Economic Foundations and Risk Analysis in Investment Management

Article excerpt


The risk management systems now used in investment analysis are based on Markowitz mean-variance optimization. Successful application depends on the accuracy with which market returns, risk, and correlation are predicted. Forecasting methods now commonly employed for this purpose rely on time-series approaches that generally ignore economic content. This article suggests that explicitly incorporating economic variables into the forecasting process can improve the ability of such systems to manage risk by providing a delineation between risk associated with changes in economic activity and that attributable to other shocks and discontinuities.


Over the Last decade, investment management evolved from its infancy to become one of the fastest growing segments of the new service economy. This is evident in the tremendous growth of assets held in mutual funds, self-directed retirement programs, hedge funds, and in derivative instruments, as well as by the massive accumulation of assets held by pension funds and endowments.

While some of this growth is attributable to structural changes in the financial system, such as the transition from defined-benefit to self-directed retirement schemes, a more important factor is the 1990s decline in global market risk. Some of the specific developments that facilitated systematic risk reduction included more responsible monetary policy, which reduced inflation in developed economies over the last two decades; the end of the Cold War; less capricious market intervention by governments; and European economic convergence in advance of the introduction of the euro.

A major consequence of lower market risk is that it has allowed investment managers to increase allocations to equities, thereby increasing portfolio returns. That said, shocks still occur and significant risks remain in the world economy. This was made clear with the Asian currency crisis in 1997, the Russian debt default, and the bailout of Long Term Capital Management (LTCM) in 1998. Thus, risk control remains a critical aspect of the investment-management process.

In this article, I analyze the role of risk in investment management, focusing particularly on the critical importance of economics to the process. My frame of reference is the classic Markowitz portfolio model, which requires return, risk, and correlation prediction for successful implementation. I first briefly review the nature of Markowitz mean-variance risk management, emphasizing its advantages and shortcomings. I then discuss typical methods for forecasting risk and the other inputs used in the approach. I then illustrate how augmenting standard time-series techniques with economic content improves results. Finally, I compare mean-variance risk management to the VAR approach now in vogue.

Key Elements of Portfolio Optimization and Investment Risk Management

Early influential studies by Brinson, Hood, and Beebower (1986), Brinson, Singer, and Beebower (1991), and a more recent update by Ibbotson and Kaplan (1999), show that the asset allocation decision is the key determinant of portfolio returns. These studies also conclude that individual security selection is of limited importance. For this reason, asset allocation decisions represent the key intellectual challenge for investment managers.

Most portfolio allocation decisions made by professional managers today rely on Markowitz (1959) mean-variance optimization or variants of the method. This approach explicitly compels the use of comprehensive risk management because probability is required in portfolio construction. The results of mean-variance analysis are often presented in the context of the efficient frontier, which shows expected portfolio return as a strict function of risk (Figure 1). The approach relies on three quantitative inputs--asset returns, measurable asset risk, and correlation between different assets. …