Analysis and valuation of insurance companies

Analysis and Valuation of Insurance Companies Industry Study Number Two Center for Excellence in Accounting and Security Analysis Columbia Business School established the Center for Excellence in Accounting and Security Analysis (CEASA) in 2003 under the direction of Professors Trevor Harris and Stephen Penman. The center aims to be a leading voice for independent, practical solutions for financial reporting and security analysis, promoting financial reporting that reflects economic reality, and encouraging investment practices that communicate sound valuations. CEASA’s mission is to develop workable solutions to issues in financial reporting and accounting policy; produce a core set of principles for equity analysis; collect and synthesize best thinking and best practices; disseminate ideas to regulators, analysts, investors, accountants, and management; and promote sound research on relevant issues. Drawing on the wisdom of leading experts in academia, industry, and government, the center produces sound research and identifies best practices on relevant issues. CEASA’s guiding criterion is to serve the public interest by supporting the integrity of financial reporting and the efficiency of capital markets. Located in a leading university with a mandate for independent research, CEASA is positioned to spearhead a discussion of issues, with an emphasis on sound conceptual thinking and without obstacles of constituency positions. More information and access to current research is available on CEASA’s website at http://www. gsb. columbia. edu/ceasa. CEASA is supported by its generous sponsors–General Electric, IBM, and Morgan Stanley–and gratefully acknowledges the support of these organizations and their recognition of the need for excellence in accounting and security analysis. Industry Study Comment These studies are intended to provide readers with a comprehensive review of the pertinent accounting conventions, academic literature, and approaches to security analysis. This paper does not necessarily reflect the views of the center’s advisory board or the center’s sponsors. Center for Excellence in Accounting and Security Analysis ANALYSIS AND VALUATION OF INSURANCE COMPANIES Doron Nissim; Ernst & Young Professor of Accounting and Finance, Columbia Business School Center for Excellence in Accounting & Security Analysis November 2010 Table of Contents Overview………………………………………………………………………………………………………………………………………………….. 2  1. Business ……………………………………………………………………………………………………………………………………………….. 3  1. 1 Activities and Organization ……………………………………………………………………………………………………………….. 3  1. 2 Products and Services ……………………………………………………………………………………………………………………… 20  1. 3 Distribution Channels ……………………………………………………………………………………………………………………… 25  1. 4 Competition……………………………………………………………………………………………………………………………………. 28  1. 5 Regulation ……………………………………………………………………………………………………………………………………… 30  1. 6 Taxation ………………………………………………………………………………………………………………………………………… 40  1. 7 Risks and Risk Management …………………………………………………………………………………………………………….. 44  2. Financial Reporting and Line-Item Analysis ………………………………………………………………………………………… 68  2. 1 Common-Size Financial Statements…………………………………………………………………………………………………… 71  2. 2 Insurance Reserves and Related Expenses …………………………………………………………………………………………. 76  2. 3 Revenue and Related Accruals………………………………………………………………………………………………………….. 95  2. 4 Deferred Policy Acquisition Costs and Related Expenses …………………………………………………………………….. 97  2. 5 Reinsurance …………………………………………………………………………………………………………………………………. 100  2. 6 Investment Assets ………………………………………………………………………………………………………………………….. 103  2. 7 Separate Accounts…………………………………………………………………………………………………………………………. 110  2. 8 Debt…………………………………………………………………………………………………………………………………………….. 111  2. 9 Derivatives …………………………………………………………………………………………………………………………………… 115  3. Valuation ………………………………………………………………………………………………………………………………………….. 120  3. 1 What Drives Value? ………………………………………………………………………………………………………………………. 120  3. 2 Profitability ………………………………………………………………………………………………………………………………….. 123  3. 3 Accounting Quality ……………………………………………………………………………………………………………………….. 131  3. 4 Growth ………………………………………………………………………………………………………………………………………… 134  3. 5 Cost of Equity Capital……………………………………………………………………………………………………………………. 138  3. 6 Macro, Industry-Wide, and Line-Specific Drivers ……………………………………………………………………………… 143  3. 7 Valuation Models ………………………………………………………………………………………………………………………….. 145  Conclusion ……………………………………………………………………………………………………………………………………………. 162  References ……………………………………………………………………………………………………………………………………………. 163  Acknowledgments …………………………………………………………………………………………………………………………………. 178 1 Overview During 2008 and 2009, the insurance industry experienced unprecedented volatility. The large swings in insurers’ market valuations, and the significant role that financial reporting played in the uncertainty surrounding insurance companies during that period, highlight the importance of understanding insurers’ financial information and its implications for the risk and value of insurance companies. To facilitate an informed use of insurers’ financial reports, this manuscript reviews the accounting practices of insurance companies, discusses the financial analysis and valuation of insurers, summarizes relevant insights from academic research, and provides related empirical evidence. The paper contains three sections. The first section describes the insurance business, including activities and organization of insurance companies, products and services, distribution channels, competition, regulation, taxation, and risks and risk management. The second section discusses how insurance activities are reflected in financial reports. Specifically, for each key line item from insurers’ financial statements, the study provides evidence on the economic significance of the item, reviews the related US accounting principles, discusses earnings quality issues, describes analyses and red flags that inform on the item’s quality, reviews selected research findings, and describes the primary differences between International Financial Reporting Standards (IFRS) and US GAAP. Building on the discussion and analyses in the previous two sections, the third section addresses the valuation of insurance companies. The section starts by discussing the primary drivers of insurers’ intrinsic value, including profitability, growth prospects and cost of equity capital, as well as accounting quality indicators that inform on the reliability of the measured drivers. It then describes relative and fundamental valuation models that translate those fundamentals into value estimates. Finally, in the context of fundamental valuation models, the study presents a template for forecasting the key financial statement line items of insurance companies. This document is rather long and its efficient use, therefore, requires an understanding of the structure and content of the different sections. The first two sections of the document are mostly descriptive, while the final section is primarily prescriptive. All three sections discuss academic papers, often with significant details. To increase the usefulness of the literature review, the papers are discussed in separate categories by main focus. However, many of the studies provide evidence relevant to multiple categories. The subsections containing detailed discussions of academic research usually follow a summary of the main findings and can generally be skipped without loss of continuity. 2 1. Business This section describes the business of insurance. It is divided into seven subsections: the primary activities and organization of insurance companies (subsection 1. 