Measuring bad governance.

• Author: Osborne, Evan
• Position: Bad economic policy and poor economic growthi in developing countries

One of the most compelling features of postwar economic history is the continued prevalence of global poverty. Its intransigence is striking given the fact that development has been one of the main items on the international agenda since the end of World War II. The United Nations has numerous branches devoted entirely to reducing poverty, and both the World Bank and the International Monetary Fund have spent billions of dollars to promote development. Many wealthy nations also have government departments devoted to development assistance, and the amount of cash and in-kind transfers these countries have disbursed over the years is substantial.

After over half a century of such efforts, success has been scattered and slow, especially given the optimism that was so prevalent in the 1950s. If a dollar a day or less in income is used as the standard of poverty, the United Nations (2002) reports that the percentage of poor people in all developing countries fell between 1987 and 1999 from 28 to 24 percent, but very unevenly. A recent estimate suggested that almost half the world's population lives on less than two dollars a day (Duraippah 2000), and progress has varied tremendously by region. East Asia, particularly China, has showed dramatic progress, and India has advanced modestly since the early 1990s. But much of Latin America and sub-Saharan Africa has shown very little improvement. In addition, many nations that were once thought to be potential stars of development, such as Argentina and Brazil, have struggled mightily during most of the postwar period, even as other nations such as Singapore, Mauritius, and Chile have had unexpected success.

What might account for this pattern of long-term, widespread failure combined with isolated but dramatic success? While the 1990s saw the emergence of the "Washington consensus," the belief that good economic policy was essential to growth, there is still a widespread belief that the failure to grow is often due to factors partially or substantially out of control of developing-country governments. Geography and biology (Diamond 1997), political and social instability (Chen and Feng 1996), and harmful cultural norms (Lal 1988) are among the contrary explanations offered for poor economic performance. But these variables are difficult to change unless they do so endogenously as a result of growth, and so a fatalistic attitude toward poverty alleviation suggests itself when they are emphasized as causes of poverty. But if bad economic policy--more broadly thought of as bad governance--is largely to blame, there is room for hope.

This article investigates the effects of government policy on growth. Its goal is to see whether a prima facie case can be established that bad policy is responsible for much of the poor economic performance in the developing world in the post-colonial period. The findings provide a contrast to Easterly (1993), who argued that luck is more important than policy in determining who grows and who does not. Easterly (2001) also contends that good policy has had disappointing effects on growth, while Klein and Luu (2003) and Burnside and Dollar (2000) argue that sound policy is critical. The literature that diminishes the role of policy does not try to estimate its full effect on growth, an omission this article remedies.

Obtaining Estimates of the Determinants of Growth

The approach taken here is similar to that of Barro (1991, 1997), who adopts the neoclassical growth model as his starting point and argues that at any moment an economy has potential as well as actual output. Potential output is defined by the available production technology, including the ability of human capital to augment the productivity of physical capital, by government policy choices, and by factors beyond any government's control. A government that provides the most productive public goods, enforces property rights, and controls externalities--while avoiding distortions and the costly disincentives of excessive or inappropriate taxation--will have a higher level of potential output, as will a country that can trade with others on favorable terms. These factors are merged with the neoclassical growth model, which suggests that physical capital should be accumulated until the production frontier is reached, subject to the above constraints. Actual output is thus a function of the stock of productive factors as well as the factors determining potential output. Economic growth (i.e., the movement toward potential output) is a function of the accumulation of physical capital.

To measure policy effects, I use the following version of the standard cross-country growth regression:

(1) GROWTH = [a.sub.0] + [a.sub.1] INVGDP + [a.sub.2] HUMCAP + [a.sub.3] PREMIUM + [a.sub.4] GOV + [a.sub.5] INFLATION + [a.sub.6] OPENNESS + [a.sub.7] INSTABILITY + [a.sub.8] DEMOCRACY + [a.sub.9] TERMS + [a.sub.10] PCGDP.

Since the data are in panel form, random- and fixed-effects models were estimated. The results for the random-effects model are a much better fit than the fixed-effects model and almost indistinguishable from ordinary least squares. Thus, OLS is used for all estimations. Unless otherwise specified, the data are from the updated Barro and Lee data set, and all income groups and geographic regions are represented. The left-hand variable is average annual growth in real per capita gross domestic product (GDP) over five-year intervals from 1960-64 to 1990-94. INVGDP is the average ratio of investment to GDP during the interval, and measures the addition of physical capital. HUMCAP is the country's average life expectancy times the average years of schooling of its population at the beginning of the interval and is a proxy for the stock of human capital (Barro 1991). Both variables are expected to have positive signs.

INSTABILITY is the Barro-Lee measure of political instability, a weighted average of the number of assassination attempts and coups during the interval. By eroding the stability of policy and property rights, instability is expected to harm growth. TERMS is the change over the interval in the price index of the country's exports relative to its imports--that is, its terms of trade. When a country's terms of trade deteriorate, the gains from foreign trade and hence growth should be lower. These changes are assumed to be beyond the control of policymakers. DEMOCRACY is the average of the Barro (1997) measure of democracy. Theory and previous empirical work have not resolved the question of how democracy should affect growth. On the one hand, it can promote more economically efficient governance, assuming the population values that goal (Wittman 1989). On the other hand, it can lower the cost of, and hence promote, costly re-distributive activities. PCGDP is per capita GDP at the start of the interval. It is included because neoclassical growth theory predicts that as countries accumulate physical capital they should approach the technological frontier, and hence growth should slow.

There are four variables that are strongly influenced or completely determined by economic policy: PREMIUM, INFLATION, GOV, and OPENNESS. PREMIUM is log(1 + BMP), where BMP is the average black market premium on the country's official exchange rate during the interval. This variable is often used as a proxy for the total level of government distortions in the economy--restrictions on foreign exchange holdings, discriminatory taxes and subsidies, and the like. The hypothesis is that such distortions change relative prices, which in turn promotes inefficient resource use and rent seeking, thereby slowing growth. Nations that have no black market premium, including those that use floating exchange rate systems, generally have higher quality policy.

INFLATION is the average inflation rate over the interval, and there are a host of macroeconomic reasons why higher inflation might work against growth, such as noise introduced into relative prices when inflation is incorporated into particular markets at different speeds or the opportunity costs of coping with inflation. GOV stands for two measures of government spending and taxation: Specification (a) follows Barro (1991) by including the average ratio of government consumption spending--other than on defense and education--to GDP. This specification, which will be called pure government consumption (PUREGC), indicates the extent to which the government is causing damage by spending on less essential functions and promoting social conflict over the division of government spoils (Hirshleifer 1991)...