Here at Economic Forces, we like to write about externalities. One reason that I like to write about externalities is that I think we often miss important aspects of externalities in our standard textbook examples. Even what seems like a simple example isn’t always that simple.
Let’s consider vaccines, for example. (This example violates my two rules about writing posts about current events and writing posts about things that people care about, but we all make sacrifices.) This seems like it would be something that is easy to understand. People who know even the slightest bit about vaccines seem to have some intuitive sense that vaccines have something to do with externalities. However, I think that there are factors that largely go unnoticed in the discussion.
Suppose that some new virus just unexpectedly appears one day. (It could happen. I’ve seen it in movies.) A vaccine is developed. Now assume that the way the vaccine works is by reducing the transmissibility of the virus. Anyone can still transmit the virus to others, but people who are vaccinated are less likely to transmit it. People who are unvaccinated are more likely to transmit it. People who are vaccinated are also less likely to get sick from the virus.
Now, consider the framing of the issue. On the one hand, you could think of vaccines in the following context. Whenever anyone goes in public, they are creating a cost for themselves because they are increasing the likelihood that they get the virus. However, they are also creating a cost to others by making it more likely that others get sick. If they fail to internalize the cost to others, then people will tend to go out in public too much and more people will get the virus than is socially optimal.
If you view things this way, those who are unvaccinated are imposing a greater social cost than those who are vaccinated. To eliminate this additional social cost, you would want to tax the unvaccinated for going in public.
On the other hand, you could think of it this way. By getting vaccinated, people are reducing the likelihood that the virus spreads. The benefit of their actions exceeds the private benefit. There is some additional social benefit beyond the private benefit. If people fail to internalize this additional social benefit of vaccination, then there will be too few people vaccinated.
If you view things this way, those who are unvaccinated fail to see that their actions convey a benefit to others in addition to themselves. In order to achieve this public benefit, you would want to subsidize people to get vaccinated.
I could make an argument here that the only difference between these two ways of viewing things is in the framing. In both scenarios, what you are trying to achieve is some optimal level of vaccination. In both scenarios, the number of vaccinations is too low. A sufficiently high tax on the unvaccinated will convince enough people to get vaccinated and to increase vaccinations to the socially desirable level since the way to avoid the tax is vaccination. A sufficiently high subsidy will increase vaccinations to a socially desirable level because people are being compensated for getting vaccinated.
However, the framing is not the only difference here. In a world with zero transaction costs, these scenarios might be just two sides of the same coin, and both result in the socially optimal level of vaccination. In a world with transaction costs, these scenarios differ.
Consider the costs of implementation. In order to get to the optimal level of vaccination, the taxation scheme requires taxing the unvaccinated in proportion to the social cost that they impose. The government cannot simply impose a surcharge on the tax bill of the unvaccinated. Doing so would tax the unvaccinated person who never leaves the house at the same rate as the unvaccinated person whose habits and behavior are largely unchanged after the appearance of the virus. This is not consistent with optimal policy.
One could attempt to put a tax on the in-person consumption of the unvaccinated, but this is also costly. Firms would have to check the vaccination status at cash registers. Furthermore, doing so would incentivize fraud for those unvaccinated people who view the cost of getting vaccinated as greater than the cost of getting a counterfeit vaccination record and the corresponding expected costs of punishment. It is also unclear how one would administer such a tax. A proportional tax on consumption isn’t necessarily accurate since the social cost created depends on how much interaction the consumer has had with other consumers, not how much the consumer spent. (The person who bought a big screen television likely spends less time in the store and comes into contact with fewer people than the person doing grocery shopping.) And that ignores the social and political costs that might be associated with alienating a particular group of people.
The subsidy, on the other hand, promotes voluntary compliance. No one has to check any vaccination records. No resources are wasted obtaining counterfeit vaccination records. This effectively eliminates enforcement costs. However, it creates a cost in terms of the direct expenditure required to pay the subsidy.
These two scenarios are not identical. The first scenario requires significant enforcement costs but does raise revenue (the magnitude of revenue, however, is questionable). The second scenario avoids the enforcement costs but creates an expenditure. One way to assess whether to use the tax or the subsidy would be to consider which scenario achieves the optimal result at the lowest cost.
Of course, if we think about this example in the context of current events, neither of these options is what we have actually observed. The closest thing that has happened is President Biden’s proposed vaccine mandate. This generates significant enforcement costs without any corresponding revenue. Requiring vaccines to be a condition of employment requires the business owner to verify vaccination status. This imposes a cost on the employer. In addition, this might also force the business owner to terminate a valued employee. In some instances, the business might not want to comply because they don’t want to lose these employees. Other firms might not comply because they have some other philosophical or legal objection to the mandate. Other firms might not want to enforce it simply to avoid the costs of enforcement. This means that the enforcement costs are not confined to the firm. The government will have to pay for the cost of enforcement as well.
This doesn’t seem anywhere close to optimal policy.
The astute reader might note that in a previous post, I actually argued that sometimes the optimal thing for the government to do is to use quantity-based regulation rather than taxes and subsidies. But does vaccination satisfy the criteria for quantity-based regulation? No. The case for quantity-based regulation is that when there is one imposer and many victims. When that is the case, the imposer actually has an incentive to increase the social cost to drive the victims out of the market. Since the optimal tax is proportional to social cost, if the victims leave the market, the optimal tax is zero. In this case, quantity-based regulation is the optimal policy. But vaccines don’t fit this criteria. There isn’t one imposer and many victims of the social cost.
Maybe they should try the subsidies.