1), the products and services offered by insurance companies (1. 2), distribution channels (1. 3), competition (1. 4), regulation (1. 5), taxation (1. 6), and risks and risk management (1. 7). 1. 1 Activities and Organization Insurance provides economic protection from identified risks occurring or discovered within a specified period. Insurance is a unique product in that the ultimate cost is often unknown until long after the coverage period, while the revenue–premium payments by policyholders–are received before or during the coverage period. Insurance contracts are classified as either property and casualty (PC) or life and health (LH) policies: PC insurance — contracts providing protection against (a) damage to or loss of property caused by various perils, such as fire, damage or theft, (b) legal liability resulting from injuries to other persons or damage to their property, (c) losses resulting from various sources of business interruption, or (d) losses due to accident or illness. LH insurance — contracts that pay off in lump sums or annuities upon the insured’s death, disability, or retirement. Some insurance policies, primarily health-related policies, have both PC and LH characteristics and can therefore be classified as either PC or LH. Most insurance companies specialize in either PC or LH insurance, but some have significant operations in both segments. In addition, while many insurers underwrite reinsurance policies (insurance sold to insurers), some focus on reinsurance as their core activity. Insurers increasingly offer products and services that involve little or no insurance protection, such as investment products and fee-based services. The industry also includes companies that provide insurance brokerage services (sourcing of insurance contracts on behalf of customers). Reflecting this variation in activities, the Global Industry Classification (GIC) system classifies insurance companies as follows: Life and Health Insurers (40301020) — Companies providing primarily life, disability, indemnity or supplemental health insurance. This category excludes managed health care companies, which are included in the Health Care sector. Examples include MetLife Inc. (MET), Prudential Financial (PRU), AFLAC Inc. (AFL), Lincoln National Corp. (LNC), Unum Group (UNM), and Torchmark Corp. (TMK). Property and Casualty Insurers (40301040) — Companies providing primarily property and casualty insurance. Examples include Berkshire Hathaway Inc. (BRK-A&B), Allstate Corp. (ALL), The Travelers Companies Inc. (TRV), Ace Limited (ACE), The Chubb Corporation (CB), Progressive Corp. (PGR), and CNA Financial Corp. (CNA). Multi-line Insurers (40301030) — Companies with diversified interests in life, health, property and casualty insurance. Examples include American International Group Inc. (AIG), Hartford Financial Services Group Inc. (HIG), and Assurant Inc. (AIZ). 3 Reinsurers (40301050) — Companies providing primarily reinsurance. Examples include Reinsurance Group of America Inc. (RGA), Everest Re Group Ltd. (RE), PartnerRe Ltd. (PRE), Arch Capital Group Ltd. (ACGL), Transatlantic Holdings Inc. (TRH), and ReinsuranceRe Holdings Ltd. (RNR). Insurance Brokers (40301010) — Companies providing insurance and reinsurance brokerage services. Examples include AON Corporation (AON), Marsh & Mclennan (MMC), Willis (WSH), Arthur J Gallagher & Co. (AJG), and Brown & Brown Inc. (BRO). The primary purpose of the insurance business is the spreading of risks. Because the risks associated with different policies are not perfectly correlated, the total risk of a portfolio of policies is smaller than the sum of the policies’ risks. Thus, insurance functions as a mechanism to diversify PC and LH risks, similar to the role of mutual funds in diversifying investment risks. In fact, because insurers accumulate substantial funds in conducting their business, they also diversify investment risks for their stakeholders by investing in diversified portfolios. The activities of insurance companies include underwriting insurance policies (including determining the acceptability of risks, the coverage terms, and the premium), billing and collecting premiums, and investigating and settling claims made under policies. Other activities include investing the accumulated funds and managing the portfolio. Investing activities are particularly important for LH insurers; for many LH insurers, the spread between the return on investments and the interest cost of insurance liabilities is the primary source of income. 1 Investment income is also significant for PC insurers. PC insurers accumulate substantial funds due to the time gap between the receipt of premiums and payment of claims, and they invest and manage these funds to generate investment income. This income contributes to earnings and so affects the pricing of insurance policies. The time gap between the receipt of premiums and payment of claims, which creates the so-called float, consists of four components. The first is the time interval between the receipt of premium and the occurrence of insured events. In most cases this component is relatively small, because the duration of PC policies is usually short, six-months to a year. This component of the float is reflected in the financial statements in the balance of the unearned premium liability. The other three components, which vary in importance across PC lines, relate to the gap between the occurrence of insured events and the subsequent payments. Some insured losses are discovered many years after the event (e. g., exposure to asbestos), and in many cases the claim settlement process extends over several years (e. g., medical malpractice litigation). Also, in some cases insurance payments are made over extended periods of time (e. g., workers’ compensation). These three components of the float are reflected in the financial statements in the balance of the reserve for losses and loss adjustment expenses, which insurers are required to accrue when insured events occur. Accordingly, the analysis of the float often focuses on unearned premium (first source of float) and, primarily, the loss reserve (other three sources of float). PC contracts involve greater uncertainty than LH contracts because both the frequency and magnitude of PC claims are more volatile than LH claims. PC losses are highly sensitive to catastrophic events such as hurricanes, earthquakes and terrorism acts, events which typically As discussed in Section 2, the primary liabilities of LH insurers are the liability for future policyholder benefits and policyholder account balances. Interest cost is accrued on both liabilities, although for future policyholder benefits it is included in the benefits expense. 