Steve's Blog

The Simple Economics of Welfare vs. Universal Basic Income

A warning to the reader: I take 2016 tax information and poverty line levels from government websites. The rest of the "facts" are "to the best of my recollection and knowledge." I think they're pretty accurate, but you might want to check back occasionally to see if I've confirmed any of them!

Recently the political left has been gathering support for two "universal welfare programs," single-payer health insurance and universal basic income. There is an important reason why these programs are opposed by many, especially on the right: to be effective, they involve large transfers of resources received in market-based transactions from richer members of the polity to poorer ones. If you believe those transactions represent "earned income," there's a moral argument for keeping it, and thus for opposing programs that involve transfers. However, if some programs to address the problems these programs are design for become politically imperative, the right should give strong consideration to supporting these among other alternatives on the grounds of efficiency.

In this essay, I provide the basic economic tools for understanding the argument for the efficiency of universal basic income (below abbreviated to UBI) as an income support program. Quantitative analysis is beyond our scope here, but in 1979 my professor taught that a flat tax of 23% (with an equal negative income tax rate and a standard deduction large enough to imply a UBI of 50% of individual poverty level) was deficit-neutral for the federal government. Presumably that would imply substantial savings for conventional state and municipal welfare programs. For comparison, in 2016 the 25% tax bracket begins at $37,651 after deductions for individuals, and $75,301 after deductions for married joint filers. Add $10,350 for standard deduction and personal exemption for individuals to get $48,001, and at least $20,700 for families to get $96,001 as approximate gross income (paycheck before deductions). Everyone above those incomes would see a decrease in their tax rate on the last dollar they received.

Surely there's a rat in the barn here. Here it is: "flat" means flat: no deductions except the standard deduction. At the lower levels of the 25% tax bracket, most use the standard deduction; their tax burdens could rise or fall, but likely not much. This is a really important point, politically: the UBI schemes that could conceivably pass are unlikely to affect middle class incomes one way or the other.

For substantially higher gross incomes (before adjustments), various income sheltering schemes become very important, and deficit neutrality depends on taking those shelters away. As an aside: this is why tax reform is so difficult. The very people who complain most about the economic disincentives of the current tax system are the ones who benefit most from its complexity. This is not to say they are all hypocrites; a few are. Many others would be willing to pay higher effective tax rates because they believe economic growth would make us all much richer, enough to outweigh the tax changes for them. Still others honestly admit they like the system as it is, but intellectually they agree there are problems. The net effect is that executives and investors as a group are giving speeches for tax reform while their financial advisors and lobbyists as a group are doing their best to kill it.

So let's look at how UBI and conventional income support programs (below, CIS programs) compare. For this purpose, forget about the flat tax we could use to finance these programs. There are no bills in Congress to use as a specific example, and tax reform is hard. We'll also ignore the fact that for UBI to really be effective, support levels would have to be similar to CIS levels, yet being universal, the aggregate amount would be much larger. We'll presume that whatever financing scheme is adopted is going to be net near zero for most people.

Another aside: this is why quantitative analysis of proposals is so hard: it is really difficult to get accurate information about the plans. Even if there's a bill in Congress, these bills tend to be hundreds, even thousands, of pages long, and the detailed "earmarks" (spending) and tax breaks can be as large as the main purpose of the bill. Pity the poor CBO, expected to evaluate "Top Secret" McConnell bills in a few hours!

The Easy Cases

There are two kinds of person whose responses to the substitution of UBI for CIS are trivial to guess. First, those mired deep in the welfare system, who are convinced they cannot get work or are actively gaming the system to avoid having to work at all. Nothing changes for them, except perhaps the amount of income transfer they receive. They're not going to go to work. Statements about "welfare queens" and others gaming the system are completely irrelevant to the question of the effects of UBI on employment.

The fact that so many opposed to, or dubious of, UBI bring up "welfare addicts who want to avoid work" suggests a serious misunderstanding of the economics of the current welfare system. Such people are not at all the target of UBI, UBI can't affect them!

The second kind is those whose income is derived from non-income-based transfers or passive investments. They don't work either, and won't start. In fact, we can pretty much ignore such "unearned" income for people with employment income as well. The main effect is that the larger "unearned" income is, the smaller the "income effects" of introducing UBI mentioned below will be. Neglecting them weakens the case for UBI.

