There are numerous criticisms of the theory that Benjamin and Kochin have developed. The first comes from Smyth, who claims that the theory is far too ad hoc and does not measure unemployment rate in a rigorous and controlled way. His main issue is that the regression which has been run by Benjamin and Kochin include both a labour demand factor (i.e. Y and Y*) and a labour supply factor (B/W). This means that it is not measuring unemployment solely from one direction, such as from a labour supply-side view or from a labour-demand side view, but rather a view that has been seemingly cobbled together from both approaches. This is known as an “identification problem” Hatton also wrote a paper which debated whether there was accurate evidence in historical records for the levels of search unemployment which Benjamin and Kochin are proposing here. If we were to graph the unemployment level against the number of vacancies, we would expect to see a rise in the number of vacancies if there was a rise in the unemployment level, as people left their jobs in order to find new ones whilst making use of the high level of benefits. However, Hatton’s historical research instead found that there was little change in the number of vacancies when the unemployment level rose, which implies that the jobs were not available after the previous incumbents left, so the job position must have been closed. This was also contested by Eichengreen, who used data from the New Survey of Life and Labour to find that rather than 50% of the unemployed being caused by search unemployment, the true figure is closer to 20% and so the effects of benefits on unemployment is minimal. This was because Eichengreen found that it was mainly secondary workers (i.e. women and children who were trading the work for the increased benefits), so the overall unemployment levels were not particularly affected. However, this data only came from Greater London, which is not where the focus of the unemployment levels was. Much higher levels of unemployment were seen in the North and East, so it is possible that there are different relationships there as well which may aggregate to Benjamin and Kochin’s model.
Ornerod and Worswick’s objection to Benjamin and Kochin’s model is that it is not still applicable if small changes to the time periods and parameters are made. For example, when 1920 is removed from their model, the relationship breaks down. This may be because there was a large jump in benefits which were available to the public after the National Insurance Act was extended to 12 million people, rather than 2.25 million. Output and unemployment also decreased in the same year, due mostly to the return to gold and the ensuing contractionary monetary policy. However, it is fairly certain than these two events happened independently of each other and so using them in a causal relationship is incorrect. This means that this relationship does not exist to a large extent and that it must be other factors, such as the demand-side effect of the return to gold and its subsequent hysteresis, which are the primary causes of changes in unemployment levels.
Collins claims that the relationship also does not hold when you break the economy down on an industry by industry basis. When the data for each sector is inputted individually, the figure generated for the replacement ratio is statistically insignificant for nine of the sectors, particularly those who have high levels of unemployment and thus would affect the unemployment rate term in the equation greatly, such as shipbuilding.
Howson and Hatton believe that the rise in benefits only seems to create a rise in unemployment because the unemployment is measured by those claiming for benefits. However, this is not an accurate measure of the changes in unemployment. For example, the rise in benefits will also persuade people without jobs who currently do not see the point of claiming or are too proud to claim for benefits to now start to register as unemployed in order to receive the rise in benefits. Therefore, the rise in unemployment seen is not a true rise in unemployment but mostly a change in the registration rate.
This system would seem to imply a failure on the part of the Government in dealing with the number of people on the benefits scheme. As it was intended to support people until they have a job, if it was being used for people just to switch jobs because they could, this would seem like a failure. This would therefore imply an issue in the administration and registration for benefits, even though historical evidence shows that this is not true. For example, the number of rejections for benefits per 100 applicants increased over this period as the administrators improved at finding false applications.