The National Bureau of Economic Research today announced that its Business Cycle Dating Committee had officially determined a peak in economic activity at December 2007, which signals the start of the recession. I am a member of the committee. Though I speak only for myself, not the committee, I offer my views on two questions of possible interest:
(1) Who needs the NBER Business Cycle Dating Committee (BCDC) anyway?
(2) Why did we pick December 2007 as the starting month of the recession?
(1) We sometimes hear the question “Who needs the NBER Committee anyway?” This question most often comes in one of two forms: (1a) Everyone in the real world has known that the economy has been in a serious recession for some time. In past cycles, media reports have sometimes taken the line “Ivy Tower Eggheads Finally Figure Out What Everybody Else Has Known All Along.” The implicit critique is that the committee takes too long after the event – typically almost a year — to make its declaration. One short answer is that our job is to be definitive, not fast. GDP and other government statistics are often revised after the fact, for example. We don’t want to have to revise our dating of the peaks and troughs later, in part because it would sow confusion among those who rely on them (from econometric researchers to political speechwriters). We leave it to others –pundits, forecasters, consulting companies, financial newsletters, and so on – to try to get there first. We deliberately get there last. (1b) The other form taken by the question “Who needs the NBER committee?” runs as follows: “The rule of thumb is simple: two consecutive negative quarters of GDP growth. Why complicate things?” The Frequently Asked Questions segment of the BCDC announcement answers this in detail. For now, observe simply that questions (1a) and (1b) are inconsistent with each other. As of December 1, 2008, the US economy has not yet experienced two consecutive negative quarters. So an argument that we should wait for two consecutive quarters (critique 1b) is the opposite of the critique that we should have acknowledged a recession before now (critique 1a). (2) The more important question is: Why did we pick December 2007 as the start of the recession? As is the case surprisingly often, different economic indicators give very different answers to the date of the peak.
Of the monthly indicators to which the BCDC gives primary attention, the most important is jobs, more specifically Payroll Employment (from the Labor Department’s Bureau of Labor Statistics). It peaked in December 2007, and has been declining ever since. My personal favorite indicator is Total Hours Worked (which is closely related, because it is number of people employed times the average number of hours per worker). Hours Worked also peaked in December, as shown in the graph below.
Of the quarterly indicators, the most important is aggregate economic activity, more specifically, Output. The Commerce Department’s Bureau of Economic Analysis computes two measures of output: Gross Domestic Product (GDP) and Gross National Income (GNI). The two should be the same in theory, but differ in practice due to measurement errors. GDP receives far more public attention, but in fact has no claim to be a more accurate measure of output than does National Income. The statistics currently available show that GNI peaked in Quarter 3 of 2007, whereas GDP peaked in Quarter 2 of 2008. A simple-minded average of the two peak dates would seem to point to midnight of New Year’s Eve, December 2007, as the peak. Another (comparably unsatisfactory) way of forcing the output data to cough up a precise month is to look at Personal Income, which is available monthly. The BCDC’s computed measure of real personal income less transfers peaked in December 2007. It would be wrong to claim that all roads arrive at the same destination, December 2007. Other indicators point to other dates, some earlier, some later. If we are very lucky, revisions that the BEA makes in July 2009 will help resolve the discrepancy between the GDP and GDI measures somewhere in the middle. But perhaps the best characterization of the output measures is that they show a rough plateau from the fall of 2007 to the summer of 2008. That the employment statistics speak more clearly allows them to have the predominant say.
Originally published at Jeffrey Frankel’s weblog and reproduced here with the author’s permission.