Benjamin and Kochin (1979) argue that “search unemployment” characterizes the high levels of unemployment in the Interwar period. Explain how they reach this conclusion and assess whether they are correct.
Benjamin and Kochin’s 1979 work was an attempt to explain the high turnover and hence the documented high unemployment levels between 1920 and 1938 through a supply-side effect, rather than the demand-side effects that had dominated discussion of the unemployment trends of this era. They claimed that the workers of this period were individually opting out of supplying labour in order to find other, higher-paying jobs, hence the term “search unemployment”. They claimed that this search unemployment was funded by the increased level of benefits and unemployment insurance that was available to a wider spectrum of workers in this period. Their famous quote from their article over this topic is that “the army of the unemployed standing watch in Britain at the time of the publication of the General Theory was largely a volunteer army”.
Over this period, the nature of unemployment benefit changed severely. The first time it was really introduced was in 1911, when the National Insurance Act offered an unemployment benefit of 7 shillings for the first 15 weeks of unemployment, but only 2.25 million workers were covered by this scheme mainly as it was restricted to certain trades, such as dock workers. However, this was expanded greatly in 1920 and 1921 when the level of people covered by the scheme rose to 12 million and unemployment benefits rose to 15 shillings, plus 5 shillings if they had a wife and 1 shilling for every extra child. This was mainly as a reaction to the unemployment seen after World War I, but Benjamin and Kochin’s argument is that this accentuated the issue. Benefits then rose again in 1934 (after a short fall in 1931, which also introduced means-tested benefits) and the scheme coverage expanded further.
Their methodology for reaching this conclusion was firstly by running a regression. Their estimate over the period from 1920 to 1938 was that the unemployment rate was given by , where U is the unemployment rate, B/W is the replacement ratio (which is given by the level of benefits available at the average full time wage divided by the average full time wage), Q is the current output level and Q* is the long term trend level of output. They then calculated the replacement ratio over the period using pre-existing data on wage levels and historical data on benefits. The figure that they calculated came to 0.49 on average over the Interwar period with a high of 0.57 in 1936. They thus claimed that the rise in benefits, B, create a rise in the unemployment rate due to this endeavour to move into search unemployment. They also contend that B/W did not even need to be equal or greater than 1 (i.e. benefits pays the same or more than wages) for the worker to move into search unemployment, as the workers will be willing to give up some income for the other benefits of unemployment, namely increased leisure time. This, coupled with the fact that, after the initial six day registration period, unemployment benefits could be claimed in the very short-term (sometimes for only one day) and could be extended pretty much indefinitely, meant that people were much happier to live on benefits and take longer over selecting employment, rather than moving straight into a job.
Overall, Benjamin and Kochin’s conclusion was that in a counter-factual universe when benefits had stayed at the same level as it did before World War I, when the replacement ratio was about 0.27, then the unemployment rate would have been a third lower than it actually was. They also drew a graph showing the actual rate of unemployment over time on the same graph as their own upper and lower estimates of unemployment if benefits had not risen. This showed roughly the same trend for both lines, but both the upper and lower limits were a lot lower than the actual rate at all times. Taking the average of the upper and lower limits, the rise from the lowest level of unemployment in 1920 to the highest level in 1933, the counter-factual unemployment rise is only 13.2 percentage points, from 3.9% to 17.1%, when the true unemployment rise is 17.2 percentage points, from 3.9% to 22.1%. They find that the unemployment rate would have been one third lower over the periods of high unemployment (i.e. 1927-1929 and 1936-1938) if the unemployment rate was the same as the pre-World War I period.