We make space for so many vanity projects, UBI is not impossible to implement: Economist Abhijit Banerjee

Abhijit Banerjee, a professor of economics at the Massachusetts Institute of Technology, is advising the Congress on the feasibility of its proposed minimum-income guarantee scheme. Saumya Khandelwal/Hindustan Times/Getty Images
11 March, 2019

On 28 January, just days before the Bharatiya Janata Party-led central government announced its interim budget, the Congress president Rahul Gandhi declared that if voted to power, his party would implement a minimum-income guarantee scheme for the poor, “to help eradicate poverty and hunger.” The scheme would entail guaranteed cash-transfers to specific sections of India’s population and is an adaptation of Universal Basic Income, or UBI. Modern economics defines UBI as “periodic cash payment, unconditionally delivered to all on an individual basis, without work requirement.”

As state policy, UBI first gained traction in India after Arvind Subramanian, the former chief economic adviser, mooted the idea of a quasi-UBI in his Economic Survey of 2016-2017—an annual flagship document compiled under the aegis of the ministry of finance that provides detailed statistical data and analysis of the country’s economic performance and outlook. Since then, a number of states, including Odisha and Telangana, have announced variants of UBI in the form of direct cash-transfers to poor farmers, while Sikkim has proposed introducing a universal income guarantee by 2022. In February this year, the central government rolled out a similar scheme for small and marginal farmers, which promised Rs 6,000 per annum in three instalments.

Recently, Thomas Piketty, a French economist, confirmed that he and Abhijit Banerjee, a Ford Foundation professor of economics at the Massachusetts Institute of Technology will be advising the Congress on its scheme. Banerjee and Piketty are also closely associated with the World Inequality Database, or WID, an open-source database on the historical evolution of income and wealth distribution.

Rohit Inani, a journalist based in Delhi, interviewed Banerjee. They discussed the feasibility of minimum-income guarantees in India, including the fiscal and social challenges of implementing the scheme. For the scheme to work in the long term, Banerjee said, “We will have to re-capitalise the banks, so that they stop being zombies, and start doing their job.” He continued, “The question that the political system needs to answer is—would you rather have some poor people left out or would you have some poor get more money?”

Rohit Inani: You and Thomas Piketty are advising the Congress on its minimum-income guarantee scheme. How did this come about and how closely are you engaging with the party on this?
Abhijit Banerjee: Just a day after Rahul Gandhi announced this scheme, Praveen Chakravarty, [the head of the Congress's data-analytics department] who is an old friend, emailed me saying, “Look, this is the announcement. Can you help us get a sense of the magnitude of this commitment?” We were not involved in their planning. The WID is set up for exactly this purpose, which is to provide information to whoever needs it to deal with inequality, poverty. And Thomas Piketty and I are involved with [the WID] together, and we provided them with our guess on what the numbers should be. The Congress is seriously pursuing this idea, asking questions about its feasibility, et cetera.

RI: Arvind Subramanian proposed the idea of a quasi-universal basic income that would guarantee Rs 7,260 per year to 75 percent of rural households. Have you suggested a similar number to the Congress?
AB: Our numbers are not based on that. What we were doing was taking the Congress’s number and suggesting what fraction of the GDP [gross domestic product] would be required for such a scheme. I have my views on this, but this is not what we were asked.

The main point here is that we should start with a smaller number and scale up. First, we should establish the credibility of the delivery system, and if it really is efficient then we can use it to phase out a bunch of these costly and distortionary subsidies that we have. The goal is to reduce the number of schemes over time and replace them with a transfer, with one efficiently run system. So, my personal view is that on day one, the goal would be to have a significant amount of money with people. It could be Rs 1,200 a month, Rs 1,500 or even Rs 1,800.

RI: UBI, as an idea, has been heavily backed by Silicon Valley and also has considerable support among European socialists. But it has never been tried in the developing world on the scale that is being proposed in India. We have very little evidence of its effectiveness.
AB: I disagree. If you think of PROGRESA [a conditional cash-transfer scheme in Mexico], then I would say it is kind of the size of the program it would be in India. Even Brazil’s Bolsa Familia [a direct cash-transfer scheme for the poor]. These are conditional cash-transfer schemes—with eligibility conditions such as primary education—but the conditions are not incredibly onerous; they are pretty light. In fact, the Congress is talking about something that is targeted to the poor, a minimum-income guarantee, and this is something that many countries deliver effectively in different forms, including some African countries. So, it is not like we are going to be the first country to adopt an income transfer to the poor on a large scale.

RI: Apart from the Congress’ proposal, the BJP’s recent interim budget provides for what can be called a targeted income-transfer to the poor. From the evidence we have, which is so far inconclusive, would a targeted scheme be as effective as Subramanian’s proposed quasi-UBI?AB: These are not similar schemes. What Arvind was arguing for is a universal scheme, which is very different. He said that do not target, do not try to find who lives on less than Rs 1,000 a month or who has less than two hectares of land. These are complicated measurement issues and there is a trade-off here; either you can give more money to fewer people or you can give less money to more people, but in a more efficient way. Arvind went for less money to more people.

