Thursday, March 5, 2020

How to Spend a Trillion Dollars

By Michael Abramowicz - March 05, 2020 at 07:34PM

Suppose that a government has a trillion dollars that it wishes to use to subsidize private spending toward some objective. For example, it might want to launch a massive campaign against the spread of COVID-19. There might be many ways to spend the money, on items large and small–say, research, education, salary for sick people who skip work, quarantine centers, ventilators, hand soap, and so on. The government could create a large administrative agency, writing rules about who is entitled to government checks and performing claim adjudication. But it would be challenging to scale such an agency even under the best of circumstances to handle processing of tens of millions of claims. The challenge is especially acute if we assume that the core goal is difficult to translate into rules, for example because a large number of factors are relevant to assessing the social utility of particular spending projects. If the process is to be governed by standards, there will be inconsistency based on who makes the decision.

In a new article, I describe and defend a novel approach that the government can use to distribute money at scale, without creation of a large bureaucracy and without enacting extensive rules. The paper provides applications focused on climate change rather than on COVID-19, but the approach can be applied to any massive governmental spending program. Here's how it would work: Anyone who claims to have contributed to the specified goal could file a claim. To discourage frivolous claims, a small fee might be applied. Rights to payment on claims could be sold. The government would commit to randomly select a small number of claims, say 1,000. An agency would then estimate social benefits produced for each of these claims, using panels of multiple decisionmakers and considering expert evidence where appropriate. It would then distribute the entire trillion dollars to the claims' owners, in proportion to the measurement of social value for the corresponding claims.

Because claims are transferable and only a tiny percentage of claims will be eligible for reimbursement, intermediaries would aggregate diverse portfolios of claims. This will allow intermediaries to bear the risk associated both with the random selection of only a small percentage of claims and with the unpredictability of the government's assessment of claims randomly selected. An intermediary will pay more for a claim that it expects will be worth more on average, if randomly selected for consideration. A claim is thus worth what it will fetch in the market. An individual or entity might perform actions to meet the government's objective and then create a claim, or it might sell a claim via a contract in which it promises the intermediary that it will invest the money provided by the intermediary in a particular way.

The principal virtue of the system is that it requires very little bureaucratic infrastructure, even if the government is distributing an enormous sum of money to a very large pool of claimants. All the government needs to do is randomly select a very small number of claims and perform adjudications where it estimates the associated social benefits. Moreover, the government need not create detailed rules. The adjudications can be based on a vague standard, such as "estimated social benefits in reducing the spread of COVID-19." Use of a standard means that there will be uncertainty, and this is the primary drawback of the system. But diversified intermediaries can bear the risk of that uncertainty relatively cheaply. Standards should be much more tolerable than in a typical administrative regime, because uncertainty will not impose risk on regulated individuals (who may offload the risk onto intermediaries), and because uncertainty will not increase adjudication costs (because the same number of claims will be adjudicated regardless of the total number filed). As usual, a standard avoids the overinclusiveness and underinclusiveness of rules, thus reducing the danger that funds will be spent inefficiently.

My claim is not that this system is necessarily better than traditional approaches to distributing government funds. My claim is simply that this is a new approach and that it might have advantages in certain contexts. Whether this makes sense for COVID-19 prevention or for any other application depends on how good a job one thinks the government can do with a more traditional, centralized system for spending money directly or choosing private projects to receive government money. This evaluation depends in part on whether one believes that the government can make its assessments relatively free of political considerations, and in part on how expensive it will be for the government to make these determinations. A more traditional system will be preferable when there are relatively few claims so achieving scale is not an issue, and when the purposes of the program can be efficiently translated into rules.

Emergency spending (whether of a trillion dollars or a mere eight billion) is potentially a good application of the market-based approach, because the government may not be equipped to make large numbers of high quality decisions extremely quickly. Eventually, government decisionmakers will need to evaluate spending in a few cases, and bad decisions ex post are possible. The system's performance, however, must be evaluated not based on the actual ex post valuations that the agency will produce, but on the market's ex ante expectations of these ex post valuations. Even if the agency is likely to make many errors ex post, the ex ante expectations might track social value reasonably well. The market process itself will impose costs, as intermediaries will seek to make a profit, but competition will tend to reduce these by driving up the amount that intermediaries offer. Because intermediaries do not need to provide due process, their costs of assessing claims may be less than the costs of relatively formal governmental adjudicative processes. Even if the government is slow, claimants will be able to receive payment quickly from intermediaries, instead of queuing while awaiting administrative determinations.

The law review article describes the functioning of the market and of the government agency in much more detail and responds to objections. I'll look at the comments for the strongest and most recurrent objections and will address these in a subsequent post.

from Reason Magazine Articles