April 9, 2016
A. Mitchell Polinsky, Stanford University and NBER, and Steven Shavell, Harvard University and NBER
The theory of insurance is considered here when an insured individual may be able to sue another party for the losses that the insured suffered and thus when an insured has a potential source of compensation in addition to insurance coverage. Insurance policies reflect this possibility through so-called subrogation provisions that give insurers the right to step into the shoes of insureds and to bring suits against injurers. Polinsky and Shavell show that subrogation provisions are a fundamental feature of optimal insurance contracts because they relieve litigation-related risks and result in lower premiums financed by the litigation income of insurers. This income includes earnings from suits that insureds would not otherwise have brought. The researchers also characterize optimal subrogation provisions in the presence of loading costs, moral hazard, and non-monetary losses.
Adriano A. Rampini and S. Vish Viswanathan, Duke University and NBER, and Guillaume Vuillemey, HEC Paris
Rampini, Viswanathan, and Vuillemey study risk management in financial institutions using data on hedging of interest rate risk by U.S. banks and bank holding companies. Theory predicts that more financially constrained institutions hedge less and that institutions whose net worth declines due to adverse shocks reduce hedging. The researchers find strong evidence consistent with the theory both in the cross-section and within institutions over time. For identification, they exploit net worth shocks resulting from loan losses due to drops in house prices. Institutions which sustain such losses reduce hedging substantially relative to otherwise similar institutions. The researchers find no evidence that risk shifting, changes in interest rate risk exposures, or regulatory capital explain hedging behavior.
Michael Geruso, University of Texas at Austin and NBER, and Timothy J. Layton, Harvard University
Upcoding manipulation of patient diagnoses in order to game payment systems has gained significant attention following the introduction of risk adjustment into U.S. insurance markets. Geruso and Layton provide new evidence that enrollees in private Medicare plans generate 6% to 16% higher diagnosis-based risk scores than they would generate under fee-for-service Medicare, where diagnoses do not affect payments. The researchers' estimates imply upcoding generates billions of dollars in excess public spending annually and significant consumer choice distortions. They show that coding intensity increases with vertical integration, reflecting a principal-agent problem faced by insurers, who desire more intense coding from the physicians with whom they contract.
Benjamin L. Collier and Erwann Michel-Kerjan, University of Pennsylvania; Daniel Schwartz, University of Chile; Howard Kunreuther, University of Pennsylvania and NBER
Collier, Schwartz, Kunreuther, and Michel-Kerjan examine whether households' risk preferences differ for small and large stakes losses. The researchers develop a structural model and estimate household loss distributions to analyze decisions for a continuum of risk using flood insurance data comprising 16 million policies. Each household makes two contract decisions: the deductible amount (ranging from $500 to $5,000), which indicates its attitudes toward small losses; the coverage limit (the maximum the insurance would pay in a claim), which indicates its attitudes toward the risk of large losses. Testing several value functions and allowing for the possibility that households distort probabilities, the researchers find that households' risk preferences are inconsistent for decisions involving small versus large stakes. For example, households' deductible and coverage limit decisions imply different coefficients of relative risk aversion. Both decisions are marked by overweighting of small probabilities and diminishing sensitivity to losses. However, households exhibit greater diminishing sensitivity to losses and overweight small probabilities more when selecting a deductible than when selecting a coverage limit. The authors conclude that despite making these low and high stakes decisions concurrently, households treat them as separable choices toward which they have different risk attitudes.
Juan Pablo Atal, University of California at Berkeley
Long-term health insurance contracts have the potential to efficiently insure against reclassification risk, but at the expense of other limitations like provider lock-in. This paper empirically investigates the workings of long-term guaranteed-renewable contracts, which are subject to this tradeoff. Individuals are shielded against premium increases and coverage denial as long as they stay with their initial contract, but those that become higher risk are subject to premium increases or coverage denials upon switching, potentially leaving them locked-in with their original network of providers. Atal provides the first empirical evidence on the importance of this phenomenon using administrative panel data from the universe of the private health insurance market in Chile, where competing insurers offer guaranteed-renewable plans. He fits a structural model to yearly plan choices, and jointly estimates evolving preferences for different insurance companies and supply-side underwriting in the form of premium risk-rating and coverage denial. To quantify the welfare effects of lock-in, Atal compares simulated choices under the current rules to those in a counterfactual scenario with no underwriting. The results show that consumers would be willing to pay around 13 percent more in yearly premiums to avoid lock-in. Finally, Atal studies a counterfactual scenario where guaranteed-renewable contracts are replaced with community-rated spot contracts, and finds only minor general-equilibrium effects on premiums and on the allocation of individuals across insurers. He argues that these small effects are the result of large levels of preference heterogeneity uncorrelated to risk.
Anna V. Chorniy, Princeton University, and Daniel Miller and Tilan Tang, Clemson University
Chorniy, Miller, and Tang examine horizontal mergers amongst Part D insurers with the aim of decomposing market power, cost efficiency, and bargaining power effects. The researchers apply a differences-in-differences identification strategy to panel data on plans offered between 2006 and 2012 to analyze the effects of mergers on plan premiums and coverage characteristics related to drug access and drug pricing. The results reveal significant market power raising premiums, but only in markets where the merging firms overlap. Mergers alter the bargaining process with drug suppliers, invoking a tradeoff between drug access and drug pricing. Merging firms realize large bargaining gains when they restructure the firm by consolidating plans. Plan consolidation also stimulates cost efficiencies, even when carried out organically by non-merging insurers. Otherwise, mergers have no cost efficiency effects.
