November 2009 Unemployment

The government released two reports on unemployment that both indicate that the employment picture is improving and the economy is recovering. The 4 week moving average of initial unemployment claims of unemployment continued its thirteen week declining to 481,250 (Unemployment Insurance Weekly Claims Report), and unemployment for November 2009 fell .2% to 10.0% (Employment Situation Summary).

Given that the 4 week moving average peaked at 658,750 on April 4th of this year, it was likely that unemployment would peak soon as well. Robert J. Gordon did research looking at the relationship between the 4 week moving averages of initial unemployment claims and found that recessions often bottom out shortly after the 4-week moving average of initial unemployment claims peaks.

The three graphs compare the current recession which began late in 2007 and one of the three previous recessions which began in 1981, 1990 and 2001. The graphs are not forecasts. I have not modeled the data with the goal of predicting future levels of unemployment. But the raw data does provide insights.

The first compares the current recession to the recession that began in 1981. At this point, the two look similar in depth and pace of recovery. The 1981 recession began a rapid recovery in the months immediately outside of the graphs' time horizon. Just eyeballing the graphs (and this is scientifically dangerous) suggests that if the 4 week moving average truly is a leading indicator of future employment trends, that the economy may experience a robust recovery.

The 1991 and 2001 recessions are somewhat different, both shorter from the peak of the previous economic boom to the highest 4 week moving average. In both cases,the unemployment rate declined slowly. Next months data will show unemployment beginning to fall approximately 15 months after the 4 week moving average peaked. During the 2001 recession, unemployment will peak 3 months beyond the graphed time frame, 20 months after the 4 week moving average peaked. If unemployment during the current recession has peaked, in will occur only 7 months after the 4 week moving average peaked.

The unemployment data suggests that unemployment begins to fall after the 4 week moving average peaks and between 84 and 128 weeks after the recessions began. It does not suggest that the 2007 will follow suit, or why the recessions ended. Did the recoveries occur because of smart policy or because markets recover on their own? Does policy help or hinder a recovery? If policy is successful, how were the policies employed to fight recessions similar and different?

My guess is that economists will consider these questions for a generation.