1 4 have limited effect on LH claims. In addition, the required payment for PC insurance claims depends on the insured’s loss (subject to limits), while for LH insurance it is often the face value of the policy. Because PC reserves involve greater uncertainty than LH liabilities, PC insurers hold larger equity cushions and generally invest in less risky assets compared to LH insurers. They also reinsure significant portions of their exposure, issue insurance-linked securities, and arrange contingent capital facilities. In addition, because the timing of PC claim payments is less predictable and generally nearer than that of LH benefit payments, PC insurers invest in more liquid and shorter maturity (and therefore less interest rate sensitive) assets, particularly securities. In contrast, LH insurers often invest significant amounts in long term mortgages and risky securities. Some insurers obtain thrift or banking charters and use the charters to cross-sell related products to insurance clients. LH insurers often use thrift or banking charters to provide trust services which complement life insurance and retirement and estate planning activities. For example, life insurance policies can be used to fund trusts, retirement funds may be direct deposited into checking accounts, and certificates of deposit may be incorporated into asset diversification plans for retirement or estate planning purposes. PC insurers use thrift or banking charters for retail activities such as home mortgages and auto loans, which complement the auto and homeowner lines of insurance offered. Insurance companies are also classified based on their form of ownership, where the primary forms are stock and mutual companies. Mutual insurers, which are owned by their participating policyholders, can issue debentures and similar financial instruments but not common stock. Stock companies are owned by stockholders and can issue debentures, common stock and a wide variety of related financial instruments. Most insurers are stock companies. Examples of mutual insurers include NY Life, Massachusetts Mutual Life, and State Farm (PC). A relatively new, hybrid form of ownership involves a mutual company converting into a mutual holding company with a subsidiary stock company that can issue stock to the public. This form of ownership is allowed in only some states and is uncommon. Two examples are Liberty Mutual Holding Company (PC) and Pacific Mutual Holding Company (LH). Academic Research on Activities and Organization Numerous studies investigate various aspects of insurers’ operating, investing, and financing activities. Another prolific area of research explores differences across organizational structures, primarily stock versus mutual companies. Studies have also looked at issues relevant to the insurance industry overall, including the value of and demand for insurance, the problems of adverse selection and moral hazard in insurance, and the underwriting cycle (PC insurance). I discuss these studies in separate categories by main focus. However, many of the studies provide evidence relevant to multiple categories. Efficiency and Profitability This area of research concerns the success of insurance companies in conducting their operating activities, primarily in terms of efficiency and profitability. Studies examining efficiency consider several dimensions, including cost efficiency, technical efficiency, allocative efficiency, and revenue efficiency. Cost efficiency measures the insurer’s success in minimizing costs by 5 comparing the costs that would be incurred by a fully efficient firm to the costs actually incurred by the firm. Cost efficiency can be decomposed into technical efficiency and allocative efficiency. Technical efficiency measures the firm’s success in using its inputs to produce outputs. 2 Allocative efficiency measures the firm’s success in choosing the cost minimizing combination of inputs conditional on output quantities and input prices. To be fully cost efficient, a firm must operate with full technical and allocative efficiency. Revenue efficiency measures the firm’s success in maximizing revenues by comparing the firm’s actual revenues to the revenues of a fully efficient firm with the same quantity of inputs. Primary factors that affect revenue efficiency include product-line diversification and geographic diversification. Summary of Studies Cummins and Xie (2008a) examine efficiency, productivity and scale economies in the US PC insurance industry over the period 1993-2006. They find that the majority of firms below median size in the industry are operating with increasing returns to scale, and the majority of firms above median size are operating with decreasing returns to scale. However, a significant number of firms in each size decile have achieved constant returns to scale. Over the sample period, the industry experienced significant gains in total factor productivity, and there is an upward trend in scale and allocative efficiency. However, cost efficiency and revenue efficiency did not improve significantly over the sample period. Regression analysis shows that efficiency and productivity gains have been distributed unevenly across the industry. More diversified firms, stock insurers, and insurance groups were more likely to achieve efficiency and productivity gains than less diversified firms, mutuals, and unaffiliated single insurers. Higher technology expenditures increase the probability of achieving optimal scale for direct writing insurers but not for independent agency firms. Cummins, Weiss, Xie and Zi (2010) investigate economies of scope in the US insurance industry over the period 1993—2006. They test the conglomeration hypothesis, which holds that firms can optimize by diversifying across businesses, versus the strategic focus hypothesis, which holds that firms optimize by focusing on core businesses. Scope economies can originate from cost complementarities (including the sharing of inputs such as customer lists and managerial expertise), earnings diversification (which permits the firm to operate with higher leverage ratios), and revenue complementarities (“ one-stop shopping” opportunities for consumers that reduce search costs). On the other hand, operating a conglomerate may increase management and coordination costs, exacerbate principal-agent conflicts, and create cross-subsidization among subsidiaries due to inefficient internal capital markets. The authors test for scope economies by estimating cost, revenue, and profit efficiency using data envelopment analysis (DEA), which measures efficiency by comparing each firm in an industry to a “ best practice” efficient frontier formed by the most efficient firms in the industry. LH outputs are measured using real invested assets, the real value of incurred benefits, and additions to reserves for individual life, individual annuities, group life, group annuities, and accident-health insurance. PC outputs include real invested assets and the present values of real losses incurred for short and long-tail personal and commercial lines. The same inputs are used for each category of insurers — administrative labor, agent labor, materials and business services, and financial equity capital, all measured in real dollars using the Consumer Price Index. The authors regress efficiency scores on control variables and an indicator for strategic focus. They find that PC insurers realize cost scope economies, but these are more than offset by revenue scope diseconomies. LH insurers realize both cost and revenue scope diseconomies. Hence, they conclude that strategic focus is superior to conglomeration in the insurance industry. Elango, Ma, and Pope (2008) investigate the relationship between product diversification and firm performance in the US PC insurance industry using data for 1994-2002. The extent of product diversification shares a complex and nonlinear relationship with firm performance. The findings suggest that performance 2 Technical efficiency is related to, but not identical to, to X-efficiency. X-efficiency is the effectiveness with which a given set of inputs are used to produce outputs. A company is considered X-efficient if it produces the maximum output it can given the resources it employs. 