That's because historically, the unemployment rate, which is derived from a survey of the population, has been an excellent "real-time gauge" of the economy's strength, which has pretty much held up in later revisions. By contrast, in the early stages of a recovery, payroll estimates, which are based on a survey of companies, are often revised sharply upward as more data come in.
UPWARD REVISIONS. That's exactly what happened in the wake of the early 1990s recession. After joblessness peaked at 7.8% in June, 1992, payrolls in the next six months were originally reported to grow by an average of only 42,000 a month.
But since then, they have been revised up to an average gain of 121,000.
Similarly, the average monthly gain in payrolls in the first half of 1993 was originally estimated at 132,000 but is now pegged at 212,000. Meanwhile, the jobless rates for the period have hardly been revised at all.
In sum, Morris says, the recent fall in unemployment suggests that the labor market has been considerably stronger than is commonly believed. By Gene Koretz in New York