Stochastic models in life insurance. Translation from the 2nd German edition. (Q665529): Difference between revisions

Stochastic modeling plays an indispensable role in actuarial mathematics. It permeates in various areas in both life and non-life insurance mathematics. The roots of stochastic modeling in life insurance mathematics may be traced back to the fundamental work of Thorvald Nicolai Thiele at the end of the 19th century, where Thiele's differential equation for mathematical reserve in a Markov chain model of life contingencies was established. This work is known as the foundation of modern life insurance mathematics. The roots of stochastic modeling in non-life insurance mathematics, particularly in risk theory, may be traced back to the revolutionary works done in 1909 and 1930, respectively, by Ernst Filip Oskar Lundberg and Harald Cramér. It was a healthy surprise that the work of Ernst Filip Oskar Lundberg appeared in the same decade as the work of Louis Jean-Baptiste Alphonse Bachelier titled ``Théorie de la spéculation'' published in 1900, which is a classic work in mathematical finance. Another healthy surprise is that the works of Thorvald Nicolai Thiele, Ernst Filip Oskar Lundberg and Louis Jean-Baptiste Alphonse Bachelier were done before the foundations of the modern theory of stochastic processes were laid down by Andrey Nikolaevich Kolmogorov. This monograph, as indicated by its title, is concerned with stochastic modeling in life insurance mathematics. Particular attention is given to applications of Markov chains to modeling life insurance policies, valuing and reserving life insurance products. The philosophical attitude of this monograph is to integrate both the theory and practice of stochastic modeling in life insurance mathematics. Concrete examples are used to motivate and illustrate the use of Markov chains in various domains of life insurance mathematics. The monograph also presents a sound mathematical basis for life insurance mathematics. For those who have a good knowledge of measure theory and functional analysis would feel comfortable with the mathematical style of presentation of the monograph. For those who are practically oriented, the monograph also provides numerical examples and computer codes to illustrate the practical implementation of Markov chain models for actuarial uses. Both discrete and continuous-time models are treated, where the former has an advantage for an econometric analysis and the latter provides a solid theoretical basis. The monograph is definitely an asset for both research and practical actuaries. Another distinctive feature of the monograph is that it presents an application of an abstract valuation theory, which plays a fundamental role in mathematical finance, to some contemporary research areas in actuarial science and risk analysis, namely Asset-Liability Management and Solvency II. For those who are enthusiastic in exploring the interplay between actuarial mathematics and mathematical finance, the monograph is a valuable source of reference. In the sequel, a snapshot for each of the chapters of the monograph is provided. The monograph starts with an introductory chapter (Chapter 1) on various products in the life insurance market. Three types of life insurance, namely insurance on life or death, disability insurance and health insurance, are discussed. Chapter 2 presents a review on stochastic processes with emphasis on Markov chains. Interest rate modeling plays a central role in actuarial valuation. Chapter 3 discusses both discrete and continuous-time stochastic interest rate models. Some stochastic interest rate models used in both actuarial theory and practice, such as an autoregressive model, the Vasicek model, the CIR model and the Markov chain model, are reviewed. Cash flows are one of the fundamental concepts in modeling life insurance contracts. Chapter 4 discusses the modeling of both deterministic and stochastic cash flows of life insurance products as well as the evaluation of mathematical reserves of the products using Markov chains. Recursive formulas for mathematical reserves based on their integral representations are derived. Chapter 5 discusses the distributions and higher-order moments of mathematical reserves. Firstly, Thiele's differential equations for mathematical reserves are derived. Then differential equations for the higher-order moments and distribution of mathematical reserve are also derived. Chapter 6 presents examples to illustrate applications of a Markov chain model to calculate mathematical reserves for pensions and disability insurance. The impact of stochastic interest rate on insurances is also studied. A classical theorem in life insurance mathematics is Hattendorff's Theorem, which was established by Karl Hattendorff in 1868. This theorem states that the losses on a life insurance contract in different years are uncorrelated and have mean zero. The concept of martingales plays an important role to understand the result of this theorem better. Chapter 7 first describes Hattendorff's Theorem in a general setting and then the theorem in a Markov model. The martingale calculus is used to discuss the theorem. Chapter 8 discusses the valuation of unit-linked policies using the modern technology of option pricing theory, in particular, the Black-Scholes-Merton model. A key feature of unit-linked policies is that benefits underlying the policies depend on the performances of underlying risky financial assets. This chapter ends with presenting an extension to Thiele's differential equation by taking into account the presence of an underlying risky financial asset. Chapter 9 discusses extensions to stochastic interest rate models. Particular attention is given to the Vasciek stochastic interest rate model. Thielie's differential equation is extended in a stochastic interest rate environment. Chapter 10 is concerned with the technical analysis of life insurance polices. Techniques such as profit testing and the calculation of the present value of the future profits are discussed in a Markov chain environment. Chapter 11 presents an abstract valuation theory based on the idea of stochastic discounting. This theory forms the basis of the Valuation Portfolio and is a consequence of the Riesz representation theorem of continuous linear functionals of Hilbert spaces. Applications of the theory to Asset Liability Management and the calculation of the required risk capital for Solvency II are discussed. The final chapter (Chapter 12) introduces the concept of policyholder bonus and investigates its effects. The appendix contains a review on stochastic integration and semimartingale calculus which presents an orientation for the modern French theory of random processes developed by Paul-André Meyer and his co-workers in the `Strasbourg School'.