Remarks, Annual World Bank Conference on Development Economics, Washington, DC, April 18, 2000

In these remarks I propose that development strategies for the 21st century cannot continue to overlook the middle class. I begin with comments linking global governance to a new style of development assistance. This new style, of which Collier and Rodrik are leading proponents, is based on a revisionist view of the fundamental development challenge. This revisionist view sees home-grown institutions as the weak link in the development process, the scarce resource critical to countries? development success in a global market. In the advanced economies with democratic systems, it has generally been the middle class that has invented and supported these homegrown, market-sustaining institutions. But who in fact are the middle class in developing countries? I provide some information on middle income households in developing countries, and conclude with some implications for development policy and for the conduct of development assistance programs.

Global governance and imperfect representation

The demonstrations in Washington D.C. this past weekend were fundamentally about global governance. They remind us that the Bretton Woods and other international financial institutions are caught in a squeeze. On one side are those who insist they become supranational vehicles for the good and the just ? be it labor standards, a cleaner environment, or uncorrupted and more transparent governments. In this camp are also those who would have these global institutions become the major force for more efficient and stable global markets ("making the world safe for capital flows" is Rodrik?s apt characterization) ? for example by enforcing disclosure of national financial and macroeconomic indicators. (That puts the social activists and IMF management on the same side of this question at least.) On the other side are those who decry the leverage of these institutions on policy decisions in independent nations, the breakdown of national autonomy their leverage implies, and their lack of adequate accountability to the citizens of countries affected by their decisions.

The IMF and the World Bank are global institutions with, inevitably, imperfect representation. Representation of member governments is imperfect, with limited voice for borrowing member governments most affected by the institutions? policies. Representation of the citizens of member countries is more imperfect, since most member governments are themselves, even if democracies, highly imperfect ones. Transparency, disclosure, and the involvement and pressures of civil society groups in member countries are partial, but only partial solutions to these problems of imperfect representation.

Paul Collier and Dani Rodrik are leading proponents of new thinking about the business of development assistance, given this system in which power and money are poorly aligned with representation of and accountability to affected parties. They both take as a starting point that development assistance will be more effective where it eschews policy leverage (aka "conditionality") and instead concentrates on supporting national autonomy, local institutions, and the construction of a sustainable local consensus. This is because, as Rodrik argues, the key to past development success, once the worst macroeconomic imbalances have been corrected, has not been any particular recipe ? certainly not a particular trade or industrial strategy, and not necessarily or primarily integration into world trade and capital markets. The key has been effective and home-grown market-sustaining public institutions that, as Rodrik puts it, are able to manage social conflicts, the constant demands of changing markets, and openness itself; and, as Collier puts it, build and sustain a political consensus around policy reform. Because local institutions must be built on history, geography and pre-existing political and social conditions, they are likely to be different across countries and within countries over time. Recipes and remedies on the part of foreign development advisors may thus do more harm than good.

But where will the market-sustaining institutions for managing change and building consensus come from? Beyond a hands-off tolerance for diversity and national autonomy, what principles might guide development assistance efforts to support such institutions? Surely a "middle class" needs to play a role; in the advanced Western economies, after all, it is the middle class that has been the backbone of democracy and of market-sustaining institutions.

The middle strata in developing countries: new opportunities and new anxieties

The development literature and economists in general have been virtually silent on the middle class. Economists have no simple, cross-societal and time-invariant definition of the middle class. But consider the middle "strata" in developing countries, for example those households with income per capita between 75 and 125 percent of the median. Households in Brazil and Poland have about one-tenth the income (in purchasing power parity terms) of households in this group in the U.S. In Brazil adults in these households have fewer than five years of education, and in Costa Rica fewer than seven years, compared to about 13 years in the U.S. Compared to rich countries, the income levels of middle strata households in Latin America (and probably in Africa and parts of Asia) are far below "average" or mean incomes in their countries, because of greater income concentration at the top. Differences in education between high and middle strata households are also substantially greater. The ratio of years of education in the top 5 percent of households relative to the middle strata is about 2 to 1 in Latin America compared to 1.2 to 1 in the U.S.

In other words, middle strata households are much poorer, in absolute terms and relative to their richer fellow citizens than are their middle class counterparts in rich countries. In fact, many are past or future members of the poor: as many as 30 percent of nonpoor households in Indonesia and 22 percent in Peru are likely to become poor in the next three to six years ? based on past movements. It is not surprising, therefore, that surveys find the middle strata in many countries are an anxious and literally unhappy group ? even when and where they have prospered.

Anxiety is in part the outcome of the greater uncertainty of more market-driven economies operating in a more integrated and volatile global economy. In Latin America, the return to growth in the 1990s brought modest income gains on average, but increasing economic insecurity for middle-strata households. The loss of secure jobs in government and state-owned enterprises was not made up by increases in private sector employment, and even those who kept secure public sector jobs lost ground in terms of wage growth compared to private sector and informal sector workers. "Unprotected" jobs, in which workers have no written contract nor social benefits, increased, probably affecting the middle strata most given that the truly poor never enjoyed protected jobs in the first place. Rodrik in other work (1999) shows that there was an increase in the volatility of the real average wage in the 1990s, and that the increase was mainly driven by macroeconomic shocks associated with more open capital markets.

In the countries of the former Soviet Union, the middle strata were hit hard by the transition process. Even in Poland, where average annual per capita growth was positive between 1986 and 1995, the share of population in the middle strata (75 to 125 percent of median household income per capita) shrank more than 15 percent, and its income share shrank proportionately. In Hungary between 1991 and 1994, the middle group?s population share shrank by 13 percent and its income share by 17 percent.

In East Asia, the financial crisis hit hard those in an emergent middle class who had enjoyed several prior decades of continuous and reliable income gains. Declining employment and real wage hurt middle-income salaried workers, and heavy reliance on high interest rates to defend falling currencies hurt business owners. The rural poor and urban households dependent on work in the low-productivity informal sector were obviously vulnerable to the economy-wide recessions. But a second group, including many households with higher wage jobs in construction and manufacturing ? what might be called market-friendly urban strivers -- was also hit. The crisis revealed their economic exposure, if not to dire poverty, certainly to sudden and severe income and welfare losses.

Easterlin has pointed out that at low income people care most about absolute gains in income, while at higher income people care more about their income relative to those above them. Households around the median and above in developing countries are more likely to suffer resulting "middle class stress" compared to households in richer countries. The gaps in income and education between the rich and the middle noted above are one indication of such stress ? no doubt exacerbated by the increasing globalization of consumption standards (Nike sneakers the goal of teens in Rio?s favelas). Consider another measure of middle class stress: the ratio of the median income of those households commanding 50 percent of all national income, to the median household?s income. That ratio is between 1.5 and 2 in most advanced economies and in Taiwan and Poland and probably in South Asia. But it ranges from 3 to 8 in Latin America and is probably in that high range in the Philippines, Thailand and much of Africa.

The point is not to bemoan the effects of market reforms and globalization on the middle strata or emerging middle class. In fact, many of those now in the middle strata are there because market reforms generated new opportunities to escape poverty. It is to underline how little consideration we have given to the determinants of the size and income share of the middle; to the effects on the middle of market reforms and the integration of economies into global markets; and to the role of this group in the political and economic discourse and the institutional underpinnings of their countries? policies.

Nurturing a market-friendly middle class

Is a market-friendly middle class relevant to development policy and to the conduct of the development assistance business? What are the causes and consequences of a market-friendly middle class? I want to note three points.

First, some market reforms ? privatization, trade liberalization, tax reform -- have not been easy on the traditional middle class. However the design and implementation of the so-called second generation of reforms can be explicitly oriented to creating new middle-class stakeholders in the market economy. The middle class in today?s advanced Western economies was built on the tripod of work, education and property ? encouraging high levels of highly productive investment by households in small farms and businesses, and in their own and their children?s human capital. In developing countries, better banking, better schools, the rule of law, contract enforcement and a labor market organized around collective bargaining are the kinds of changes that will consolidate the stake of the emerging middle class in a new market system.

