Testimony of R.J. Lehmann to House Financial Services Subcommittee on Housing, Community Development and Insurance Hearing
Chairman Clay, Ranking Member Stivers and members of the subcommittee,
My name is R.J. Lehmann, and I am editor-in-chief and senior fellow with the International Center for Law & Economics. ICLE is a nonprofit, nonpartisan research center that works with a roster of more than 50 academic affiliates and research centers from around the globe to develop and disseminate academic output and build the intellectual foundation for rigorous, economically grounded public policy.
I am a recent addition to the ICLE staff. My own background is that I have spent the past 17 years as a journalist and public-policy analyst specializing in the business of insurance. That includes running the insurance policy program at the R Street Institute, which I co-founded in 2012.
The COVID-19 outbreak has triggered unprecedented interruption in the operations of businesses across the country and around the world. While roughly 37 percent of U.S. businesses maintain insurance policies to cover the loss of business income due to direct physical damage to a business property, such policies are not designed to insure revenue loss resulting from a pandemic, even where closure is required by a civil authority order. Indeed, many policies contain explicit endorsements clarifying that viruses and bacteria are excluded as causes for business interruption and loss-of-use coverages.
Earlier this year, Congress sought to address the disruption caused by COVID-19 through the Paycheck Protection Program and there have been various efforts to extend further relief to affected employers and employees. But it is understandable that many seek a more permanent solution and look to insurance markets as offering the framework to provide it.
I agree entirely with the analysis that the pandemic has highlighted a massive protection gap in commercial insurance products. I also agree that it is a problem that almost certainly calls for a governmental solution. I would, however, raise the threshold question of whether insurance is actually the best means to accomplish the public policy goals in question.
Insurance is a system of risk transfer, not a system of economic relief. Even if private insurers could provide this coverage—on their own or with government support—it is not clear their incentives would align with public health goals or with the aims members of Congress likely have in mind.
I would urge the subcommittee and Congress generally to proceed deliberatively before erecting structures that may not prove to be well-suited to the crisis we are currently experiencing, much less unforeseen future crises whose nature and scope we cannot know. In sum, do not legislate for the next pandemic when we are still in the midst of dealing with the current one.
I. Are Pandemics Insurable?
In the business of insurance, there are certain general characteristics that determine whether it is possible, in theory, to insure a given risk. These include having a large number of similarly exposed individuals and having losses that are reasonably predictable. A textbook example of an uninsurable risk would be intentional acts, such as arson. You could not transfer to an insurer the risk that you will burn down your own home, because that risk is fully within your control.
Business interruption caused by a pandemic is not uninsurable in the same sense that intentional arson is uninsurable. There were insurance products available to cover loss of business income due to viral contagion before COVID-19 hit our shores, although clients’ interest in those products was reportedly fairly limited. There are still products that offer such coverage now, although the price of coverage has gone up significantly. On the micro level, for any given insurer and any given insured, viral business interruption is an insurable risk.
The problem is at the macro level. There is only a limited amount of capital that insurers would be willing to devote to a risk like pandemics. Some insurers will write some coverage. They might, for instance, cover a restaurant’s risk of food spoilage resulting from an extended shutdown. But they will not and should not gamble their entire balance sheets. And the capacity that the global insurance and reinsurance industry would ever be willing to devote to this risk cannot possibly match its unique scale.
In this macro sense, for a risk to be insurable, it must be possible to manage it through careful underwriting and diversification. Global pandemics make that impossible. They hit every business sector and every geographical region simultaneously. They even degrade the invested assets insurers use to back up their promises. In a scenario where half the global economy shuts down overnight, there is no world in which the insurance industry can single-handedly carry the other half on its back.
The only entity with the financial resilience, the balance sheet and the risk tolerance to offer such assistance is the federal government itself.
II. Moral Hazard and Government Insurance
When I have appeared before this committee in the past, it has been to warn about the dangers of moral hazard that frequently accompany government intervention in insurance markets. The 50-year-old National Flood Insurance Program is a prime example of this danger. By providing insurance coverage to all comers at rates insufficient to match the level of risk, the NFIP encourages development in disaster-prone and environmentally sensitive regions.
While I remain disposed to skepticism about government insurance programs, I do not believe any of the proposals discussed here today—such as Rep. Maloney’s Pandemic Risk Insurance Act (PRIA) or the joint-trades’ Business Continuity Protection Plan (BCPP)—pose much, if any, risk of moral hazard. With or without insurance and with or without government support, there is likely nothing at all a business owner could do to avoid the impact of a pandemic. Indeed, the greater threat is a business that would go out of its way to keep its doors open, despite the dangers that could result.
Which is not to say moral hazard is irrelevant to the pandemic or to how insurance responds to it. Business interruption is far from the only insurance coverage implicated by viral contagion. Most obviously, employers in nearly every state must provide, on a no-fault basis, workers’ compensation coverage for illnesses contracted on the worksite or in the usual course of job duties. Businesses also obtain various commercial liability coverages that could be triggered if they breach a duty of care or otherwise recklessly cause foreseeable harm by exposing a customer or other third party to the virus.
