Estimating market values for non-publicly-traded U.S. life insurers

In the insurance industry, the vast majority of individual insurers are not publicly traded and thus their market values, i.e., market values of common stocks, are not available for financial research. Lack of market values is a serious limitation because of the importance attached to market-determined value in the theory of finance. As a surrogate, statutory book value (capital and surplus) has been used extensively in previous empirical research. Some researchers have tried to develop other proxies. Instead of looking for better surrogates, this paper develops methodologies for directly estimating the market values of thousands of non-publicly-traded U.S. life insurers contained in the NAIC annual statement databases. To this end, three distinct classes of life insurers are identified. Class 1 insurers are publicly traded. Their market values are known. Class 2 insurers are not themselves publicly traded but are owned or controlled by publicly traded holding companies. The market values of Class 2 insurers are estimated by allocating the market value of the holding company among its subsidiaries by a methodology that we call the “matching-allocation method”. Class 3 comprises all remaining insurers. Neither publicly traded nor affiliated with a publicly traded company. Class 3 insurers are treated as having missing market values and their market values are estimated by an established statistical methodology called multiple imputation. Instead of producing a single estimate of market value for each insurer, multiple imputation produces multiple estimates. Use of estimated market values in a finance study introduces a source of uncertainty, and multiple estimates permit assessment of that uncertainty. In this study, 30 replicated estimates are produced for each Class 3 insurer. The major tangible product of this study is a collection of 30 datasets, each of which contains a complete set of estimated market values of all individual U.S. life insurers in our sample over 1998-2001. This research is an important first step in attempts to use market values of firms to study general financial issues in the insurance industry.