All-ETF Portfolios From Agile Investing Yield Strong Risk-Adjusted Return (ETFs: IVV, IVW, IJH, IYH, EFA, SHY, IYM; APL)

One of the largest sources of value from Monte Carlo tools is the ability to analyze portfolios, project risk and return, and look critically at the likely future performance. One of the ways that we test Quantext’s portfolio management tools is by looking at a range of model portfolios that have been published and running them through the Quantext Monte Carlo portfolio management software.

For this study, we have analyzed three portfolios proposed and published on ETF Investor by Agile Investing, an ETF advisor and investment manager:

[View larger version]

While these specifications (conservative growth, moderate growth, and aggressive growth) are qualitative, the Monte Carlo analysis allows us to determine exactly what this implies. The components of this portfolio include the usual common elements, such as large cap, mid cap, international funds, and bonds. The portfolio also includes some focused funds. Focused funds include U.S. healthcare, commodities, precious metals, and energy.

We had to change the portfolio components somewhat, as shown below, due to short data records for a number of these ETF’s:

When we first put all of these tickers into Quantext Portfolio Planner and retrieved historical data for the past three years, it was immediately evident that many of these ETF’s have been around for a short period of time—some less than two years and six of them less than three years. We found close substitutes for the short-lived components (shown above). We replaced GLD with VGPMX, Vanguard’s Precious Metals and Mining fund. GLD has only been trading since November 2004 which provides scant data for a risk analysis. We have also replaced PCRDX, a commodities fund with a short history, with iShares materials ETF, IYM. While PCRDX is designed to follow a more sophisticated strategy with respect to commodities, it has tracked fairly closely to IYM over the past two years. That said, the results started to diverge in mid and late 2005.

The Quantext Portfolio Planner (and Retirement Planner) takes the input portfolios, retrieves total return data for a specified history, calculates statistical parameters for each position, and then simulates the entire portfolio for years into the future. The simulation is a Monte Carlo process, which means that the software generates hundreds of possible future outcomes and then analyzes across the possible outcomes. A key feature of Monte Carlo models is how they generate the all-important assumptions about future uncertainty in returns. We use a process called risk-return balancing (described in a number of papers on our site) to estimate future returns and risk such that the projected future is not simply a rehashing of recent history. The analysis accounts for fees (except for loads) and assumes reinvestment of dividends.

Case 1: Conservative Growth Portfolio

Monte Carlo Analysis of Conservative Growth Portfolio

The conservative growth portfolio (above) has generated an average annual return of 11.3% per year over the past three years with a standard deviation in annual return of 5.34% (see Historical Data above). This is impressive, but must also be considered in light of the fact that the S&P500 has returned 13.4% per year, with standard deviation in annual return of 8.97% (also shown above). The Beta for the total portfolio is 50.7%, so this portfolio tends to be quite insensitive to the S&P500.

When we look at the Monte Carlo projection into the future, we see (under Portfolio Stats above) that this portfolio is predicted to generate an average annual return of 7.24% with a standard deviation of 8.26%. For these results, we have assumed that the S&P500 will return an average of 8.3% per year, with a standard deviation of 15% (see Market Index above). This portfolio is projected to generate average return that is 1% per year less than the S&P500 but with slightly more than half the risk (as measured by standard deviation). This portfolio shows a good application of strategic asset allocation—the use of offsetting risks between positions to manage risk while maintaining return.

Case 2: Moderate Growth Portfolio

Monte Carlo Analysis of Moderate Growth Portfolio

The moderate growth portfolio (above) has generated average returns over the past three years that are very close to the S&P500 (13.54% vs. 13.4%, above), but with less risk (volatility) than then S&P500. The portfolio Beta is 65%, still quite low.

This portfolio is projected to generate 8.39% per year with a standard deviation of 10.34%, as compared to the assumed conditions for the S&P500 with an average annual return of 8.3% per year and a standard deviation of 15.07%. In other words, the moderate growth portfolio is projected to generate returns equal to or slightly higher than the S&P500, but with only 2/3 of the portfolio volatility (10.3% vs. 15.07%, as above).

Case 3: Aggressive Growth Portfolio

Aggressive Growth Agile Portfolio

The aggressive growth portfolio (above) follows the other two portfolios, making good use of strategic diversification. The projected average annual return is 9.45% per year, with a standard deviation of 12.7% per year. This performance beats the projected returns on the S&P500 by 1.15% (9.45% vs. 8.3%), but with less risk than the S&P500. The standard deviation in return on this portfolio is 12.7% per year, about 4/5 of the projected standard deviation on the market as a whole.

Summary

These portfolios are good examples of effective use of strategic asset allocation to preserve returns while managing risk. One of the most notable features that a potential investor should consider is that the projected future returns on all three of these portfolios are markedly less than the returns over the past three years. The Monte Carlo analysis suggests that these are all good portfolios, but the last three years have been particularly high gainers and it would be a bad idea to assume that such high returns (and such low volatility) will continue into the future.

This analysis has assumed that the future average annual return on the S&P500 will be 8.3% per year, with a standard deviation of 15% per year. These are consistent with a range of expert estimates, but the projected future returns are (I hope) a bit conservative. The projected performance of these portfolios are dependent on these assumptions via correlations between these assets and the broader market.