May 1999

Inequality and poverty eradication - toward a Post-Washington Consensus

This growing focus on poverty has been accompanied by the emergence, consolidation and diffusion of an economic paradigm - known as the Washington Consensus - that emphasizes macroeconomic stability, domestic liberalization, privatization and a search for market solutions in the provision of public goods. The Washington Consensus has also emphasized policies (such as removal of barriers to international trade and liberalization of short-term portfolio flows) that have helped accelerate the pace of globalization of the world economy.

This approach, which has deeply marked national policy-making worldwide, is claimed to reduce "rent-seeking," improve competition, offer major opportunities for export and growth to developing countries, and promote the convergence of the living standards of poorer countries with those of the advanced nations - thus helping to reduce the incidence of global poverty. It is also asserted that the within-country distributive impact of policies based on this approach is, on the whole, neutral, that long-term income distribution is broadly stable and that there is no clear association between inequality and growth.

Several of these claims, however, lack solid support. During the past two decades, for example, with the exception of a few East Asian countries, little or no convergence has been achieved at the global level; for the majority of developing and transitional economies, the North-South and East-West income gaps were greater in the late 1990s than in 1980 or 1960. Further, a growing polarization among countries has been accompanied by a rise in inequality within most nations, adversely affecting their growth and poverty alleviation efforts.

The recent surge in inequality has been triggered, at least in part, by policies that are part of the Washington Consensus. It therefore seems clear that, to deal successfully with poverty, nations must not only tackle the traditional sources of inequality but also introduce macroeconomic and structural policies that avoid the distributive distortions of the Washington Consensus.

Inequality trends and causes

Most studies of global income distribution trends have been concerned with "between country" inequality. These studies suggest that while over the past 30 years China (particularly its coastal regions) and a few East and South-East Asian countries grew fast enough to converge towards the income per capita of the OECD group, the majority of the developing countries and all European transitional economies have further diverged from it. A 1997 United Nations Conference on Trade and Development (UNCTAD) study covering 124 countries (representing 94 per cent of the world population) found that the income share of the world's richest one-fifth rose from 69 per cent in 1965 to 83 percent in 1990. The report concluded that increasing global inequality has been a persistent feature of this period, with a noticeable worsening over the 1990s, and that there is little evidence that this tendency has since been reversed.

Other studies that have attempted to account for changes in per capita income both between and within countries suggest that global inequality has risen during the past three decades mainly because of an increase in between-country inequality. A 1999 study of 88 countries, for instance, found that between 1988 and 1993, three-quarters of the observed increase in the world Gini coefficient (a standard measure of inequality) was attributable to growth in between-country inequality, and just one-quarter to within-country inequality.

Inequality is affected by many factors. The so-called "traditional causes" of inequality include land concentration, dominance of natural resources, unequal access to education, and urban bias (in pricing policies, allocation of public expenditure and investment, and so on).

While such traditional factors were clearly responsible for the high income concentration observed in the 1950s through 1970s and their persistence at a high level in the subsequent two decades, they cannot (with the possible exception, in some regions, of educational inequality) explain the widespread surge in inequality observed over the past twenty years. Instead, several "new" factors are held by many to have had more relevance to the recent rise in inequality.

These factors can be classified as technological changes, with "new technologies" generating a demand for skills and an earnings distribution more skewed than that emanating from "old technologies," and policy-related changes. Among the more prominent policy-related factors are:

• distributionally adverse stabilization and structural adjustment policies (such as those developed to implement the numerous adjustment programmes introduced in the 1980s and 1990s with the assistance of the IMF and World Bank);
• erosion of labour institutions (such as greater wage flexibility, reduced regulation and minimum wages, and dilution of union bargaining power);
• a policy of (until recently) high interest rates and financialization of the economy;
• rapid international trade and (especially) financial liberalization;
• the faulty design and rushed implementation of massive privatization programmes; and
• erosion of the state's redistributive role (through regressive changes in tax systems and in the level and composition of public expenditures).

A new look at income inequality

Some recent mainstream studies have argued that long-term income inequality is stable, and thus inferred that growth is the only realistic option for poverty alleviation. These findings, however, run counter to a host of country and regional studies, and an examination of the analysis procedures suggests the conclusions were biased by the methodology used.

To re-estimate inequality trends following an unbiased methodology, UNU/WIDER created the World Income Inequality Database (WIID), which provides adequate time series data (often up to the mid-to-late 1990s) for 77 countries. These time series on inequality were interpolated using both linear and non-linear functions to achieve best-fit estimates.

The results found that in only 16 of the 77 countries analyzed is there evidence of a decline in income concentration over the long-term. In fact, inequality rose in 45 of the countries, and stopped declining over the long-term in 4 countries. (In 12 countries, no statistically significant trend was identified.)

If these results are weighed by population size and GDP-PPP (purchasing power parity), the conclusions are strengthened. Inequality has risen or stopped declining in nations accounting for 79 per cent of the population and 77 per cent of the GDP-PPP of the sample countries. The 16 countries in which inequality fell were mainly small and medium-size nations whose total population and GDP-PPP are only 16 and 20 per cent, respectively, of the total sample.

Thus, while inequality declined in many (but not all) countries during the first three decades of the post-WW II period, this trend was reversed with increasing frequency over the last two decades. Of the 45 countries with rising inequality, the trend reversal occurred during the 1980s in 18 cases and since 1990 in 14 cases. The rise of inequality was universal in the former Soviet Bloc, almost universal in Latin America, common in the OECD countries (including in the Nordic countries), and frequent in South, South-East and East Asia.

This trend reversal seems to reflect, to an important extent, a gradual "one-off" policy shift towards the Washington Consensus paradigm that took place during the 1980s and 1990s. In countries where such a shift in policy regime is still underway, inequality may surge in the future. In those where the shift is completed, income inequality is expected to remain stable at a level higher than the before the policy shift.

