The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent

Volokh Conspiracy

Markets in Not-Quite-Everything: Post-Pandemic Insurance


Suppose that you own a restaurant. Because you have had to close your business, the COVID-19 pandemic has imposed on you a great loss. It will be too late to buy insurance against the pandemic, of course, and attempts to expand coverage retroactively (as some legislators seek to do with business interruption insurance) are likely to be fruitless. Still, there remains great uncertainty about your circumstances. If your restaurant must remain closed for another year, then your loss will be far greater than if your restaurant will be able to open next month. You can estimate your expected loss by multiplying the probability of various losses by their probabilities and summing, but you face high variance. If you are risk-averse, then you would like to be able to reduce the variance substantially, even if that means a slightly higher expected loss. Thus, you might benefit from some kind of insurance product.

So far as I can tell, there is no readily available product. There is talk about what kind of insurance might be offered for the next pandemic, but little if any discussion of insurance products that one could buy to reduce the impact of this pandemic. Perhaps you could hedge by shorting the S&P 500 or some set of securities hit hard by the pandemic. A portfolio with such unfortunates as Carnival Cruise Lines (CCL) and American Airlines (AAL) might work, or maybe you would short only stocks most related to the restaurant industry. But this is a crude approach. The fortunes of large restaurant chains and those of small restaurants may be only loosely linked, and it doesn't account for the possibility of geographic disparities that might manifest if some states allow restaurants to open before others. It would be awfully hard for the average investor to figure out just how much to buy to hedge optimally. Meanwhile, you might not have the liquidity that would allow you to take short positions.

Alternatively, you might try the opposite strategy of obtaining a personalized cash infusion specific to your own business. For example, you might sell equity in the restaurant to someone capable of assuming some risk, or you might take out a loan. Either way, the counterparty would be betting on your avoiding bankruptcy. The transactions costs of such a transaction, however, are fairly high. The investment may be more personalized than you need. You don't want a minority owner looking over your shoulder, and, though a loan is not an insurance transaction, the prospect of adverse selection might mean that the bank would insist on an extra-high interest rate for your loan. You might prefer a product at an intermediate level of generality, one whose payouts will be inversely correlated with the prospects of business like yours but that will not depend on precisely what happens with your specific business.

For example, one can imagine a derivative based on a measure of the number of restaurants that go out of business in some period in a state or on the number of OpenTable reservations over some period or on any number of other metrics of restaurant business health. Any individual business owner can have only a small effect on this index, so moral hazard is minimal. Meanwhile, the issuer does not need to know any details about any particular business, so transactions costs could be low. If such a product existed, a restaurant owner could reduce risk by investing in these securities, producing an effect similar to shorting the regional restaurant market. It should be easier to figure out how much to invest in such a market than in securities much less connected to the applicable market. If liquidity is an issue, a business owner might still need to sell equity other assets to be able to buy such securities, but less than if the restaurant owner were selling equity solely to reduce risk.

So, why do such markets not exist?

A supply-side explanation is that transactions costs are not so low after all. Conceptually, they should be low, because it isn't all that hard to come up with a measure and define a contract. But in fact there are large legal expenses associated with the creation of such derivative products. There are probably some benefits to scale--selling derivatives corresponding to many different industries and regions--but there might be a large startup cost overall and a not insignificant cost associated with each additional offering. If this is so, the answer might be legal reform to reduce transactions costs. Securities of this sort should not require extensive regulation, so long as there is some decisionmaking scheme specified in the event of dispute, such as arbitration. Perhaps the greatest concern is ensuring that the offeror can pay, but such regulation should be possible on a bulk basis, rather than by focusing on each issuance separately.

The demand-side story is that sellers aren't confident that investors want such products. Robert Shiller, the Nobel winner, has long touted the idea that individuals might invest in bespoke financial products, and he has even succeeded in creating securities based on housing prices in particular metropolitan areas. Lee Fennell has thought deeply about how we might structure such markets to insulate homeowners from fluctuations in local real estate prices. I don't have statistics available, but I suspect that very few homeowners hedge by betting against their local real estate market. Even narrowing down to homeowners who have a high chance of moving a great distance, likely few have purchased such products. If homeowners don't want to hedge local real estate values, why would restaurant owners want to hedge the post-pandemic local restaurant sector?

But why isn't there interest? In principle, in a variety of contexts, local real estate values and pandemic responses among them, derivatives offer insurance-like protection without the need for underwriting based on individual risk assessments or for claim settlement based on individual loss assessments. Of course, there may be good reasons that customers sometimes want to buy insurance products tailored to their own individual risk. I want to protect against the risk that my own home floods, not just the risk that a high level of flooding occurs in my neighborhood. But when the difficulties of crafting such individualized insurance mean that insurance markets don't exist, derivatives seem like they could be a useful substitute. And yet homeowners don't seem to be rushing to the Chicago Mercantile Exchange to bet against their local real estate markets, and restaurants probably wouldn't rush to buy the derivatives of the sort that I describe.