The whole point of a universal basic income is that it would be targeted, at most, on gender—it is not targeted on anything else. There is evidence from our research—where we carried out lot of targeting in Indonesia and other places—that targeting is often very inefficient. You exclude a lot of poor people and include a lot of non-poor. So, one of the things the Economic Survey does—and I was quite involved in conversations about it at the time with Arvind—was to say that let us not bother targeting. But the question is, is it so bad that we give up on targeting?

RI: Have you taken up the possible inefficiency of such a targeted cash-transfer scheme with the Congress?
AB: I have told them that targeting is going to be a problem. Which is not to say that I think it is impossible to implement. I think they have made a choice and it is based on the idea that they would rather give more money to less people.

The question that the political system needs to answer is—would you rather have some poor people left out or would you have some poor get more money? It is a decision based on the cost of such a scheme. And it is for political judgement. You have to go with the less-worse option here.

RI: Even if it is a universal or a targeted income scheme, how would such a move be effective in the long term? Some experts have cautioned that it could lead to a distortion in the labour markets in the long term.
AB: We [at MIT] conducted seven randomised control trials [experiments to weed out biases in conclusions] on government-run cash-transfer programmes in six countries, where we handed cash to people to see if they would stop working. We published our findings on this myth of the “lazy welfare recipient.” We found no evidence for such behaviour.

In India, people are generally aspirational. Take the case of people who are involved in manual scavenging to earn a living. I would rather have the government invest in machine-cleaning than have people humiliated in what they have to do to earn their living. I am okay with certain labour market changes, but there is no evidence to claim that it would lead to a large-scale withdrawal of labour from the market.

RI: In one of your papers, you have explored the possibilities of UBI also functioning as a pro-growth policy in the long term. Do you think this is possible?
AB: We [at GiveDirectly, a not-for-profit organisation] are conducting a very large-scale trial in Kenya on this and we will be getting the first set of results in a few months. At this point, I do not think anybody has an answer to this question of larger positive effects of UBI. According to me, there is no evidence that suggests that incomes will go down, but will it go up a lot? We will have to wait and watch.

In the short term, there is mixed evidence—there are ten studies that find positive effects and six that show there is no evidence of impact on earnings. I want to make it clear that there is nothing called “uniformly defining.” The results are varying at the moment. For instance, in rural Ghana there was evidence of sales revenue increasing to 15 percent a month.

RI: How can we create the fiscal space for such a scheme as we grapple with high fiscal deficits? Should targeted cash transfers be a replacement for welfare programmes such as the Public Distribution System and the Mahatma Gandhi National Rural Employment Guarantee Act, or MGNREGA, or an add-on to these?
AB: We find ways to make fiscal space for all too many vanity projects, such as high-speed rails, and various subsidies for the rich, loan write-offs, et cetera. An economy growing at 7 percent should be able to put aside one extra percent per year for three years to get to 3 percent of GDP for a good cause. In particular, we need to ramp-up taxation and given that we have a burgeoning number of billionaires, we could tax the rich more. In terms of substituting for other welfare programmes, I think we should go there, but not before making very sure that we can deliver what we promise.

Politically, the idea of taking something away from people before you have an efficient replacement is, according to me, unrealistic. This is another reason why starting small—1 to 1.5 percent of GDP in year one—is a good idea.

RI: So the scheme should replace multiple state subsidies but in a phased manner?
AB: Correct. First we have to establish that what you are putting in place works. And then you are in a much better position to proceed. For instance, you can tell the farmers that instead of providing subsidies on fertilisers, we are giving you this; or in place of loan waivers, we would offer you this. Establishing a clear mechanism for delivering to a large population, without a lot of glitches, is extremely valuable and important to move on from our incredibly complex subsidies.

RI: The Congress recently waived farm loans in Chhattisgarh and made similar election pitches in other states. They are trying to implement a minimum income guarantee scheme along with fiscally heavy loan waivers. How do these two balance out in the larger and immediate scheme of things? Have you had any discussions with the Congress on this?
AB: I have had no discussions with any political party about the loan waivers. I see why they happen though—there is just a lot of pain on the ground and loan waivers, while inequitable and inefficient, deliver the goods fast. This is why we need an efficient pipeline for making transfers, ideally combined with a database where we know who is a farmer, how much land he has, et cetera. We have been under-investing in statistical and informational infrastructure, in part because we do not like what the data shows, and this is why we cannot design better transfer programmes than the loan waivers.

RI: Another fear among policymakers is that if we implement a cash-transfer scheme on a pan-India basis and it fails, it will be difficult to roll it back in the future.
AB: Unless growth radically slows down—which could happen but is not the likeliest scenario—the fiscal pressure can be handled in the medium run. We are growing at 7 percent now, and if we spend even 3 percent of the GDP on the scheme, in ten years it will be just 1.5 percent of the GDP. The real issue is in the very short run.

