This is an article I wrote for SnoQap. The list of references can be found on their site.
Gloomy reports and predictions that technological advances will beget massive job displacements throughout the worldwide economy in the upcoming years have caused disquiet among many individuals. Technological advances will enhance productivity and lower prices, at the cost of high unemployment and consequently poverty caused by the lack of income to those unemployed. Artificial intelligence and robotics are expanding into transportation, manufacturing, retail, medical diagnosis, translation services, legal research, banking, financial services, and many other areas. A paper by the University of Oxford predicts that about 47% of contemporary US jobs will be automated out of existence in the near future (Frey & Osborne, 2017); and McKinsey reports that one out of three American workers are at risk of losing their jobs to new technologies (Manyika et al., 2019). This is not only a domestic issue, but a global one. That same McKinsey study asserts that 800 million jobs globally are at risk; and according to The National Bureau of Economic Research, increased adoption of robots in the US decreases employment and earnings for foreign workers as well (Kugler et al., 2020). Undoubtedly, technology will disrupt employment. Ideas on how to fight back are developing among academic circles and political parties. The most prevailing idea is that of guaranteed income. We shall examine this idea. According to economic literature, a guaranteed income program has (i) a positive financial, emotional, and physical impact, (ii) but a negative labor force impact. Further, forms of guaranteed income programs vary greatly in design, ranging from a Minimum Income Guarantee (MIG) to a Universal Basic Income (UBI), to a Negative Income Tax (NIT). We shall define each program, identify if they are useful in fighting technological unemployment and poverty, and determine which would be the most effective.
As I lay forth the definition for the three types of guaranteed income, let us assume that all of the programs’ goals are to alleviate poverty and technological unemployment. Further, let us say that the programs alleviate poverty by providing individuals a subsidy so that no individual has an annual after-tax income that is less than $1 above the poverty line. That is $12,881 per year according to US Federal Poverty Guidelines (2021). Lastly, we are going to assume any citizen 18 years of age or older qualifies.
An MIG program is designed in a way such that any individual with an income below a certain threshold is subsidized by a sum sufficient to make them reach that level of income. If we assume that the government deems that no citizen should have less than $12,881 per year, then anyone with an income under that level would be given the difference between $12,881 and their income. For example, an individual with an income of $7,500 would be provided with a subsidy of $5,381 ($12,881 - $7,500). This program is tailored to directly aid only people with an income below $12,881.
A UBI program is designed in a way that every citizen receives the same lump-sum subsidy to their income. Assuming that the government deems that no citizen should have less than $12,881 per year, then every citizen would be entitled to a lump-sum subsidy of $12,881, regardless of their level of income. For example, an individual under the poverty threshold that is unemployed currently earning $0 would end with an income of $12,881 per year; an individual with an income of $957,834, on the other hand, would also receive the subsidy of $12,881, bringing his total income to $970,715. This program has an egalitarian perspective, wherein everyone is entitled to the same sum of money per year.
Lastly, we have the NIT, perhaps the hardest to comprehend as it is slightly more complex. Under a Negative Income Tax, each individual would be entitled to the same tax deduction when filing their taxes. If the individual’s income is less than the tax deduction, the individual would end up with a negative income; he would be entitled to receive a rate, say 50%, of his negative income as a subsidy. If we assume that the government deems that no citizen should have less than $12,881 per year, then every citizen would be entitled to a tax deduction of $25,762 ($12,881 x 2), regardless of their income level. On one hand, an individual with an income of $0 would have an income of negative $25,762 after the tax deduction, consequently, he would be entitled to receive a subsidy of 50% of the negative income of $25,762: $12,881. On the other hand, if an individual has a gross income equal to the subsidy ($25,762), he would have a taxable income of $0 ($25,762 - $25,762), that is, he would break even, receive no subsidy and pay no taxes, leaving him with a total income of $25,762. On the third hand, someone with an income greater than the subsidy would have to pay taxes; an individual with an income of $957,834, for example, would be left with a taxable income of $932,072 ($957,834 - $25,762) after the deduction. This program is tailored to aid people with an income under the poverty line while phasing out the amount of the subsidy as their income increases.
We now know how each different program of guaranteed income is defined. Note that the requirements, levels of subsidies, and tax rates all could be altered to whatever level is desired by the policymakers. But insofar as how they function, what is outlined above is enough to comprehend and analyze them at a fundamental level.
The first question is, would a form of guaranteed income be a good policy to implement in the fight against unemployment and poverty brought about by machines? Many trials of guaranteed income have been tried in a variety of countries: South Africa, India, Kenya, the United States, Canada, Finland, Namibia, Switzerland, Netherlands, Scotland, Spain, Uganda, and others (Gibson et al., 2018; Hoynes & Rothstein, 2019). What does the research say?
