Anthony Dardis and Vinh Loi Huynh
Abstract
We present some rudimentary concepts on asset/liability management and describe an approach to asset allocation modeling for institutions that invest to meet liabilities. The traditional risk/reward framework of financial economics is used as a starting point. The definitions of risk and reward are then refined with regard to the institution under consideration. A simple model of a U.S. life office is examined. We assume that the only investments available are domestic stocks and long-dated government bonds. Stochastic simulation is used to create a large number of future investment scenarios using historical total return data for these asset classes. The ability of the institution to meet its liabilities under each simulated scenario is examined. We construct optimal risk/reward profiles, and hence the optimal asset allocation strategy, and show that they can vary considerably by liability profile.
Key words and phrases: asset/liability management, Monte Carlo simulation, risk, reward, solvency
Anthony Dardis