Phantoms on the labor market

Written for

Having trouble finding a job? Blame the phantoms.

If you ever spent some time looking for a job online, you have most likely stumbled upon a listing for a position that has already been filled, and most people would agree that it is very annoying. In their recent paper, Bruno Decreuse and Arnaud Chéron call these job listings phantoms. It is not hard to imagine how phantoms directly create inefficiency in the labor market. Job seekers who apply to phantom listings lose their time by calling and employers who leave their ads online even after they have filled the advertised position lose their time by responding to job seekers and explaining them that the position has already been filled. Yet, many employers just leave their phantoms out there. The authors estimate that around 37% of job listings on Craigslist (a major job board in the US) advertise for already filled vacancies.

There are several reasons why employers may leave phantom job listings online: keeping active ads is usually free on websites such as Craigslist and deleting ads takes time and is thus costly in itself. Another reason is that some firms and state-run agencies are legally obliged to publish job listings despite already knowing who they will hire. Further, some employers may choose to keep outdated job listings as insurance until the new hire has started working. It is not easy to identify the presence of phantoms, but the authors found a way – the table below shows the average age (in days) of job listings in 23 major US cities. The maximum age of a job posting is 30 days, after which Craigslist automatically deletes it. The authors argue that if outdated ads were destroyed by employers, the density of job listings by age would be decreasing, and the quartiles of the distribution would be separated by an increasing number of days. However, as the table shows, this is not the case.

The authors then develop an extended version of the canonical continuous-time equilibrium search unemployment model by embedding a matching technology that permits the existence of phantom job listings. This model allows them to analyze the consequences of this particular market friction. They conclude that in the long run, there exists a unique steady state ensured by two mechanisms working against each other. When firms expect the matching process to be very efficient, they have strong incentive to post vacancies. This reduces the proportion of phantom job listings. Then this phantom proportion grows, and there is a reversal in the supply of vacancies, which ultimately determines the steady state.


Cheron, A., & Decreuse, B. (2016). Matching with phantoms. The Review of Economic Studies, rdw032. Available here.

Leave a Reply

Your email address will not be published. Required fields are marked *