I got lots of very helpful responses to my recent post on the search theory of unemployment, here and at Crooked Timber. But it has occurred to me that I haven’t seen any answer to one crucial question: How many offers do unemployed workers receive and decline before taking a new job, or leaving the labour market? This is crucial, both in simple versions of search theory and in more sophisticated directed search and matching models. If workers don’t get any offers, it doesn’t matter what their reservation wage is, or what their judgement of the state of the market. Casual observation and my very limited experience, combined with my understanding of the unemployment benefit rules, is that very few unemployed workers receive and decline job offers, except perhaps for temporary work where the loss of benefits outweighs potential earnings. Presumably someone must have studied this, but my Google skills aren’t up to finding anything useful.
And, on a morbidly humorous note, it’s a sad day for conservative politicians when efforts to bash the unemployed actually cost them support. But that seems to be the case for the LNP government in Australia with their latest plans, both expanded work for the dole and the requirement for 40 job applications a month. I’ll leave it to Andrew Leigh to take out the trash on work for the dole (BTW, his new book, The Economics of Almost Everything is out now).
The 40 applications requirement has already been the subject of some amusing calculations. I want to take a slightly different tack. Suppose (to make the math simple) that the average job vacancy lasts a month. There are roughly five unemployed workers for every vacancy, so meeting the target will require an average of 200 applications per vacancy. The government will be checking for spam, so lets suppose that all (or a substantial proportion) of the applicants take some time to talk about how they would be a good fit with the employer and so on. Dealing with all these applications would be a mammoth task. One option would be to pick a short list at random. But, there’s a simpler option. In addition to the 200 required applications from unemployed people, most job vacancies will attract applications from people in jobs. A few of them may be looking for an outside offer to improve their bargaining position with their current employer (this is a big deal for academics), but most can be assumed to be serious about taking the job and in the judgement that they have a reasonable chance of getting it. So, the obvious strategy is to discard all the applications except for those from people who already have jobs. What if there aren’t any of these? Given that formal applications are going to be uninformative, employers may pick interviewees at random or may resort to the informal networks through which many jobs are filled already.
Trying to relate this back to theory, the effect of a requirement like this is to negate the benefits of improved matching that ought to arise from Internet search. By providing strong incentives to provide a convincing appearance of looking for jobs for which workers are actually poorly suited, the policy harms both employers and unemployed workers who would be well suited to a given job.
Update I found the following quote widely reproduced on the web
On average, 1,000 individuals will see a job post, 200 will begin the application process, 100 will complete the application,
75 of those 100 resumes will be screened out by the Applicant Tracking System (ATS) software the company uses,
25 resumes will be seen by the hiring manager, 4 to 6 will be invited for an interview, 1 to 3 of them will be invited back for final interview, 1 will be offered that job and 80 percent of those receiving an offer will accept it.
Data courtesy of Talent Function Group LLC
Visiting the TFG website, I couldn’t find any obvious source. The numbers sound plausible to me, and obviously to those who have cited them. But, if the final number (80 per cent acceptance) is correct, then it seems as if the search theory of unemployment is utterly baseless. Assuming independence, the proportion of searchers who reject even three offers must be minuscule (less than 1 per cent).