Summary of the Ph.D. Thesis
AN EMPIRICAL ANALYSIS OF
CAPTIVE INSURANCE COMPANIES
AND RISK MANAGEMENT
January 2006
YUJI MAEDA
Supervisor Professor Yuji Nakano
Supervisor Professor Jituo Goto
Supervisor Professor Isao Ogawa
Graduate School of Economics
Risk Research in Economics and Management
(Doctor Course)
An Empirical Analysis of Captive Insurance Companies and Risk Management
1. Research Motivation about Captives
In the international risk management community, risk management has never been more strategically treated than today. Also, the modern management style concerns risk finance and risk controls more than ever before. Along with the development of risk management, risk finance techniques have also been advanced.
With the development of risk finance and risk control techniques, companies can now tailor their risk management programs to precisely satisfy their needs. One of the advanced risk finance techniques is to utilize “Captive.” Even though risk management is a strategically important and critical for an organization, only recently in Japan have there been no serious discussions about risks and risk management within companies until very recently in Japan. As similarly late response has occurred regarding the establishment of Captive.
A captive is a strategically beneficial risk management tool. Reportedly, more than 5,000 captives exist worldwide. Presently, it is reported that more than 5,000 captives exist worldwide with US $50 billion in annual premiums, US $230 billion in assets and US $202 billion in capitals and surplus, representing a huge market. Most of these captives have been established by U.S., U.K. or European companies. For this new area, Japan is still at the infant stage as we can anticipate it from the facts that the captives owned by Japanese companies count only around eighty.
Almost all academic research, therefore, concerns the U.S., U.K. and European company captives. In these countries, academic research has importantly contributed to the practice by theoretically justifying captives from the viewpoints of the corporate finance, investment, insurance, risk management and tax treatment. We expect that future academic research into this new area will contribute to the business practice in Japan.
Under such circumstances, in 2003, Nago City of Okinawa declared itself to be the first captive domicile in Japan and has been intensively promoting its captive
domicile to Japanese corporations. The promotion of captives spread not only to the multinational corporations but also to the domestic companies. Accelerated by the deregulation of the Japanese insurance industry, captives may have a bright future in this country.
2. Objectives of Captive Research
In reality, little empirical research has ever been conducted on captives owned by Japanese companies. We also observe little theoretical support which leads to the establishment of captives in Japan. Yet, it does not make sense at all to relate the U.S. or U.K. captive research to the Japanese owned captives.
While many research questions need to be answered to justify the captive establishment for the Japanese companies, this research especially concerns the questions that need to be answered urgently.
In response to the urgent demand for the captive research in Japan, we attempt to identify the actually threatening risk to the Japanese companies and model the risk based on the actual data. The research applicability and practicability is thus examined with the real data. Also, we focus the research themes under real risky conditions for the Japanese companies and the results will have effective influence on practice by providing the practically applicable answers.
The research attempts to use the real loss data of a risk and model the risk with several forms of stochastic process models. We virtually form a captive covering the risk identified by the model, by which we compare the captive domiciled in Japan, namely in Nago City, with the same captive domiciled in foreign countries. By examining each captive’s pro-forma financial statements, we demonstrate the distribution of the year-end capital and surplus, and analyze the probability of bankruptcy by the distribution. In this way, we can discuss the qualitative advantages
or disadvantages of the captive in Nago in comparison with those in foreign domiciles to provide the practically applicable answers for the feasibility of the captive in Japan.
3. Corporate Needs for Captives
The dissertation consists of seven chapters. Chapter 1 provides the overviews of the dissertation.
Chapter 2 discusses the economic theories of risk, insurance and risk management, on which our following discussions are founded.
Chapter 3 attempts to reveal the advantages of having captives and then to examine the hurdles for forming captive insurance companies from the viewpoint of Japanese corporations. The chapter listed the numerous advantages of captives to the corporations. They include reducing their total cost of risks, meeting exactly their risk transfer needs, accumulating emergency funds in a tax efficient environment, having future profit opportunities among others.
