Abstract
The paper examines efficiency, productivity and scale economies in the U.S. property-liability insurance industry. Productivity change is analyzed using Malmquist indices, and efficiency is estimated using data envelopment analysis. The results indicate 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. More diversified firms and insurance groups were more likely to achieve efficiency and productivity gains. Higher technology investment is positively related to efficiency and productivity improvements.
Similar content being viewed by others
Notes
Previous studies of financial institutions show that although scale economies do exist, the benefits cease to be meaningful beyond a threshold that many companies have already exceeded. Berger and Humphrey (1997) provide a review of the financial institutions efficiency literature, including the literature on scale economies. The insurance efficiency literature is reviewed in Cummins and Weiss (2012).
Cummins and Nini (2002) estimate scale efficiency as part of their study of capital utilization in the P-L industry using data from 1993 to 1998. Choi and Weiss (2005) utilize scale efficiency estimates as regressors in their analysis of market structure in the P-L insurance industry, for the sample period 1992–1998. There also have been several studies that estimate scale economies for P-L insurers in countries other than the U.S. These include Toivanen (1997) (Finland), Fukuyama and Weber (2001) (Japan), Mahlberg and Url (2003) (Austria), Hirao and Inoue (2004) (Japan), Jeng and Lai (2005) (Japan), Cummins and Rubio-Misas (2006) (Spain), and Fenn et al. (2008) (Europe).
This is sometimes referred to as the corporate veil rule because creditors cannot proceed against the corporation or brother-sister subsidiaries unless they succeed in “piercing the corporate veil,” which usually requires showing that the corporation has engaged in fraud or other illegal activities.
The alternative to DEA is parametric analysis, which requires the specification of a functional form for the cost and revenue frontiers and also distributional assumptions about the regression error terms. Although there was some debate in the earlier efficiency literature about whether the parametric or non-parametric approach was more appropriate, research by Cummins and Zi (1998), Elling and Luhnen (2010), and others shows that parametric and non-parametric approaches generally produce consistent results.
The tail refers to the length of the loss payout period, as defined by Schedule P of the National Association of Insurance Commissioners (NAIC) regulatory statements.
We utilize constant maturity Treasury yields obtained from the Federal Reserve Economic Data (FRED) database maintained by the Federal Reserve Bank of St. Louis. Yields are obtained by linear interpolation for maturities where constant maturity yields are not published in FRED.
Let \( P_{i} \) denotes the price of insurance output i, \( PE_{i} \) denotes the real premiums earned of insurance output i, and \( LLE_{i} \) denotes the real present value of losses and loss adjustment expenses incurred of insurance output i, then \( P_{i} = [PE_{i} - LLE_{i} ]/LLE_{i} \).
See Cummins and Xie (2008) for the detailed description of the smoothing process.
Some studies (e.g., Cummins and Nini 2002) use home state wage rate for administrative labors, and state-weighted average weekly wage rate for agent labors. Cummins et al. (1999b) conduct a robustness check for alternative types of wages rate for the U.S. life insurance industry and conclude that using the alternative labor price variables do not materially affect the results.
Equity capital is equal to statutory policyholders’ surplus plus reserves required by statutory accounting principles (SAP) but not recognized by generally accepted accounting principles (GAAP). This includes liability items such as the provision for reinsurance, a reserve required by regulators for transactions with non-U.S. regulated reinsurers.
Such firms are highly specialized and are not participating in the mainstream property-liability insurance market.
The Department of Justice horizontal merger guidelines define a market as moderately concentrated if the Herfindahl index is between 1,000 and 1,800 and concentrated if the index exceeds 1,800 (U.S. Department of Justice 1997). Choi and Weiss (2005) provide evidence consistent with the efficient structure hypothesis for the U.S. P-L insurance industry, i.e., the hypothesis that efficient firms compete more effectively by charging lower prices. Their results do not support the structure-conduct-performance hypothesis that concentration and larger firm size lead to market power and anti-competitive conditions.
