Poverty & Inequality

DUET: Towards employment as a universal right

  • Blog Post Date 11 September, 2020
  • Print Page
Author Image

Debraj Ray

New York University


Debraj Ray contends that we should push forward with the agenda of employment as a universal right, and DUET would move that needle. He discusses two aspects of the proposal – keeping track of workers, and keeping track of projects.


Jean Drèze has recently proposed a "Decentralised Urban Employment and Training" Scheme, or DUET for short. In his words, "DUET could act as a step towards urban employment guarantee." The essential idea is for state governments to issue job stamps to “approved public institutions”, who would use these to pay wages to suitably registered workers. Workers would present job stamps and a work certificate (from the institution) for reimbursement directly into a bank account. I refer you to Jean's proposal for more detail.

I personally believe that we should push forward with the agenda of employment as a universal right, one that transcends rural versus urban. DUET would move that needle. With the rising tendency to displace labour – a trend that will only be heightened by the pandemic – such a right provides protection, at least to some minimal degree. So I support this proposal.

That said, there are many issues to be discussed, among them the funding of such a programme, its impact on rural-urban migratory balance in India, and the overall philosophy of employment as a right. I have views on these matters but leave these to my colleagues in this symposium. My comments pertain to two aspects. The first has to do with keeping track of workers. The second has to do with keeping track of projects.

The proposal requires workers to be drawn from a (presumably) local pool of registered workers. This takes me back to a long-held fascination with the framing of Indian policy, which often – if not invariably – involves a myriad regulations, some of which appear to have no real underlying objective, and most of which can either be gamed or create insuperable difficulties. Take this business of worker registration. It will invariably raise (for instance) the question of where millions of seasonal migrants really live. Presumably, the spectre of them double-dipping into MNREGA (Mahatma Gandhi National Rural Employment Guarantee Act) and DUET, will raise its churlish head. We will then see bureaucrats making arbitrary decisions, horrendous quantities of paperwork to be filed, bonafide applications denied, and suspect applications accepted for political reasons.

If this were up to me, I would say: let India be India. Let a citizen of India work wherever s/he wants. Yes, to obtain income from the government, you need to register, as MNREGA job card holders now do, and a bank account has to be either provided or opened. You would then be permitted x days of government work anywhere in India; you, not your household. Say x is 100 days per person. If 40 of those 100 days get used up in a DUET project in Bangalore, then the remaining quota is 60, which – for all I care – can be spent on MNREGA or another DUET project in Kolkata. Of course that would not happen for the simple and sane reason that no one will want to shuttle between Bangalore and Kolkata to work on DUET projects. But let there not be a silly law – which will require all sorts of paper checks on local residence and accompanying scams and audits – that prohibits it.

If we want to nevertheless maintain the distinction between employment as a rural right and not as an urban right, then a MNREGA card can be given India-wide rights, while a DUET card holder can only be given city rights. A simple bank account with a yearly cap on government payments can easily monitor the rest. That would also take care of seasonal migrants who work in both village and city. However, that said, I would follow MS Dhoni and “keep it simple”.

The above arguments rely on the constraint that no more than a single bank account should be attached to any job card holder. Otherwise it is impossible to verify the yearly cap. This may require a link to the Aadhaar card, a potential bone of contention.

The second issue concerns projects. Groups of workers could still collude with organisations or employers, and pocket job stamps without doing any work, while the ‘employers’ pocket cash payments from the workers. Jean mentions "approved institutions – schools, colleges, government departments, health centres, municipalities, neighborhood associations, urban local bodies, etc." He also mentions "placement agencies" that will oversee the projects and presumably weed out scams. Sadly, it is turtles all the way down. I can see each of these ‘august’ bodies – or the relevant functionaries in these bodies – handing handing out job stamps in exchange for cash payments. A net injection of funds – as DUET surely calls for – spawns these financial opportunities by definition.

While I agree that the use of a placement agency could weaken collusive bonds, I have enough faith in the ingenuity of all human beings that I would not bet on it.

This is a hard problem. Collusion could be avoided by a system of ex-post project approval, but this will weigh heavily on honest employers. While approval is given after the project is completed, workers would have to be paid up front: they have gone home already (or should have) with their stamps. Then, must honest employers take on the risk of not having their project approved, perhaps by some government functionary who is holding them up? Who pays the workers then?

Some of my colleagues might feel that this is a second-order problem. Well, the graveyards of Indian public policy are littered with such second-order problems. The least we can do is try and avoid adding more skeletons. Here is a two-pronged suggestion:

Randomly audit projects using an oversight board, with stiff fines for fake projects – that much is obvious – but I suggest we reward the auditor with a proportion of those fines.

Yes, of course, that immediately raises the problem of harassment – that genuine projects will be touted as fake by auditors. There must be guidelines for the documentation of satisfactory project completion that cannot be contested by auditors: for example, photographic proof of trash removal, construction jobs, whitewashing of buildings, and so on. This creates a restriction, alas, on the kinds of projects that can be taken up. Specifically, the projects will have to be relatively long on measurable physical output and relatively short on hard-to-measure services. That could be a problem, but less of a problem than relying on the goodwill of some oversight authority.

In this context, I end with a comment on subsidies to existing business. Imagine awarding job stamps to existing employers – perhaps non-profits – to subsidise a fraction of labour costs. As Jean observes, this could have a lower employment impact in the net – that would depend on how much capital-labour substitution a firm is willing to do, while at the same time a given budget of stamps goes further.

All that is worth thinking about, but the point I want to make is that the fake-worker problem is attenuated with cost-sharing. Say the subsidy rate is 50%; then an employer with one employee could effectively get a second for free. If that second employee has zero marginal product, then the scheme is subject to the same level of gaming as before. But if the employee has a positive marginal product, the incentives for gaming substantially decline. In this sense, when costs are shared on existing business activities and are not fully borne by job stamps, there is a lower overall incentive to game the system. The potential disadvantages are: the output of that business may not be what we want to promote, and the net employment gain could be lower.

I hope these comments are viewed as constructive and not at all negative, because this DUET is worth playing.

No comments yet
Join the conversation
Captcha Captcha Reload

Comments will be held for moderation. Your contact information will not be made public.

Related content

Sign up to our newsletter