On October 14, Esther Duflo, Abhijit Banerjee and Michael Kremer were announced as the winners of the 2019 Nobel Prize in Economics. In this extract from their forthcoming book, Good Economics for Hard Times, Banerjee and Duflo argue that a Universal Ultra Basic Income would be more suited for a developing country like India with its massive population size and lower-income levels.
So, there is no evidence that unconditional cash transfers lead to a life of dissolution. What does that tell us about the design of welfare policy? In developing countries, where lots of people are at risk of finding themselves destitute from time to time and where the safety nets, however imperfect, that exist in rich countries (emergency rooms, shelters, food banks) are missing, the value of having an assured fall back option like UBI could be enormous, both in dealing with bad luck and in making it easier to try something new.
One of the most common ways in which people safeguard themselves against income risk in many parts of the developing world is by holding on to land. We discussed the reluctance to migrate in chapter 2, and one reason for that is that those who migrate risk losing their land rights. Interestingly, most rural land-owning households these days in India, for example, get a majority of their income from something other than agriculture. But land ownership is still valuable because it comes with the assurance that if all else fails, they can grow their own living.
This has the consequence that areas with a large fraction of small-holders tend to find it difficult to industrialize. This is partly due to the design of land reform—when the poor are given land rights, it is often inheritable but not saleable. But there is also a strong resistance to sell among the farmers themselves. In the Indian state of West Bengal, when the communists came to power after winning the 1977 elections, their first priority was to give the tenant farmers permanent rights on the land they tilled. The right could be inherited but not sold.
Thirty years later, the same communist government, conscious of the lack of industrial development in the state, tried to buy out the farmers (including the tenant farmers). That met with such furious resistance that the plans were shelved, and the communists ended up getting booted out after massive protests against the land expulsions, and the bloody repression it was met with.
The one thing the farmers in West Bengal wanted in compensation for giving up their land was a promise of a job, a stable source of income. Perhaps if there had been some kind of UBI to provide this income, the resistance might have been much less and it might have been easier to move arable land into industrial use. In chapter 5, we mentioned that poor use of land is a major source of misallocation in India, probably responsible for a significant loss of economic growth. If UBI alleviated the need to stick to your land at all costs, it would reduce this misallocation. It may also reduce labor misallocation by making it more palatable for the landed to sell their land and move to where there are better labor market opportunities.
India, however, does not have anything like UBI right now. The current scheme proposed by the government applies only to farmers and is nowhere near a living. The Minimum Income Guarantee proposed by the opposition is more akin to the negative income tax credit. The plan is that it should be targeted to the poor, and progressively taxed away as incomes grow.
In fact, very few countries have anything like a UBI, which is guaranteed to everyone and is not taxed away. If they have anything, they have transfers targeted to the poor that can be conditional or unconditional. But targeting the right people in the developing world tends to be especially difficult because most people work in agriculture or in tiny firms, and it is almost impossible for the government to know how much they are earning, which makes it very hard to isolate the poor and target them with the extra income.
The alternative to targeting is self-targeting; India’s National Rural Employment Guarantee Act (NREGA) is the largest of these self-targeted programs (and perhaps some sort of model for the Federal Job Guarantee that has been proposed in the US). Every rural family is entitled to 100 days of work per year at the official minimum wage, which is higher in most places than the actual wage. There is no screening except that there is the requirement to work, usually in construction sites, which screens out anybody who has something better to do than standing in the sun for eight hours a day.
The program is popular with the poor, so popular that the government decided not to fight with it head-on after they won the election, despite having campaigned against it. One advantage of a workfare program like NREGA is that it substitutes, at least partly, for a minimum wage in places minimum wage cannot be enforced. Workers can use the NREGA wage to bargain with private sector employers, and there is evidence that they do. Moreover, one study found that private employment actually went up even though salaries went up: by colluding to pay too little, employers were actually reducing the number of jobs, perhaps because some people were unable or unwilling to work for very little money.
The main sticking point with any workfare program is that someone has to create millions of jobs. In India this is meant to be the responsibility of the village governments (the panchayats). But there is a lot of mistrust between the central authorities and the village governments, with each side, often with some reason, accusing the other of corruption.
The result is the kind of red tape and inefficiency that often arises when there is a lot of emphasis on fighting corruption. Approving a project proposal and getting work going takes several months and quite some effort by the head of the panchayat. This means that the program is unable to respond effectively to sudden changes in need, say an unexpected drought. It also means that if you happen to live in a village where the panchayat leadership has decided that these projects are too much trouble, you are out of luck. In Bihar, India’s poorest state, less than half of those who want work through NREGA get it.
The program ends up being rather prone to corruption, because the very people involved in monitoring the program can use their power to block payments and extract bribes. Cutting the number of layers of bureaucrats involved in the monitoring of the program reduced the wealth of the median NREGA functionaries by 14%. And even when people get work, it often takes months to get paid.
