Adam Smith, the patron saint of free markets, had a clear-eyed analysis of inequality. "Servants, laborers and workmen of different kinds," he wrote, "make up the far greater part of every political society. But what improves the circumstances of the greater part can never be regarded as an inconveniency to the whole."

What a contrast with the dominant view in our time! Today the leading economists tell us that real wages must be cut, to compete with those whose wages are lower still. They say that minimum wage laws, and trade unions, cost jobs. They say that my own country, the United States, is a model of flexible labor markets of full employment achieved by accepting poverty. They teach that a global order of privatization, deregulation, free trade and open capital markets the Washington Consensus will produce great new gains for all the world"s population.

In Europe, the modern wisdom holds that high unemployment owes to a legacy of socialism, social democracy and the welfare state. Yet if you examine European countries one by one, you discover that those with the strongest egalitarian traditions (think Norway and Denmark), also enjoy the lowest unemployment. And it is in those countries who most recently left fascism behind (think Spain), where unemployment has been highest.

There is no general case, in Europe or elsewhere, of countries achieving sustained new prosperity by cutting real wages or accepting increased poverty for parts of their population. Despite great efforts, the economists have failed to show that minimum wage laws or unions cost jobs. And the United States, a country with strong public pensions, 90 percent public school enrollment, vast public (as well as private) health and university sectors, many municipal enterprises and extensive regulation and where the minimum wage rose as pay inequalities and poverty and unemployment all fell in the late 1990s is no model for the Washington Consensus.

In development studies, the great pioneer Simon Kuznets argued long ago that as industrialization proceeded, an urban, democratic middle class would normally bring economic inequalities down, as happened from the mid-19th through the mid-20th centuries in both the U.K. and the US. But today, development economists have mostly set Kuznets aside, in favor of views that emphasize the supposed efficiencies of market process.

One viewpoint, favors an initial equalization of opportunity, particularly in education and land tenure. But then, the argument goes, markets should be allowed to determine pay, and if the final result is more unequal, so be it. Another viewpoint holds that inequalities per se foster growth, perhaps because saving requires accumulation, and this is something that can be entrusted only to the rich.

But, don"t the economists know about all this by now? Unfortunately the most widely used measurements of inequality published in 1996 by the World Bank have only confused the issues. The numbers are often implausible showing Indonesia more equal than Australia, just for instance. And the coverage is so limited that one cannot gauge developments through time. Amazingly, you cannot even say for sure, from these measures, that worldwide inequality has risen in the past twenty years. But does anyone seriously doubt it?

To fill the gap, a small group called the University of Texas Inequality Project now offers systematic, nearly annual measures of inequality in industrial pay for over 150 countries, going back to the early 1960s. With a narrow focus on pay measured accurately and consistently through time we are able to cast new light on the crucial relationships between economic inequality and economic growth.

What do we find? First, worldwide inequality has been rising sharply under globalization. While the increase did begin in the United States, back around 1970, it has been more dramatic in later years in Latin America, Central Europe, China, and above all Russia. Among OECD countries, New Zealand experienced the largest increases, Australia somewhat less. In a few countries, notably in Scandinavia, pay inequalities have not risen much or at all.

Second, in broad terms Kuznets was right: inequality usually does decline as total incomes grow high inequalities are more prevalent in poor countries than in rich. Part of the problem in the world economy has been a slowdown in average growth the upcoming world slump will almost surely make pay inequalities worse. (There is no support in our measures for either variant of the "modern view.")

Third, the global trend in inequality was stable in the 1960s, slightly downward in the 1970s, and up sharply after 1981 and ever since. What accounts for this? Technology? Trade? We think not. Rather, the timing points mainly at the quasi-violent financial regime change of the early 1980s: rising real interest rates, debt crisis, and the triumph of private global finance.