Sunday, March 17, 2019

Universal Basic Income in the Developing World

Banerjee, Niehaus and Suri (February 2019, wp22598) have a new NBER working paper out on the what we know and don’t know so far about universal basic income, particularly 
  1. what recipients would likely do with the incremental income
  2. whether this would unlock further economic growth
  3. the potential consequences of giving the money to everyone (as opposed to targeting it)
They consider UBI as an incremental anti-poverty intervention and argue that UBI is unlikely to be cost-effective at achieving any particular narrow policy goal (health, education, nutrition, etc) or to relax any particular constraint (financial constraints such as credit or insurance market imperfections, psychological constraints such as hopelessness, or limited bandwidth). 

This blog post summarizes the main takeaway from their working paper

What recipients would likely do with the incremental income?

What we know
  • Little evidence of UBI’s effect in developing countries. 
  • Three schemes closely align with UBI: one for two years in nine villages in the Indian state of Madhya Pradesh (2010-2011), one in two villages in Namibia (January 2008-December 2009), and Iran’s nation-wide cash transfer introduced in 2011 to offset the withdrawal of food and fuel subsidies. 
  • Existing transfers have not been universal but rather targeted, both to subsets of households (through means testing, ordeals, conditions, etc.) and to specific adults within those households (often the female head). Existing transfers typically last for relatively short time periods, as opposed to the long-term commitment envisioned by UBI advocates.
  • Many transfers (particularly in South and Central America) were paid out conditional on certain conditions being met, but many others (particularly in Africa) were not. Pensions are also transfers and have structure (size, frequency and duration) quite similar to UBI payments. As of 2018, there were 552 million people living in the developing world who received some form of cash transfer.
  • Evaluations generally have not found the negative impacts that many feared. Transfers had on average reduced expenditure on temptation goods. There is no systematic evidence that transfers discourage work. 
  • Evaluations have also shown a great diversity of positive impacts on income, assets, savings, borrowing, total expenditure, food expenditure, dietary diversity, school attendance, test scores, cognitive development, use of health facilities, labor force participation, child labor migration, domestic violence, women’s empowerment, marriage, fertility, and use of contraception, among others.
What we do not know
  • Examining impacts on entire populations is important to understand the effects of a UBI. Universality could change impacts if there are interactions between the effects of one’s own and one’s neighbors (basic) income. We can say little about this at the moment as very few studies have simultaneously varied individual treatment status and the overall intensity with which communities are treated, and then tested for interaction effects. 
  • There is relatively little evidence to date on how transfers affect local markets through general equilibrium effects (transfers affect prices and wages too).
  • Delivering a UBI to each adult in a household (which is what many basic income proposals contemplate) could have different effects from delivering the same amount of money to a single adult representative of the household (which is how most existing cash transfer programs function). What evidence exists does suggest that the identity of the recipient matters, but not that one type of recipient is unambiguously better” than others.
  • Committing to deliver basic incomes for a long period of time could have effects that differ from the effects of shorter-term transfers themselves. To understand whether and how duration matters, a key issue is separating the effects of current transfers from the effects of anticipated future transfers.

Can UBI unlock economic growth?

UBI is unlikely to be the most cost-effective way to alleviate any one of the underlying constraints on investment. But we currently know little about which constraints bind for whom, or how in practice to target interventions to the specific people who need them. This is what makes UBI, which does not try to step on any specific lever, interesting as a pro-growth policy.

Several strands of research suggest that lack of access to capital constrains some more than others. Where they do, a UBI potentially provides a source of capital to relax the constraint. Firm marginal products are much more dispersed in India and China than in the US. There are various possible reasons for such results, but the most obvious theory is that some firms are (differentially) credit constrained and therefore underinvest relative to others.

At the micro level, an extensive literature on the returns to capital suggests that they are widely dispersed and very high in some cases, over 100% a year. There is also strong prima facie evidence of inefficiencies in the financial sector in most developing countries. Borrowing interest rates in the non-banking sector are often extremely high. These data co-exist with estimates from randomized trials of microcredit showing that borrowers on average do not use loans to grow their businesses. Grants also seem to have (widely varying) effects on investment.

For some, lack of insurance markets may be a binding constraint. Entrepreneurs may shy away from investing borrowed or owned capital because they want to avoid exposing themselves to business risk. Evidence that uninsured risks distort investment and production is mixed.

External constraints such as missing markets for credit or insurance have been development economist’s traditional bread and butter. But what if some people face constraints within their own minds? For some, poverty may simply tax the mental and emotional bandwidth needed to think through important decisions. 

UBI could be transformative for people bound by internal constraints. Not having to worry about making ends meet could free up the mental and emotional bandwidth needed to focus on getting ahead, or re-set hopes and beliefs about the future. Whether this is true or not is, of course, still to be seen.

Targeted or universal transfer?

A central question about UBI is whether universality is in fact efficient. For any given budget, is it better to spread those resources evenly or to give larger amounts to the poorest? There has been a large literature on different approaches and their effectiveness, ranging from by far the most common, a proxy means test (PMT), to community targeting to self-selection. Universality has several under-appreciated benefits, and targeting several underappreciated limitations. 

Work quantifying the relationship between impacts and targeted characteristics is limited, as is work on the potential disincentive effects of targeting. It is also unclear how effectively targeting poor households succeeds in targeting poor people, taking into count the unequal distribution of resources within households and redistribution of resources across them.

Universality could reduce administrative costs. Targeting well requires data collected repeatedly, given significant shares of people in developing countries change their poverty status from year to year. The administrative costs of universality are likely to be far lower, especially given the ongoing investments emerging market governments are making in digital ID and payment systems.

Universality could improve the “political economy” of redistribution. Government capacity to implement nuanced targeting schemes is often limited, particularly so in the poorest area where it is most important to get it right. In cases like these, making eligibility universal may have a modest effect on the realized incidence of benefits while at the same time substantially reducing the scope for corruption and other abuses of power. 

Broad eligibility could also help build political bases of support for sustained redistribution.

**********
Given the discussion of costs, it is worth asking about the administration or implementation costs of a universal system, i.e. the costs of distributing the transfers themselves. To avoid double counting, it will be important for developing countries to have a well-functioning universal identification system, preferably digital, though this will be a one-time cost. When state capacity to discipline front-line bureaucrats is limited, targeting can also create opportunities for corruption and other abuses of power. 

The per-person costs of delivering transfers are falling rapidly in many places due to advances in last-mile digital payments infrastructure. All else equal this will tend to further increase the appeal of broad or universal targeting.