Actively Passive? A Longer Look

his article follows up on an article in the September, 2003 issue of Financial Planning entitled "Actively Passive?" The main thrust of the original piece was to examine the difference between actively managing passive (i.e. index) funds versus passively managing passive funds.

Index funds are typically viewed as a passive asset. It naturally follows that one would ask: should portfolios that are comprised of passive assets be managed passively or actively? In other words, if passive is a good approach for the management of individual assets (i.e. funds) is it therefore a good approach for the management of multi-asset portfolios? There is no definitive answer to such a direct question; however this current article seeks to broaden the analysis originally presented last September.

The active/passive dichotomy is generally applied to how individual assets are managed, i.e. an actively managed fund vs. a passive index fund. This article attempts to extend the discussion to how equity funds themselves (whether passive or active) are managed within a portfolio. Active portfolio management suggests frequent reviews potentially leading to fund changes or reallocations among the funds. Passive portfolio management would engage in portfolio reviews/reallocations less often (or not at all), and thus represents a "buy-and-hold" approach.

This article explores only two boxes in the Active/Passive Grid (Figure 1) both of which are in the Index Funds column, namely Index / Passive and Index / Active. Studying active and passive portfolio management using actively managed funds at a macro level is virtually impossible due to an enormous number of possible combinations of active funds. Indexes which attempt to replicate the return of prominent asset classes are much fewer in number and thus more feasibly analyzed.

In this study ten assets were studied over a 25 year period from 1978 through 2002. The annual returns for each asset are shown in Figure 2. The color coding in Figure 2 will be explained in a moment. Annual return data were obtained from Morningstar's Principia software. The assets (i.e. indexes) examined in this study included:

1. 3 month Treasury Bill
2. Lehman Brothers 1-5 Year Government Bond Index
3. Lehman Brothers Long Term Government/Credit Index
4. MSCI Europe Index (return in U.S. dollars)
5. MSCI Pacific Index (return in U.S. dollars)
6. Wilshire Large U.S. Growth Equity Index
7. Wilshire Large U.S. Value Equity Index
8. Wilshire REIT Index
9. Wilshire Small U.S. Growth Equity Index
10. Wilshire Small U.S. Value Equity Index

Figure 1. The Active/Passive Grid

The ten assets were assembled into a portfolio which was subjected to 18 different management scenarios. The baseline portfolio management option was a buy-and-hold approach which simulated Passive Portfolio Management. The buy-and-hold option involved an initial investment of $1 into each of the ten assets at the beginning of 1978. At the end of 2002 the value of the 10 asset portfolio was calculated by summing together the ending account value for each asset. The average annual return (i.e. geometric mean) was then calculated as was the arithmetic standard deviation of annual return.

The remaining 17 simulations represented a variety of Active Portfolio Management scenarios. Scenario 2 called for annual rebalancing. At the end of each year the 10 assets were rebalanced so that 10% of the total portfolio account value was represented in each asset (or index).

Scenario 3 involved shifting all portfolio assets at the start of each year into the asset which had the highest return in the prior year. Scenario 4 shifted 80% of the portfolio assets into the prior year's best asset and spread the remaining 20% equally among the remaining 9 assets. Scenarios 5-7 followed the same protocol, but with ratios of 60/40, 40/60, and 20/80.

Scenario 8 invested all portfolio assets at the start of each year into the asset that had the lowest return in the prior year. Scenario 9 shifted 80% of the portfolio assets into the prior year's worst asset and spread the remaining 20% equally among the remaining 9 assets. Scenarios 10-12 followed suit but using ratios of 60/40, 40/60, and 20/80.

Scenario 13 moved 80% of the portfolios assets into the best fund from the prior year and the remaining 20% of total assets into the prior year's worst asset. Scenario 14 used a ratio of 60% best asset/40% worst asset, scenario 15 was 40% best asset/60% worst asset, and scenario 16 was 20% best fund/80% worst asset.

Finally, scenario 17 shifted all portfolio assets equally among the three assets with the highest return during the prior year. Scenario 18 allocated all portfolio assets equally into the three assets with the lowest annual returns from the prior year.

Now we will explain the color coding in Figure 2. The orange colored boxes represent the fund with the lowest return in the prior year. The turquoise blue boxes represent the fund with the highest return in the prior year.

Figure 2. Annual Returns

Results of the 25 year analysis are shown in Figure 3. The single asset class with the highest average annual return was the Wilshire Small Value index. It also had the highest Return-to-Risk coefficient (.56) of the 10 individual equity assets. The Return-to-Risk coefficient is simply an adaptation of the Sharpe Ratio, but in this case the excess return is divided by the amount of excess risk (i.e. standard deviation of return). Excess return is defined as the return above the return of cash, in this case the return above T-bills. Excess risk is defined as the amount of standard deviation of return in excess of the standard deviation of T-bill annual returns. On a technical note, comparing the Return-to-Risk Ratio of equities to non-equity assets is problematic; hence we will confine the Return-to-Risk comparisons to equity assets only. The higher the Return-to-Risk Ratio the better.

