Market-Based Social Distancing Regulation

When governments end lockdowns, they must determine what set of rules to impose. For example, governments must decide whether some types of businesses must remain closed, whether the maximum number of people in a gathering must be limited, whether to limit the number of people using mass transit, whether to require people to wear masks, whether to mandate periodic testing, whether and how to enforce quarantine of those who test positive, whether to give greater freedom to those who report positive antibody tests, and so forth. Further, governments must decide how these rules must change should circumstances change, for example if there is a local outbreak. Governments, unfortunately, must make decisions under great uncertainty, and we cannot have much confidence that the political considerations facing leaders will give strong incentives to make decisions that would pass a hypothetical cost-benefit calculus.

One policy option is liberation, not imposing any rules at all. This theory might be justified even given an assumption that social distancing is critical to containing the pandemic. Absent requirements, many individuals will still choose to distance socially, because they do not want to become infected with COVID-19, because they do not want to infect others, and because they prefer not to violate social norms. The international progress in fighting COVID-19 has been impressive, but it is difficult to disaggregate the extent to which success has occurred because of governmentally imposed rules and the extent to which success has occurred because of voluntary compliance.  In principle, voluntary social distancing might be optimal or even greater than optimal, especially if social norms are sufficiently powerful. The contrary economic argument would be that voluntary action will be insufficient, because individuals do not fully internalize the costs of their actions on others. This argument seems especially powerful when a single case of COVID-19 transmission may lead to an outbreak affecting many others.

An alternative to the command-and-control and no-regulation opposites would be to have a market-based system of compliance. The goal would be for third parties like insurance companies to have incentives that would lead them to price insurance in a way that will encourage appropriate precaution and discourage the activities most likely to cause transmission. Suppose (to assume a can opener) that the full COVID-19 transmission graph could be perfectly traced. Now suppose that all businesses may open, subject to purchasing a special type of insurance. Under the market-based system, when someone dies or suffers injury from COVID-19, responsibility would be divided among the covered businesses (if any) at which transmission occurred, and the corresponding insurers would collectively pay a fine to the government, a monetary translation of the loss suffered. Municipalities might themselves be required to purchase insurance to cover transmission in public spaces outside of covered businesses, and individuals might be required to purchase insurance if they wished to host neighbors. The basic insight is that in this hypothetical world, the responsibility for devising rules and deciding on how to relax lockdowns would pass from governments to insurers, and higher-risk activities would carry higher prices. Activities with the highest cost and lowest benefits would shut down altogether. Insurers would have some incentive to conduct statistical studies that would allow them to improve their pricing (though perhaps less incentive than is optimal, given the possibility of free-riding on competitors' studies). In a highly dynamic environment, insurers would likely enter into relatively short-term contracts, so that they could update prices based on new information.

There are many regulatory contexts that work approximately like this (consider, for example, John Rappaport's analysis of how insurers effectively regulate the police). But such systems usually arise because there is a perceived need for insurance coverage, not because there is a view that insurers might serve as effective regulators. It seems unlikely that such a system will naturally develop for COVID-19 or, barring a change in legal culture, for future pandemics. We do not impose liability on a business establishment that is a venue where infectious disease transmission takes place, certainly not a strict-liability basis. If such liability existed as a matter of course, then perhaps insurers would take an active role in risk classification here. But it does not, and they do not.

The most obvious practical problem with market-based social distancing regulation then is the can-opener assumption that I made above, that the infection chain can be traced. Can this problem be overcome? In principle, I think that it could be. As part of a test-and-trace program, some reasonable proportion of the public might agree to have their movements traced, with some privacy protection but in a way that permits the government to perform aggregate statistical studies. We might be able to perform statistical studies assessing how individual choices affect the risk of infection. For market-based regulation to work, we need predict only the likelihood that someone becomes infected based on how much time the individual spends in locations insured by different insurance providers. With such a model, statisticians could (pursuant to methods announced in advance) estimate the proportion of COVID-19 costs attributable to each insurer's insured locations and impose fines on insurers accordingly. These fines could then be redistributed to the government, which in turn might use them to subsidize businesses. Insurers of course would earn profits in such a regime, but they would also be performing the service of risk-classification, with far more powerful incentives than governments have to set appropriate taxes.

This system is likely not politically enactable. Maybe that is in part because I have left many details unspecified, and the devil's in the details. But the devil will be in the details of just about any other approach as well. Rather, the system is not politically enactable because it's just not how we generally think about regulation. We generally assume that political actors will regulate directly, rather than relying on actors with strong market incentives to make the appropriate regulatory distinctions. But there are some regulatory areas where market-based mechanisms are common (say, radio frequency allocation and environmental regulation), and an important project is to consider other domains in which market-based mechanisms might operate. Perhaps eventually, market-based regulation will seem less weird and accordingly will become more common.

Still, COVID-19 or other pandemics may be one of the most difficult possible applications of a market-based mechanism. Nonetheless, COVID-19 illustrates a policy process that at its best is only very loosely evidenced-based. In general, market-based regulation may enhance both the collection of evidence and the application of rules based on the presence of risk.