How theories shape policies

Determining the impact on growth of the inequality changes observed over the past two decades requires a brief discussion of distribution theories. The "old theories of distribution" suggested that inequality is pro-growth, and that public policy, therefore, should seek to promote a skewed distribution of income. Keynesian theories, for example, posited that profit earners have a higher propensity to save than wage earners, while monetarists argued that an increase in interest rates is associated with faster savings and growth. In both cases, however, supporting empirical evidence is weak or contradictory.

The 1990s has seen an explosion of "new theories of distribution" that arrive at radically different policy conclusions. The first group of such new theories, which focuses on political-economy considerations and imperfect capital markets, argues that high initial inequality in the distribution of assets (e.g., land) is detrimental to growth. Thus, from a policy perspective, these models recommend an initial asset redistribution and improved access to financial markets. Income redistribution via taxation, in contrast, is seen as having harmful effects on growth. One fault with this approach is that it neglects the recent factors that have raised the inequality of the initial distribution of market incomes and focuses only on redistribution.

A second group of new theories focuses on the impact of inequality on incentives, social conflicts, transaction costs and property rights. In this model, incentive erosion occurs at both very low and very high inequality levels. At very low levels (as in some socialist economies in the 1980s), growth is affected negatively as too compressed a wage distribution may not adequately reward different capabilities and efforts, and thus erode work incentives. Similarly, when the gap between the rich and the poor widens excessively, the work incentives of the "asset-less" poor wane, leading to higher labour shirking and supervision costs.

Security of property rights, minimum social cohesion and political stability tend also to be affected by high rates of inequality, unemployment and deprivation. Indeed, these theories suggest that high levels of vertical inequality (i.e., among undifferentiated households) and horizontal inequality (i.e., between social, class or ethnic groups) can create political instability and social tensions that, in turn, erode the security of property rights, augment the threat of expropriation, drive away domestic and foreign investment and increase the cost of business security and contract enforcement.

In contrast to the other models mentioned above, which posit monotonic linear relations between income or asset inequality and growth, the "distribution theories" that focus on incentives and transaction costs suggest a concave, asymmetric relationship between inequality and growth, with both "too low" and (especially) "too high" inequality hampering growth. In this distributive model, growth remains broadly invariant within a given "efficient inequality range."

This distribution model thus arrives at fundamentally different policy conclusions than the "old Keynesian theories" of distribution or the "new political economic models." In this model, already at moderately high levels of inequality, the rich have an incentive not only to reduce asset inequality and improve financial markets, but also to increase public spending on education, raise minimum wages and adopt other distributive and redistributive measures that lower transaction and monitoring costs and increase the security of property rights.

Policies for poverty reduction

By stifling growth and hindering incentives, the widespread increase in inequality observed over the last 20 years has proved detrimental to the achievement of the poverty reduction objectives adopted by the international community since the late 1980s. The World Bank's World Development Report 1990 projected that the total number of poor (people surviving on less than PPP$1 per day) would have fallen from 1.125 billion in 1985 to 825 million in 2000. Yet, recent assessments by the same institution (for years both before and after the Asian crisis) indicate that such target will be missed by a large margin. The number of poor worldwide is now expected to approach 1.3 billion in 2000. While part of the increase is due to population growth, another is certainly explained by a raise in inequality.

Thus, unless the Washington Consensus evolves in a distributionally favourable manner in the years ahead, high inequality is likely to continue to depress growth and reduce its poverty-alleviation elasticity, thus preventing the achievement of the DAC poverty-reduction target cited earlier.

As has been noted, while the traditional causes still explain a lot of the variation in cross-country inequality, they cannot account for the recently observed "increase over time" of inequality. Commitment to pro-poor growth, therefore, requires a two-pronged approach: continued focus on the "structural causes of inequality" but also the introduction of alternative macroeconomic, liberalization and redistributive policies with a more favourable distributive impact. Such new policies must be incorporated in a new Post-Washington Consensus approach to development.

First of all, as noted, the old sources of inequality must of course be addressed. This can be accomplished through traditional measures of asset and income redistribution (such as well-designed land reforms), reallocation of public spending (on education, health and infrastructure) and the creation of broad employment opportunities to eliminate social, regional, urban/rural and ethnic bias.

But secondly, the new sources of inequality must be addressed. In its rush to promote ill-designed privatization and premature liberalization of financial markets in the presence of weak regulatory capacity, the Washington Consensus has contributed to rising income and asset inequality. Traditional measures, such as those described above, will not be able to contain the poverty upswings caused by these factors. To accelerate poverty reduction, therefore, the Post-Washington Consensus must explore alternative macroeconomic and adjustment reforms.

Among these reforms might be:

• introduction of stabilization and structural adjustment programmes that minimize the present distributional bias;
• creation of a healthy balance of power between capital and labour, through appropriate wage policies and the design of appropriate labour market regulations;
• international mechanisms to dampen the volatility of commodity prices and short-term portfolio flows;
• greater attention to the institutional, efficiency and distributive impact of privatization in transition economies, and greater caution in its use; and
• in emerging economies, a reduction in the output volatility associated with financial contagion (including, where needed, a short-term reversion to capital controls).

Many more policy suggestions could be added to the list. The significant point is that discussion has shown that equity is not necessarily in conflict with efficiency. Indeed, there is room for well-designed macroeconomic, structural and redistributive policies with a fair distributive impact, which can shift economies towards an optimal combination of growth, inequality, and poverty reduction. If rapid poverty reduction is the over-arching goal of the international community, it is essential that more thought and attention is given to the elaboration and introduction of such Post-Washington Consensus policies.