Readers of TheStreet.com know me mainly as a Fed basher . But I have another life entirely. In it, I pursue, for the pure pleasure of discovery, the measurement of economic inequalities. There are many different kinds: inequalities of pay, of income, across industries and economic sectors, across regions, in the U.S. and around the globe. I measure them, graph them, map them and compare them to measures of other economic variables, like unemployment, gross domestic product and other people's inequality measurements.

Related Stories

Tech-Driven Productivity Gains=U.S. Surplus=Tax Cuts, the Thinking Goes

M&A Madness Rings in 2001 With Deafening Quiet

What a Week: Bulls Shrug Off Bad News, Embrace Al's Testimony and Advance the Market

This work -- on display at the University of Texas Inequality Project Web site -- has been well received by professional economists, giving rise to papers with impossibly long titles, dissertation topics for my students and a small career for myself as a traveling evangelist for certain statistical techniques. But it also produces, on occasion, information of larger importance.

In particular, the monthly employment and earnings data sets produced by the Bureau of Labor Statistics can be used to measure inequality in pay in the U.S. manufacturing sector, and this calculation can be made just as soon as the data are released. Because employment and earnings data come out every month, monthly inequality measurements can be just as up to date as unemployment measurements, one of the most sensitive indicators of current economic performance that we have.

An Early Warning?

This measure of inequality has been very closely related to unemployment and to economic performance in general over a very long time; month to month, inequality and unemployment match as far back as 1939. The inequality measure may actually lead unemployment by a bit, giving an early warning of downturns. When the economy slows and unemployment is about to rise, firms first cut the hours and overtime of their hourly workers. This shows up as a fall in relative pay (monthly earnings) for the lowest-paid, and the inequality measure ticks up. Unemployment itself tends to rise a bit later, when firms are (as a rule, reluctantly) forced to eliminate whole shifts, shut factories and let workers go.

Here's the headline: This particular inequality measure had been falling smartly since 1994, pretty much in line with the unemployment rate. But it stopped falling in early 1999. Since June 2000, the measure has started rising sharply, with particularly large jumps in October and November of last year. A chart showing the entire series, all the way back to 1947, illustrates this. Just last week, Vidal Garza Cantu, a UTIP veteran who is now an assistant professor at the Monterrey Institute of Technology in Monterrey, Mexico, produced this chart.

Wage Inequality in the U.S. When the Inequality Measure climbs, recession tends to follow

Source: calculated by Vidal Garza Cantu with data from Bureau of Labor Statistics U.S. Dept. of Labor

As you can see, the start of recessions in 1957, 1960, 1970, 1974, 1980 and 1990 were all marked by distinct upturns in this series. It's a sensitive indicator, in other words, of larger economic conditions. And, given what it has started to do over the past year, this is not good news.

It is risky to lean heavily on a single measure -- particularly in this case, when the weight of manufacturing in employment has been falling over the decades, as it has been. But on the other hand, it would be foolish to ignore a bit of evidence, when a lot of other evidence also points the same way.

The Politics of Policy

Policy implications? Let's concede a point to Dick Cheney. The risk of recession is a serious one, as he said in December. But let's also face the implications. The Fed should cut interest rates, sure. And Congress should cut taxes (as well as increase spending on public services, like education). But what taxes? Not, surely, taxes that mainly fall on creditors, savers and the wealthy. Not, surely, tax rates four, five and six years from now. So leave the income tax, the capital gains tax and the estate tax alone. We are seeing a fall in income, relatively speaking, of the working American families with the lowest incomes, not the highest. That is what our inequality measure shows. And we are seeing it today. In fact, it has been going on for more than a year.

So, let's cut taxes on working Americans now. How? The best idea I've seen so far has been proposed by Randall Wray of the University of Missouri at Kansas City and my fellow fellow at the Jerome Levy Economics Institute.

Wray points out that Social Security payroll taxes now exceed benefits, this year, by around $150 billion. A cut in payroll tax rates, sufficient to bring the trust fund into balance for a year or two, would usefully put money directly into the pockets of working Americans this year where it is most needed. If the measure stabilizes the economy, the long-term net effect on the Social Security trust fund would likely be quite small.

Obviously some Republicans, determined to cut taxes for the wealthy and to privatize Social Security over their own dead political bodies, will oppose this simple, effective proposal. But others, who may recognize what a long and steep recession will do to them politically in any event, might just line up with public-spirited Democrats and, for once, save the country. Or am I dreaming?