Employment surveys point to a recession that's already begun

Hopes that the United States could avoid a recession were renewed this past week by evidence that the government's rebate checks were stimulating spending, at least for now.

But two recession indicators based on employment statistics offer evidence that a recession may have begun months ago.

While both indicators are based on declining job markets, they differ in that each comes from a different survey conducted by the Bureau of Labor Statistics. Those surveys — one of business establishments and the other of households — sometimes offer contradictory views of the employment picture. But now they both indicate an economy that has slowed significantly.

The business establishment survey reflects changes over 12-month periods in the number of private sector jobs, before seasonal adjustments.

Normally, that total increases even if the economy is not strong, simply because the working-age population is rising. But when the number of jobs falls over a 12-month period, that is an indication that a recession is already under way.

Since 1953, there have been 10 periods when the figures fell into negative territory. The first nine came months after a recession had already started, although that was never clear until later. The 10th period began in May.

The lags in that indicator — the time from the beginning of the recession until the time that the year-over-year job figures enter negative territory — have ranged from two to 12 months, with an average of five to six months.

The job figures from the establishment survey will be modified several times before the final number for any month is reported, and it is possible the revisions may change the recession signal, making it come earlier or later or not at all.

The household survey, however, is not revised, so that recession indicator will remain. The indicator is based on a 1 percentage point increase in the unemployment rate over a 12-month period. As with the other indicator, each time that has happened since 1953 a recession had already begun three to 10 months earlier. It also signaled a recession in May.

The popular definition of a recession is two consecutive quarters of decline in real gross domestic product, but that is not the definition used by the National Bureau of Economic Research, which will eventually determine if there was a recession.