Academic journal article Journal of Risk and Insurance

The Risk-Sharing Implications of Disaster Insurance Funds

Academic journal article Journal of Risk and Insurance

The Risk-Sharing Implications of Disaster Insurance Funds

Article excerpt


We study the risk-sharing implications that arise from introducing a disaster insurance fund to the cat insurance market. Such a form of intervention can increase efficiency in the private market, and our design of disaster insurance suggests a prominent role of catastrophe reinsurance. The model predicts buyers will increase their demand in the private market, and the seller will lower prices to such an extent that their revenues decrease upon introduction of disaster insurance. We test two predictions in the context of the Terrorism Risk Insurance Act (TRIA). It is already known that the introduction of TRIA led to negative abnormal returns in the insurance industry. In addition, we show this negative effect is stronger for larger and for low-risk-averse firms--two results that are consistent with our model. The seller's risk aversion plays an important role in quantifying such feedback effects, and we point toward possible distortions in which a firm may even be overhedged upon introduction of disaster insurance.


Severe catastrophes frequently fuel the discussion about the necessity of public intervention in the market for catastrophe insurance. One form of public intervention that has emerged in several forms is disaster insurance programs. Popular examples in the United States include the National Flood Insurance Program (NFIP), the California Earthquake Authority (CEA) program, the Florida Hurricane Catastrophe Fund (FHCF), and the Terrorism Risk Insurance Act (TRIA). But the existence of such funds is more diverse across the globe, as recently shown by Cummins and Mahul (2008) in an international survey.

The goal of this article is to illustrate the risk-sharing implications of such disaster insurance funds. If markets were entirely frictionless then an additional risk sharing mechanism would be redundant. However, this should not be expected in the context of catastrophe insurance. There are several market frictions that could be associated with inefficient risk sharing, as summarized in Gollier (2008), enforcing the question of public intervention to begin with. We assume that prices in the private market (absent of intervention) are set at strategic levels, allowing us to abstract from specific market frictions such as solvency issues, transaction costs, or liquidity constraints.

Why study the risk-sharing implications in such a case? The introduction of a disaster insurance fund effectively creates a nontradable asset whose payoffs are correlated with the payoff space offered in the private market. The fund has, however, a different risk and return trade-off. This stems from the possibility that a governmental agency is in a position to time-diversify even catastrophic risks, and can provide government funding should the net balance of the fund be negative, as argued in Lewis and Murdock (1996). Hence, we should expect a feedback effect in the private market with respect to price and quantity, and the first contribution of this article is to quantify this feedback mechanism. This feedback effect is possible since market participants do not face a single, but an entire set of pricing operators they are willing to consider. Therefore, our exercise also belongs to the research topic about adding risk-sharing technologies to an incomplete market setting. For example, Detemple and Selden (1991) and Dieckmann (2011) study market completion in the presence of heterogeneous beliefs, and Bhamra and Uppal (2009) study the effect of adding a derivative in the presence of heterogeneity in risk aversion. All these papers have in common that the introduction of an asset has an effect on already-existing securities, but all agents act as price takers.

The key players in our model are insurance buyers and a seller, as well as a risk-neutral government entity. Our proposed fund has the form of an ex ante program, as argued for by Kunreuther and Pauly (2006). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.