Michael Dueker is Head Economist for North America at Russell Investments and a member of the Blue Chip forecasting panel. In February of 2008 he warned Econbrowser readers that it appeared unlikely that the economy was going to escape the slowdown without a recession. In December of 2008, he predicted in this forum that the recession would last until July or August of 2009, but that employment growth would not resume until March of 2010.

With that track record, we were very interested to learn the latest macroeconomic predictions stemming from Russell’s Business Cycle Index, subject to the disclaimer that the content
does not constitute investment advice or projections of the stock market or any specific investment.

The latent variable behind the NBER recession indicator in a Qual VAR model of recessions provides a useful business cycle index, where the distance above zero indicates the strength of an expansion or the distance below zero indicates the depth of a recession. The Qual VAR model (or this link) is designed to identify and characterize recessions in real time, based on the incoming financial and macroeconomic data, and it provides a framework for forecasting future business cycle developments. The particular data used in the estimation of the model are personal consumption expenditures, CPI inflation, the slope of the yield curve, the spread between 3-month commercial paper and Treasury bills, the spread between Baa and Aaa corporate bond yields and nonfarm payroll employment.

Starting this month, Russell Investments has adopted this business cycle measure as the Russell Business Cycle Index (BCI) and will post updated estimates of the Russell BCI and forecasts on the Helping Advisors site at Russell Investments. A real-time history of the business cycle index and its forecasts will be available in a spreadsheet.

For now, we can evaluate past performance of the business cycle model by noting that December’s forecast of a July or August 2009 trough on Econbrowser appears to have been correct and by examining last December’s forecasts of payroll employment. The figure below shows the path of payroll employment changes predicted by the Qual VAR model in the forecast posted on Econbrowser in early December 2008 along with the actual data since then. To date the actual pattern of job losses matches the forecast from December 2008 quite closely. In particular, the forecast last December was that the economy would lose about 4.9 million jobs between December 2008 and August 2009, whereas the data to date show a loss of about 4.5 million jobs.

Source: St. Louis Fed’s FRED database for actual employment data and author’s calculations. These macroeconomic forecasts do not constitute a projection of the stock market or of any specific investment.

Nevertheless, the job losses projected last December for the period September 2009 to March 2010 might be smaller than anticipated earlier. We examine this question with a chart of a current employment forecast below.

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. For recent observations that belong to a period where the eventual NBER classification is not yet clear, the model can be run without imposing a sign and the data can speak regarding the yes/no recession classification. Starting with the June 2009 data, we stopped imposing a negative sign on the business cycle index and let the data determine it. As of the August 2009 data, the first month where the probability that the business cycle index went below 50 percent was August. This result is the basis for our call that the economy was out of recession by August 2009.

One key point of discussion at present is whether the economy faces a danger of sliding back into recession for the second part of a double-dip recession. The business cycle index and its accompanying payroll employment forecasts can help assess the risk of a double-dip recession.
The chart of the business cycle index below illustrates that some backsliding in business cycle conditions is projected early in 2010.

The distance from zero indicates the depth of a recession or the strength of an expansion. Source: author’s calculations. These macroeconomic forecasts do not constitute a projection of the stock market or of any specific investment.

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 standard deviations from zero after 1984. The depth of the 2008-09 recession poses a strong counterargument to the Great Moderation hypothesis, however. The business cycle index also illustrates how some economic expansions were sufficiently strong and sustained to bring the unemployment rate down to low levels, such as 3.8 percent by April 2000 and 4.4 percent by October 2006.

To see if the projected backsliding in business cycle conditions in early 2010 represents a serious threat of a double-dip recession, we can look at the forecasted recession probabilities for future months. The chart below shows that the recession probabilities do not reach the threshold between 40 and 50 percent where we would call a double-dip recession.

Source: author’s calculations. These probabilities are derived from simulations of a Qual VAR model with the specific variables mentioned above.

The current employment forecast shows some promising job gains between October and December 2009 and then some backsliding, with a small negative forecast for March 2010. If this occurs, there will be discussion about the possibility of a double-dip recession, although that is not the scenario projected here. Instead of a double-dip recession, the forecast presented here is one where it takes a long time for the economy to achieve consistent triple-digit job gains, which are not expected until 2011.

Source: St. Louis Fed FRED database for actual data and author’s calculations. These forecasts are derived from simulations of a Qual VAR model with the specific variables mentioned above.

Last 5 posts by James Hamilton

• Factors in local house price declines - November 22nd, 2009
• Receiver operating characteristics curve - November 18th, 2009
• Commodity inflation - November 15th, 2009
• Will rising oil prices derail the recovery? - November 10th, 2009
• Consequences of the Lehman failure - November 7th, 2009