How Far Would You Travel for Work? Disentangling the Effects of Geography, Job Search and Employment

After the financial crisis, the number of unemployed workers in the United States reached nearly 15 million in 2009 and the unemployment rate rose almost 10 percent in the same period. The slow recovery of the economy, and particularly of employment, has led to debate over the reasons for this trend and the appropriate policy responses. One hypothesis for explaining this behavior is that the recession caused a mismatch between job vacancies and unemployed workers. Thus, there is a misalignment between the location of the vacancies and the geographic preferences of those who are unemployed.  A new study from Ioana Marinescu and Roland Rathelot discusses whether the geographic mismatch hypothesis is useful in understanding the high level of unemployment during the most recent economic recession.

The authors take advantage of a large and unique dataset to measure existing unemployment mismatch, which avoids the commonly used mismatching indices that impose constraints regarding the level of geographic mismatch. The authors construct a merged dataset from one of the largest U.S. employment websites, CareerBuilder.com. First, they obtained a random sample of registered unemployed users with active accounts between April and June 2012. Next, the authors took a sample of vacancies published on the website during the same period. They then merged the previous dataset in order to determine the jobs each candidate applied to. For job seekers (registered users) and the job openings (job postings) the authors have the location at the ZIP-code level.

Controlling for fixed effects, the results reveal that job seekers are less likely to apply to job vacancies farther away from their ZIP code. In particular, they found that a job seeker is 35 percent less likely to apply to a vacancy that is 10 miles away than to a vacancy that is in the job seekers’ ZIP code. Additionally, the authors proposed a new mismatch index to measure the degree of geographical mismatch based on a search and matching model of the labor market. This new index has less restrictive assumptions and allows job seekers to apply for jobs in every location, whereas previous measures  assumed that applicants apply to jobs near where they live. Using their approach to a mismatch index, the authors find that if job seekers were relocated to areas with job vacancies, aggregate unemployment would be reduced by 5.3 percent at most. Considering these results, the authors conclude that geographic mismatch is not a major contributor to U.S. unemployment.

The findings have significant policy implications. The theory of mismatch has generated debate over how to reduce the unemployment rate. Some policymakers suggest that government should encourage companies to set up factories and stores in areas with high unemployment, while others suggest that helping workers move to areas where there are job openings might have a more significant impact. Evidence from this particular study suggests the latter might only have a minor effect on aggregate unemployment. Even though the evidence shows geographical mismatch may not have a significant impact on unemployment, other types of mismatch in the labor market–such as mismatch in skills and working conditions–are issues that both workers and employers struggle with and should be investigated further.

Article source:  Marinescu, Ioana and Roland Rathelot. “Mismatch Unemployment and the Geography of Job Search.” National Bureau of Economic Research Working Papers Series: No. 22672, September 2016.

Featured photo: cc/(rvlsoft, photo ID: 74754955, from iStock by Getty Images)