The Target Case

The important aspect of UBI for economic analysis is the "U". "Universal" means that it will not be denied, regardless of what you do. Your net income, that you can spend for your purposes, is guaranteed to be above a certain level. The net effect of taxation is that, if you choose to do so, you can do accounts in such a way as to cancel the UBI portion out with enough tax paid. But you can also do accounts so that it is never decreased. The latter is more accurate for the purpose of comparison to CIS. What we want to consider is how "the rules of the game" affect "take-home" pay given gross employment compensation. The "easy cases" were easy precisely because there is no employment compensation at all.

Let's consider those who currently on welfare who are not adamantly opposed to working in principle. An important aspect of CIS programs is eligibility: above a certain level of income, you are not eligible for income support. From the point of view of the individual, this functions exactly as an income tax: more money is deducted from your total income as your employment income increases. And the tax rate is huge. Current federal income tax rates are 10% on taxable income for people below the poverty line. After deductions, the average tax rate is near 0%. On the other hand, the eligibility cutoff is typically around the poverty line, while support levels are often more than 50% of the poverty line. In other words, the average "reduction rate" for welfare recipient currently unemployed who is considering taking a job at the poverty line level of income is far greater than the tax rate of a bank executive receiving millions in compensation (39.6%). If high tax rates are considered a "disincentive" for people who have so much money they need to hire people to figure out what to do with their money, how much greater is the disincentive for the poor? In some areas, a $15/hr minimum wage is proposed. For CIS recipients, that looks like $6-$7/hr after losing their benefits.

On the other hand, with UBI, the same person will pretty much have exhausted the standard deduction as soon as they start working. At the poverty level of employment income, they'll be paying at a tax rate of 10% (maybe a little more for "single" taxpayer status). UBI provides a much stronger financial incentive to work than CIS does. That proposed $15/hr minimum wage is $13.50 after deductions.

Note that the concrete calculations of net income or wage are rather inaccurate, as they ignore state and local income taxes, as well as Social Security taxes. Still, this gives a feel for the magnitude of the disincentive to work generated by CIS vs. UBI. The latter is much smaller, perhaps negligible for current CIS recipients. This is an extremely important consideration, given the very plausible theory of a generation-spanning "welfare trap".

Grey- and Black-Market Employment

Another easy case is that of people who work "off the books" in order to preserve eligibility for CIS benefits. These jobs provide no employment security, and in some cases involve criminal activity, or work for criminals, and in some cases involve elevated incidence of work-related violence or victimization by criminals. As off-the-books employment means zero deductions by government, it is not clear that there is much stronger incentive to work "on the books" under UBI than CIS, but it is certain that there is stronger incentive. UBI lowers incentives for criminal and grey-market activity.

Poverty-Level Employment Income

Those whose employment income puts them at or near the poverty level, and makes them ineligible for CIS benefits, have no stronger incentive to quit work under UBI than under CIS, unless UBI is greater than CIS benefits. This seems unlikely, for political reasons, for financial reasons, and for economic reasons. The financial reason is that the broader eligibility means far greater nominal outlays. The economic reason is that the much stronger incentives for at least part-time employment under UBI suggest that its level can be substantially lower than CIS benefit while providing a higher standard of living to the poor.

It is true that there is an income effect. Many people who have some bargaining power over total hours worked may prefer to reduce those hours if they can maintain the same or higher levels of income. However, introspection suggests that this disincentive is not so great, especially in the long run. If it were large, we would expect very few to be working near the poverty line: they would prefer to receive no employment income and nearly as much in CIS benefits. However, the benefits of work, which include future improvements in conditions such as wage raises and improved hours, seem to be large enough that there is no such sharp dropoff.

Moreover, there are some countervailing effects. For example, many of those are likely to be individuals who will spend the time on caring for families, both young children and family members incapacitated by extreme age or degenerative diseases. Others who might decide to stop working entirely include students who could not otherwise afford to go to school (either due to saving for high tuition or due to family responsibilities they decide to serve first). This is not to deny that some will take the opportunity to "just goof off." I merely wish to point out that the employment disincentive may be partly an investment incentive, or due to engaging in non-market productive work.

Substantial Wage Income

The effects of introducting UBI on those with substantial wage income are similar to those with poverty-level incomes, except smaller because the UBI would be a smaller fraction of total income, and therefore expected to have a smaller effect on desired working hours. Also, a larger fraction of those with larger wage incomes are probably under contracts specifying that they have little bargaining power with regard to hours worked. Overall, it seems very unlikely that introduction of UBI would have any substantial effect on aggregate incentive to work for this class of persons.