The adjacent-year comparisons provide more complete information about the industry as a whole than the three endpoint to endpoint comparisons (1993-to-2009, 1993-to-2006, and 2006-to-2009). Firms had to be present only in the two adjacent comparison years in order to be included in the adjacent-year analysis, whereas firms in the endpoint comparisons had to be present both in the beginning and the ending year of the period. Thus, firms entering or exiting the industry during the comparison period were not included in the endpoint indices. Thus, there is a degree of survivor bias in the endpoint comparisons.
The multiplicative calculation simply multiplies the sixteen adjacent year indices, whereas the additive calculation sums the sixteen indices and subtracts sixteen. Taking the sixteenth root of the multiplicative number gives the geometric mean, and dividing the additive number by 16 gives an arithmetic mean. These calculations are meant to be summary statistics because technically Malmquist indices do not “chain.” Nevertheless, these calculations provide useful information about average productivity growth for the sample period.
These decompositions are exact for individual firms but do not hold exactly for the averages shown in the table.
Two sub-period comparisons were conducted—(1) splitting the overall period roughly in half (1993–2000 vs. 2001–2009), and (2) comparing the pre and post-financial crisis periods (1993–2006 vs. 2007–2009).
Similar results are found for the other years and for the entire sample period of 1993–2009.
To be consistent with the Malmquist approach, we measure cost, scale and revenue efficiency change using ratios as well, i.e. cost efficiency change (t, t + 1) = cost efficiency (t + 1)/cost efficiency (t), and analogously for other types of efficiency. This approach is also used in Cummins et al. (1999b) and Cummins and Xie (2008).
The Breusch-Pagan Lagrange multiplier test showed that unit effects are present in our data so that either fixed or random effects estimation should be used (Greene 2003). Hausman tests clearly reject the null hypothesis that the unit effects are orthogonal with the regressors, implying that random effects estimation would be inconsistent. Accordingly, the regressions are based on two-way fixed effects, with dummy variables for firms and years included in the models. Ordinary least squares estimation produces consistent estimators when used with two-way fixed effects in a panel data model (Greene 2003). Regressions omitting the fixed effects support the same conclusions.
The quantity of equity capital is an input in our efficiency analysis. However, the ratio of premiums-to-equity capital does not measure the quantity of equity but rather its relationship with the amount of premium revenues generated by the insurer. Hence, we do not believe that using equity in both contexts is likely to bias the results. As a robustness test, we also conducted the regressions excluding the premiums-to-equity ratio. The signs and significances of the other variables in the regressions remained unchanged.
The signs and significances of the other variables in the regressions were not materially affected by the omission of the squared size variable.
Specifically, we utilize the sum of admitted and non-admitted electronic data processing (EDP) assets plus furniture and equipment divided by total assets. This variable is suggested in Garven and Grace (2001).
As an alternative technology variable, we tested the ratio of technology expenses to non-commission expenses. This variable is positively related to productivity and efficiency change but is statistically significant only for total factor productivity and revenue efficiency.
The marginal effects are estimated as: \( \frac{{\partial p_{ki} }}{{\partial x_{ji} }} = p_{ki} \left( {\beta_{kj} - \sum\nolimits_{k = 0}^{K} {p_{ki} \beta_{kj} } } \right) \), for k = 0, 1, 2, where \( p_{ki} \) represents the multinomial logit model right hand side and \( \beta_{kj} \) is the coefficient of variable \( x_{j} \) for returns to scale type \( k \). The marginal effects depend on the point of valuation, and their sign can differ from that of the coefficients \( \beta_{kj} \).
