Sustained economic growth: do institutions matter, and which one prevails?

• Author: Mijiyawa, Abdoul' Ganiou

In 1965, the growth rate of per capita GDP in Niger and Nigeria was 2.1 percent and 4.2 percent, respectively, and 2.9 percent in Botswana. From 1966 to 1969, however, Niger and Nigeria recorded a negative growth rate, while Botswana continued to experience a positive growth rate over the same period. In 1990, the growth rate of per capita GDP was 1 percent in Ghana and 5.2 percent in Nigeria. Yet, from 1991 to 1994, the growth rate was negative in Nigeria and positive in Ghana. Why does the trajectory of economic growth episodes differ among countries? In other words, why is economic growth more sustainable in some countries than in others?

There are at least two policy-relevant justifications for answering these questions. First, durable poverty reduction requires sustained economic growth. Second, in the absence of sustained growth, policymakers constantly reexamine their policies. In this situation, private investors also continually reexamine their investment projects, which increases the risk of poor economic performance. Thus, policymakers need to identify a framework that allows them to make economic growth sustainable as soon as they succeed in generating it.

The thesis that I propose in this article is that sustained economic growth (SEG) requires "good" institutions. By "good" institutions I mean those institutions that guarantee private investors relatively lower costs for their investments and ensure that private investors appropriate the fruit of their investments. Good institutions enable private investors to take advantage of favorable business opportunities. Indeed, private investors prefer lower costs for their investments, since this allows wealth creation. They also want to be able to reap a significant share of the return of their investments when they invest. These two conditions are satisfied by the presence of good institutions. In the absence of good institutions, investors may pass on favorable business opportunities, reducing the probability of sustained economic growth.

Rodrik (2005) also argues that good institutions are necessary for SEG. The first objective of this article is to empirically test this assumption and to identify the most important institutions for SEG. To do so, I analyze the combined, separate, and simultaneous effects on SEG of democratic, private property rights, and economic regulatory institutions. The second objective of this article is to identify the mechanisms by which good institutions could affect SEG. In this case I suggest that good institutions, through their favorable effects on private investment, involve an increase in total factor productivity (TFP) that increases economic competitiveness, which is necessary for SEG.

This article tackles the general question of the role of institutions for economic performance, treated by Acemoglu, Johson, and Robinson (2001) and Hall and Jones (1999) among others. Contrary to those authors who are interested in the effect of institutional quality on the level of per capita income, in this article I am interested in the effect of institutional quality on SEG, which seems to be one of the best measurements of economic performance. Indeed, the more a country's economic growth is sustainable, the higher its per capita income will be, especially when the growth rate is strong. SEG is thus the necessary input for the determination of a country's income level, which is determined by both the rate and the sustainability of economic growth.

By focusing on SEG, I also take into account Pritchett's (2000) claim that economies experience various phases of growth over time and that by calculating the average of growth rates over a long period, scholars lose useful information. As a result, while analyzing SEG I do not calculate average growth rates over a long period, but I observe the evolution of growth rates over five consecutive years and I investigate whether the durability of economic growth episodes could be due to institutional quality.

Empirically, this article builds on recent studies by Hausmann, Pritchett, and Rodrik (2004, 2005) and Jerzmanowski (2006). While these authors are interested in the changes of economic growth regimes, this article is interested in the durable character of growth, regardless of whether or not this growth characterizes a change in economic growth regimes. Hausmann et al. focus on political institutions and find a positive and significant effect of these institutions on growth accelerations. Jerzmanowski, however, concentrates on economic institutions and finds a positive and significant effect of economic institutions on the occurrence of favorable and sustained changes in growth regimes. This article reconciles the approaches followed by Hausmann et al. and Jerzmanowski by testing the effect on SEG of a composite index of political and economic institutions. Also, contrary to previous studies, my analysis examines the mechanisms of transmission of the effect of institutions on SEG. Lastly, to my knowledge Hausmann et al. and Jerzmanowski do not tackle the endogeneity issue in their articles, whereas I try to overcome this kind of shortcoming by using the generalized method of moments (GMM) method of Blundell and Bond (1998).

The remainder of the article is organized as follows. Section 2 presents the various characteristics of SEG from 1960 to 2003. Section 3 expounds the theoretical arguments of this article, while section 4 is devoted to empirical analysis. Section 5 presents the results and section 6 concludes.

Characteristics of Sustained Economic Growth

Looking at Table 1, it appears that SEG over the period 1960-2003 is not a rare phenomenon. Indeed, the probability of a representative country in my sample to experience sustained growth during this period is 0.36. Yet, the probability of high sustained growth is only 0.21 during the same period, suggesting that sustaining high economic growth rate is relatively more difficult. (1)

For the whole sample, the period preceding the oil crises of the late 1970s and early 1980s is most favorable for SEG. During the oil crises, the probability of SEG in a country in my sample is reduced almost by half relative to the previous period. Soon after the oil crises, the number of countries having experienced sustained growth immediately increased, before diminishing during the first five years of the 1990s. At the end of the 1990s, the probability of SEG reached again its value of the period preceding the oil crises, whereas it was not the ease for high SEG.

This overall picture of the evolution of SEG masks differences among groups of countries. Indeed, although the period preceding the oil crises is more favorable for SEG for all the countries, it appears that a developed country generally is more likely to experience sustained growth than a developing country. Moreover, the trajectory of the probabilities of SEG in developed and developing countries reveals differences which are especially stark after the oil crises.

In developed countries, the five-year term following the oil crises was marked by an increase in the number of countries having experienced sustained growth, whereas the period 1990-94 was marked by a reduction in this number. Thus, developed countries quickly recovered from the oil crises, but their recovery was not durable because of disturbances in European financial and exchanges markets in the late 1980s and early 1990s, as well as the Gulf War. During the last two five-year terms, the probability of SEG in developed countries reached its value of the period preceding the oil crises, but we can observe a small decline in the value of this probability during the last five-year term.

In developing countries, the recovery from the oil crises was not immediate and did not take place until 1990-94. This recovery was incremental and improved during the last five-year term, when the probability of SEG reached the same value it held during the period preceding that of the oil crises. Thus, there is a difference in the cycles of SEG of developed and developing countries. Even among developing countries, however, there also are differences in the cycles of SEG.

The countries in Asia and the Pacific, compared to the other developing countries are atypical in terms of SEG, because in general the probability of SEG for countries in Asia and the Pacific is always higher than that of the representative country of nay sample.

The evolution of SEG probability for North African and Middle Eastern countries shows that for this group of developing countries, the most favorable period for SEG was that of the first oil crisis. In this region, the recovery from the second off crisis was delayed and undertaken in incremental way. There was a clear improvement in the probability of SEG during the last five-year term, especially due to the performance of North African countries.

In sub-Saharan Africa, the last two five-year terms were the most favorable period for SEG. It is possible that this is the materialization of the effects of economic reforms such as structural adjustment, and the devaluation of the CFA franc, as well as the beginning of democratization in sub-Saharan Africa during the 1980s and 1990s. The cycle of SEG immediately following the oil crises in sub-Saharan Africa is similar to that of developed countries, which to a certain extent, reflects the tight connection between this region's economies and those of developed countries. Indeed, the upturn of economic activities in sub-Saharan Africa following the second oil shock stopped five years later, just as in developed countries.

In Latin America and the Caribbean, the period preceding the oil crises was most favorable for SEG. The region's growth was marred by the second oil shock, during which time none of the countries in this region experienced SEG. In Latin America and the Caribbean, the recovery from the oil crises was immediate and was characterized by the...