different countries for the years 1960 to 1985 to run a cross-country growth
regression. A growth regression can be thought of as a sequence of questions,
about what factors influence growth. Barro asked: How strong and reliable is
the influence of education on economic growth holding all other factors
constant? The question was repeated for investment, then government spend-
ing and so on. Barro found that poorer countries (measured by average
incomes in 1960) tended to grow faster than richer countries. This suggested
that over time there would be a process of convergence (or narrowing of gaps)
in the average standard of living between countries. This finding was surpris-
ing, given the facts of the Great Divergence which show that exactly the oppo-
site has happened over a much longer period than covered by Barro’s data.
Barro found that investment (private, not government) had a strong positive
impact on growth. Government consumption as a share of national income had
a negative impact. These two findings supported one of the big economic ideas
of the 1980s and 1990s: the relative role of the state, whether through govern-
ment investment or more interventions, should be reduced (and that of the
private sector increased). Barro’s statistical results were weaker for Latin
America and Sub-Saharan Africa, indicating that important aspects of the
growth process in both regions remain unexplained. Case studies, such as
those in this book, provide a more focused approach that tries to find out why
the growth process may differ between regions or countries.
This type of statistical analysis has allowed researchers to compare the relative
influence of all these various drivers of economic growth. A common finding in
the many studies over the years since Barro’s pioneering effort is that girls’ educa-
tion has a much stronger and more reliable impact on growth than does trade
policy (such as reducing taxes on imported goods). This provides strong evidence
in support of policy-makers spending less time taking part in international trade
negotiations and more time trying to improve the education of girls. We can only
speculate then why ministers have spent so much time in recent years travelling
to World Trade Organization trade conferences in Bali (Indonesia), Geneva
(Switzerland), Cancun (Mexico), and Hong Kong and not to the dusty and distant
villages of their home countries to open schools for poor, rural girls.
Although research inspired by Barro identified more than ninety determi-
nants of economic growth over the following decade, this research effort feels
stale and confusing, and the inherent problems of the method are becoming
increasingly evident now that twenty years have passed. These problems are
discussed in more detail in Chapter 2 but a few are introduced here.
There is generally a significant problem in interpreting causation in such
studies. Whilst education and growth are clearly related, does education cause
economic growth by making workers more productive and better able to utilize
new technologies or does economic growth cause education to expand by
providing the government with tax revenue and parents with income to pay for
education and by creating well-paid jobs that motivate people to acquire more
education? Researchers have also questioned the value of the statistical results
produced by Barro and others as guides for policy-makers. The original idea
8 Introduction