25. Decentralized Risk Management for Global Property and Casualty Insurance Companies

25.1 Overview of DFA (centralized) Dynamic financial analysis (DFA) provides a tool to analyze various business strategies and risk/return structures within enterprise-wide planning systems. A DFA aims at maximizing the shareholder value and tracking the free cash flow over time. Leading insurance and reinsurance companies have begun applying DFA to increase profitability, reduce enterprise risks, and identify the optimal capital structure of the firm. A DFA process should analyze the financial status of an insurance enterprise, namely, the ability of the firm's capital and earnings path to adequately support its future operations in light of stochastic external factors affecting the enterprise. A DFA model should combine the asset/liability structure of the enterprise and dynamic optimization of the strategies together with headquarters decisions. A DFA system consists of three major elements: a stochastic scenario generator (also see [14] for generating scenarios over a stochastic programming tree and see [4] and [9] for different scenario-generation methods), a multiperiod simulator, and an optimization module; see Figure 25.1. Linking the assets and liabilities in a consistent fashion requires modeling the driving factors; e.g., see Figure 25.2. The factor models are well placed to support DFA. DFA is described further in [10]. Considering these external dynamic factors, a scenario tree is built up (Figure 25.3). Today, there are two practical approaches for optimizing a multiperiod DFA system. The first involves stochastic programs. An alternative to stochastic programming involves developing a set of rules or policies to guide the company across the planning period at each decision node; this approach is called policy optimization.