Has Okun's Law changed over time, with emphasis on the 2007 - 09 recession?

Has Okun’s Law Changed over Time, with Emphasis on the 2007-09 Recession?

By

Fatma Abdel-Raouf, Ph.D.

Associate Professor of Economics and Finance

Goldey-Beacom College

4701 Limestone Road

Wilmington, DE 19808

Phone: (302) 225-6212

raouff@gbc.edu

With the persistent high unemployment rate that the 2007-09 recession is witnessing, it is important to reexamine the strength of the relationship between the change in the unemployment rate and the output growth. Arthur Okun in 1962 quantified this relationship where he estimated that for every 3 percentage point increase in the U.S. output growth rate, the unemployment rate decreases by one percentage point. Since then, many economists have examined this relationship and no consensus has been reached about the extent of this relationship.

In this paper, I’m going to examine whether Okun’s law has changed over time, with emphasis on the recent economic downturn. Two specific events are examined: The Great Moderation and business cycles. This research examines whether the Great Moderation has a significant effect on Okun’s coefficient and whether Okun’s coefficient has changed over the business cycles. In doing so, I’m going to use a rolling regression technique to enable me to see whether there has been change over time or not.  The analysis is done using two versions of Okun’s law: The difference version and the dynamic version. The difference version is estimated using the following equation:

                 = a + b X_t + e_t

Where                    : the change in the unemployment rate

                                X_t: Real output growth

The dynamic version of Okun’s law allows the change in the unemployment rate to depend on current and past output growth as well as past unemployment rate. It is estimated using an autoregressive distributed lag ARDL (2, 2) model:

                 = a + b₁ X_t + b₂ X_t-1 + b₂ X_t-2 + c₁ u_t-1 + c₂ u_t-2 + e_t

Where                  X_t-1: output growth lagged one quarter

                                X_t-2: output growth lagged two quarters

                                u_t-1: change in unemployment lagged one quarter

                                u_t-2: change in unemployment rate lagged two quarters.

This topic is important for policy makers in trying to reduce the unemployment rate. The accurate relationship between change in the unemployment rate and the output growth rate helps policymakers in setting their goals to reduce unemployment rate. In that regard, the autoregressive distributed lag (ARDL) model is useful in forecasting the growth rate needed to reduce the unemployment rate over different periods of time. ARDL models have two main uses that are important as policy tools. They are forecasting and multiplier analysis. Both of them are applied to Okun’s law and examined in this paper.

The paper starting with re-estimation of Okun’s relationship using the same time period he used and then proceeds to extend the length of time until the second quarter of 2010. The results show that Okun’s law has not been stable over time and is sensitive to the state of the business cycles. Okun’s coefficient has varied over time along with the output growth needed for stable unemployment.  Moreover, in the 2007-09 recession, Okun’s coefficient in absolute value has increased significantly.