References
Aly HY, Grabowski R, Pasurka C, Rangan N (1990) Technical, scale, and allocative efficiencies in U.S. banking: an empirical investigation. Rev Econ Stat 72:211–218
Banker RD (1993) Maximum likelihood, consistency and data envelopment analysis: a statistical foundation. Manag Sci 39:1265–1273
Banker RD, Natarajan R (2008) Evaluating contextual variables affecting productivity using data envelopment analysis. Oper Res 56:48–58
Berger AN, Humphrey DB (1992) Measurement and efficiency issues in commercial banking. In: Griliches Z (ed) Output measurement in the service sectors. University of Chicago Press, Chicago
Berger AN, Humphrey DB (1997) Efficiency of Financial Institutions: international survey and directions for future research. Eur J Oper Res 98:175–212
Berger AN, Mester LJ (1997) Inside the black box: what explains differences in efficiencies of financial institutions. J Bank Financ 21:895–947
Berger AN, Cummins JD, Weiss MA (1997) The coexistence of multiple distribution systems for financial services: the case of property-liability insurance. J Bus 70:515–546
Cho D (1988) Some evidence of scale economies in workers’ compensation insurance. J Risk Insur 55(2):324–330
Choi P, Weiss MA (2005) An empirical investigation of market structure, efficiency and performance in property-liability insurance. J Risk Insur 72(4):635–673
Comment R, Jarrell GA (1995) Corporate focus and stock returns. J Financ Econ 37:67–87
Cooper WW, Seiford LM, Zhu J (2004) Handbook of data envelopment analysis. Kluwer, Boston
Cummins JD, Danzon P (1997) Price, financial quality, and capital flows in insurance markets. J Financ Intermed 6:3–38
Cummins JD (1999) Efficiency in the U.S. life insurance industry: are insurers minimizing costs and maximizing revenues? In: Cummins JD, Santomero AM (eds) Changes in the life insurance industry: efficiency, technology and risk management, Kluwer, Norwell
Cummins JD, Nini GP (2002) Optimal capital utilization by financial firms: evidence from the property-liability insurance industry. J Financ Serv Res 21:15–53
Cummins JD, Phillips RD (2005) Estimating the cost of equity capital for property-liability insurers. J Risk Insur 72(3):441–478
Cummins JD, Rubio-Misas M (2006) Deregulation, consolidation, and efficiency: evidence from the spanish insurance industry. J Money Credit Banking 38:323–355
Cummins JD, Weiss MA (1993) Measuring cost efficiency in the property-liability insurance industry. J Bank Financ 17:463–481
Cummins JD, Weiss MA (2012) Analyzing firm performance in the insurance industry using frontier efficiency and productivity methods. In: Dionne Georges (ed) Handbook of insurance, 2nd edn. Kluwer, Boston
Cummins JD, Xie X (2008) Mergers & acquisitions in the US property-liability insurance industry: productivity and efficiency effects. J Bank Financ 32:30–55
Cummins JD, Xie X (2009) Market values and efficiency in U.S., property-liability insurer acquisitions and divestitures. Manag Financ 35:128–155
Cummins JD, Zi H (1998) Comparison of frontier efficiency methods: an application to the U.S., life insurance industry. J Product Anal 10:131–152
Cummins JD, Grace MF, Phillips RD (1999a) Regulatory solvency prediction in property-liability insurance: risk-based capital, audit ratios, and cash flow simulation. J Risk Insur 66:417–458
Cummins JD, Tennyson S, Weiss MA (1999b) Consolidation and efficiency in the US life insurance industry. J Bank Financ 23:325–357
Cummins JD, Weiss MA, Zi H (1999c) Organizational form and efficiency: the coexistence of stock and mutual property-liability insurers. Manag Sci 45:1254–1269
Elling M, Luhnen M (2010) Efficiency in the international insurance industry: a cross-country comparison. J Bank Financ 34:1497–1509
Fare R, Grosskopf S, Norris M, Zhang Z (1994) Productivity growth, technical progress and efficiency change in the industrialized countries. Am Econ Rev 84(1):66–83
Farrell MJ (1957) The measurement of productive efficiency. J R Stat Soc A (General) 120:253–290
Fenn P, Vencappa D, Diacon S, Klumpes P, O’Brien C (2008) Market structure and the efficiency of European insurance companies. J Bank Financ 32:86–100
