Wednesday, May 16, 2018

A Nudge carried too far?

I am a strong believer in the use of nudges in development. They are cute and simple. But the same cuteness and simplicity can blind us to the relevance of the specific nudge (in the specific context) as a serious enough development tool.

Martin Abel and Co have a paper which examines a nudge to increase the effectiveness of job search efforts by unemployed youth in S Africa. The youth are grouped with peers, encouraged to make daily job search plans, and are sent weekly SMS reminders. The youth in the treatment group (in the RCT evaluation) had 30% more job offers and were 27% more likely to get a job.

This is undoubtedly a really cute nudge, a useful addition to the body of knowledge on nudges in development. But is this something which would excites policy makers? Very very unlikely. Let me illustrate. 

The hypothesis is that nudging job seekers into being more effective with their job searches will increase their likelihood of getting jobs - more active searches, more offers, and greater likelihood of a placement. Given all these jobs will anyways be filled, the intervention will only add one more person (the one treated with this intervention) into the long line of active job seekers. We cannot claim any social return on investment (SROI) since for each person who is nudged ahead of his competitors to get the job, another equally deserving person (who ironically was enterprising on his own with the job search and did not need the nudge to be active with job search) is denied the job. 

In other words, the net partial equilibrium social impact of the intervention is virtually zero. 

In the general equilibrium, one can of course say it has contributed its tiny bit to improving the efficiency of labour market matching (for e.g., assuming the treated person was more capable than the one who would have otherwise got the job).

But should a deeply capacity constrained government get into such things, with the certainty of this activity displacing effort from something else more likely more important?

And we have not even talked about the state capacity challenge in implementing this with acceptable fidelity in these countries. Take this description of the treatment,
Study participants are randomly assigned to one of three treatment arms: Control (pure control), Workshop (workshop only), and Workshop Plus (workshop + plan making). A random subset of job seekers in the Workshop Plus group were additionally asked to identify a peer. Participants across the Workshop and Workshop Plus groups were further randomized to receive text-message reminders. 
This is not as simple as it appears. Making plans, weekly reminders, peer-support engagements etc in scale requires reasonably strong state capacity. It is inconceivable that this can be done in scale without being routinised by the implementing bureacracy to a degree that makes it ineffectual. Do we really think that the S African State (in any province) can do this in scale without significantly compromising the quality of implementation as to render it virtually ineffectual?

We are also talking about countries where for every decent job opening that becomes available, there are potentially thousands go job seekers. It is very unlikely that a job will remain unfilled because of lack of applicants or lack of awareness among job seekers (in any case with such cases the more appropriate intervention, by orders of magnitude, is to support placement agencies). For sure, some individual job seekers will be better off if they can be nudged into being more effective at jobs search (though at the cost of making some others, who would have got the job without this intervention, worse-off). 

The best that can be said about this is that, IF implemented in scale, it can very marginally (or more appropriately trivially, considering all the factors that constrain S African labour market and placing this intervention in perspective) increase the efficiency of S African labour market. And compare this with the competing areas for policy attention on the labour market side - education of kids, trainings to make the educated kids employable, facilitate job creation by businesses, enhancing productivity of those jobs, mechanism for efficient matching job seekers with jobs and so on.

No comments: