Keywords

The “bottom line” is the central question in any policy debate: an overall evaluation of that policy in both empirical and ethical terms. Does it work? Should we do it? Because experiments can’t address ethical questions, researchers will have to focus on the does-it-work question in light of the ethical issues also under discussion. This chapter attempts to identify an overall bottom-line question and to understand how to frame smaller bottom-line questions for specific issues. The following chapter goes on to identify more-specific claims that are important to the discussion, setting up the subsequent discussion of the extent to which experiments can directly or indirectly address each of those claims.

1 Identifying and Overall Bottom Line

Identifying the bottom line is more difficult than it might appear. The question does UBI work is too vague for a social science experiment, partly because whether something “works” depends on controversial ethical questions such as what goals it is supposed to accomplish and how tolerable are potential side effects. Social scientists tend to translate the does-it-work question into the cost-effectiveness question: how cost-effective is it? This question sounds very scientific and neutral, but it still requires a resolution to controversial ethical questions. Which effects of UBI morally count as costs? Which count as benefits? What relative weights do we put on benefits X, Y, and Z and on costs A, B, and C? Whether something (such as a decline in average labor hours among low-wage workers ) is considered a negative “side effect” or a positive “effect” often depends on controversial ethical issues. If citizens and policymakers could resolve all of these issues and hand empirical researchers an index to weigh costs and benefits, researchers would have a purely empirical question to examine. But no one can resolve these deep moral controversies in advance of a study.

Empirical researchers are, therefore, forced to impose some controversial judgments on their evaluation process. They should warn readers what these judgments are in an attempt to create a shared set of background assumptions. But doing so can sound as if it merely adds yet another caveat. Perhaps, they should go farther and examine several different moral weighting systems to provide information for people with differing ethical positions.

Empirical economists sometimes ignore ethical background assumptions in their evaluative tools. Many economists look at costs exclusively in dollar terms and cast cost-benefit calculations in efficiency terms, with little or no discussion of the debate over whether these measures should have ethical priority over other options. For example, although a dollar lost to anyone is an efficiency loss, citizens might have good ethical reasons to value a dollar used to cure poverty more than several dollars used to provide luxuries for the already wealthy.Footnote 1

In the absence of a national resolution to the ethical controversies that create this problem, researchers will have to impose something, but they should avoid presenting their resolution to moral issues as if it were uncontroversial. It is better to be open about the moral judgments necessary to frame the empirical issues. It’s also valuable to recognize the different moral perspectives that are relevant in the local political context and present evaluations relevant to each. This book cannot resolve this issue and won’t dwell on it.

I attempt to state the cost-effectiveness question in broad terms as:

Is a fully implemented national UBI a cost-effective method to benefit people in the short and long run in the ways UBI supporters claim it does, assuming cost-effectiveness is judged relative to other methods of achieving similar benefits for the same people?

Many of the things UBI supporters claim UBI can do (see Chap. 13) require a generous UBI in the context of an extensive welfare system doing the things UBI cannot do. Although some aspects of the welfare system can be replaced by UBI (most notably policies designed to maintain the incomes to a level sufficient for normally abled people), other aspects are not so replaceable. Exactly what that extensive welfare system should involve is controversial even among UBI supporters, but it might include education, healthcare, childcare, eldercare, disability care, a higher-than-basic income for people with greater-than-normal needs, family leave, infrastructure, transportation, public safety, an affordable housing policy, and so on.

Testing a full UBI in that context might not be possible, and it is reasonable for researchers to test only a small step in the direction of UBI supporters’ vision. But, if we test only that step, we are not testing the UBI that inspired the movement. Sometimes small steps work when big leaps fail (such as toward the end of a dock). Sometimes big leaps work when small steps fail (such as over a ditch). Whatever version of UBI (or whatever UBI-like policy) we test, researchers should clearly explain how it differs from other versions and the extent to which this test’s findings do or do not have implications for other versions of UBI.

It might, therefore, be useful in some circumstances to state the bottom-line question in slightly more incremental terms:

What policy (basic income, the current system, or any other alternatives to be tested) produces the greatest increase in recipients’ welfare per unit of cost (both in terms of tax cost and efficiency loss), in the context of a long-term, fully implemented national policy?Footnote 2

Obviously, these statements of the bottom line can be shortened if some of their constraining phrases can go without saying. I hesitate to do so because of the amount of misunderstanding these issues have caused in the past.

I suggest that one of these cost-benefit questions—or something like them—should be considered the bottom line for UBI experiments. Experimental evidence cannot definitively answer the bottom-line question, but experimenters can relate experimental findings to it: how does this research improve our understanding of the bottom line?

These specifications of the bottom line impose answers to some moral questions. I’ve tried to reduce this problem by phrasing the question in relative terms—relative to supporters’ claims about its benefits and relative to other ways of achieving those benefits. It intentionally leaves open what the claimed costs and benefits are.

I’m concerned with overidentifying any claim as “the” goal of UBI in any political context. The UBI movement is diverse, as is the opposition to UBI. Some see UBI as a way to eliminate the threat of poverty for everyone. Some see it as a way to make alternative lifestyles possible. Some see it as a way to simplify and streamline the tax and benefit system. And so on. I doubt there is any political context in which virtually everyone who discusses UBI is interested only in a very limited range of issues.

Phrasing the cost-effectiveness question in relative terms does not eliminate moral controversies. For example, even if nearly everyone might agree that a central goal of UBI is to “increase recipients’ welfare” (as used above), any effort to define “welfare” is controversial. Popular welfare measures might leave out some of the concerns that are important to the UBI discussion. Researchers should not simply stop using these measures, but they can supplement them by discussion of how UBI affects important items that can’t be incorporated into the index.

The important points are not that the bottom line is phrased as I suggest, but that the experiments have a bottom line, that it is a broad question, that it compares costs and benefits, that it refrains from distracting attention from things experiments cannot measure, and that it addresses what people need and want to know to evaluate UBI as a potential policy in their country or region.

The overall bottom line is important for two reasons. First, virtually any empirical research question can and should be understood as some part of the answer to this general question. Second, it is what citizens and policymakers ultimately need or want to learn from empirical policy research. The more they know about the cost-effectiveness of UBI, the more fully informed they will be as they discuss and make the decision whether to implement UBI.

If citizens and policymakers believe many of the media reports on the launch of experiments, they not only want but expect a bottom-line answer. This expectation is an important reason to relate findings to the bottom line. Experiments have a much narrower objective. Experiments divide people into control and experimental groups, observe whatever differences they can, and test those differences for statistical significance. If experimental reports are limited to explaining what these differences are, they stop far short of any effort to find what people are looking for.

2 Issue-Specific Bottom Lines

Many issues can be usefully addressed in isolation. But no one has a direct interest in the simple comparison between the control and experimental groups for any observational variable. They have an interest in a long-term estimate for the impact of a national UBI on that variable. And they have an interest in viewing it in the context of cost-benefit analysis relative to other policies. Therefore, in addition to the overall bottom-line question, each variable can have a mini bottom line of its own. The bottom line for any particular variable is the cost-effectiveness of a long-term national UBI on that variable.

The calculation of the long-term impact of UBI on any variable involves considering community effects, the difference between a short-term study and a permanent policy, the ways in which the sample succeeds or fails to be representative of the entire population, and so on. For some variables, researchers might be able to use simulation techniques to calculate that answer. For others, they might have to bring in more qualitative information or simply have a qualitative discussion. Even if they lack data to make a reasonable estimate, they can explain the differences between what they found and what we really want to know. They can also discuss the missing factors necessary to get closer to the bottom line.

One example of an issue-specific bottom line is whether a step in the direction of universality can free people living on benefits from the poverty trap. This question, which seems to be important in the Finnish and Dutch experiments, is worth looking at even in isolation as long as the difference between it and an overall evaluation of UBI is clear.

Calculation of the overall bottom line requires a comparison of the bottom line for each particular variable estimated in the experiment and probably also with estimates for other variables the experiment could not examine. This effort, again, might be achieved with simulation techniques; it might instead require more qualitative techniques, or it might involve admitting why the effort falls short of that goal.