Fukuyama H (1997) Investigating productive efficiency and productivity changes of Japanese life insurance companies. Pac-Basin Financ J 5:481–509
Fukuyama H, Weber WL (2001) Efficiency and productivity change of non-life insurance companies in Japan. Pac Econ Rev 6:129–146
Garven JR, Grace MF (2001) The profit efficiency of information technology investment: a comparison of alternative distribution systems in the US property-liability insurance industry, working paper. Georgia State University, Atlanta
Greene WH (2003) Econometric analysis. Prentice Hall, Upper Saddle River
Greene WH, Segal D (2004) Profitability and efficiency in the U.S. life insurance industry. J Product Anal 21:229–247
Grosskopf S (1996) Statistical inference and nonparametric efficiency: a selective survey. J Prod Anal 7:161–176
Hanweck GA, Hogan AB (1996) The structure of the property/casualty insurance industry. J Econ Bus 48:141–155
Hirao Y, Inoue T (2004) On the cost structure of the Japanese property-casualty insurance industry. J Risk Insur 71:501–530
Hughes JP, Mester LJ (1998) Bank capitalization and cost: evidence of scale economies in risk management and signaling. Rev Econ Stat 80:313–325
Jeng V, Lai GC (2005) Ownership structure, agency costs, specialization, and efficiency: analysis of keiretsu and independent insurers in the Japanese nonlife insurance industry. J Risk Insur 72(1):105–158
Johnson JE, Flanigan GB, Weisbart SN (1981) Returns to scale in the property and liability insurance industry. J Risk Insuranc 48(1):18–45
Joskow PL (1973) Cartels, competition and regulation in the property-liability insurance industry. Bell J Econ Manag Sci 4(2):375–427
Katrishen FA, Scordis NA (1998) Economies of scale in services: a study of multinational insurers. J Int Bus Stud 29:305–323
Kittelsen S (1995) Monte Carlo simulations of DEA efficiency measures and hypothesis test, Memorandum, No 09/99, Department of Economics, University of Oslo
Klumpes PJM (2004) Performance benchmarking in financial services: evidence from the U.K. life insurance industry. J Bus 77:257–273
Kneip A, Park BU, Simar L (1998) A note on the convergence of nonparametric DEA estimators for production efficiency scores. Econom Theory 14:783–793
Mahlberg B, Url T (2003) Effects of the single market on the Austrian insurance industry. Empir Econ 28:813–838
Martin JD, Sayrak A (2003) Corporate diversification and shareholder value: a survey of recent literature. J Corpor Financ 9:37–57
McAllister PH, McManus D (1993) Resolving the scale efficiency puzzle in banking. J Bank Financ 17:389–405
Pasiouras F, Tanna S, Zopounidas C (2009) The impact of banking regulations on banks’ cost and profit efficiency: cross-country evidence. Int Rev Financ Anal 18:294–302
Ray SC, Desli E (1997) Productivity growth, technical progress, and efficiency change in industrialized countries: comment. Am Econ Rev 87:1033–1039
Rogers KE (1998) Nontraditional activities and the efficiency of U.S. Commercial banks. J Banking Financ 22:467–482
Schlesinger H (2000) The theory of insurance demand. In: Dionne G (ed) Handbook of insurance. Kluwer, Boston, pp 131–151
Shephard RW (1970) Theory of cost and production functions. Princeton University Press, Princeton
Simar L, Wilson PW (1998) Productivity growth in industrialized countries, (Discussion paper 9810). Université Catholique de Louvain, Belgium
Taylor G (2000) Loss reserving: an actuarial perspective. Kluwer, Norwell
Toivanen O (1997) Economies of scale and scope in the finnish non-life insurance industry. J Bank Financ 21:759–779
United States, Department of Justice (1997) Horizontal merger guidelines, Washington, DC
Weiss MA (1991) Efficiency in the property-liability insurance industry. J Risk Insuranc 58(3):452–479
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Cummins, J.D., Xie, X. Efficiency, productivity, and scale economies in the U.S. property-liability insurance industry. J Prod Anal 39, 141–164 (2013). https://doi.org/10.1007/s11123-012-0302-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-012-0302-2