The ultimate objective of research on economic growth is to explain cross‐country differences in growth experience and to determine whether such explicable differences carry any useful implications for raising poor countries’ standards of living to those of richer nations. One can make a case that Young (1928) and Rosenstein‐Rodan (1943; chapter 5) launched development economics as we know it today. Drawing on and carefully invoking Alfred Marshallʹs distinction between internal and external economies, Young introduced the analytics of growth based on increasing returns. This provided the central concepts that underpin the seminal works of what Krugman (1993) called “high development theory” of the 1940s and 1950s. The balanced growth model of Rosenstein‐Rodan (1943), Myrdal’s (1957) “circular and cumulative causation”, Nelson (1956) and Leibenstein’s (1957) “low level equilibrium trap”, Scitovsky’s (1954) external economies, Nurkse’s (1952, 1953) “big push” theory, and later formalizations of these and related strands of the growth literature relied on the pecuniary externalities and inter‐firm or inter‐sectoral complementarities to generate growth processes characterized by multiple equilibria, more popularly known today as “poverty traps”.
This also gave rise to more careful attention paid to backward and forward linkages among sectors, initially and eloquently explored by Hirschman (1958) and expanded upon by others, such as Johnston and Mellor (1961; chapter 45), who emphasized the particular importance of the agricultural sector due to its extraordinary linkages to other sectors. The ideas introduced by Young and applied to development problems by Rosenstein‐Rodan and his successors fell dormant for decades, however. Krugman (1993) argues that the “high development theorists” inability or unwillingness to formalize their insights in mathematical form helped enable the more precise formulations of Solow (1956) to outcompete them. Solow’s model, based on key assumptions of constant returns to scale, competitive equilibrium, exogenous technological change and no externalities, refined and extended the familiar Harrod‐ Domar formulation. In the Harrod‐Domar formulation that dominated analyses of economic growth throughout the 1940s and 1950s, the capital‐output ratio was assumed fixed and growth rates could be readily manipulated by changing savings rates. Solow endogenized the capital‐output ratio, letting it vary naturally with per capita availability of capital in the economy and the resulting marginal returns to capital under the assumption of diminishing returns to any single factor of production.
The implication of the Solow growth model were powerful: accumulation of physical capital then affects not the steady‐state growth rate in an economy, only the level of income, and technological progress became a key, exogenous driver of growth. Solow’s growth model dominated the field for over 30 years, even though it treated savings and technological progress as completely exogenous. The neoclassical Solow model’s predictions of convergent growth processes were nonetheless hard to square with the empirical evidence that suggested instead “divergence, big time” (Pritchett 1997). Empirical research by Quah (1996; chapter 50), Durlauf and Johnson (1995; chapter 51) and Hansen (2000) developed more advanced econometric methods capable of identifying high‐order nonlinearities and threshold effects in macro‐level growth processes.
The growing body of empirical evidence against the convergence hypothesis reinforced advances in what came to be known as “endogenous growth theory” (Romer 1986, 1990, 1993, Lucas 1988, 1993; see chapters 48 and 49). These macro models of growth, like leading micro models of poverty traps (e.g., Loury 1981, Murphy, Schleifer and Vishny 1989, Azariadis and Drazen 1990, Banerjee and Newman 1993, and Galor and Zeira 1993, Mookherjee and Ray 2000; see chapters 6, 52 and 53), revived some of the internal and external economies of scale, pecuniary externalities and complementarities notions of the early development theorists, but now with the mathematical formalism required of contemporary economic theory. The literature has gone back to the future in rediscovering powerful ideas that lay dormant for many years.
The emergence of detailed micro‐level panel data sets has sparked a recent empirical literature of micro‐level studies of growth, many of which find suggestive evidence of poverty traps and the persistent effects of shocks on the poor (Dercon 1998, 2004, Lybbert et al. 2004, Barrett et al. 2007).2 The basic idea behind these recent explorations of micro‐level poverty traps is analogous to that in the more macro literature: locally increasing returns to scale may generate multiple equilibria and thwart the ability of initially poor households to catch‐up and converge with their wealthier neighbors. Carter and Barrett (2006; chapter 8) develop the reasoning behind such phenomena, discuss the empirical evidence, and offer a relevant extension of the familiar Foster‐ Greer‐Thorbecke class of poverty measures to the dynamic realm in the presence of potential asset thresholds associated with poverty traps.