6 benefits associated with product diversification are contingent upon an insurer’s degree of geographic diversification. Gardner and Grace (1993) estimate hybrid translog cost functions for 561 life insurers using data for the period 1985-1990. The resulting residuals are used to determine the relative efficiency of insurers in the sample and are tested to see if they are related to so-called X-efficiencies because of internal and external monitoring, or to other factors related to rent-seeking. Results show a large degree of persistent inefficiency seems to exist among sample firms, the inefficiencies relate to some internal or external monitoring, and rentseeking may be occurring. Liebenberg and Sommer (2008) develop and test a model that explains insurers’ performance as a function of line-of-business diversification and other variables using a sample of property-liability insurers over the period 1995-2004. The results indicate that undiversified insurers consistently outperform diversified insurers. In terms of accounting performance, the diversification penalty is at least 1 percent of return on assets or 2 percent of return on equity. Using a market-based performance measure (Tobin’s Q) the authors find that the market applies a significant discount to diversified insurers. The existence of a diversification penalty (and diversification discount) provides strong support for the strategic focus hypothesis. The authors also find that insurance groups underperform unaffiliated insurers and that stock insurers outperform mutuals. Ma and Elango (2008) investigate the relationship between property-liability insurers’ international operations and their risk-adjusted returns using cross-section and time-series data for the years 1992 through 2000. The findings indicate that the relationship between international operations and performance is contingent upon the degree of product diversification. Insurance companies with focused operations in terms of product lines achieve higher risk-adjusted performance as they increase their exposures to international markets. However, insurers who are highly diversified across product lines face declining returns with greater exposure to international markets. Eling and Luhnen (2010) conduct an efficiency comparison of 6, 462 insurers from 36 countries. They find a steady technical and cost efficiency growth in international insurance markets from 2002 to 2006, with large differences across countries. Denmark and Japan have the highest average efficiency, whereas the Philippines is the least efficient. Regarding organizational form, the results are not consistent with the expense preference hypothesis, which claims that mutuals should be less efficient than stocks due to higher agency costs. Only minor variations are found when comparing different frontier efficiency methodologies (data envelopment analysis, stochastic frontier analysis). Using a sample US property-liability insurance companies that had an IPO during the period 1994 to 2005 and a benchmark sample of private insurers, Xie (2010) finds that the likelihood of an IPO significantly increases with size and premium growth. IPO firms experience no post-issue underperformance in efficiency, operating profitability, or stock returns; they register improvement in allocative and cost efficiency; and they reduce financial leverage and reinsurance usage. Moreover, IPO firms are active in follow-on SEO issues and acquisition activities. The findings are mostly consistent with the theory that firms go public for easier access to capital and to ease capital constraints. New business for PC insurers generates high loss ratios that gradually decline as a book of business goes through successive renewal cycles. Although the experience on new business is initially unprofitable, the renewal book of business eventually becomes profitable over time. Within this context, insurers need to manage their exposure growth in order to maximize long run profitability. D’Arcy and Gorvett (2004) use a Dynamic Financial Analysis (DFA) model, which utilizes Monte Carlo simulation, to determine optimal growth rates of a PC insurer. Economies of Scale Operating efficiency–the focus of the previous section–is affected by the scale of operations. Thus, studies examining efficiency often provide evidence on the relationship between performance and size. This topic also has important implications for mergers and acquisitions, 7 another area of research that has received significant attention. I discuss M&A studied in a separate section below; here I focus on studies that specifically address economies of scale. Summary of Studies Cummins and Weiss (1993) investigate the efficiency of PC insurers by estimating stochastic cost frontiers for three size-stratified samples of property-liability insurers over the period 1980—1988. A translog cost function and input share equations are estimated using maximum likelihood techniques. The results show that large insurers operate in a narrow range around an average efficiency level of about 90 percent relative to their cost frontier. Efficiency levels for medium and small insurers are about 80 and 88 percent in relation to their respective frontiers. Wider variations in efficiency are present for these two groups in comparison with large insurers. Large insurers slightly over-produce loss settlement services, while small and medium-size insurers under-produce this output. The small and intermediate size groups are characterized by economies of scale, suggesting the potential for cost reductions from consolidations in the industry. Toivanen (1997) studies economies of scale and scope in Finnish non-life insurance. The production process is separated into cost and portfolio management functions. Firms expand their branch network to either gain market power or informational advantages. There are diseconomies of scale at firm level and economies of scale at branch level, and economies of scope in production. Large firms in the non-life insurance industry pay a substantial premium to gain market power via branch networks. The retained premiums-curve of portfolio management is U-shaped and a positive function of the number of branches. Investments Although investment income constitutes a large share of insurers’ income (see the statistics in Section 2. 1), relatively few studies investigate the investing activities of insurance companies. This is likely due to the fact that insurers’ investment activities are not particularly different from those of other financial institutions. I review here research that explores investment policies specifically relevant for insurers. Studies that examine accounting issues related to insurers’ investments are discussed in Section 2. 6. Summary of Studies Heyman and Rowland (2006) point out that investment officers of publicly held PC companies wrestle with the question of how best to contribute to shareholder value. One approach is to manage the investments independent of the insurance operations, as if they were a closed-end investment company that happens to be funded by insurance underwriting. Another approach is to invest funds primarily to defease the firm’s liabilities and thus support the insurance operations of a company whose principal value derives from its insurance activities. The authors emphasize the second approach. They argue that the investment policy of most insurance companies should have two primary objectives: (1) immunizing insurance reserves with a fixed-income portfolio and (2) earning “ abnormal returns” on surplus in “ a responsible and disciplined” way. The latter goal means adhering to an asset allocation approach that takes account of the risk-reward tradeoffs presented by a broad variety of investment types as well as the accounting treatment of investment income. The authors further argue that net investment income (“ NII”) is the best benchmark of performance and that active management and portfolio approaches that aim to produce a growing, but relatively stable NII would maximize market value. Pottier (2007) examine the determinants of private debt holdings in the life insurance industry. The results suggest that larger insurers, insurers with higher financial quality, mutual insurers, publicly traded insurers, insurers facing stringent regulation, and insurers with greater cash holdings are more prevalent lenders in the private debt market. 