Naoki Aizawa, University of Minnesota, and You Suk Kim, Federal Reserve Board
Aizawa and Kim study impacts of advertising as a channel of risk selection in Medicare Advantage. The researchers show evidence that both mass and direct mail advertising are targeted to achieve risk selection. They develop and estimate an equilibrium model of Medicare Advantage with advertising to understand its equilibrium impacts. They find that advertising attracts the healthy more than the unhealthy. Moreover, shutting down advertising increases premiums by up to 40% for insurers that advertised by worsening their risk pools, which further reduces the demand of the unhealthy. The researchers argue that risk selection may make consumers better off by improving insurers' risk pools.
Sabrina T. Howell, New York University
Should government provide risk management services? Howell exploits a natural experiment in which the government fully insures highway paving firms against oil price risk. The data permit the first firm-level study of risk management among private firms. Surprisingly, the risk shifting policy reduces costs. Capital requirements and financial frictions prevent firms from accessing derivative markets. Howell shows that private firms value hedging more than public firms, but that family owned firms are no more risk averse than non-family owned firms. Howell finds that the cost of managing risk is especially high for high credit risk and undiversified firms. The results are consistent with financial constraints and distress costs leading firms to value hedging.
Viral V. Acharya, Thomas Philippon, and Matthew P. Richardson, New York University and NBER
Acharya, Philippon, and Richardson describe a framework to measure the systemic risk of a financial firm that allows for externalities arising from (i) the contribution of the firm to aggregate capital shortfall, which results in a loss of going-concern values; and, (ii) the undertaking of fire-sales by the firm which destroys the value of assets in place. The researchers use this framework to conclude that since the traditional insurance business involves holding highly diversified assets and is not subject to short-term creditor runs, it is unlikely to be as systemically risky as the banking sector. However, the increasing concentration of insurance firms' portfolios in corporate bonds and structured fixed-income products as well as the increasing proportion of their liabilities that are subject to withdrawals suggests a reconsideration of the insurance sector's systemic risk relative to the banking sector. The researchers discuss empirical implications of their framework that can be used to assess these conclusions.
Darius N. Lakdawalla, University of Southern California and NBER; Julian Reif, University of Illinois at Urbana-Champaign; and Daniel Bauer, Georgia State University
Economic models of risks to life and health are widely employed to assess the costs and benefits of public policies. Yet, these models fail to explain a number of common phenomena, such as the large fraction of spending that occurs near the end of life and a positive correlation between fatality risk and the willingness to pay for life-extension. These and other anomalies have led some authors to abandon the standard framework and posit a variety of ad hoc alternatives. However, Lakdawalla, Reif, and Bauer show that the standard framework accounts naturally for these phenomena, as soon as one relaxes its restrictive and unrealistic assumption of complete annuity markets. This generalized life-cycle model predicts that a fixed survival gain is worth more to individuals facing bleaker survival prospects, and vice-versa. This insight yields a number of novel policy implications. First, the commonly observed practice of spending disproportionately more resources on individuals facing limited life expectancies may be efficient, not problematic. Second, conventional methods currently used by public and private health insurers undervalue life-extension for severely ill patients compared to the moderately ill, because the value of a statistical life-year varies with the level of mortality risk. Third, public annuity programs like Social Security are strong complements for investments in retiree healthcare, because they increase the value of remaining life for the elderly. For example, annuitization roughly doubles the value of statistical life for a typical 85-year-old American. Finally, holding life-extension benefits constant, treatments are more valuable than preventive technology, because they extend life for people that value it the most.
Pietro Tebaldi, Stanford University
The recent U.S. health care reform (Patient Protection and Affordable Care Act; ACA) instituted a subsidy program that provides discounts on health insurance premiums for low-income households. Any evaluation of the ACA or other related programs necessitates the estimation of consumer demand and the way subsidies affect insurers' costs and market power. To do so, Tebaldi combines data from the first year of the Californian ACA exchange where 90% of buyers receive federal subsidies with an equilibrium model of insurance demand and imperfect competition. Taking advantage of ACA pricing restrictions and regional variation in market composition, the author identifies and estimates demand and cost, and then assess outcomes under the current and alternative subsidy designs. Tebaldi finds that, compared to a voucher program, the current design leads to a substantial increase in market power: Price increases do not translate one-for-one to premium increases, and this leads to a less elastic demand response, hence higher markups. The researcher also shows that varying subsidies by age could be an attractive possibility: Increasing participation of younger individuals estimated to be cheaper to cover and more price sensitive reduces market power and average cost. Tebaldi estimates that this implies that participation is 30% greater and per-insured public expenditure is 15% lower when compared to the current subsidy design.
Amanda Starc and Robert Town, University of Pennsylvania and NBER
Starc and Town show that profit-maximizing firms alter product design in the market for Medicare prescription drug coverage to account for underutilization by consumers. Using plausibly exogenous variation in coverage, the researchers examine prescription drug utilization under two different plan structures. They document that plans that cover all medical expenses spend more on drugs than plans that are only responsible for prescription drug spending, consistent with drug spending offsetting some medical costs. The effect is driven by drugs that are likely to generate substantial offsets. The authors' supply side model confirms that differential incentives across plans can explain this disparity. Counterfactuals show that the externality created by stand-alone drug plans is $405 million per year. Finally, the researchers explore the extent to which subsidies and information provision can mitigate the externality generated by under-consumption.