|
The National Bureau of Economic Research, the nonpartisan group charged with making the call, announced Monday that U.S. economy entered a recession in December 2007. AIER’s statistical indicators, in contrast, warned about a recession months before its official start date and more than a year and a half before NBER’s official announcement. We first wrote that the recession was likely, based on the analysis of our statistical indicators of business-cycle conditions, in April 2007. By July 2008, we wrote that the recession had probably begun. Last December, the official peak of economic activity that marked the start of the recession, AIER had already begun to warn that the recession was likely, despite the seemingly upbeat economic data available at the time.
AIER’s methodology aims at forecasting the reversals in business-cycle trends just before or shortly after they happen, when there is still time to alter plans in preparation for an economic downturn. Given the complexity of measuring economic activity accurately, it is rare to forecast the business-cycle reversals exactly. A discrepancy of a few months is inevitable, but AIER forecasts usually lead the official start date of a recession by several months. The current recession, which is the 11th since World War II, ended an expansion that lasted 73 months. The table below shows when AIER first said a turn in the business cycle was likely compared to the official dates for business-cycle peaks and troughs. | AIER “Forecasts” of Business-Cycle Turns | | ----------------- | Peaks | --------------- | ----------------- | Troughs | ------ | | NBER Date | AIER Forecast | | NBER Date | AIER Forecast | | | Jul. 1953 | Jun. 1953 | (–1)* | May 1954 | Jun. 1954 | (+1)* | | Aug. 1957 | Jan. 1957 | (–7) | Apr. 1958 | Jun. 1958 | (+2) | | Apr. 1960 | Feb. 1960 | (–2) | Feb. 1961 | May 1961 | (+3) | | Dec. 1969 | Aug. 1969 | (–4) | Nov. 1970 | Mar. 1972 | (+16) | | Nov. 1973 | Nov. 1973 | (0) | Mar. 1975 | Jun. 1975 | (+3) | | Jan. 1980 | Jul. 1979 | (–6) | Jul. 1980 | Sept. 1980 | (+2) | | Jul. 1981 | Oct. 1981 | (+3) | Nov. 1982 | Apr. 1982 | (–7) | | Jul. 1990 | Jan. 1990 | (–6) | Mar. 1991 | Jun. 1991 | (+3) | | Mar. 2001 | May 2001 | (+2) | Nov. 2001 | Mar. 2002 | (+4) | | Dec. 2007 | Apr. 2007 | (–8) | | | | | * Lead (–) or lag (+) in months from business-cycle peak or trough. | To receive AIER’s detailed monthly analysis of business-cycle conditions, plus articles and books on a wide range of economic issues, become a member.
|
Our methodology aims at warning about a recession somewhat ahead of time, while the information can still be usefully employed in adjusting one's plans for a coming downturn. Therefore, I would consider warning about a recession 8 months in advance of its "official start date" to be quite good an outcome.
You bring up a valid point that there is a difference between warning that a recession is imminent and asserting that the recession has already started. We did say, in the Research Reports issue from July 7, 2008, that our analysis suggests business contraction has begun (at that point the analysis was based on the available data, which only provided information though May of 2008). Thus, we did “call the recession” in early July, while NBER made their determination only this week.
No forecasting procedure is perfect, and ours is not an exception. It would be great to find a procedure that allowed us to predict the business cycle turns more precisely than we currently can, but we have not found one yet. Our current methodology is better at warning about an imminent recession that at timing the date of the downturn exactly. Our hope is that this information is still useful to our reader.
Respectfully,
Polina Vlasenko