Growth accounting
Growth accounting gives life to the formula for the proximate determinants of
growth by calculating the impact of factor inputs and TFP but severe empirical
problems are associated with this method. The measurement of TFP, for exam-
ple, depends critically on assumptions about how the various factors of
production are translated into output, and there are problems with accounting
for the quality of inputs. A further problem is aggregating all the roads,
computers, buildings and tractors in which individuals, households and firms
have invested into one measure of the capital stock and then accounting for
depreciation.
Growth accounting is not a formal test of any economic theory. This book
makes the simplifying assumption that the accumulation of factors of produc-
tion or productivity growth (TFP) causes economic growth, but there are other
potential arguments. This chapter has discussed a number of studies that
suggest economic growth stimulates the incentive to acquire greater levels of
education (human capital), so that the causal relation can work the other way
round.
The Solow Surprise
One of the earliest attempts at growth accounting was by Nobel Prize winner
Robert Solow who argued in a 1957 paper that the key source of long-term
growth is technological change. He found that TFP growth accounted for
nearly 90 per cent of US growth per worker over the first half of the twentieth
century. This finding undermined many early Development Economics theo-
ries that physical investment was the key to boosting economic growth. More
recent work has tended to support this finding (Klenow and Rodriguez-Clare,
1997; Easterly and Levine, 2000). Mankiw et al.(1992), exceptionally,
disagreed with the Solow Surprise. They argued that 80 per cent of the interna-
tional variation in country per capita incomes could be explained using only
population growth and investment in physical and human capital. This finding
was based on two controversial assumptions: first, that investment rates are not
related to the level of income; and second, the human capital variable only
measures the variation in secondary schooling (and ignores primary school-
ing) so exaggerates the variation in human capital across countries. When
correcting for these Klenow and Rodriguez-Clare find that the Mankiw et al.
model only explains around half of the variation in incomes and does leave a
central role for technology and productivity.
Growth accounting and developed countries
We can use growth accounting to draw some interesting comparative histori-
cal conclusions. The general story for the UK, US and Japan in the modern era
is that growth has been based on productivity gains (improving factors of
production rather than accumulating more of them). Growth has been based
on a better educated workforce, using more and better capital equipment and
Growth in the Modern World Economy since 1950 55