Economic growth and economic development 48

Introduction to Modern Economic Growth
general patterns they show are typically consistent with evidence from a variety
of different sources. Nevertheless, there are points of contention. For example, as
Figure 1.12 shows, Maddison’s estimates show a slow but relatively steady growth
of income per capita in Western Europe starting in 1000. This is disputed by many
historians and economic historians. A relatively readable account, which strongly
disagrees with this conclusion, is provided in Pomeranz (2001), who argues that
income per capita in Western Europe and China were broadly comparable as late as
1800. This view also receives support from recent research by Allen (2004), which
documents that the levels of agricultural productivity in 1800 were comparable in
Western Europe and China. Acemoglu, Johnson and Robinson (2002 and 2005)
use urbanization rates as a proxy for income per capita and obtain results that are
intermediate between those of Maddison and Pomeranz. The data in Acemoglu,
Johnson and Robinson (2002) also confirms the fact that there were very limited
income differences across countries as late as the 1500s, and that the process of rapid
economic growth started sometime in the 19th century (or perhaps in the late 18th
century).
There is a large literature on the “correlates of economic growth,” starting with
Barro (1991), which is surveyed in Barro and Sala-i-Martin (2004) and Barro (1999).
Much of this literature, however, interprets these correlations as causal effects, even
when this is not warranted (see the further discussion in Chapters 3 and 4). Note
that while Figure 1.16 looks at the relationship between the average growth of
investment to GDP ratio and economic growth, Figure 1.17 shows the relationship
between average schooling (not its growth) and economic growth. There is a much
weaker relationship between growth of schooling and economic growth, which may
be because of a number of reasons; first, there is considerable measurement error in
schooling estimates (see Krueger and Lindahl, 2000); second, as shown in in some
of the models that will be discussed later, the main role of human capital may
be to facilitate technology adoption, thus we may expect a stronger relationship
between the level of schooling and economic growth than the change in schooling
and economic growth (see Chapter 10); finally, the relationship between the level of

schooling and economic growth may be partly spurious, in the sense that it may be
capturing the influence of some other omitted factors also correlated with the level of
34