14 thoughts on “Recoveries and unemployment”

1. Arnold, to agree or disagree with your statements in the last 3 paragraphs, I’d like to know your opinion on the statistical methods Hall&Kudlyak used to reach their conclusions.

   I have yet to read the paper and I can’t assess their method by comparing it with other statistical methods that have been used to analyze unemployment. Indeed, it’s surprising that their method could have supported the second conclusion. It’s funny because if we were to accept it, as you point out, we could argue the irrelevance of all government intervention (a new Lucas critique). Is their method good enough to argue that all occasional sharp upward or downward movement in any variable in time of crisis (for example, the much larger number of deaths at the beginning of the pandemic) may be followed by an inexorable glide until the variable reaches its normal value?

   • Arnold, Hall&Kudlyak circulated two other papers this year, complementing the one you refer to. They are NBER WPs 27234 and 27886. These are the abstracts:

     WP 27234
     It is a remarkable fact about the historical US business cycle that, after unemployment reached its peak in a recession, and a recovery began, the annual reduction in the unemployment rate was stable at around 0.55 percentage points per year. The economy seems to have had an irresistible force toward restoring full employment. There was high variation in monetary and fiscal policy, and in productivity and labor-force growth, but little variation in the rate of decline of unemployment. We explore models of the labor market’s self-recovery that imply gradual working off of unemployment following a recession shock. These models explain why the recovery of market-wide unemployment is so much slower than the rate at which individual unemployed workers find new jobs. The reasons include the fact that the path that individual job-losers follow back to stable employment often includes several brief interim jobs, sometimes separated by time out of the labor force. We show that the evolution of the labor market involves more than the direct effect of persistent unemployment of job-losers from the recession shock— unemployment during the recovery is elevated for people who did not lose jobs during the recession.

     WP 27886
     Potential workers are classified as unemployed if they desire to work but are not working. The unemployed population contains two groups–those with jobs and those without jobs. Those with jobs are on furlough or temporary layoff. They wait out periods of non-work with the understanding that their jobs still exist and they will be recalled. We show that the resulting recall unemployment dissipates quickly following a shock. Those whose jobs no longer exist constitute what we call jobless-unemployment. Shocks that elevate jobless-unemployment have much more persistent effects. The unemployed without jobs often circle through short-term jobs, spells of unemployment. and spells out of the labor force, before finding stable employment. Historical major adverse shocks, such as the financial crisis in 2008, created mostly jobless unemployment and consequently caused extended periods of elevated unemployment. The pandemic starting in March 2020 created a large volume of recall-unemployment, most of which dissipated by August. It also created a bulge in jobless-unemployment, which is lingering.

2. It’s worth remembering that this statistical model explains the economy’s behavior under a specific set of circumstances, and that changing the circumstances (e.g. getting rid of stimulus) may result in different statistical behavior.

   • You are right if the statistical analysis was intended to find regularities but not to seek evidence for a theory. It’s important to understand the statistical methods and know the purpose of their use.

     • I have just found an example. Read

       https://pjmedia.com/instapundit/414007/

     • Even if you’re seeking evidence for a theory, the context of the evidence sets limits on how confident you can be. For example, if the data used was pre-Covid and there was nothing like Covid in the dataset, you don’t have much evidence that your theory will hold throughout 2020. Other out-of-context events (war with China, supervolcano eruption, alien invasion) may also change the underlying dynamics.

       If you have multiple studies showing that the model holds under a wide variety of circumstances, then I get much less skeptical of your model. OTOH, my training was in chemistry so my expectations for data-to-model fit may be unrealistic in economics.

3. Does the explanation still work in terms of changes in wages or for the employment to population ratio?

4. “Instead, the entrepreneurial trial and error that is needed to create new patterns of specialization and trade seems to take place at a persistent, steady rate.”

   Hmmm….seems quite a stretch. Even assuming that entrepreneurial activity is not influenced by changing incentives, taxes, regulation, available capital, demographic changes and aging, available wages, available labor supply, random technological changes, cost of energy, foreign industrial and economic planning and subsidies etc., there is no reason why we would assume employment level changes are controlled by entrepreneurial activity. Consider that Foreign direct investment in the USA bounces around from year to year but generally is on an upward path. Consider government debt has been growing steadily as has the share of the workforce in government. From 1960 to 1990, the number of state and local government employees increased from 6.4 million to 15.2 million.

   But indices of actual entrepreneurial activity don’t show steady rate . See
   page 10 at: https://www.kauffman.org/wp-content/uploads/2019/09/2017_Kauffman_Index_Startup_Activity_National_Report_Final.pdf

   Cosmopolitan indifference to the need for improved domestic competitiveness while maintaining the status quo for the elites in tax exempt industries is no way to go forward.

5. So let’s imagine for a moment that there are a number of things that take people ‘out of the market’ and that those are accounted for (school age children, say) and that everybody else who isn’t employed is ‘unemployed.’ It makes sense that there’s a jiggling kind of process (like annealing) for them to find their place in the global, national, and local market; and the better integrated they become with the local market, the higher the cost of new dislocation becomes to them and their primary partners. They gain tacit knowledge, interpersonal and organization specific skills, etc. As a result, they should anticipate sharing the benefits with progressively higher productivity and wages. A personal dislocation would hurt them and their partners individually; a global dislocation hurts many people at once, erases signals of fit, makes signals unreliable, and so on.

6. I’ll answer your rhetorical question:

   Stimulus protects the existing wealth distribution. Its beneficiaries are the powerful, people with access to now near infinite leverage. We are all paying for it with seemingly permanent low real rates.

7. What if macro level policies in the last contraction had not undermined residential investment? Then 2 million construction workers wouldn’t have needed to depend on PSST adjustments and they could have gone back to building homes.

   • So fewer jobs would have been lost, but the rate at which jobs came back would have been the same, at least if the results of the paper apply

     • Yes.

8. Just curious: Isn’t the amount of stimulus intertwined with monetary policy, since the latter is going to respond differently depending on both the amount of the stimulus and its observable success?

Comments are closed.