There are two small issues here. First, there may be a tiny negative effect due to people who hate their jobs, or have a strong thirst for education, and so are willing to temporarily leave the labor force, depend on their UBI, and invest in better employment in the future. Second, self-employment becomes more attractive, as described below.

Substantial Salary Income

Since, as a first approximation, for salaried workers hours worked are defined by the job not the amount of income, we should expect introduction of UBI to have no effect on employment for salaried workers.

That said, it is probably obvious that the two negative "small issues" described for high-wage-income workers apply here, too.

Self-Employment and Entreprenuership

Finally, one theoretical countervailing effect is on self-employment and small-business entrepreneurship. Because businesses have both a high rate of failure compared to unemployment in general, and have much riskier incomes, self-employment in a small business has substantial risk. In general, the presence of UBI will tend to smooth income (smaller annual percentive deviations from one's lifetime average), making self-employment more attractive.

However, the effect on incentives for entrepreneurship will be much greater. The UBI means that one need only finance the business's startup expenses. Under CIS, one has to finance one's own standard of living as well in most cases. This means either a longer period of saving in advance, or a larger bank loan (if you can even get one), or support from a venture capitalist. Also, new businesses are at very high risk of failure (I believe the rule of thumb is that 2/3 go out of business within 3 years of being founded). For my purposes, "entrepreneurs" are a special class of self-employed who are determined to found a business. UBI means that they can engage in serial entrepreneurship at a substantially higher rate, even if their first businesses do fail.

This effect is worth considering because foundation of small- and medium-sized businesses has historically been the engine of employment growth in boom times. On the other hand, big businesses tend to hire somewhat steadily both in periods of faster growth and in recession, but have large layoffs only in recessions. Thus big business tends to account for a larger fraction of new unemployment in recessions, even though business failures go up in those periods.

Conclusion, and the Parable of the Bear

I suppose it's obvious that I support substitution of UBI for CIS programs if deficit-neutral. I also think it's a good idea on other grounds. Even so, I've presented the employment incentives and disincentives as fairly as I know how. I think there's good reason for conservatives as well as progressives to support substitution.

A final aside: I don't really understand the reasons why so many conservatives oppose UBI, given that for about 50% of the population it's basically equivalent to the standard deduction, and for a majority of the rest (children and retired) most people already accept a social responsibility for their support in the form of education and retirement benefits. However, it seems to me that many have a moral objection to people receiving income that's not proportionate to their contribution to society, at least as averaged over their lifetimes. The frequency with which "welfare queens" and other "deadbeats" come up is some evidence for this moral position.

Unfortunately for that moral position, there's no basis for the accompanying assumption of an equivalence of market income received to social contribution that is frequently observed. High employment incomes certainly are correlated with ability and effort, but the level of that income depends on the amount of capital associated with the position. (The same is true in reverse for capital income: large amounts of capital income are mostly related to association with large amounts of "ordinary" labor.)

In fact, the truth of income distribution is mostly "being in the right place at the right time". But when "the right time" is simply "first" and "the right place" is "in college with a bunch of CEOs' kids", small differences in ability (and good choice of parents) can account for very large differences in income. It's like the guy whose friend screamed to him, "the bear is catching up!", and he replied, "but I don't have to outrun the bear, just you!" In what world does a difference of a couple tenths of a second in a 50-meter sprint morally justify the death penalty? None that I can think of. As far as I can see, morally, it's much the same for the huge income disparities we see. (In the economics of labor markets, there are other reasons that come up for the income distribution we see, and it may not be a good idea to mess with those incentives. But I see that as separate from the position that market income is deserved, and that therefore income transfers are morally wrong.)

I suppose that there is plenty of "surplus" that is currently going to the decision makers just because they can make that decision. I expect that it's sufficient to cover that part of UBI that would go to people with income lower than the standard deduction. Morally, I see no reason why that should go to rich people rather than poor people, but your mileage may vary.

So now you know everything I've left out of my main argument, to present it as "objective." I hope that you will still agree with the main economic argument, and I don't see good reason why you might not.

Final Summary

Even if you disgree with my moral position, the main point stands: UBI almost certainly creates much smaller employment disincentive than CIS programs do. Everybody should prefer UBI to CIS if that's the choice.


Tsukuba, Japan. v1.1: July 25, 2017

Tsukuba, Japan. v1.0: July 25, 2017

<< Previous COUNT=10 Next COUNT=10 >>


About this site:

This site is running Django now!

Stephen J. Turnbull