Thursday, June 7, 2012

Cleveland Fed Throws Keynesian (and Bernanke) Notion Overboard

In a new paper, Federal Reserve Bank of Cleveland researchers Brent Meyer and Murat Tasci raise questions about the usefulness of Okun's law, a Keynesian spinoff rule of thumb, that suggests that there is a stable relationship between the growth rate of economic output and the unemployment rate.

According to the researchers, Okun's law would suggest that output growth above 3.4 percent is needed to see a decrease in the unemployment rate. However, in 2011, the unemployment rate fell from 9.1 percent to 8.3 percent, while real GDP grew only 1.6 percent.

Meyer and Tasci tested several versions of Okun's law and found that the relationship between output and unemployment is not stable over time. This instability makes Okun's law a joke that is unusable in assessing past movements, and forecasting future movements, in unemploymen, M & T concludet.

BTW, this is one of the wacky econometric type correlations that Bernanke buys into. According to Andrew Abel and Bernanke, estimates based on data from more recent years that about a 2% decrease in output correlates with every 1% increase in unemployment (Abel and Bernanke, 2005).

Here's Abel and Bernanke writing the pseudo law in pseudo sophisticated mathematical jargon.

The gap version of Okun's law may be written (Abel & Bernanke 2005) as:
(\overline{Y}-Y)/\overline{Y} = c(u-\overline{u}), where
  • \overline{Y} is potential GDP
  • Y is actual output
  • \overline{u} is the natural rate of unemployment
  • u is actual unemployment rate
  • c is the factor relating changes in unemployment to changes in output
In the United States since 1955 or so, the value of c has typically been around 2 or 3, as explained above.
The gap version of Okun's law, as shown above, is difficult to use in practice because \overline{Y} and \overline{u} can only be estimated, not measured. A more commonly used form of Okun's law, known as the difference or growth rate form of Okun's law, relates changes in output to changes in unemployment:
\Delta Y/Y = k - c \Delta u\,, where:
  • Y and c are as defined above
  • \Delta Y is the change in actual output from one year to the next
  • \Delta u is the change in actual unemployment from one year to the next
  • k is the average annual growth rate of full-employment output
At the present time in the United States, k is about 3% and c is about 2, so the equation may be written
\Delta Y/Y = .03 - 2 \Delta u.\,
The graph at the top of this article illustrates the growth rate form of Okun's law, measured quarterly rather than annually.


  1. I won't believe this until Krugman has blessed (rejected) this finding?

  2. Since they are continually tampering with how "unemployment" and "GDP" are calculated one would need to go to more fundamental "untamperable" underlying metrics. Any rule based on tampered variables will be invalid, working on the false assumption there exists any true rules (as you point out).

  3. Isn't Okun's law what is used to claim that the "Long Depression" experienced high unemployment since employment data isn't actually available?

  4. How many different ways can the tea leaf readers re-state the "Phillips Curve," have the same theory proven wrong, then tag another idiot's name to the same alleged chicken entrail pattern? Does this cycle have to be repeated every generation?