Second, Rodrik?s emphasis on managing social conflict and Collier?s argument to get beyond conditionality to consensus both in the end require the existence of an informed middle class in a reasonably democratic setting. The institutions they call for require a middle class not only with voting rights, but with the liberties that protect the rights of speech and association. This seems particularly important in settings where ethnic, racial, cultural, religious and linguistic differences mean decisions about economic priorities can only be sustained if they are the outcome of a political process, not a technical analysis. Moreover, an informed middle class may be necessary to sustain the democratic processes through which difficult economic choices are made.

Third, the middle class is at the heart of any effective and sustainable social contract. To be sustained politically, the social safety net programs built into a social contract must reflect the outcome of a political bargain that meets the needs of the middle class as well as the poor. In the West, the poor benefit more from a small portion of massive middle-class driven social insurance programs than from a large portion of targeted services. Thus it may be that an early focus on "targeting" the poor, if it implies a new burden on an insecure middle class, will be less politically acceptable and sustainable than a more encompassing medium term strategy ? an example of the need for local choices over design. A social contract, moreover, extends far beyond the narrow notion of the safety net; it includes the political freedoms and rights of participation governing a broad range of social and economic decisions. The design and maintenance of this broad social contract depends ultimately on the allegiance and intelligence of a stakeholding middle class.

* * *

Some will argue that the middle class is endogenous ? that it is an outcome not an input to healthy institutions. This may be true. But my point is not to propose an analytic framework or a new magic bullet. It is simply that we would do well to incoporate the idea of the middle class into our thinking about development.

Finally, a focus on the middle class does not imply a neglect of the poor. In the end, the sustainable growth on which poverty reduction depends requires adequate institutions. History tells us that the middle class is the bedrock of those institutions, at least in the democratic and open market systems to which so many people in the developing world justifiably aspire. Poverty reduction may ultimately require a market-friendly middle class.



Aghion, Philippe, Eve Caroli and Cecilia Garcia-Penalosa. 1999. "Inequality and Economic Growth: The Perspective of the New Growth Theories." Journal of Economic Literature 37:4, December.

Birdsall, Nancy and Stephan Haggard. Forthcoming. "A New Social Contract in East Asia: Politics Matters Too". Washington D.C.: Carnegie Endowment for International Peace Monograph.

Birdsall, Nancy, Carol Graham and Stefano Pettinato. 2000. "Stuck in the Tunnel: Have New Markets Muddled the Middle Class?" Carnegie Endowment for International Peace. Washington D.C.

Collier, Paul. "Consensus-Building, Knowledge and Conditionality." Paper presented at the World Bank Annual Conference on Development Economics, April.

Graham, Carol and Stefano Pettinato. 1999. "Assessing Hardship and Happiness: Trends in Mobility and Expectations in the New Market Economies." Brookings Institution Center on Social and Economic Dynamics Working Paper No. 7. Washington D.C.

Easterlin, Richard A. 1995. "Will Raising the Incomes of All Increase the Happiness of All?", Journal of Economic Behavior and Organization 27(1).

Nelson, Joan. 2000. The Politics of Social Sector Reforms. Overseas Development Council. Washington D.C.

Pritchett, Lant, Asep Suryahadi and Sudarno Sumarto. 2000. "Quantifying Vulnerability to Poverty: A Proposed Measure with Application to Indonesia". World Bank. Washington D.C.

Rodrik, Dani. 2000. "Development Strategies for the Next Century." Paper presented at World Bank Annual Conference on Development Economics, April.

Rodrik, Dani. 1999. "Why is there So Much Economic Insecurity in Latin America? John F. Kennedy School of Government. Cambridge, Mass.