Barriers to Search and Hiring in Urban Labour Markets

In this final section, we discuss four important points on the returns to search interventions.

First, search interventions tend to be very cheap and hence the marginal return to public funds spent on successful search interventions is very high. The following comparison is instructive. Kreft (2023) finds that the average treatment effect of a certification intervention is a 3 percentage point increase in employment. For comparison, the average employment treatment effect of youth vocational training is a 6.5 percentage point increase in employment (Ghisletta et al. 2021). At the same time, the certification or subsidy interventions reported in McKenzie (2017) have costs ranging from 7.5 USD to 25 USD per person treated (with the exception of one study in Jordan with costs of 203 USD), while the training interventions have costs ranging from 330 USD to 1722 USD per person treated (with the exception of one study in India that only costed 13 USD per treated respondent). In other words, if we exclude the two outliers, the most expensive search intervention is more than ten times cheaper than the least expensive training intervention. Prima facie, the marginal return to spending public funds in a certification intervention is thus vastly greater than the return to training. 

Second, the most important caveat to this conclusion is that we have very limited evidence to quantify the potential negative employment impacts of interventions on untreated individuals – the so-called displacement effects. Displacement effects would arise if treated and untreated individuals compete for the same pool of vacancies and, prior to the intervention, most vacancies are filled with a recruit. It is plausible that, at least in the short run, displacement effects can be large, since anecdotally firms in low and lower-middle income countries do not appear to have a large number of unfilled vacancies. Displacement effects can also limit the effectiveness of training interventions. Whether training interventions and search interventions have different displacement effects remains unclear. If the key short-run constraint is the presence of unfilled vacancies, it is plausible that the displacement effects of these two types of interventions may be similar. More research on these issues is urgently needed.

Similarly, it would be important to collect more systematic evidence on whether these interventions generate productivity gains for firms (e.g. gains arising from a better allocation of talent). These gains are at the core of the economic argument for having search interventions, but the evidence on their magnitude is unfortunately limited. 

Third, we also have close to no evidence to assess the general equilibrium (GE) impacts of these interventions when they are offered at scale.[1]1. Field et al. (2023) use data from a job search and matching platform in Pakistan to show that increases in application volumes do not generate negative spillovers on interview invitations, although they do not observe employment.  One plausible GE impact of search (and training) interventions is an increase in labour demand, since firms cite the inability to find and identify talent as a key constraint to their growth. But there are at least two types of negative GE impacts. First, interventions that increase search can generate congestion and raise average search costs. Second, interventions that generate signals about jobseeker skills can harm low-skilled workers – an effect that would only emerge when skill certificates become expected by employers and hence low-skilled workers are forced to reveal poor skill assessment results. We flag the area of the GE impacts of search interventions as a key area for research.

Finally, it is important to note that there exists an equity rationale for offering search interventions, since they are designed to support individuals who would otherwise face search barriers (e.g. liquidity constraints or lack of credentials) and have weaker labour market outcomes as a result of this. There is indeed initial evidence suggesting these interventions are most effective for groups with weaker expected labour market outcomes. This raises the question of whether there would be gains from targeting these interventions more effectively. Caria et al. (2023) estimate some gains from targeting search interventions in the context of an adaptive experiment in Jordan, but much more research is needed on this topic.

References

Caria, A S, G Gordon, M Kasy, S Quinn, S O Shami, and A Teytelboym (2023), “An adaptive targeted field experiment: Job search assistance for refugees in Jordan,” Journal of European Economic Association, jvad067.

Field, E, R Garlick, N Subramanian, and K Vyborny (2023), “Why don’t jobseekers Search More? Barriers and Returns to Search on a Job Matching Platform,” Working Paper.

Ghisletta, A, J Stöterau, and J Kemper (2021), “The Impact of Vocational Training Interventions on Youth Labor Market Outcomes: A Meta-Analysis,” Working Paper.

Kreft, B (2023), “Information, skills and job search: A Bayesian hierarchical analysis of the impacts of reducing information frictions for unskilled jobseekers in developing countries,” Manuscript, University of Oxford.

McKenzie, D (2017), “How Effective Are Active Labor Market Policies in Developing Countries? A Critical Review of Recent Evidence,” The World Bank Research Observer, 32(2): 127–154.