Research in Namibia found that child malnutrition and poverty decreased under a UBI pilot (Banerjee et al., 2019). Research in Canada found that among the subjects in an NIT trial, accidents, injuries, mental health, and high school enrollments were all positively impacted (Forget, 2011). Research in India found “impressive improvements in education, healthcare and nutrition” from a Universal Basic Income pilot (Davala et al., 2015, p. 187). Research in the United States found that a UBI program is an efficient way of reducing the poverty rate, especially child and elder poverty (Berman, 2018). Other research has also found positive effects on guaranteed income programs (Garfinkel et al., 2006; Hanna & Olken, 2018; Painter, 2016). In short, “findings are generally positive that [guaranteed income] programs alleviate poverty and improve health and education outcomes” (Hasdell, 2020, p. 18). Further, implementing a safety net, such as a form of guaranteed income, has indirect as well as direct effects. The direct effect is obvious, individuals receiving the transfer payments are benefiting; the indirect effect, somewhat less obvious, is that the citizens who are not receiving any form of payment still benefit in risk reduction, i.e. they feel marginally financially secure and able to take on more risks, given that they know that if they fail they will be “caught” by the net — Forget (2011) describes guaranteed income as an “insurance policy” in the sense that individuals benefit even if they never collect from it. Safety nets not only provide financial security but also allow individuals to learn new skills, such as coding, that enable them to rejoin the labor force and become productive members of society. Based on the empirical economic literature, it seems that implementing some form of a guaranteed income program might be a sound policy in fighting technological unemployment as well as poverty.
Given that we would like to employ a guaranteed income policy, which one (UBI vs NIT vs MIG) would be preferred? Put another way, which one would be most efficient? Theoretically, all three programs (MIG, UBI, NIT) realize their goal: to ensure that no one in our society has less than a certain predetermined income — in our example, $12,881 per year. It is visible that all programs realize that goal. However, what is invisible, and what we also need to take into account, is how each program affects work incentives.
We shall notice that the NIT and UBI are superior to a MIG. The fundamental difference between MIG and UBI/NIT is that for people being subsidized by the government under a Universal Basic Income or under the Negative Income Tax, an extra dollar earned is always extra money that the individual can spend. Under a Minimum Income Guarantee, conversely, an extra dollar earned does not give the individual any extra disposable income. To put it in a concrete example, suppose we have an individual with an income of $0 that would end up with a total income of $12,881 after the subsidy. And suppose that this individual were to find a job that paid, say, $11,000 a year. How would the UBI/NIT program differ from MIG? Under the UBI program, the government would provide him with a subsidy of $12,881, bringing his total income to $23,881 ($11,000 + $12,881), an income $11,000 greater than he would have had had he decided not to work. Under the NIT program, the government would provide him with a subsidy of $7,381 ([|$11,000 - $25,762|]*0.50), leaving the individual with a total income of $18,381 ($7,381 + $11,000), an income $5,500 greater than he would have had had he decided not to work. Under the MIG program, however, the government would subsidy him the difference between $12,881 and his income, a subsidy of $1,881 ($12,881 - $11,000), leaving the individual with a total income of $12,881 ($11,000 + $1,881), exactly the same income he would have had had he decided not to work. Ergo, it should not be surprising that under MIG, an individual considering whether or not to take up an $11,000 job would be strongly disincentivized to do so.
In short, UBI and NIT have an advantage over the MIG in that they do not eliminate entirely the incentive to work. Yes, people will have a base salary of $12,881 if they choose not to work, but under Universal Basic Income and the Negative Income Tax, an extra dollar earned is always extra money that can be spent. Theoretically, therefore, a UBI or NIT would be preferred over MIG.
We must note that there are some proponents of guaranteed income that purport that a guaranteed income program would not disincentivize people to work. We have already laid the groundwork theoretically to disprove that statement, or at least to be highly skeptical of it. But empirical economic research disagrees with the statement as well. Plenty of studies on trial programs have consistently found that recipients of a form of guaranteed income, such as UBI, NIT, or MIG, lower their labor supplied to the market (Gilbert et al., 2018; Heffernan, 1972; Keeley & Robins, 1979; Robins, 1985). The drop in labor supplied tends to be modest, but it must be noted that most of these studies suffer from transitory bias, where the results from a temporary experiment do not necessarily reflect what would happen under a permanent one — I expect that under permanent programs, when individuals feel a greater sense of financial security, the labor supply would be negatively impacted by a greater amount. Additionally, many proponents of guaranteed income, such as Andrew Yang, the 2020 Democratic presidential candidate that made UBI the center of his campaign, claim that the financial security provided by UBI would encourage individuals to spend their time engaging in more artistic work, such as an “expansion of painting, making music, shooting videos, playing sports, writing, and all of the creative pursuits many Americans would love to try, but can’t seem to find the time for today” (Yang, 2018, p. 231). Yet, ceteris paribus, the only way to spend more time exploring one’s artistic nature is by spending less time on the job — that is, by lowering the labor supplied to the market. We must, therefore, bear in mind that any form of guaranteed income policy implemented will almost certainly have a negative impact on the labor supply.
With MIG discarded, an NIT or UBI seem to be the top policy options. Although these two policies tend to be similar, there are fundamental differences between them. Differences range from costs to moral philosophies, but those differences go beyond the scope of this paper.
In this paper, we have discussed the positive effects of implementing a form of guaranteed income: positive financial, emotional, and physical impacts on the citizenry, and we have noted that we must not forget about incentives: negative labor force impacts. However, the benefits are only half of the equation — the other half, and perhaps the more important half, are the costs of implementing such policies. Further research is required: in Part 2, I will be exploring the costs of such policies and concluding on their feasibility. But for now, it seems that economic literature, research and theory, inform us that a form of guaranteed income, such as UBI or NIT, might be a good way of alleviating poverty and fighting against the unprecedented levels of technological unemployment that awaits us.