8 Liebenberg, Carson, and Hoyt (2009) examine life insurance policy loan demand in terms of four hypotheses that have been put forth in the literature. In contrast to previous studies that examined aggregate data, the authors use detailed data from the Survey of Consumer Finances that allow an alternative and in some cases more direct examination of policy loan demand based on individual household circumstances. Unlike prior studies, the authors find a significantly positive relation between loan demand and recent household expense or income shocks. By observing actual life insurance holdings and policy loan data for families, the authors provide evidence in support of the policy loan emergency fund hypothesis. The findings are particularly relevant for insurers since the results provide evidence of an increase in policy loan disintermediation following expense and/or income shocks at the household level, conditions that are particularly prevalent during recessionary times due to unemployment or reduced work hours. Such a finding is fundamental for insurers as they account for the effects of economic conditions in their estimates of policy loan demand. The results also suggest that credit scores may be useful predictors of loan demand, and thus insurers may be able to improve their estimates of future policy loan demand by using credit scores or other credit information. Chen, Yao, and Yu (2007) find that active equity mutual funds managed by insurance companies underperform peer funds by over 1% per year. The lower returns of insurance funds are not due to less risky investments; instead insurance funds have lower risk-adjusted returns, and their fund flows are less sensitive to performance when they perform poorly. Across insurance funds, those with heavy advertising, directly established by insurers, using parent firms’ brandnames, or whose managers simultaneously manage substantial non-mutual-fund assets, are more likely to underperform. The authors conclude that insurers’ efforts to cross-sell mutual funds aggravate agency problems that erode fund performance. Governance and Compensation While the use of governance and compensation characteristics as control variables is quite common in studies analyzing insurance companies, a few papers specifically investigate issues related to governance structures and compensation schemes in the insurance industry. Summary of Studies Ke, Petroni, and Safieddine (1999) investigate the relation between CEO compensation and accounting performance measures as a function of ownership structure. Publicly-held property-liability insurers are used to consider the relation for firms with diffusely-held ownership; privately-held property-liability insurers are used to consider the relation for firms with closely-held ownership. The authors find a significant positive association between return on assets and the level of compensation for publicly-held insurers but, consistent with optimal contracting theory, no such relationship for privately-held insurers is found. Their results suggest that within closely-held firms, CEO compensation is less based on objective measures like accounting information and more on subjective measures. Mayers and Smith (1992) compare compensation levels of executives of mutual and stock life insurance companies. They find that (1) the compensation of mutual executives is lower than that of stock executives, (2) the compensation of mutual-subsidiary executives is lower than that of stock-subsidiary executives, and (3) the compensation of mutual executives is less responsive to firm performance than that of stock executives. This evidence is consistent with the existence of differences in corporate investment opportunity sets and resulting differences in required managerial discretion between mutual and stock life insurance companies. Extant research on non-financial service firms indicates that board size is a key determinant of firm performance. PC insurers, however, face a different set of agency costs and a more intense regulatory environment than most non-financial firms. Both of these factors were reinforced by the implementation of the Financial Services Modernization Act in 2000. Pacini, Hillison, and Marlett (2008) document a significant inverse relation between publicly traded PC insurer performance and board size in the post-Financial Services Modernization Act period. Publicly traded PC insurer performance, measured by market-to-book ratio, return 9 on revenues, and the operating ratio, was enhanced for firms with smaller board sizes in 2000 and 2001. The authors find that publicly traded PC insurers on average increased board size in 2000 and 2001. In a postFinancial Services Modernization Act environment, board size appears to be related to publicly traded PC insurer performance, but more research is necessary to develop a complete understanding of its role in P&L insurer corporate governance. Monitoring by outside board members and incentive compensation provisions in executive pay packages are alternative mechanisms for controlling incentive problems between owners and managers. The control hypothesis suggests that if incentive conflicts vary materially, those firms with more outside directors also should implement a higher degree of pay-for-performance sensitivity. Mayers and Smith (2010) provide evidence consistent with this control hypothesis. They document a relation between board structure and the extent to which executive compensation is tied to performance in mutuals: compensation changes are significantly more sensitive to changes in return on assets when the fraction of outsiders on the board is high. Ke (2001) investigates how taxes affect managerial compensation for a sample of privately held insurers whose managers own a large percentage of the firm’s stock during 1989-1996. Shareholder/managers receive two types of income from the firm they own: compensation income as employees, and investment income as shareholders. Although compensation income is taxable to employees and deductible by employers, investment income is subject to double taxation. Thus, the mix of the two is an important tax-planning decision for management-owned insurers. The author finds that as individual tax rates increased relative to corporate tax rates from 1989-1992 to 1993-1996, shareholder/managers paid themselves less tax-deductible compensation relative to a control sample of non-management-owned insurers (i. e., privately held insurers with no managerial ownership). Mergers and Acquisitions Merger and acquisition activity in the insurance industry has been very significant for many years. According to the Insurance Information Institute, 3 the number of merger and acquisition transactions in the US insurance industry during the years 1999 to 2008 has ranged between 255 and 522 annually, with average total annual transactions value of about $40 billion. Similar to other industries, the motivations for M&A transactions in the insurance industry are to increase geographical reach and product range in order to benefit from scale and scope economies, and to obtain financial synergies including benefits due to diversification, size, debt capacity and tax effects. Given the proliferation of M&A activity in the insurance industry, it is not surprising that a large number of studies have examined these activities, including the following. Summary of Studies Akhigbe and Madura (2001) investigate how the market revalues the acquirer, target, and rival insurance companies in response to merger announcements. They find that the target, acquirer and rival insurance companies experience favorable valuation effects at merger announcements. These findings support the signaling hypothesis (i. e., that rivals are also potentially undervalued and may be acquired). The intra-industry effects of insurance company mergers are more pronounced for insurance companies that have a similar size and are located in the same region as the target insurance company. The reasons for mergers and acquisitions in the insurance industry are usually not disclosed by regulators, investors, or managers. BarNiv and Hathorn (1997) explicate that accounting and financial information can explain merger or insolvency decisions in the industry. The study emphasizes that a timely merger can serve as a viable alternative to insolvency. A logit analysis of solvent and insolvent insurers is performed to generate the probability of insolvency for each merged insurer. Timely mergers serve as an 3 Insurance Information Institute, The Insurance Fact Book 2010. 