All of this suggests that there are many very good reasons to consider moving to a UBI in many developing countries. The problem, of course, is money. Most developing countries need to tax more, but that won’t change quickly. Initially, most of the money will have to come from shutting down other programs, including some of the big and popular ones such as power subsidies. Cutting the number of programs potentially has the added advantage of concentrating the limited government capacity on just a few things.
The government of India has hundreds of programs on its books. Many of these have essentially no funding but they have an office dedicated to them and some staff who accomplish very little. Manish Sisodia, the Deputy Chief Minister in the Delhi Government, once joked that when he came to power, there was a line item in the budget for opium purchases. He discovered that this was the remnant of a long-defunct program to help refugees from Afghanistan who had once come to Delhi and were addicted to opium.
Any universal income that governments of poor countries can afford will be ultra-basic. Hence UUBI. The Economic Survey of India proposed something like that in 2017. It estimated that an annual transfer of 7,620 rupees ($430 at PPP) to 75 percent of India’s population would push all but India’s absolute poorest above the 2011– 12 poverty line. INR 7620 is very little by Indian standards (less than what several economists have proposed for an Indian UBI), but perhaps enough to survive on. The survey puts the cost of such a scheme at 4.9 percent of India’s GDP. In 2014–15, India’s major fertilizer, petroleum, and food subsidies cost 2.07 percent of GDP, while the ten largest central welfare schemes cost 1.38 percent, so cutting these existing programs entirely would pay for about two thirds of the UUBI.
This proposal assumes that it would be fairly easy to exclude 25% of people from the program. It may indeed be possible to introduce a mild form of self-selection. Requiring that each beneficiary visit an ATM every week and put their biometric ID into the system, whether or not they take out the money, would have the dual advantage of eliminating ghost claimants and making it too much of a hassle for the wealthy to want to claim the benefit. There should be fallback options that allow the disabled to get their money, or in case the technology malfunctions. But with the right framing (“come and get some extra money when you need it” ) a mild requirement like having to visit an ATM once a week could potentially discourage more than 25% of the population at any given time, while making sure that those who really need it still get it.
While we are in favor of a UUBI based on what we know so far, there is no data yet on its longterm impact. Most of our evidence so far is from relatively short-lived interventions. We cannot be sure how people will react to being assured a basic income for ever; when the novelty of the extra income wears off, will they go back to being discouraged and work less, or aspire higher and try harder? What will be the long-term impacts on their families of being assured of an income?
This is what a large-scale RCT of a UBI in Kenya that Abhijit is involved in will hopefully find out. In 44 villages, every adult has been guaranteed 0.75 dollars per day for twelve years. In 80 villages, every adult will receive the transfer for two years. In 71 villages, every adult will receive one one-time payment of $500 per adult. Finally, in 100 more villages no one is guaranteed any income, but data will be regularly collected. In total, almost 15,000 households are involved in the experiment. We will start seeing results in early 2020.
We can however already see long-term evidence from the conditional cash transfers that have been in place for many years in several countries. These programs started in the 1990s, and those who were children at the time are now young adults. There seems to be enduring positive effect on their welfare.
For example, in Indonesia, in 2007, the government introduced PKH, a conditional cash transfer program in 438 sub-districts across Indonesia (randomly selected from a pool of 736 subdistricts) to a total of about 700,000 households. The program had the standard features of most conditional cash transfer programs: households received a monthly transfer if they sent their children to school and obtained preventive care. Villages that were enrolled into the program in 2007 continue to receive the benefits even today, but the government never expanded the program to the control villages.
Comparing treated and control village shows some large persistent gains on health and education: there is a dramatic increase in births attended by a health professional and a halving of the number of children not in school. Over time, the program also affected the stock of human capital: there has been a reduction of 23% in the number of stunted children, and school completion increased.
However, despite these gains in human capital and the transfers themselves, households are not measurably richer. This is an important warning about the long-term effect of purely financial transfers: it may be the case that small sums of money are too small to make a real difference to incomes (and the cost of large transfers is too much for the system to bear).
Given all this, the best combination may be a UUBI that everyone can access when they need it, and larger transfers, targeted to the very poor and linked to preventive care and children’s education. The conditions for receiving transfers do not need to be very strictly enforced. In Morocco, we saw that a “labelled cash transfer” that merely encourage to use the money to help with education costs, but without enforcement, seems to have been just as effective at changing behavior as a traditional conditional cash transfer program.
Similarly, the PKH program in Indonesia did not strictly enforced conditionalities: in this sense it also was “labelled cash transfer”. This lowers administrative costs and avoids excluding the most fragile families. Targeting can also be done relatively cheaply, by focusing on poor regions, and relying on some identification by community leaders and readily available data. There will be errors. But as long as we are willing to be liberal in the application of the tests (so that those who need help don’t get thrown out, even at the cost of giving it to some people who do not need it) , and as long as the UUBI is there to provide a minimum, we might get the best of both worlds.
Excerpted from Good Economics for Hard Times: Better Answers to Our Biggest Problems, authored by Esther Duflo and Abhijit Banerjee (winners of the Nobel Prize in Economic Sciences 2019), published by Juggernaut Books.