One of the oldest unresolved dilemmas in economics is whether market economies are naturally stable, or whether they need to be stabilized by policy. Classical and neoclassical models argue that a market economy can look after itself and that, in the absence of egregious government interference, it would gravitate naturally to full employment, greater innovation, and higher growth rates. However, Keynes emphasized the flimsiness of the expectations on which economic activity in decentralized markets is based. While the dominant intellectual paradigm since the 1970s has adhered to the idea that a market economy gravitates toward full employment, the historical record shows that the regime of the 1950s and 1960s was more successful than what followed afterwards. Indeed, other than the exceptions of the People's Republic of China (PRC), and perhaps a few more developing countries recently (e.g., India, Viet Nam), economic growth was faster and much more stable in the Keynesian golden period; its fruits were more equitably distributed and social cohesion and moral habits better maintained.

Modern analysis of economic growth started with the so-called Harrod– Domar model ( Harrod 1939, Domar 1946). This model has Keynesian and classical (mostly Ricardian) features. It is important to understand this model because subsequent modern growth models provide different solutions to the so-called Harrod– Domar knife-edge problem (discussed below). Moreover, I believe that, despite being “old,” it is still useful because it contains a fundamental feature of the experience of many developing countries, including those in developing Asia: the instability of growth. Indeed, a review of the world experience leads to the conclusion that growth is far from stable. For this reason, I consider this model an important thinking tool. Some may disagree with my view that the world experience at large is characterized by the instability of growth. For example, the United States (US), and in general the developed world, has had few downturns since 1979–1980, and these have been both short and shallow. Likewise, growth in the developing world has increased and African and Latin American economies are performing better. But the reality is that not all developing countries have done well. Indeed, episodes of high and sustained growth are not the norm across the developing world. Rather, these are concentrated in a few countries.