10 alternative to insolvency for 20% to 46% of the merged insurers, which is higher than that found in other industries. Attributes are identified that distinguish merged distressed insurers from insolvent insurers. It is concluded that investors in firms that acquire distressed insurers earn significant negative returns and earn significantly lower returns than investors in firms that sell distressed insurers. Berger, Cummins, Weiss, and Zi (2000) provide evidence on the validity of the conglomeration hypothesis versus the strategic focus hypothesis for financial institutions using data on US insurance companies. They distinguish between the hypotheses using measures of the relative efficiency of joint versus specialized production, which take both costs and revenues into account. The results suggest that the conglomeration hypothesis dominates for some types of financial service providers and the strategic focus hypothesis dominates for other types. This may explain the empirical puzzle of why joint producers and specialists both appear to be competitively viable in the long run. Boubakri, Dionne, and Triki (2008) examine the long run performance of M&A transactions in the property-liability insurance industry. Specifically, they investigate whether such transactions create value for the bidders’ shareholders, and assess how corporate governance mechanisms, internal and external, affect such performance. The results show that M&A create value in the long run as buy and hold abnormal returns are positive and significant after 3 years. While tender offers appear to be more profitable than mergers, the evidence does not support the conjecture that domestic transactions create more value than cross-border transactions. Furthermore, positive returns are significantly higher for frequent acquirers and in countries where investor protection is weaker. Internal corporate governance mechanisms, such as board independence, and CEO share ownership, are also significant determinants of the long run positive performance of bidders. Chamberlain and Tennyson (1998) investigate the prevalence of financial synergies as a motive for merger and acquisition activity in the property-liability insurance industry. Two hypotheses are developed and tested based upon theories of information asymmetries and firm financing decisions (Myers and Majluf 1984): (1) that financial synergies are a primary motive for insurance mergers in general and (2) that mergers motivated by financial synergies will be more prevalent in periods following negative industry capital shocks. The hypothesis that financial synergies are a motive for mergers following negative industry capital shocks receives strong support. Cummins, Tennyson, and Weiss (1999) examine the relationship between mergers and acquisitions, efficiency, and scale economies in the US life insurance industry. The authors estimate cost and revenue efficiency over the period 1988-1995 using data envelopment analysis (DEA). The Malmquist methodology is used to measure changes in efficiency over time. The authors find that acquired firms achieve greater efficiency gains than firms that have not been involved in mergers or acquisitions. Firms operating with nondecreasing returns to scale (NDRS) and financially vulnerable firms are more likely to be acquisition targets. Overall, mergers and acquisitions in the life insurance industry have had a beneficial effect on efficiency. Deregulation of the European financial services market during the 1990s led to an unprecedented wave of mergers and acquisitions (M&As) in the insurance industry. From 1990-2002 there were 2, 595 M&As involving European insurers of which 1, 669 resulted in a change in control. Cummins and Weiss (2004) investigate whether M&As in the European insurance market create value for shareholders by studying the stock price impact of M&A transactions on target and acquiring firms. The analysis shows that European M&As created small negative cumulative average abnormal returns CAARs) for acquirers (generally less than 1%) and substantial positive CAARs for targets (in the range of 12% to 15%). Cross-border transactions were value-neutral for acquirers, whereas within-border transactions led to significant value loss (approximately 2%) for acquirers. For targets, both cross-border and within-border transactions led to substantial valuecreation. Cummins and Xie (2008b) analyze the productivity and efficiency effects of mergers and acquisitions (M&As) in the US property-liability insurance industry during the period 1994—2003 using data envelopment analysis (DEA) and Malmquist productivity indices. The authors seek to determine whether M&As are value-enhancing, value-neutral, or value-reducing. The analysis examines efficiency and productivity change for acquirers, acquisition targets, and non-M&A firms. They also examine the firm characteristics associated with becoming an acquirer or target through probit analysis. The results provide 11 evidence that M&As in property-liability insurance were value-enhancing. Acquiring firms achieved more revenue efficiency gains than non-acquiring firms, and target firms experienced greater cost and allocative efficiency growth than non-targets. Factors other than efficiency enhancement are important factors in property-liability insurer M&As. Financially vulnerable insurers are significantly more likely to become acquisition targets, consistent with corporate control theory. M&As are also motivated to achieve diversification. However, there is no evidence that scale economies played an important role in the insurance M&A wave. Cummins and Xie (2008c) examine the relationship between firm efficiency and stock market reaction to acquisitions and divestitures in the US property-liability insurance industry during the period 19972003. The authors use data envelopment analysis (DEA) to estimate firm cost and revenue efficiency. Abnormal returns are measured using the standard market model event study methodology. The authors then conduct multiple regression analysis with cumulative abnormal returns as dependent variables and efficiency and control variables as regressors. The results show that efficient acquirers and targets have higher cumulative abnormal returns but inefficient divesting firms have higher cumulative abnormal returns. The findings are consistent with insurance acquisitions and divestitures being driven primarily by value-maximizing motivations and also show that frontier efficiency provides relevant information for value-maximizing managers. Elango (2006) examines the shareholder wealth effects of 52 international acquisitions in 24 countries undertaken by US firms in