The Federal Reserve Just Published a Paper Comparing UBI and UI

The Federal Reserve Bank of St. Louis just released a November 2014 report comparing universal basic income and unemployment insurance, to determine which is better under varying conditions.

Before I go into their findings and how they arrived at their conclusion, here's their conclusion:

Conclusion

In this paper, we provide a quantitative comparison of an optimal universal basic income policy and an optimal unemployment insurance program in an economy with idiosyncratic shocks. While an unemployment insurance program is better equipped to respond to employment shocks, it suffers from moral hazard and is costly to manage. Monitoring costs are not trivial and add to the social cost of administering the policy. So does moral hazard, by allowing a fraction of agents to abuse the unemployment insurance program. A universal basic income policy, however, has no such social costs and is very simple to manage. . We test the conjecture that a universal basic income may, under moral hazard and in the presence of monitoring costs, perform better. We conduct a wide variety of experiments and compare the social desirability of these policies. . Our results show that an optimal UBI is feasible. Nevertheless, the UI policy is socially robust to the introduction of UBI in the presence of idiosyncratic shocks. It takes empirically implausible monitoring costs and shirking success probabilities for the optimal basic income policy to dominate in terms of welfare the unemployment insurance policy in the economy calibrated to the 2011 United States labor market. With idiosyncratic shocks of smaller amplitude (such as those characterizing the 1990 US labor market dynamics), UBI may represent a more reasonable alternative, even if the UI policy remains socially preferred. . The superiority of UI is anchored in its ability to help those who are most in need. Even a very crude UI system with no asset tests and indefinite eligibility is able to easily beat UBI, which distributes funds blindly and must be financed through distortionary taxation. . Of course, this paper limits the study of universal basic income to its comparison with a UI policy in a particular model. It does not claim to have the last word on UBI systems. Other paths could be explored. We could consider different skills among the population; in this case we have argued that UBI is likely to reinforce the dominance of UI. One could study the impact of transitions. Indeed, we have only compared steady states so far. Transitions can induce large costs in the short-term that can outweigh long-term advantages. Given that the status quo in industrialized countries is typically a UI policy, it is very unlikely that a transition to UBI would prove to be beneficial overall.

First of all, it appears that the definition of "superiority" applied here, is giving the least amount of total money using the least amount of taxation. This appears to represent free-market principles of preferring UI to UBI.

I think it's entirely possible however to fund a UBI in ways that "distort" the market in ways we actually want, like in reducing inequality to levels considered better for GDP growth, reducing financial speculation, making it more expensive to pollute, etc.

So the conclusion of this paper, even though it appears to also hold up UBI fairly highly for its administrative savings and lack of moral hazard, should be recognized as making such an analysis through the above viewpoint of less market intrusion as being inherently better, without any regard for how it intrudes.

As for the real meat of this paper, in regards to the model it uses and its assumptions, this paper is a great example of why economics is called the dismal science.

They use a model that assumes everyone has the same odds of having or not having a job, aka a lottery model. We know this kind of assumption does not work in the real world. A blind/deaf economist would use this model to claim that because everyone has the same odds of being unemployed at a rate of 12.6%, everything must be fine, while unbeknownst to him, riots have broken out around him because the African American population which comprises 12.6% of the population has an unemployment rate of 100%. In this situation a UBI would be far superior to a UI, because no one who has black skin can ever get a job at all. In other words, the real world does not reflect that everyone has the same odds of getting a job, and therefore a model that assumes this, ignores reality.

They use a model built purely on theory and not available evidence even where real data exists and contradicts their assumptions. Their model assumes that a UBI as small as 5% of normal income would lead to a voluntary unemployment rate of 4%. So four percent of the entire population of workers would voluntarily stop working if they received a few thousand dollars. They therefore set the "optimal" UBI at around $2,000. WTF? We already know from testing work disincentive effects in the US and Canada, that not only do people not entirely drop out of the workforce at all, but that even looking just at the slight work reductions found, it would require a UBI set at 150% of the poverty level, e.g. $18,000 to start seeing worrisome and possibly problematic work reductions. They assume these same problems start as low as $2,000 which is entirely false, and that the result is not working at all instead of reducing hours, which is also entirely false.

One interesting assumption in their model is based off of numbers from Oregon, which is that the cost of administering UI is $500 per person per year. That's 4% of a $12k UBI. This number matches other estimates I've seen of the minimum cost of administration, and therefore the amount we could convert to cash and just give instead of wasting on administration.

Basically my big problem with this paper is that they completely ignored available evidence for work disincentives and instead chose to just use an equation that assumes massive work disincentive effects. It ignores what we already know, and pretends to discover what we already nullified as a valid hypothesis over thirty years ago.

It's great to see The Fed looking at UBI, but it's really disappointing to see them building models entirely lacking in available evidence to inform the model.

Did you enjoy reading this? Please click the subscribe button and also consider making a monthly pledge in support of my daily advocacy of basic income for all.