Predicting the trough and a jobless recovery

Michael Dueker is a senior portfolio strategist at Russell Investments and formerly was an assistant vice president in the Research Department at the Federal Reserve Bank of St. Louis. Michael is also a member of the Blue Chip forecasting panel. In early February 2008, Michael submitted a piece to Econbrowser that correctly predicted the onset of the current recession, using a model-based forecast. We are pleased that that he is now presenting forecasts from the same Qual VAR model concerning the recession’s trough date and the magnitude of a jobless recovery to follow, subject to the disclaimer that the content is the responsibility of the author and does not represent official positions of Russell Investments and does not constitute investment advice.

Current business cycle forecasts see a July or August 2009 trough and a jobless recovery until March 2010 by Michael DuekerIn analyzing the current recession, it is useful to be able to predict the trough date and its depth and to compare it with previous recessions. To do so, I use the Qual VAR model (or this link) because it is designed to identify and characterize recessions in real time, based on the incoming macroeconomic data, as well as forecast future business cycle developments. The announcement from the National Bureau of Economic Research on December 1, 2008, of a cyclical peak in December 2007 meant that the recession onset date I forecast earlier was realized, despite the fact that the current recession did not begin with at least two consecutive quarters of negative GDP growth (at least not in the GDP data available to date).In general, the Qual VAR model espouses Jim Hamilton’s view, that recessions correspond to a switching of gears in the economy, rather than simply a loss of speed due to a few randomly missed strokes within a single gear. As Hamilton has noted, if economic cycles do not signify more than the loss of speed, then the amount of speed the economy needs to lose to signify an economic downturn becomes completely arbitrary, as does the meaning of an economic downturn. In other words, the Qual VAR treats recessions as distinct events in the economy, akin to switches between high and low gears. Inferences as to whether the economy is in an expansionary or recessionary state add important perspective to interpretations of GDP numbers, such as the 2.8 percent GDP growth rate in 2008Q2.One attribute of the Qual VAR approach is that, during periods when the NBER classification is well-established, one can use those classifications to determine the sign of the business cycle index. Until the NBER’s announcement on December 1 of a peak, however, the NBER classification of 2008 was not yet certain. Throughout 2008, I was able to run the model without imposing either a positive or negative sign on the index’s 2008 values. During this period, I could use the model to infer the probability that a given month pertained to a recession. Throughout 2008, the model’s estimates suggested that a recession was underway, although the data for July and August 2008 indicated that the probability of recession had fallen to about 50 percent prior to the failure of Lehman Brothers. After that, the business cycle index began a sharp decline throughout autumn 2008.The release on December 5 of the November 2008 payroll employment data made the projected path of the recession go deeper than it went in previous iterations of the Qual VAR forecasts. One handy feature of the Qual VAR is that it provides separate forecasts of the end of the formal recession and the date at which employment growth is expected to become positive. Thus the model can project the extent to which we can expect a jobless recovery to follow the formal recession. For example, after the 2001 recession ended in November 2001, negative employment growth continued unabated until October 2002 and payroll employment did not reach its cyclical bottom until August 2003.The figure below shows the model-implied business cycle index, where its distance from zero indicates either the depth of a recession or the strength of an expansion. The history of the business cycle index illustrates the so-called Great Moderation in the U.S. economy after 1984. Until the current recession, the business cycle index stayed within a comparatively narrow range between -0.5 and 1.5 after 1984. The projected depth of the current recession poses a strong counterargument to the Great Moderation hypothesis.` fig1BCI.jpg

Nevertheless, one could argue that the U.S. economy was experiencing a run-of-the-mill recession until the failure of Lehman Brothers, with the business cycle index bouncing around -0.5. After that shock, financial market conditions sent the economy sharply downward. As for a snap-back after the end of the formal recession, the projected path of the business cycle index is not as vertical as one might hope between zero and one in the second half of 2010. In fact, the cumulative area below +0.5 in the current downturn between August 2007 and January 2011 is quite large. Again this event calls into question the permanence of the Great Moderation in the U.S. economy after 1984.

Given that the past two recessions have been followed by jobless recoveries, where employment continues to fall and remain sluggish well after the end of the formal NBER recession, it is interesting to look at the Qual VAR’s forecasts of employment growth. The second figure shows that the forecast is for a jobless recovery to follow the NBER recession again, with payroll employment declining until March 2010 and not returning to trend growth until July 2010.

` fig2empl.JPG

The most similar recession to the current one is 1973-75 when the unemployment rate increased about 4-1/2 percentage points in about a year and a half. In the double-dip recession in the early 1980s, the unemployment rate increased by 5 percentage points but this increase took 3-1/2 years. The stock market decline this year also resembles the death by a thousand cuts in 1974, when the U.S. stock market declined by about 30 percent without hitting a particular crash point.

A few details about the model: Here I have focused on a monthly version of the business cycle model that uses nonfarm payroll employment as the measure of economic activity. In addition, the model uses data on core CPI inflation, the spread between the interest rate on one-month commercial paper rate and Treasury bills, the rate spread between corporate and Treasury bonds and the slope of the yield curve. These quality spreads among interest rates are particularly important in the current context where the failure of the investment bank Lehman Brothers led to a severe contraction of lending in financial markets. For example, issuance of three-month commercial paper essentially ceased.


Originally published at Econbrowser and reproduced here with the author’s permission.