A passive management approach utilizing 10 different "passive" indexes was simulated through a buy and hold approach (Scenario 1). This approach generated an average annual return of 11.5%, and standard deviation of return of 16.4%, and a Return-to-Risk ratio of .34. The next 17 portfolios represented active management approaches to asset allocation. As a preface to the results, it is understood that active portfolio management has the potential to increase tax exposure. On the other hand, active portfolio management may focus on tax efficiency by harvesting losses. In either case, the analysis herein assumes that the portfolio management took place in a tax-sheltered environment.

Scenario 2, annual equal rebalancing, generated a slightly higher return than a buy-and-hold approach but lowered the standard deviation of return by 34%, from 16.4% to 10.7%. Scenario 2 also had a very attractive Return-to-Risk ratio. The next several scenarios (3-7) were variations on a common theme of focusing on the best asset from the prior year. Scenario 3 reallocated all assets into the best fund from the previous year and nothing to the other nine. This approach, while radical, produced an average annual return higher than scenario 1 or 2. However, the standard deviation of return was the 2nd highest among the 18 scenarios. In comparison to scenario 2, the relatively slight increase in average annual return (30 basis points) does not warrant the sizeable increase in standard deviation of return. As the commitment to the prior year's best asset was reduced by increments of 20%, the average annual return held steady and the risk (i.e. deviation of return) steadily dropped. The two best portfolios (out of the entire set of 18), on the basis of the Return-to-Risk Ratio, were scenarios 6 and 7. Scenario 6 probably gets the nod as the best choice due to the fact that a client would likely rather experience 30 extra basis points of return versus reducing risk by 30 basis points.

Scenarios 8-12 simulate the interesting approach of focusing on last year's worst performing asset. Clearly, this approach taken to the extreme (as in scenario 8) is a bad idea - at least when using the 10 assets selected for this study. The average annual return was 160 basis points below cash and the standard deviation of return was the highest among all 18 scenarios. As the commitment to last year's asset was scaled back (from 100% to 80%, then from 80% to 60%, and so on) the return increased and the risk level declined such that scenario 12 generated the 6th highest Return-to-Risk ratios. Overall, the notion of placing large bets on last year's loser did not get strong support from this analysis.

Scenarios 13-16 essentially modeled a two-asset portfolio. The only scenario worth considering seriously appears to be 13, in which 80% of the portfolio is reallocated to last year's winner and the remaining 20% to last year's loser. Nevertheless, none of these four scenarios produced an attractive Return-to-Risk ratio.

The highest active management return was observed in scenario 17. In this scenario the entire portfolio was annually reallocated (in equal portions) into the three funds with the highest one-year returns from the prior year. This approach also had one of the highest Return-to-Risk ratios of .64. The mirror image approach was scenario 18, using the prior year's 3 worst funds. That approach produced a much smaller return with a higher level of volatility compared to scenario 17.

The green boxes in figure 3 highlight the best performers (of all 18 scenarios) from the respective categories of average annual return, standard deviation of return, return in excess of cash, and risk in excess of cash. The blue boxes in Figure 3 highlight the strategies which produced a return-to-risk ratio equal to or higher than that of the Willshire Small U.S. Value, the individual equity asset with the highest return-to-risk ratio.

This analysis suggests that an active account management approach which over-weighted the past year's best asset (i.e. scenario 6 or 7) produced superior risk-adjusted equity returns. However, an equal-weighted annual rebalancing protocol had nearly identical levels of return and risk. Only by engaging in more aggressive active management (scenario 17) was the 25 year average annual return significantly boosted.

One significant observation: moving from passive portfolio management (scenario 1) to active portfolio management (scenario 2) produced very little gain in return but a significant reduction in the volatility of return. A similar effect is observed when comparing the results for scenarios 3 through 7 -- very little difference in return but large reductions in standard deviation of return.

The aim of this study was not to recommend a particular asset allocation strategy, but rather to illuminate some of the possibilities of active management and the potential consequences of such compared to a passive approach. We believe that there are potential pre-tax benefits associated with active management, whether through simple rebalancing or performance-based reallocations. The benefit, in fact, may not be so much an increase in return, but rather a potentially significant reduction in volatility of return. It may be that clients have a renewed interest in tactical management approaches that provide significant risk reduction.

Figure 3. Results over 25 Years (1978 - 2002)

This much we also know: leave last year's loser alone!

(Next month's article will present additional portfolio management scenarios.)

____________________________________________________________________________________
Craig L. Israelsen is an Associate Professor in the Department of Consumer and Family Economics at the University of Missouri-Columbia (http://www.missouri.edu/index.cfm) where he teaches courses in Personal Finance and Family Living. He holds a Ph.D. in Family Resource Management from Brigham Young University. He received a B.S. in Agribusiness and a M.S. in Agricultural Economics from Utah State University. Primary among his research interests is the analysis of mutual funds. He writes monthly for Financial Planning magazine.

Thomas J. Dohogne is an undergraduate student at the University of Missouri majoring in Economics.