We will have to re-capitalise the banks, so that they stop being zombies, and start doing their job. That might cost even 7 to 8 percent of the GDP eventually, and we have to do it quickly because we will lose growth otherwise. This will require making some hard fiscal choices, because that is money we will not have if we want to simultaneously meet the FRBM targets. [The Fiscal Responsibility and Budget Management Act of 2003 outlines measures for institutionalising financial discipline and macroeconomic management.] This is why the National Democratic Alliance government has let the problem fester, but it is probably not a good idea to continue to do so.

At some point we need to bite the bullet and pay for it. This means that the economy will face fiscal stress in the short run, and adding 3 percent of the GDP to that is going to be a challenge. If we are willing to sell a bunch of equity of the public-sector banks, it will probably improve their governance—though ICICI Bank seems to have messed up quite badly—and relax the fiscal pressure.

RI: There is criticism that rolling out such targeted cash transfers to the poor would make a case for the state to withdraw basic public services to all. How would you answer that critique?
AB: The fact is that nobody wants to go to government schools. Every year the fraction of people going to private schools is going up. We have to just take a hard look at it. When the current NDA government came to power in 2014, they promised a wholesale reform of the education sector, but even the report has not been tabled in the parliament. So, primary education in the public sector is in the doldrums and to some extent in the private sector too. It makes no sense to think that the government is delivering something very valuable.

And if we take the state of healthcare, it is much worse. We have to seriously consider that now there isn't much demand for government services. The reason is somewhat the quality of services and something to do with people's misperception of quality. To say that we should be spending more money on these [basic public] services, that strikes me as completely missing the point.

RI: Many people would be excluded as beneficiaries as effective targeting would be difficult due to lack of digitised land records and the challenges of measuring poverty on a year-on-year basis. Do you think it could also lead to social discontent?
AB: Discontent is not a bad thing. Our problem is that we have had too little discontent from the poor. Our discontent is middle-class and lower-middle-class people protesting about job quotas. I would rather have discontent about the design of social services. Maybe it will put pressure on politicians to increase the efficiency of transfers. I think some discontent is useful. It is what drives good political systems.

RI: In one of your papers, you note that one important aspect of these cash transfers is that it boosts the psychological well-being of the extremely poor by lessening the “cognitive burden” induced by poverty. You cited your research on cash-strapped sugarcane farmers in Tamil Nadu, who performed badly in cognitive tests just before harvest season.
AB: Anything that delivers a stable and secure minimum living-condition to the poor will make them less desperate. This is not so much about a universal or a targeted scheme, than it is about whether it is okay for a society to have desperately poor people.

RI: In the last two decades, India has witnessed thousands of farmers committing suicides. If we had a cash-transfer scheme for the poor in some predominantly agrarian states, do you think it would have had an impact on the number of suicides?
AB: It is a very good question but I do not have an answer. It is an interesting hypothesis. But if you look at exactly why people committed suicide and what will change that by a substantial amount, which is too hard [to quantify]. We cannot just think of it statistically. We do have some evidence that if you take the poorest people and give them assets, their self-reported well-being goes up. So, this is a good marker for what might prevent suicides, but we still do not really understand it.

My point is, you are asking if it is a plausible causal mechanism that could help take the mental pressure off people? Absolutely. Will it actually affect suicides, which is an extreme outcome? This is hard to say.

RI: Various experts see these schemes as a cover for successive governments’ failure at creating jobs and containing agrarian distress. Some have labelled it a “cash for votes” scheme for the coming elections. Do you think this sudden acceptance of UBI, and its variants, across the political spectrum is an election gimmick?
AB: I am all for the view that the government provides a very leaky bucket. Has MGNREGA made people better off? I think the answer is yes. There are a bunch of studies that find that MGNREGA leads to an increase in wages while at the same time poverty goes down. Has the PDS improved over the last ten years? Yes, it has.

Every party has felt that there is now a sense that people feel there is so much inequality in the country. There have been so many scams and at the same time there is so much wealth. So, why isn’t some money coming to them? We write off one lakh crore for person X, then why can’t they get any money? I think people are legitimately asking these questions. So, the idea that whenever the government makes a promise, it is a scam, strikes me as being over cynical. It is not true. And I say this across political spectrums.

RI: Do you think these income-transfer schemes by both, the Congress and the BJP, were a hasty move with an eye on the upcoming general elections?
AB: I think, yes and no. The elections are here and we should have a clear and long-term weighing of the cost and benefits. I think both parties have done it and both parties have gone ahead because of elections and hence, they do not have time to figure this out.

On the other hand, is it a sign of the times? Surely. In a sense, we are coming to terms with the fact that we are a lower-middle-income country and therefore, there is some justification in thinking that we should not leave anybody behind in desperate poverty.

This interview has been edited and condensed.