Conclusion
In this chapter, we have analyzed growth models that do not rely solely on physical capital accumulation and exogenous technical progress. Instead, they are based on external effects from the accumulation of physical capital (learning by doing), or on investment in human capital (education and training), or even on the endogenous generation of technical progress through ideas and innovations.
Under certain conditions, such models can generate endogenous growth.Endogenous growth models do not necessarily imply convergence, as do the corresponding exogenous growth models. However, the available empirical evidence from postwar international experience (see, for example, Mankiw et al. [1992] and Barro [1997a]) indicates that the issue of conditional convergence of per capita incomes of the various economies cannot be dismissed easily. Convergence can be explained by generalized models that include learning by doing, accumulation of human capital, and endogenous technical progress, but not to a degree that would completely neutralize the diminishing returns from the accumulation of physical capital.
Consequently, such generalized exogenous growth models could, in principle, explain most aspects of the process of economic growth that cannot be explained by the original Solow model or the corresponding representative household or OLG models.
The analysis of these generalized growth models helps explain additional stylized facts about economic growth, such as those recently summarized by Jones and Romer [2010, p. 242]. To quote from their conclusions:
The virtuous circle between population and ideas accounts for the acceleration of growth. Institutions may have their most important effects on cross-country income differences by hindering the adoption and utilization of ideas from throughout the world. Institutions like public education and the university system are surely important for understanding the growth in human capital.
And institutions are themselves ideas—inventions that shape the allocation of resources—and the search for better institutions is unending.The emphasis here is on the accumulation of physical and human capital, and endogenous technical progress as the proximate determinants of living standards and the growth process in the long run. But one needs to understand why some countries succeed and others fail to raise living standards and promote economic growth. The evolving literature on the political conditions and institutions that are conducive or nonconducive to long-run growth has already shed some light on these questions and continues to do so.
This chapter concludes our examination of the key macroeconomic models of long-run economic growth. We shall next turn to stochastic models and the rational expectations hypothesis, before moving on to stochastic models of consumption, investment, and money demand as a prelude to the analysis of models of aggregate fluctuations.13
1. Throughout the chapter, we mainly address effects on the efficiency of labor, but all the results would go through if instead we considered effects on total factor productivity. As we use a Cobb-Douglas production function throughout this chapter, labor augmenting technical progress (Harrod neutral) and total factor productivity augmenting technical progress (Hicks neutral) are equivalent.
2. This endogenous-growth model belongs to a class of models known as AK models from the form of the aggregate production function (8.7). The AK model with learning by doing was analyzed by P. M. Romer [1986], and we shall refer to it as the Arrow-Romer model. An alternative AK model that is widely used is due to Rebelo [1991], who demonstrates that as long as there is a core of capital goods whose production does not involve nonreproducible factors, endogenous growth is compatible with production technologies that exhibit constant returns to scale. The simplest Rebelo [1991] model is a two-sector model in which consumption goods are produced using both capital and labor, but capital goods are produced using only capital.
In this class of models, there are no externalities through knowledge spillovers, but in competitive equilibrium aggregate output turns out to be proportional to the capital stock. Because the Rebelo model is not based on externalities, its policy implications are different from those of the Arrow-Romer model.3. This market failure does not arise in the benchmark AK model of Rebelo [1991] without externalities.
4. This is demonstrated in section 8.1.10.
5. See chapter 5 for a detailed derivation of aggregate consumption in the Blanchard-Weil model.
6. For example, see Mankiw et al. [1992] and Barro [1997a].
7. For R&D models of the 1960s, see Uzawa [1965], Phelps [1966], Shell [1966], and Nordhaus [1969]. For models that build on the ideas of Romer [1990], see Grossman and Helpman [1991], Aghion and Howitt [1992], and Jones [1995].
8. Due to the increasing returns implied by ideas and innovations, the assumption of perfect competition is no longer appropriate, and one should ideally use models of imperfect competition. See Romer [1990].
9. For an empirical investigation of how population growth affects technical progress in historical perspective, see Kremer [1993].
10. See, among others, Galor and Weil [1999, 2000], Lucas [2002], Hansen and Prescott [2002], and Galor [2005; 2011].
11. North [1990] is an important contribution to this historical literature, which has been recently surveyed by Crafts and O’Rourke [2014].
12. See Acemoglu et al. [2005] and Ogilvie and Carus [2014] for surveys of this literature. Acemoglu and Robinson [2012] apply such a framework to the explanation of differences in prosperity among nations.
13. For comprehensive surveys of the extensive recent literature on economic growth and other additional models, including political economy models, see The Handbook of Economic Growth, edited by Aghion and Durlauf [2005] and Aghion and Durlauf [2014], as well as the excellent presentations in the books by Acemoglu [2009]; Aghion and Howitt [2009]; Barro and Sala-i Martin [2004]; and, on unified growth theory, Galor [2011].