Monte Carlo is Risky

Last week I discussed Monte Carlo Analysis as a tool financial planners use to model retirement scenarios for clients. The process involves applying random values to the assumptions inside a plan. With the use of modern computers we can run thousands of iterations each with a unique set of assumptions. The results can be reviewed to determine how often the client actually meets their objectives. Most advisors like to see a “success rate” of at least 90%.

Unfortunately, Monte Carlo Analysis is not without its flaws. A recent article in The Wall Street Journal (“Odds-On Imperfection: Monte Carlo Simulation”) discussed how Monte Carlo simulation may be less reliable than most advisors (and their clients) realize. The problem is that Monte Carlo assumes that stock market returns fall along a normal bell-curve shaped distribution. This would suggest that returns fall proportionately along the curve. If the average annual return of the stock market has been 10% (a reasonable view), then a return of 5% (5% below the norm) or 15% (5% above the norm) would not be surprising. However, a decline of 50% (60% below the norm) would be highly unlikely.

The reality is that stock market returns are more random than the bell-curve would suggest. Morningstar points out, for example, that the Standard & Poor’s 500-stock index has declined more than 13% in a month at least 10 times since 1926. The conventional bell-shaped curve would not predict this.

To address these limitations, some are moving away from the normal distribution and instead considering a “fat-tailed” distribution. This distribution assigns higher probabilities to extreme market events. This allows advisors and clients to anticipate how a deep bear market would affect the client’s portfolio.

Within the last decade we have experienced two stock market crashes. One was caused by the speculative bubble in technology-oriented stocks that took place from 1995 through 2001. The Nasdaq lost 39.3% in 2000, 21.1% in 2001 and 31.5% in 2002. The Dow lost 6.2% in 2000, 5.4% in 2001, and 16.8% in 2002.

We are still dealing with the aftermath of the crash began in the fall of 2007 and occurred when large financial institutions faltered under the load of packaged securities that failed to perform. The Nasdaq was up 9.8% in 2007 and down 40% in 2008. The Dow was up 6.4% in 2007 and the Dow lost 33.8% last year. Year to date the Nasdaq is up 1.97% and the Dow is down 4.4%.

The stock market is volatile. The static returns that are assumed in basic retirement plans are unrealistic and unreliable. Traditional Monte Carlo Analysis is helpful in modeling other scenarios. However, stock market returns do not follow the normal that underlies Monte Carlo simulation. Therefore, the investment advisory industry needs to develop other methods which will better enable investors to plan for retirement. Let’s hope this happens soon.