The demand- and supply-side explanations are linked. Conceivably, large demand for such products might someday exist, but only after considerable experience with such products. Consumers (even business purchasers) are fairly conventional when they purchase financial products. People buy the types of financial products that others buy. If a businessperson thought it were common to buy securities that would pay out more when the local restaurant industry faultered, then the businessperson might seriously consider buying such insurance. But even a relatively simple derivative product may be sufficiently opaque that few would want to be the first purchaser. Thus, a seller may need to sell such products initially at a loss or engage in a substantial marketing campaign.

That may seem doable. Firms often invest heavily in products where they will lose money at first in the hope of long-term profits. It might even seem that a pandemic would be an excellent time to introduce such products, given the great amount of uncertainty. But this is not a market with strong network effects or first-mover advantages. To the contrary, there may well be second-mover advantages. If Firm A is successful in creating some class of derivative securities, obtaining approvals, and selling them to customers, then Firm B can offer such securities too. The business model doesn't involve any trade secrets. Firm B might be able to offer the likely purchasers a better deal, because its expenses and risk entering the market were lower. In this scenario, Firm B may be free-riding on Firm A's legal work. It also might be free-riding on any marketing by Firm A explaining the idea behind such a product. John Duffy and I have discussed the possibility that such second-mover advantages may substantially delay market innovations.

The second-mover problem may be especially severe when it is conjectural whether consumers will ever be interested in a product. Maybe a sufficient campaign could succeed in getting restaurants or individuals to buy derivatives of this sort, but there remains a possibility, perhaps even a probability, of failure. Even if there are some first-mover advantages that would allow Firm A to retain greater market share than competitors if successful, Firm A may not be willing to enter the market if there is only a small possibility that sufficient consumer demand would exist. Even if the idea is sound, it may be hard to convince consumers of that. Eventually, the size of the market might be large enough that firms will be willing to bear the cost of being first movers, but that may be a long time from now.

Can government remedy the absence of such products? Not easily. Duffy and I discussed how various forms of intellectual property may provide protection for those who engage in market experimentation, even where the ostensible justification for such protection is different. But one certainly couldn't patent such products, given the general difficulties of obtaining patents for business methods and more fundamentally because the work of Shiller and others has long made the idea of selling such products obvious. I have described how a bespoke intellectual property regime might protect market experimentation, but we don't have such a regime. In principle, the government might have some other role in jump-starting a market of this sort, for example by subsidizing such derivatives markets. The government might well generate more bang per buck doing this than by granting forgivable loans to businesses, but it's hard to motivate public support for subsidizing markets that don't exist.

Still, discussions on insuring the next pandemic are already beginning. Perhaps derivatives markets of this sort deserve some role at least in that discussion. The problem of correlated losses makes it impractical for business interruption insurance to include coverage for pandemics, and so if pandemic insurance emerges, it will likely take some novel form and might well benefit from government subsidy. My view is that business interruption insurance is not an efficient vehicle for future pandemics. It makes sense for highly localized interruptions, where there may be disputes about whether a genuine interruption occurred. Derivative markets are cheaper mechanisms for redressing systematic interruptions that affect not only individual businesses but entire industries regionally, nationally, or globally. There is no fundamental reason that they cannot be used to protect individuals whose livelihoods are tied to the success of industrial sectors, whatever the dangers to those sectors may be.

NEXT: Judge "Begin[s] a Gender Discrimination Trial with a Presentation Highlighting the Great Achievements" of "Our Nation's Civil Rights Leaders"

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. re: "So far as I can tell, there is no readily available product."

    That's not true. All-risks business interruption insurance that covered pandemic risk was available before COVID-19 hit. Not every insurance carrier offered it but some did. Very, very few people bought it because it was expensive. Like, seriously expensive. For what we all thought was a very low-probability risk. The decision not to buy that particular coverage was a rational cost/benefit decision.

    Post-COVID, buyers will probably make different cost/benefit decisions - and that's perfectly okay. And post-COVID, insurance carriers will offer policies that they think their customers want to buy. Some new carriers will probably begin to sell all-risks business interruption insurance. It will still be very, very expensive for a low-probability risk. But it will probably be a lot better than another incomprehensible derivative-based "solution".

    1. I am aware of all-risks BI insurance. But do you know of any post-COVID policies for businesses currently experiencing business interruption?

      1. You mean "my house is on fire - I'd like to buy insurance now"? No, I don't know of any insurance policy that you can buy for any risk after the event has already started.

        If that's not what you meant, can you please restate the question?

        1. Ummm -- flood insurance. While it takes 30 days to go into effect, January snow won't melt until late March...

          1. The "event" in flood insurance is not mere snowfall. While it is true that heavy snows in January might result in heavy melt in March and that heavy melt might result in water damage to your property, it is also possible that the snow will melt more slowly, that the local aquifers will absorb it, that it will be diverted to rivers and other run-off, etc.

            But it's also true that flood insurance isn't market-based insurance anymore. The government took over that market (because politicians weren't happy that people were declining to pay the rates that carriers set for the risk) and has distorted it rather drastically. Flood insurance is at best a weak example.

    2. Although insurance doesn’t exist for everything, mitigating risk is something any business or law firm of any size can invest time and thought into.