Where a business is a potential nexus of contagion, we should want them to internalize that cost and to adjust their operations in the interest of better protecting public health. That could mean investments in mitigation, adaptation and prevention. It could mean making sure a worksite is well-stocked with personal protective equipment or that the spatial orientation is changed to reduce the risk of infection.
These casualty and liability lines of business exemplify how risk-based insurance rates can serve a regulatory function, providing price signals that encourage businesses to adopt those practices that best protect their employees and others. If Congress is to move forward with creating a federal insurance or reinsurance program to manage pandemic risk, I would strongly urge to focus tightly on the unique challenges of business interruption and not extend it to casualty and liability lines of coverage. It would be extremely unwise to extend public subsidies that could serve to encourage recklessness.
III. A Murky Role for Business Interruption Insurance
The approach proposed by PRIA is to graft coverage for pandemic risks onto the existing structure of business interruption coverage by providing a $750 billion federal backstop for insurers who choose to participate, with the industry retaining only about 5 percent of the total risk. But only a minority of businesses—a little over a third—currently maintain business interruption coverage. Given that the program would be voluntary for insurers to offer and voluntary for insureds to purchase, it is reasonable to assume less than a third would ultimately elect to carry it.
Even for those who do, there are real questions about whether the sorts of claims we can reasonably anticipate policyholders to make would actually be paid. PRIA is a good faith attempt to extend coverage and avoid the sorts of claims disputes that have prompted hundreds of businesses to sue their insurance companies. The program may well extend coverage but there are some predictable areas of conflict that will almost certainly land policyholders back in court.
Business interruption and contingent business coverages are components of commercial property insurance policies. PRIA would ask participating insurers to vitiate standard contract language that excludes claims for viral pandemics. But that would not change a more fundamental presumption of any property insurance policy: that there must be demonstrable physical damage to the insured property.
To be sure, there are legal theories—some of them currently being tested in the courts—that business closures are necessitated by viral contamination of surfaces within the covered property. But whether that is applicable in any given case is going to depend both on the nature of the virus and the nature of the property. If contamination can be easily cured by wiping down surfaces, that is going to be an extremely limited claim. The reality is, creative legal theories aside, most business closures in this pandemic have had nothing to do with potentially contaminated surfaces. They have instead sought to avoid transmission between people. That is not a risk covered by property insurance.
There is also the question of what triggers coverage. Both PRIA and the BCPP proposal tie coverage to public health emergency orders, such as mandated shutdowns. But the experience we have had in this pandemic shows why that is almost certainly insufficient. The initial wave of business closures did not come as a result of mandated shutdowns; they were in response to customers choosing to stay home. A number of states and localities never formally “shut down” businesses at all and yet still suffered precipitous drops in economic activity. As of October, after nearly all states lifted shutdown orders, airport traffic remained down 60 percent from before the pandemic and OpenTable restaurant reservations were down nearly 40 percent.
There is no “business is bad” insurance. Without some sort of external trigger, there is no cause to make a business interruption claim for a business that has merely been depressed, not interrupted.
Again, insurance is risk transfer, not economic assistance. It should give lawmakers pause that PRIA would represent a $750 billion investment of taxpayer dollars in a program that two-thirds or more of businesses will not access, where many claims will still be denied and where the kind of loss that will be most commonly experienced by businesses does not and cannot constitute a claim.
IV. Why Risk-Based Coverage May Be Bad
A central argument for a public-private partnership to support business interruption insurance for pandemics is that, while the federal government can bring its balance sheet to bear, it does not have the insurance industry’s expertise in modeling, managing and mitigating risk. I find myself in the uncomfortable position of critiquing that argument, given that it is one that I myself have made for the entirety of my career in public policy, whether the subject was flood insurance or crop insurance or terrorism insurance.
But it is important to ask: modeling, managing and mitigating the risk of what, specifically? In the case of business interruption insurance, it is not the risk of viral transmission. It is not even the risk of a pandemic, not quite. It is the risk of business closure as a result of a pandemic.
I mentioned earlier that I do not believe there is anything a business owner could do to avoid the impact of a pandemic. What they could do—what risk-based insurance might encourage them to do—is to avoid making a claim by refusing to shut their doors and by pressuring local leaders not to issue mandatory shutdown orders. From a public health perspective, that is the opposite of what we want to happen. And yet, we see it has happened. It is one reason we see the incoherent outcome that, in some cities, schools are closed while bars and restaurants are allowed to remain open.
Like any efficient insurance market, a risk-based insurance market for pandemic business interruption insurance would seek to align the incentives of the insured and the insurer. Among the ways this is generally accomplished is through deductibles, which discourage policyholders from making claims for shallow losses. More broadly, it is accomplished by matching premiums to the level of risk. For example, businesses that could continue operating remotely even in the midst of a pandemic are low-risk and would be offered the most affordable coverage.
