theoretical interest (for example using school enrolment to try and capture
human capital). In regression analysis policy variables are typically added to
regressions separately which allows for no interaction between and among
them. Economic theory suggests complementarity is important. Education, for
example, may only be related to growth if individuals earn higher wages as
engineers rather than as pirates. Cross-country growth regressions often test
the relationship between the rate of economic growth and the level of govern-
ment expenditure. Any government using Keynesian-style demand manage-
ment policies will be likely to boost government spending when GDP growth
slows. This will generate a spurious negative relation between the ‘size’ of
government and economic growth. Growth effects may be contemporaneous,
some take several years (changed investment incentives), others even decades
(incentives affecting the rate of technical change), which is difficult to capture
in statistical testing. Some policy variables may have output/growth effects at
all three horizons – cyclical, transitional and steady-state. There is no reason to
assume these are of the same magnitude or even the same sign (Temple,
1999:124). Higher taxes, for example, may reduce growth in the short term by
reducing demand in the economy but boost growth in the longer term by
providing the government with a source of revenue to invest in education and
infrastructure.
In order to run large cross-country regressions researchers are making the
assumption of universalism: that the relation between a factor like education
and growth is identical across countries so each individual country provides
evidence that can be used to help measure this one underlying universal
economic relationship. Many studies explain Africa’s slower growth as a
function of different levels of explanatory variables: that the population is less
literate,or that Africa has a more unforgiving climate or a more adverse colo-
nial history. This method seeks to explain African growth as the result of a
common worldwide growth process that begins from different levels of the
same explanatory variables. However significant regional effects remain
common in much of the empirical literature. As we saw in the Introduction,
Barro found the growth process worked differently in Sub-Saharan Africa and
Latin America from the rest of the world. The usual reaction to this finding is
that if the statistical effort is failing to explain growth in particular countries
or regions there must be missing variables which then need to be discovered.
This has led researchers to propose and test ever more variables in the hope
that growth in Sub-Saharan Africa will finally be ‘explained’. An alternative
methodology is to drop the assumption that only the levels of explanatory
variables are different and explore the idea that the growth process in Sub-
Saharan Africa works differently (McCartney, 2011). There are a limited
number of studies that suggest this latter idea may be true. Block (2001) finds
that trade openness in Sub-Saharan Africa has a much stronger effect and
fiscal policy a weaker effect on growth than in other regions. Brock and
Durlauf find ‘the operation of ethnic heterogeneity on growth is different in
Africa’(2001:264). Mosley (2000) finds that inequality only has a negative
52 Sources of Growth in the Modern World Economy since 1950