  2. Insurance against once-in-100-years events that affect every business in the country at once is unlikely to remain solvent when the event occurs. If it does remain solvent, then the premiums have to be rather high.

    Which means that for 99 years out of 100, every business not buying global pandemic insurance will have a significant advantage over every business with pandemic insurance. A guy starting a business would be ill-advised to buy it until his business got very large. And after that, the premiums would be need to be very large to cover his large business and he should probably still not buy it.

    Any derivative product to hedge the same risk will have the same issues. Would the party you bought it from be able to pay out? Maybe not. And if he could, you will have transferred your ruination onto him — resulting in no overall benefit to the society as a whole.

    Preparing for once-in-100-years catastrophic events is a highly questionable use of any resources at all. Those resources will be wasted the other 99 years. Better to create and maintain resilient social structures that can adapt. That's what you end up needing anyway, and their strength and dynamism helps every year, not just 1 year out of 100.

    1. Plus, the govt will bail you out, so you are a sucker to insure against events like these.

      1. Regardless, it's probably a bad idea to buy.

        I think you will see that the bailouts don't make the businesses whole. Businesses will still lose a lot and many otherwise viable businesses will close. Maybe I will be proven wrong about that.

    2. Wimbledon bought pandemic insurance. Just saying.

  3. Insurance is all & good, but what the slimy lawyer will suggest is forming at least a half dozen corporations, and shifting all the assets to one and the liabilities to the other. Then when someone sues, that person is fighting it out with the unpaid vendors....

  4. Hello, My house burnt down, I would like to buy coverage. Click, dial tone.

  5. What is needed is insurance against government over-reaction, not a pandemic.
    Every virus outbreak in the past was dealt with medically, not by fascism. The costs here are related to government edicts, often unconstitutional. How will an insurance company write that policy?

    1. Anti-government cranks are among my favorite conservative clingers, especially during a pandemic.

      I fear we may lose some of them to participation in Pres. Trump’s Clorox Challenge, though.

      1. Alana Goodman effectively debunked the aquarium cleaner incident.

    2. Longtobefree has neither legal knowledge nor historical knowledge. Quarantines were historically quite common in the United States in response to epidemics.

  6. What is being proposed in legislatures is not “retroactive coverage” but simple payment. For which the insurance companies will get reimbursed by state insurance commissioners.

    1. Can you read? If so, have you ever bothered to? Specifically, the exclusions section of your insurance policy where it spells out in black and white what is not covered under your policy? And does so in language that had to be approved by your state insurance commissioner first?

      If you actually have all-risks insurance and are not getting paid, by all means, enforce your rights under the contract. In court if necessary. But for everyone else, there's nothing to call it but "retroactive coverage".

      Yes, there are some proposals to reimburse the insurance companies for the retroactive coverage. Reimbursement by the state would make the law less unconstitutional. But that ignores the fact that it would be wildly inefficient and will inevitably result in unfair payouts since it ties the payments to pre-COVID business decisions and not to post-COVID needs. If the state is going to make such payments to business owners, it makes far more sense for the state to do so directly.

      The proposal to make the payments through the insurance companies is nothing more than a thinly-veiled attempt to save face by people who really did think that they could push through retroactive coverage.

      1. I view it as hiring the insurance companies to reduce fraud.

  7. "It would be awfully hard for the average investor to figure out just how much to buy to hedge optimally."

    Not any easier for the professional investor employed by an insurance company to hedge effectively. This is why they aren't offering the insurance products you'd prefer. Insurance is about quantifiable risk. They absolutely HATE uncertainty.

  8. Taking a bet on how long a known condition will last is qualitatively different from hedging on the risk of the occurrence of a future event not now occurring. It is more akin to gambling then insurance.

  9. But why isn't there interest? In principle...


  10. One day perhaps people will create prediction markets on platforms such as Augur for this kind of thing. I think the main argument you make here is valid:

    "there are large legal expenses associated with the creation of such derivative products. There are probably some benefits to scale–selling derivatives corresponding to many different industries and regions–but there might be a large startup cost overall and a not insignificant cost associated with each additional offering. If this is so, the answer might be legal reform to reduce transactions costs. Securities of this sort should not require extensive regulation, so long as there is some decision making scheme specified in the event of dispute, such as arbitration."

    This is why regulatory arbitrage is needed to circumvent these unnecessary regulatory costs. You point out correctly that:

    "the greatest concern is ensuring that the offeror can pay, but such regulation should be possible on a bulk basis, rather than by focusing on each issuance separately."

    This is exactly why Ethereum and smart contract technology was created, to guarantee payments on markets. Augur solves the issue with feeding reliable oracle data to the contracts and the markets resolve.

    Perhaps a bigger problem however is that:
    "payouts will be inversely correlated with the prospects of business." Although you may not be betting against your specific business it still feels like you are betting against oneself and people tend to avoid this type of counter-intuitive logic. Insurance IS a bet that loss will occur in the future but it doesn't feel like that. So part of solving the problem will be figuring out how to market what is in reality "a bet against your industry" as an insurance product.

    We need to use technology to lower the barrier of entry to the creation of these types of prediction markets. Hopefully Augur will be capable of this once version 2.0 launches.

Please to post comments