1088 The Econometrics of Convergence
history, the opposite extremes of savage and civilised life. Such reflections tend
to enlarge the circle of our ideas; and to encourage the pleasing hope, that
New Zealand may produce, in some future age, the Hume of the Southern
Hemisphere. (book II, ch. XXV)But another famous passage notes the permanent effects of particular events, argu-
ing that, had the Franks not defeated the Arabs at the Battle of Poitiers in 732,
“Perhaps the interpretation of the Koran would now be taught in the school of
Oxford, and her pulpits might demonstrate to a circumcised people the sanc-
tity and truth of the revelation of Mahomet” (book V, ch. LIII). Modern social
science is neither as literary nor as broad in its sweep, yet Gibbon’s remarks res-
onate with the broad questions involved in the study of convergence. Over long
epochs, will initially different societies or economies evolve towards a common
form, or will their initial conditions play a role in determining their long-run
outcomes?
The nature of interest in convergence has varied over time, referring to differ-
ent countries and different socioeconomic forms. From the vantage point of the
Cold War, one debate concerned the extent of the differences between the West
and the Soviet bloc. It was not uncommon to argue that capitalist and socialist
economies would converge over time, as market institutions began to shape social-
ist economies, while capitalism might increasingly be accompanied by extensive
government regulation and intervention, and a range of activist social welfare
policies. More recently, a new convergence debate has focused on issues related to
the persistence or transience of differences between rich and poor countries, with
parallel interest in the income differences of regions, states, and districts. Conver-
gence here is typically conceived in terms of narrowing differences in per capita
income, rather than broader socioeconomic institutions. It is this literature that
we will review in this chapter. The possibility of income convergence has been
studied more intensively than any other hypothesis in growth economics, even if
the effort to identify growth determinants, both proximate and fundamental, has
become the main area of current empirical research.
Contemporary interest in the convergence hypothesis stems from at least three
factors. The first is the enormously high levels of international inequality, and
attendant levels of human suffering in poorer societies. This inequality means that
the extent to which economies are converging or diverging, and at what rate,
forms the background for almost any discussion of globalization and the work
of international institutions and aid agencies. In popular commentary, it is often
claimed that the gap between rich countries and poor countries is widening, and
the study of convergence helps to evaluate such claims in a rigorous way.
Second, endogenous growth models, at least when based on increasing returns
to scale, often imply long-run divergence of per capita incomes. Hence the inves-
tigation of convergence came to be seen – perhaps mistakenly, as we will discuss
below – as the best way to discriminate between the new growth theories and
their neoclassical predecessors. Finally, the greater availability of internationally
comparable data for a broad cross-section of countries, primarily due to the work