On the other hand, risk-based insurance premiums for restaurants, gyms, theatres, barbers, manicurists—any environment where you have close personal contact with strangers or indoor mass congregations of people—would be punishingly expensive. Actuarial science is notoriously complex, but the basics of risk-based premiums are fairly simple: frequency times severity. The severity of a pandemic contagion, even if it happened just once a century, is so extreme that a risk-based premium could not be affordable for the overwhelming majority of small businesses—or even churches and social groups—that rely on in-person human interaction. If they were forced to buy this coverage, many could simply no longer exist. That is not a socially desirable outcome.
The saving grace—the reason we would not likely see that outcome—is itself discouraging. Because the coverage would be voluntary, these sorts of businesses almost certainly would not take it up. Thus, the very businesses who have been hardest hit by this pandemic and would likely be hardest hit in any future one would remain the most exposed.
V. Learning from the Current Pandemic
Proposals like PRIA and the BCPP initially were put forward in the early days of the pandemic. The folly of imagining that lawmakers could have the foresight to craft structures that anticipate future pandemics is just how much has changed in the few months since those proposals were debuted.
I consulted with the insurance trades on the earliest drafts of what became the BCPP. I believe it was my idea to cap the maximum coverage the program would offer at three months of business income. Back in April, that seemed like a generous benefit. Seven months later, with caseloads breaking new records every day and a vaccine at least months away from broad distribution, it seems much less so.
PRIA was originally a $500 billion proposal. It is now a $750 billion proposal. But it is also clear that that amount, while a lot of money for a federal program, is not nearly enough for the scope of the problem. Moreover, PRIA is structured as a single pot of money. Were it in place during COVID-19, it may well have been completely depleted by the earliest phases of the pandemic, when the virus was contained largely to New York and New Jersey. By the time the second wave spread across the Sunbelt in June and July, there may have been nothing left, to say nothing of the third wave we are now encountering.
Any program that Congress does establish should follow some broad principles gleaned from our experience thus far with COVID-19. But we also should be humble about how much we still do not know even about the current pandemic, much less the next one.
The program should endeavor for broad participation, with a bias toward encouraging small businesses, nonprofits and community organizations to take part. Larger enterprises already have available to them a number of insurance options that small businesses do not, from the ability to create captive insurance companies to relatively easy access to bespoke products in the excess and surplus lines market. Indeed, our experience with the Terrorism Risk Insurance Act suggests we should be particularly skeptical of how large companies might use captives to game a structure like PRIA, including for tax-avoidance purposes, with the overwhelming majority of risk passed on to taxpayers.
If there is to be a premium or a participation fee for the program, it should be flat, not risk-based. One common concern of insurance markets is the problem of adverse selection. Because an insured has more information about their own risk than an insurer does, the riskiest businesses are also the most likely to buy coverage. While that is a problem for writing insurance profitably, the public health goals of pandemic response turn that issue on its head. The businesses most at risk of shutting down are the ones to whom we most need to extend a safety net. We want them to cooperate with shutdowns, not push back.
In the spirit of broad participation, the insurance industry should not be the sole marketing force for any federal pandemic risk program. PPP was administered primarily by banks and credit unions and that appears, on balance, to have worked pretty well. There is no reason that lending institutions, payroll processing companies or credit card issuers could not help to sign up participants.
The same applies when it comes to distributing benefits. The insurance industry’s claims-adjustment force is already pushed to capacity to keep up with disasters like hurricanes and wildfires. Adjusting business interruption claims requires special training. Moreover, adjusting claims is a slow and laborious process, which conflicts with the goal of getting money out the door as quickly as possible. A parametric trigger, such as the one in the BCPP, would better accomplish that goal.
The BCPP balances the parametric structure by enumerating specific purposes for which benefits can be used, like rent and payroll. Any disbursed benefits not used for those purposes could later be clawed back. While this is how PPP worked and how the BCPP would work, it is not how business interruption insurance works. A policyholder that makes a claim for business interruption might use the money to continue paying staff, but there would be nothing requiring them to do so. Even with PRIA in place, a business owner could make a claim for interruption while simultaneously placing all his or her employees on furlough. Lawmakers should understand that.
Another question is whether it is wise to create a federal program at all. Given that public health orders are overwhelmingly the jurisdiction of state and local governments, one option would be to allow the states to create their own programs, with the U.S. Treasury partially reimbursing the cost. This would require Congress to establish some minimum guidelines for qualifying programs. But so long as the reimbursement formula was relatively transparent and applied equitably, it would permit innovation and local customization in program design, while also limiting the “run on the bank” danger that a single pot of federal money like PRIA might face.
But above all, my recommendation to lawmakers is to take your time. Perhaps more private solutions, from the insurance industry or some other source, will emerge to meet these challenges before the next pandemic. Perhaps Congress would again have to provide ad hoc assistance. It is profoundly more important that Congress do its job to get assistance to the businesses, workers and communities who need that help right now that it is to pretend to have the answers in 2020 to a crisis of unknown and unknowable dimensions that may befall us in 2025 or 2050 or 2100.
Thank you, and I would be happy to answer any questions.