It’s a weak recovery, not a jobless recovery

That is the conclusion of a fascinating new paper from the National Bureau of Economic Research titled ‘Okun’s Law at fifty’. The authors, Laurence M. Ball, Daniel Leigh and Prakash Loungani, look at a long series of historical data to assess the claim that Okun’s Law has broken down in recent times. Whilst there has been many claims that this fundamental economic ‘law’ is no longer applicable, Ball et al show that this interpretation is incorrect: the relationship between economic growth and unemployment has held up remarkably well.

What is Okun’s Law?

In the early 1960’s, Arthur Okun noted that output and unemployment had a negative short run relationship. In other words, an increase in output of one percent above potential will reduce unemployment by a certain percentage (the coefficient of Okun’s Law). Ball et al note that this is as close to an economic ‘law’ as possible in the field, so it is with great importance that we are sure if it has been violated or not.

Okun’s Law is captured by the following relationship:

Okun's law

In this equation, U refers to unemployment, Y refers to output and * variables refer to the long run levels of unemployment and output. β here is the coefficient in Okun’s Law.

How did Ball et al test this apparent breakdown?

Ball et al examined Okun’s Law on both quarterly and annual data for the US, in both the ‘levels’ version of the law as shown above and the ‘changes’ version of the law. In the levels equation, they use d the Hodrick Prescott (HP) filter to estimate the long run levels of output and unemployment. The authors take care to account for the different interpretations of recent movements of output and unemployment (whether it is structural or cyclical as discussed in last week’s post ). Regardless of the interpretation of recent events, Ball et al find that Okun’s Law fits the data well, with a coefficient of around -0.4. When this coefficient is inverted, it provides a rule of thumb that a one percent reduction in unemployment will occur when output grows by 2.5 percent above potential.

The use of quarterly data highlighted some important points about the dynamics of adjustment in unemployment. With lags included in the equation, the coefficient on the current output gap is smaller, needing the addition of the coefficients from previous periods to give a total coefficient similar to the annual only specification. This implies a modest delay in the adjustment of unemployment to output movements. Ball et al note that this may reveal a slight mis-specification on behalf of Okun, who thought the effects were entirely contemporaneous – that is, in the same period.

Why is this study so important?

This study is very important because it has become almost common language that the US is experiencing a ‘jobless recovery’. To a non-economist this sounds as if output is back to normal but the labour market has failed to recover. Yet, Ball et al believe this to be an incorrect conclusion. In fact, what the US is experiencing is a ‘weak recovery’.

Ball et al stress this point with the figures shown below. They show (the log of) output and unemployment from 2007 to 2011. They note that some commentators have claimed that GDP growth has returned to normal – as seen by the similar slope of the two output lines from 2009. Okun’s Law, however, is a statement about deviations from long run levels. As output is clearly below it’s long run (pre-recession) trend, then Okun’s Law would expect unemployment to be higher than it’s trend as well.

output & unemployment

Ball et al also look across countries to examine claims that Okun’s Law has broken down in these countries as well. The missing factor here, from the analysis of other commentators, is the length of recession. Ball et al note that when comparing across countries, Okun’s Law will hold if the length of recession in different countries is controlled for. Furthermore, when Ball et al adjust for country specific factors, the fit across countries is even more impressive.

Placing these results in a macro-economic context

It is a powerful result to state that Okun Law’s still holds and that what the US is experiencing is a weak recovery rather than a jobless recovery. So, if Okun’s Law is still intact and the real issue is weak output growth, a similar conclusion can be drawn to that of last week’s post. It seems, again, that it is not a dysfunctional labour market that is the cause of high unemployment, rather, it is weak output growth from inadequate demand.

Crucially, then, the focus shifts from looking for labour market failures to the causes of weak output growth. It seems that a coherent macro story involves the notion of financial recessions (see here ) and something like the leverage-deleverage cycle (see here ). A deleveraging economy that occurs during a financial recession such as this one will indeed exhibit weak output growth below trend, thus resulting in high unemployment. This, however, does not imply that weak output growth is inevitable in these types of recessions. The IMF (see here ) has already performed some excellent analysis about the need to address private sector debt and how some policies are much more successful than others in accelerating the deleveraging process.

This story contains a nice blend of new and old in macro-economics. One of its oldest laws still holds and is highly relevant. But a new understanding is emerging about the aftermath of financial crises and the need for additional theories is clear (see here for my attempts). Importantly, given the existence of new and old macro-economic behaviour, policy still matters and the appropriate recognition of the macro story will encourage the appropriate policies to be engaged.



Filed under Uncategorized

4 responses to “It’s a weak recovery, not a jobless recovery

  1. Okun law is only as good as the figures that are used to complete the puzzle….FACTS 2009 was height of recession by april of 2010 people starting exhausting UI benifits and the how the government derived these numbers made sure those people falling off the UI rolls was no longer counted as unemployed since the factor of under 2 years was used. In feb 2011 they changed the 2 years to 4…FACT IS calling 100 homes in each state in a time where cell phones are more predominant than home phone service is also another factor as to how the numbers that are derived are false. Example in 2009 we had 15 million unemployed americans…go 99 weeks out from there to 2011 still over 12 million unemployed…5 million total jobs created in 4 years….where did the first 15 million go?????? Numbers are false and the process on how they derive those numbers is not updated for todays society…SO FACT IS…there are over 25 million unemployed people and 40 million underemployed due to no choice of their own….using numbers for Okun law with numbers produced by government provides a false positive. PERIOD..

    • Hi There. Thanks for your response, it was interesting.

      I’m not sure that I agree with the gist of your post, but it is thought provoking, either way. I do trust government statistics, as I’ve seen no evidence not to. I also don’t know what you refer to in your phrase ‘a false positive’. My post, and the research it cites, clearly shows that unemployment is much higher than it’s long run level. I think most economists would agree that unemployment is an issue. I think another one of your points is embodied in the employment-population ratio, as I showed in last weeks post. A number of people have left the workforce, as shown by the steep fall in this chart. Again, I think most people would agree that this is not a good thing for the long run health of the US economy.

      Thanks again for your response,


  2. If you like the to derive your numbers from the BLS report allow me to point in the direction of interest where it somewhat proves my response. If you notice that in 2011 2 years into the height of the jobs crisis they change the CPS number by changing length of unemployment from 2 to 5 years….The fine print is anyone unemployed past 2 years prior to Feb 2011 IS NOT included. So therefor all those who lost their jobs in 2008, 2009(height of recession) and part of 2010 would not be included in those numbers. Back then our UI rates was in double digits…So although we have many thrown into the zone of U6 or otherwise known as out of the workforce. All those prior to that date are not included.

  3. Pingback: Inflation, recession and the labour market | globalmacromatters

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s