Thursday, October 11, 2012

Forecasting America's Unemployment Rate


After September's jobs numbers showed a drop in unemployment to below 8 percent, many Americans grew leery of the accuracy of the data used, particularly in light of the fact that the presidential election is less than one month away.  A recent article by Dr. Regis Barnichon, a junior researcher at the Centre de Recerca en Economia Internacional (CREI) in Barcelona provides us with another analysis of unemployment data and gives us an outlook for the next six months.

Let's open by looking at a chart showing the unemployment rate for the last five years from the National Conference of State Legislatures:


The October 2012 rate of 7.9 percent is the second lowest rate since January 2009 when the rate was 7.6 percent,  right at the cusp of the employment problems that began as the Great Recession became entrenched.

October 2012's data looked better than expected largely because there was a strong increase in the number of workers finding work and a decrease in workers being laid off.  The author feels that these improvements are likely transitory in nature.

The author uses the Barnichon-Nekarda model to forecast future levels of unemployment.  This model, a more robust way to forecast changes to unemployment, is based on labor force flows, looking at the relationship between the published unemployment rate derived from the number of employed and unemployed workers "u" and the rate of unemployment implied by the underlying labor force flows into and out of unemployment "u*".  When "u*" is above the actual rate "u", the unemployment rate tends to rise and vice versa as shown on this graph:


This model predicts the rate of unemployment that would prevail if the flows into and out of unemployment remain at their current rate.  An unemployment steady-state is reached when inflows into unemployment and outflows from unemployment are balanced; during a recession, inflows jump and the conditional steady-state unemployment rate also jumps.  In roughly three to five months, the actual unemployment rate converges with the new conditional steady state.  By examining the changes to the steady-state unemployment rate, the model provides information about the unemployment rate in the near future.

The author's calculations show that the current steady-state unemployment rate (the rate of unemployment implied by the aforementioned labor force flows) is around 7.9 percent as shown in the blue line on this figure:


From the Barnichon-Nekarda model, Dr. Barnichon provides us with the following unemployment forecast:


Since the unemployment rate is now very close to the steady-state unemployment rate, the author suggests that the unemployment rate is likely to stay constant over the next few months and that the unemployment level, while starting from a lower reference point, has very little room to drop further.    This is largely because the separation rate is likely to be elevated at the same time as the increase in the ability of workers to find jobs is slow to recover. His analysis suggests that, over the next six months the unemployment rate will hover around the 7.9 percent level.

With very modest economic growth being the "new norm", it would appear that Dr. Barnichon's analysis could be correct and that, for the next half year or so, very modest improvements to America's employment picture will take place.

1 comment:

  1. Employment figures are always fudged. Realistically, when we look at those figures and delineate those who are only working for $5.00/hour and only 15 hours/week, or even less, then we come to realize that many who are considered employed, can't even earn enough to pay for rent, or board and lodgings or even enough food to keep them healthy! So, why are these people even considered in the statistic? A more realistic figure is more like 25%!

    ReplyDelete