human capital
that the firms expect the workers to bring to the job. Hence, an increase in
education for a group of workers induces the firms to invest more in tangible
assets, thereby increasing the return to all workers in the economy. Through a
similar argument, the model is seen also to imply that there are “social
increasing returns” in physical capital.
In
the early 1990’s pioneering econometric studies (based on international panel
data for a widely diverse array of countries during the post-1960 era) provided
empirical support for the conclusion that human capital formation was among the
factors that significantly affected the aggregate level rate of economic
growth.
·
They
found that success in the process of catching up internationally in terms of
GDP growth was positively related to the overall social rate of human capital
formation.
·
Furthermore,
the poor countries that were tending to catch up with the higher income
economies were restricted to those that were maintaining levels of investment
in formal education which were high in relation to their respective GDP levels.
More recent econometric studies have
yielded three robust empirical findings:
·
There
is only weak empirical support for the hypothesis that changes in the human
capital stock affect growth rates.
·
There
is strong statistical support for the hypothesis that the relative level of the
stock of human capital (in relation to the labour force or aggregate output)
has a positive effect on growth rates.
·
The
magnitude of the “level effect” of the human capital stock is itself far from
uniform across the distribution of economies; the impact on growth rates does
not vary linearly with the relative size of the stock but, instead, becomes
proportionately smaller among the economies where the average educational
attainment is already high.
The broad interpretation
of these findings in the context of recent growth models is that raising the
general level of educational attainment interacts positively with other forces-
among them the accumulation of complementary physical capital and the
application of new technologies. Higher human capital intensity thus permits
countries to accelerate their productivity growth rate and narrow the relative
size of the per capita real income gaps separating them from the leading
economies.
Maintaining a high average
level of educational attainments, and correspondingly high rates of investment
in other forms of human capital (e.g. health, internal spatial and occupational
mobility), would appear to serve as a stabilising
force – although not a guarantee-
against continuing secular decline in a country’s relative per capita income
position.
Most of the theoretical literature on economic
growth focuses on the role that investment in formal education plays in modern
economies