the insurance industry during the years 1997-2003. Firms undertaking overseas acquisitions face statistically insignificant negative market returns. The market returns faced by firms during such acquisitions tend to vary by the degree of wealth of the host country, amount of bilateral trade between host and home country, extent of potential liabilities of foreignness (LOF) faced by the firm, and economies of scope. Insurers are likely to face relatively higher positive returns while seeking entry into countries with large size markets and which have extensive trade relationships with the US insurers, but are likely to face negative returns when entering markets that have potential pitfalls of LOF. The market does see higher risk in acquisitions made in countries characterized by greater differences in culture, environment, legal systems, and geographic distance. Additionally, there is limited evidence indicating that firms gaining scope economies might be able to reduce the negative impact of LOP. Fields, Fraser, and Kolari (2007) examine the viability of bank/insurer combinations for US and non-US mergers between 1997 and 2002. They find positive gains and no significant risk shifts for shareholders of bidding firms, and that higher CEO stock ownership results in less positive gains for shareholders. Floreani and Rigamonti (2001) examine the stock market valuation of mergers in the insurance industry between 1996 and 2000 in Europe and in the US. They form a sample of 56 deals in which the acquiring company is listed. Insurance companies’ mergers enhance value for bidder shareholders. Over the event window (-20,+2), the average abnormal returns for acquiring firms is 3. 65%. The abnormal returns increase with the relative size of deal value. Mergers occurring between insurance companies located in the same European country are not valued positively by the market, while cross-border deals appear to increase shareholder’s wealth. The analysis of a sub-sample of listed bidders and targets reveals that the combined insurance companies experience a 5. 27% gain over the (-20,+2) event window and, consistent with previous findings, target shareholders substantially increase their wealth. Focarelli and Pozzolo (2008) investigate what factors might help explain the internationalization strategy of banks and insurance companies, by comparing the determinants of cross-border M&As in the two sectors in a unified framework. The empirical analysis shows that between 1990 and 2003 the internationalization of banks and insurance companies followed similar patterns. Distance and economic and cultural integration are important determinants for both the banks’ and the insurance companies’ expansion abroad. Comparative advantage also has a prominent role, the more so for banks. The evidence is less supportive of the view that cross-border M&As are more frequent between similar countries, as predicted by the new trade theory. Finally, and most interestingly, the authors find indirect evidence consistent with the hypothesis that implicit barriers to foreign entry are more important in explaining the behavior of banks than that of insurance companies. 12 Organizational Form A significant number of studies examine differences between stock and mutual companies, primarily as they relate to efficiency of operations and risk-taking. Also, several studies investigate the initial and subsequent pricing and performance of demutualization IPOs. Summary of Studies Cummins, Weiss, and Zi (1999) estimate the relative efficiency of stock and mutual PC insurers using nonparametric frontier efficiency methods. Cross-frontier analysis measures the relative efficiency of each organizational form by computing the efficiency of each stock (mutual) firm relative to a reference set consisting of all mutual (stock) firms. The authors test agency-theoretic hypotheses about organizational form, including the managerial discretion and expense preference hypotheses. The results indicate that stocks and mutuals are operating on separate production and cost frontiers and thus represent distinct technologies. Consistent with the managerial discretion hypothesis, the stock technology dominates the mutual technology for producing stock outputs and the mutual technology dominates the stock technology for producing mutual outputs. However, consistent with the expense preference hypothesis, the stock cost frontier dominates the mutual cost frontier. These results are inconsistent with an earlier study by Cummins and Zi (1998), which finds that stock and mutual insurers are equally efficient after controlling for firm size. Cummins, Rubio-Misas, and Zi (2004) provide new information on the effects of organizational structure on efficiency by analyzing Spanish stock and mutual insurers over the period 1989—1997. The authors test the efficient structure hypothesis, which predicts that the market will sort organizational forms into market segments where they have comparative advantages, and the expense preference hypothesis, which predicts that mutuals will be less efficient than stocks. Technical, cost, and revenue frontiers are estimated using data envelopment analysis. The results indicate that stocks and mutuals are operating on separate production, cost, and revenue frontiers and thus represent distinct technologies. In cost and revenue efficiency, stocks of all sizes dominate mutuals in the production of stock output vectors, and smaller mutuals dominate stocks in the production of mutual output vectors. Larger mutuals are neither dominated by nor dominant over stocks in the cost and revenue comparisons. Thus, large mutuals appear to be vulnerable to competition from stock insurers in Spain. Overall, the results are consistent with the efficient structure hypothesis but are generally not consistent with the expense preference hypothesis. Fukuyama (1997) investigates productive efficiency and productivity changes of Japanese life insurance companies by focusing primarily on the ownership structures (mutual and stock) and economic conditions (expansion and recession). This research indicates that mutual and stock companies possess identical technologies despite differences in incentives of managers and in legal form, but productive efficiency and productivity performances differ from time to time across the two ownership types under different economic conditions. Jeng, Lai, and McNamara (2007) examine the efficiency changes of US life insurers before and after demutualization in the 1980s and 1990s. The authors use two frontier approaches —the value-added approach and the financial intermediary approach–to measure the efficiency changes. The results using the value-added approach indicate that demutualized life insurers improve their efficiency before demutualization. On the other hand, the evidence using the financial intermediary approach shows the efficiency of the demutualized life insurers relative to mutual control insurers deteriorates before demutualization and improves after conversion. The difference in the results between the two approaches is due to the fact that the financial intermediary approach considers financial conditions. The results of both approaches suggest that there is no efficiency improvement after demutualization relative to stock control insurers. There is, however, efficiency improvement relative to mutual control insurers when the financial intermediary approach is used. Mayers and Smith (1986) test the implications of different theories of the efficiency of mutual insurance companies by examining the changes in stock prices, premium income, and management turnover that accompany mutualization — the switch from a common-stock to a mutual-ownership structure. The sample used in the analysis includes 30 firms that went through the mutualization process over the period 1879-1968. 