Third quarter real GDP posted its largest quarterly decline since 1991 and many economists have already declared that the U.S. fell into recession during this period. Two quarterly declines in real GDP is the common definition for a recession, but in fact the National Bureau of Economic Research (NBER) looks at more than just GDP to make its judgment. Some of the major data that the NBER monitors are the index of coincident indicators - nonfarm payrolls, personal income less transfer payments, industrial production and total business sales.

The table below looks at average changes in these key indicators during recessions and expansions. All the data are monthly averages, except for GDP which is a quarterly average. Also, while payroll and production represent physical changes and are therefore "real" data, income and sales are in nominal terms, not adjusted for inflation. An inflation adjustment would make the analysis more complete, but we are using the exact format of the series that are part of the coincident indicators. GDP is adjusted for inflation, however.

The yellow lines represent periods of growth, the green lines represent recessions. For instance, the U.S. economy peaked in December 1969 and fell into a recession with a trough in November 1970. During this time period, nonfarm payrolls were declining 0.1 percent per month, on average; income (less transfer payments) was growing 0.4 percent per month; industrial production was posting average monthly declines of 0.5 percent; and total business sales were recording declines of 0.1 percent per month. Real GDP fell 0.5 percent per quarter during this period. In order to determine the depth of the weakness, one can compare this recession period to the expansion of the 1960s that lasted for 105 months. Nonfarm payrolls grew then at an average of 0.3 percent per month; income (less transfer payments) grew 0.9 percent per month; production gained 0.7 percent per month; and total business sales grew 0.8 percent per month. Real GDP grew 3.4 percent per quarter (at an annualized rate).

Instead of looking at each of the specific recessions or expansions, we can also look at the average recession and expansion over the past five cycles. The average gain during expansions is shown in the last line of the table. The average growth during this expansion (listed as "Expansion 119 months") shows generally below average growth for an expansion for each of the monthly indicators and real GDP. Incidentally, we "ended" this expansion in March, which coincided with the peak in nonfarm payrolls. If the expansion ends any later, we would actually see lower average growth for this time frame. But expansion growth is old news right now. We are concerned with what to expect in these key indicators going forward.

During recessions, nonfarm payrolls have declined 0.2 percent per month; production has dropped 0.7 percent per month; business sales have decreased 0.1 percent per month and personal income less transfer payments have increased a modest 0.3 percent on average. Real GDP has declined 2 percent per quarter at annualized rates.

The longest recessions in the postwar have lasted 16 months (1973-75 and 1981-82) and the shortest was six months (1980). The average recession is about one year. The NBER is the official arbiter of recessions and they tend to declare the peaks and troughs of recession with a lag of several months. Oddly enough, they have already announced several times that a recession had not begun earlier this year. However, they are likely to look at the data again soon given the third quarter drop in real GDP.

What to expect in the next few months
If the U.S. is indeed in recession, expect declines for a few more months in key indicators such as nonfarm payrolls and industrial production. The shortest recession was six months back in 1980, but that recession was induced by stringent credit controls and should be considered an outlier. The next briefest recession lasted for eight months in 1990-91. At that time, many economists considered it a short and shallow recession. It was short, but based on the indicators shown in the table above, it wasn't really shallow. (The data we are looking at today is not the same data that economists had access to in 1990 and 1991. Current data, supposedly more complete, reflects several years of revisions.) Some might make the case that this recession is also a bit of an outlier given that the economy was teetering on the edge of recession before being seemingly pushed in by the September 11 terrorist attacks.

We've already seen pretty healthy declines in industrial production over the past twelve months as the manufacturing sector was indeed in recession while the rest of the economy managed to hobble along. Don't be surprised with declines of 250,000 to 350,000 per month in nonfarm payrolls over the next few months since that would be in line with the average 0.2 percent monthly decline. The employment situation for October will be reported on Friday, November 2, and the consensus forecast is indicating a drop of 300,000 for the month. (An aside on this figure: since this incorporates a good chunk of the impact from the September 11 attacks, don't be surprised if the decline is larger than expected.)

What about financial data?
The stock market is considered a leading indicator of economic activity. Thus, we would expect stocks to turn around in advance of the economy. In the previous six recessions, the S&P 500 increased in advance of the economy by as little as 3 months to as long as 5 months. The average lead time was four months. If the stock market bottomed in September, we could reasonably expect an economic recovery to begin in January or February.

The bottom line
A drop in third quarter real GDP - along with the expectation that growth will also decline in the fourth quarter - suggests that the U.S. may be in recession. Once the NBER evaluates key economic data, they will decide whether the U.S. was in recession and when it started, and they will give an official peak to the previous expansion. If the U.S. is indeed in recession, the depth of the recession will be closely monitored. Typically, a deep recession is followed by a robust recovery, while a more shallow recession is usually accompanied by a more moderate recovery. The stock market leads the economy by about four months. If September marked the trough in the stock market, we can begin to look for recovery early in 2002.

Evelina M. Tainer, Chief Economist, Econoday