A couple of decades ago, the vast majority of economists tended to believe that – due to the flows of knowledge around a globalized world – poor countries had access to the exact same technologies as rich nations. Concretely, if such theory were true, an increase in financial openness would lead to foreign investment in poor countries because their deep poverty would imply low capital stocks and therefore a high marginal productivity of capital (by the law of diminishing marginal returns). Indeed, this is a standard neoclassical prediction. In the long-run, this framework would imply economic convergence between rich and poor countries, as the increase in capital per worker would lead to an increase in output per worker.
However, in May 1990, the American economist Robert Lucas Jr. added to the increasingly critical reviews of the neoclassical model by arguing, in a very influential paper – “Why Doesn’t Capital Flow from Rich to Poor Countries?” – that if such a model were anywhere close to being accurate, given large capital return differentials, then one would observe no investment or whatsoever in rich countries. In his own words, if this model indeed fitted reality, then “[…] investment goods would flow rapidly from the United States and other wealthy countries to India and other poor countries. Indeed, one would expect no investment to occur in the wealthy countries […]. The assumptions on technology and trade conditions that give rise to this example must be drastically wrong, but exactly what is wrong with them, and what assumptions should replace them? […]”. The observation that, in reality, capital does not flow from rich to poor countries is widely known as the Lucas Paradox. According to Robert Lucas, this riddle is exceedingly important for economic development. Now, why is there a mismatch between this neoclassical model and reality? Where exactly did we go “drastically wrong”?
For one thing, the assumption that countries have access to the same productivity is flawed. Many economists nowadays think of productivity as social efficiency rather than technical knowledge. Social efficiency may include elements such as trust and institutions while technical efficiency is defined narrowly as “a function of technology and management capabilities”. Specifically, a low level of social efficiency may be linked to poor education policies, low levels of human capital, excessive bureaucracies, corrupt governments, under provision of public goods etc. As a matter of fact, one of the fundamental reasons for why poor countries are poor is because of their low productivity that induces low marginal productivity of capital despite low capital stocks. Regrettably, low levels of productivity in poor countries discourage investment in these nations: the result is no longer one of economic converge as predicted by the benchmark model; once we allow for differences in productivity, poor countries will have limited gains from access to capital markets and divergence will replace convergence.
Besides differences in fundamentals, the Lucas Paradox could potentially also be explained by allowing for risk premiums, uncertainty and asymmetric information. In fact, investors may demand high risk premiums in order to invest in risky countries (such as developing countries), if so, we would actually observe less investment in developing countries and more investment in rich countries: capital would hence flow ‘uphill’. Finally, in poor countries, a large share of GDP is derived from natural resources and the capital share of GDP is small, which further decreases the marginal productivity of capital, and further discourages investment.
All in all, the Lucas Paradox is an important concept. It tells us that (sadly) capital is not expected to flow from rich to poor countries, because the former will tend to have a relatively higher marginal productivity of capital. A careful investigation of the Lucas Paradox leads to a serious claim: it is quite possible that capital markets will not contribute to significant convergence unless poor countries can find a way to increase their levels of productivity and social efficiency.
 Lucas Jr., Robert (1990) Why doesn´t Capital Flow from Rich to Poor Countries? American Economic Review.
 In his paper, Lucas compared India and the US in 1988 and found that (if the neoclassical model were true) the marginal product of capital in India should be about 58 times (!) that of the U.S.
 Cf. Feenstra, Robert C.; Taylor, Alan M.; International Economics, 3rd edition, Chapter 17, Worth Macmillan.