Especially, Japanese companies are likely to achieve a greater reduction of the current risk costs than the U.S. companies would because with captives they can arbitrage on large insurance price differences between Japan and overseas. This possibility is illustrated by the comparison of the combined ratios in Japan with those in US. Despite those various benefits Japanese corporations seem reluctant to establish captives. The reasons would be due to the comparative advantages that commercial insurance companies have in Japan, the rigid business relationship between the Japanese corporations and the stockholder insurance companies, the deficiency in risk management functions and the lack of entrepreneurship mind at the corporations.
4. Impacts of Captives on Nago and Okinawa
Chapter 4 demonstrated captives from the viewpoint of Nago City, the first captive domicile in Japan. Japanese corporations are likely to prefer Nago City to other domiciles because the city provides greater conveniences with regard to its access, the
captive controllability, communication means and the language. In addition, no premiums, loss reserves nor capital expose exchange rate risks at captives in Nago. Further, special reduced tax and other incentives must be attractive to Japanese corporations. As a captive domicile, Nago City of Okinawa can expect a significant increase in employment and the increase in the tax income.
By the economic growth and the necessity of supporting captives, the enhancement of educations, the improvement of information and communication systems, the protection of environment and maintenance of resort amenity can be expected, improving the standard of living of the local people. An economic simulation conducted in the research identified that one hundred captives would provide the new employments of 2,200 and the additional tax income of yen 233 billion generated over 10 years. Those numbers provide evidence that captives have potentially significant economic impacts to Nago City and Okinawa.
5. Risk Modeling of US Product Liability Risks for the Feasibility Study
Chapter 5 and Chapter 6 provide an analysis as to whether or not, or what extent, Nago City is feasible as the captive domicile in comparison to the other domiciles such as Bermuda and Hawaii. It is assumed that a particular Japanese manufacturer, called “Company A”, is considering establishing its captive insurance company to manage the U.S. product liability risks. Chapter 5 focused on modeling the US product liability risks which is used for the captive feasibility as the ultimate objective discussed in Chapter 6.
Based on the past five year data of the US product liability losses, the most appropriate model is chosen among the five distinct compound Poisson processes: the Inverse Gaussian (IG) type compound Poisson process, the Lognormal type compound Poisson process, the Gamma type compound Poisson process, the Pareto type compound Poisson process and the Weibull type compound Poisson process. The parameters are estimated by the maximum likelihood estimation method and the models
are compared by AIC (Akaike Information Criterion) for the goodness-of-fit to the risk. As a result, the lognormal type compound Poisson process with the maximum likelihood estimated parameters is considered the most appropriate among them to describe the US product liability risks. This model is then used for the feasibility study in Chapter 6.
6. Captive Feasibility Study and Conclusion
Chapter 6 conducted the feasibility study of a captive company for Company A. To examine the feasibility of the captive, it attempted to determine the initial capital for the captive establishment and the net premiums charged to the risk based on the probability of the captive’s bankruptcy of 10 % or lower in the case of the Hawaii domicile. The probability of bankruptcy is obtained from the distribution of capital and surplus created as a result of the Monte Carlo simulations run on the pro-forma financial statement model, in which the lognormal type compound Poisson process model is inserted as the future loss payment model.
The several simulation results identified that the captive company established with the initial capital of US $1 million charging the net premiums of U.S. $40 millions is feasible for Company A given the company’s risk tolerance of 10%. Same simulations were conducted as if the captive was established in Bermuda and in Nago. Those three sets of simulations are compared with respect to the year end capital and surplus as well as the probability of bankruptcy. The results reveal that Nago would be more favorable than Hawaii and Bermuda. Therefore, the quantitative analysis also supports that Nago is feasible as a captive domicile.
The final chapter is left for the summery of the research, the management implications, the Nago’s implications and future research opportunities as a conclusion of my dissertation.
