factors economic theory considers important in influencing growth for as large
a number of countries as possible over as long a time period as possible.
Unfortunately the results from regressions remain very poor. Even those
factors many would accept as self-evidently related to economic growth, such
as investment, fiscal policy and education, have an ambiguous empirical rela-
tion to economic growth within cross-country regression analysis. These
points are illustrated using regression results for Pakistan.
Investment can have either a direct link with growth (such as a new factory
producing more output) or an indirect link (such as a new road increasing the
incentives of farmers to expand production which can now be better marketed).
Levine and Renelt (1992) took a number of variables commonly used in
econometric growth analyses and ran them in thousands of regressions. They
found that only investment was robustly related to economic growth. Evidence
for Pakistan based on simple regression analysis supports this general finding.
Khan et al.(2005) find a positive relation between investment and growth in
Pakistan between 1971 and 2004. Khan (2005) finds a positive relation
between investment and GDP growth between 1980 and 2002. Iqbal and Zahid
(1998) find a positive relation between investment and growth between
1959/60 and 1996/97. This link is even stronger when looking at sub-compo-
nents of overall investment. De Long and Summers (1991, 1992, 1993) and
Jones (1994) found a positive correlation between investment in machinery
and equipment and productivity growth. While transport investment reflects
differences in needs caused by urbanization, geography and population density,
equipment,they argue,is more directly linked to growth in the manufacturing
sector.
There remains a problem with identifying causality in the investment–
growth relation. Blomstrom et al. (1996) find an inverse causal relation between
gr owth and investment. Growth induces subsequent capital formation (perhaps
firms creating capacity to meet expected increases in market demand) for their
sample of 101 countries between 1965 and 1985. A particular theoretical ques-
tion relevant for empirical studies of the investment–growth relation is that of
‘credibility’. Rodrik (1989) argues that only a ‘credible’ policy change will be
likely to promote growth in the private sector. Credibility can be thought of as
a policy change the private sector thinks the government are committed to and
will sustain. If, for example, private entrepreneurs believe an investment tax
incentive will be later rescinded they won’t undertake any long-term invest-
ment in response. Credibility is a very difficult concept to test using cross-
country regression techniques (and is discussed in more detail in Chapter 5).
Fiscal policy can affect the proximate determinants of growth by changing
the incentives for investment (Rebelo, 1991). Barro (1991) found a significant
negative association between the ratio of real government consumption (minus
spending on education and defence) and economic growth. In Pakistan, Tahir
(1995) found government defence expenditure had a positive, Iqbal and Zahid
(1998) found the government budget deficit had a negative and Ghani and Ud-
Din (2006) found government consumption had a positive relation with GDP
50 Sources of Growth in the Modern World Economy since 1950