Thursday, January 9, 2014

Is Extended Unemployment Insurance Progressive or Stimulative?

This is kind of half-done, but I hope it might be interesting enough for somebody.  The idea I was working on was that, as the number of weeks of unemployment insurance (UI) increases, the policy becomes less progressive, because unemployment duration increases with age.  But, to take a stab at this, I would need more detailed data about income, by both education and age, that I don't see available anywhere.  (edit:  I did find some income data, in a PS below.)

But, I can at least get a good idea of how age relates to UI.  First, I constructed estimated unemployment durations by age, using average durations for each age group and unemployment numbers (for these graphs, I used May 2013):

For young workers, there is a lot of unemployment churn, but most workers are re-employed in a few weeks.  For older workers, there is much less churn, but workers tend to take longer to become re-employed.  This is reflected in the level and slope of the lines in this chart.

From this, we can estimate the average age of unemployed workers at any given duration of unemployment.  As can be seen in the following graph, the average age of workers who have just become unemployed is about 32 years.  But, workers who have been unemployed for more than 87 weeks average more than 40 years old.

I had previously seen research that suggested, at least for older workers, workers with more education tend to have longer unemployment durations.  But, this chart, from the Cleveland Fed, shows little difference in durations between different educational categories:

My hunch is that educational effects on unemployment duration change with age, so that young workers with more education have shorter unemployment durations and old workers with more education have longer unemployment durations.  But, I haven't found any available data to confirm that.  (edit: That pattern could explain the convergence of the durations of the education levels as baby boomers have aged, which is shown in the graph.)

In any case, the point to consider here is that, while the first few weeks of UI are certainly a progressive redistribution, as UI is extended to a longer duration, the progressivity of the distribution would decline.  If the average worker who has been unemployed for 70 or 80 weeks is about 40 years old, with above average education, it is possible that the beneficiaries at the long end of emergency UI tend to be workers who were earning above median wages.

In addition, as the age and financial condition of EUI beneficiaries increases with the extension of benefits, the supposed stimulative effects of the program would also diminish, since the benefits would be likely to replace savings.  In fact, even if we accept the notion that benefits redistributed to lower income households will benefit the economy because they will be more likely to be used in consumption, the application of that idea to EUI is suspect.  Except for the poorest recipients, I would expect most households to react to unemployment by consuming with dissaving (through debt accumulation or reductions in savings), which would be replenished or repaid with previously unplanned future earnings.  So, a good portion of EUI would simply replace savings today, and would decrease future production because of the income effect.

This problem would limit the supposed stimulative effects of EUI increasingly as it is extended to longer durations.  The assertion of the stimulative effect of EUI calls for empirical support, even if the idea is accepted on its own terms.

PS.  I did find some education and income data at the US Census Bureau, here.
I made some simplifying assumptions - that unemployment duration is the same for all education levels within each age group, and that the relative unemployment rate of each education group is the same for all age groups.  This produced a single mean income level for each age group that remained stable at all unemployment durations, so that the income level is just a product of the number of unemployed workers of each age.  I realize that this is a lot of simplification, but it probably gives a pretty good first estimate for my purposes here.

Of course, the income figure reflects the income these workers would expect to receive when they are employed.  These income figures are based on individuals, not households.

Here is the result:

The mean income (when employed) of unemployed workers rises, much as the average age rises, with duration of unemployment.  But, if my simplifications and estimates are somewhat accurate, the mean expected income of unemployed individuals begins at just over $30,000, and as unemployment duration rises toward 99 weeks, starts to level off around $40,000.

So, while the average working income of beneficiaries does appear to rise as EUI is extended to longer periods, it doesn't appear to reach the level of the mean income of the average US individual who is working.

No comments:

Post a Comment