13 The results indicate that: 1. growth in premium income from policyholders remains constant, 2. policy lapse rates do not increase, 3. stockholders receive a premium for their stock, 4. management turnover declines, and 5. there is no significant change in product mix. Therefore, no group of claimholders systematically loses in the sample of firms that chose to mutualize. A subdivision of the sample by concentration of ownership prior to mutualization reinforces these findings. It is concluded that, for this sample of firms, mutualization is, on average, efficiency-enhancing. The size distribution of mutual property-liability insurers has a larger proportion of relatively small companies than the size distribution of stock property-liability insurers. Small mutuals are unlikely to offer risk-sharing advantages over conventional insurance, so these firms must offer their members other advantages. Ligon and Thistle (2005) develop a theoretical model showing that these mutuals may offer advantages over conventional insurance in addressing problems of adverse selection. When adverse selection exists, conventional insurers may coexist with small mutuals. Small mutuals may be strictly preferred by lowrisk individuals. The size of the mutuals is limited by asymmetric information problems. Using life insurer data for 1993-1999, Baranoff and Sager (2003) find that stock companies have larger financial leverage and more risky assets than mutual companies. These differences in risk-taking are a (if not the) major reason for the higher ROE of stock companies compared to mutuals. In mutuals, ownership rights are not transferable, which restricts the effectiveness of control mechanisms like external takeovers and increase the importance of monitoring by outside directors. Accordingly, Mayers, Shivdasani, and Smith (1997) find that (1) mutuals employ more outside directors than stocks; (2) firms that switch between stock and mutual charters make corresponding changes in board composition; (3) mutuals’ bylaws more frequently stipulate participation by outside directors; and (4) mutuals with more outside directors make lower expenditures on salaries, wages, and rent. Lai, McNamara, and Yu (2008) examine the wealth effect of demutualization initial public offerings (IPOs) by investigating underpricing and post-conversion long-run stock performance. The results suggest that there is more “ money left on the table” for demutualized insurers than for non-demutualized insurers. The authors show that higher underpricing for demutualized firms can be explained by greater market demand, market sentiment, and the size of the offering. Further, contrary to previous research reporting an average underperformance of industrial IPOs, the authors show that demutualization IPOs outperform non-IPO firms with comparable size and book-to-market ratios and non-demurualized insurers. The authors present evidence that the outperformance in stock returns is mainly attributable to improvement in post-demutualization operating performance and demand at the time of the IPOs. The combined results of underpricing and longterm performance suggest that the wealth of policyholders who choose stock rather than cash or policy credits is hot harmed by demutualization. Stockholders who purchase demutualized company shares either during or after the IPO earn superior returns. These findings are consistent with the efficiency improvement hypothesis. Viswanathan (2006) examines the pricing of initial public offerings (IPOs) that follow insurance company demutualizations. The study finds that on average demutualization insurer IPOs post significantly higher first-day returns than nondemutualization insurer IPOs. These gains would accrue to the initial investors and to those policyholders who receive compensation in the form of shares in the newly created stock insurer. Attractive returns are sustained for both groups of insurers during the first few years after IPO. Value of and Demand for Insurance Most products and services provide or are expected to provide ex-post benefits. Insurance, in contrast, provides ex-post benefits only if low probability events occur. Therefore, the value of insurance to customers may not be as clear as the value of other products or services. Accordingly, studies have investigated the value of and demand for insurance. 14 Summary of Studies Ligon and Cather (1997) argue that insurance reduces uncertainty regarding future wealth and so allows insureds to make better decisions regarding consumption and investment. This informational value of insurance does not require consumer risk aversion. Lines of insurance with longer resolution periods should impact relatively more decisions and have higher informational value. Empirical tests using data from the property-liability insurance market suggest that the willingness to pay per dollar of coverage (as measured by relative market demand across lines of insurance) is greater for lines of insurance with longer resolution periods consistent with a positive informational value of insurance. Li, Moshirian, Nguyen, and Wee (2007) examine the determinants of life insurance consumption in OECD countries. Consistent with previous results, the study finds a significant positive income elasticity of life insurance demand. Demand also increases with the number of dependents and level of education, and decreases with life expectancy and social security expenditure. The country’s level of financial development and its insurance market’s degree of competition appear to stimulate life insurance sales, whereas high inflation and real interest rates tend to decrease consumption. Overall, life insurance demand is better explained when the product market and socioeconomic factors are jointly considered. Li, Moshirian, and Sim (2003) study the determinants of intra-industry trade (IIT) in insurance services. The article analyzes and measures the magnitude of IIT in insurance services for the US. The empirical results of the determinants of IIT indicate that foreign direct investment in insurance services (FDI) is a significant contributor to the volume of trade in insurance services. These empirical findings confirm the new theoretical trade models that, unlike the traditional trade theory that considered trade and foreign direct investment in insurance services as substitutes, trade and FDI complement each other and hence multinational insurance companies are contributing to an increase in the volume of trade in insurance services. Furthermore, this study shows that trade intensity between the US and its trading partners leads to product differentiation in insurance services and hence an increase in consumer welfare. Cummins and Danzon (1997) develop a model of price determination in insurance markets. Insurance is provided by firms that are subject to default risk. Demand for insurance is inversely related to insurer default risk and is imperfectly price elastic because of information asymmetries and private information in insurance markets. The model predicts that the price of insurance, measured by the ratio of premiums to discounted losses, is inversely related to insurer default risk and that insurers have optimal capital structures. Price may increase or decrease following a loss shock that depletes the insurer’s capital, depending on factors such as the effect of the shock on the price elasticity of demand. Empirical tests using firm-level data support the hypothesis that the price of insurance is inversely related to insurer default risk and provide evidence that prices declined in response to the loss