Adaptive Asset Allocation: Simulation And Experimentation With ETFs

|
Includes: EEM, IWM, MDY, QQQ, SPY, TLT
by: Toma Hentea

Summary

Adaptive asset allocation between stocks and bonds enables an investor to capture higher returns with lower risk than most other investment strategies, including "Buy and Hold" and "Fixed Asset Allocation".

Since July 2013 the Chicago South Suburban Investment Club has experimented with a monthly rotation strategy between seven largest ETFs, achieving 27.82% annualized return, -3.17% drawdown.

A simulation study from July 2003 to November 2014 including the 2008 market crash shows that an investor using the rotation strategy achieved 24.16% annual return, -15.62% drawdown.

Experimentation with Adaptive Asset Allocation

The Chicago South Suburban Investment Club has been experimenting with a monthly asset rotation strategy as presented in a previous submission. On the last trading day of each month, the assets are ranked by the previous 3-month return. All funds are invested in the asset with the highest return, as long as that return is positive. If all the assets had negative returns over the previous 3 months, then all funds are moved into cash.

The six ETFs considered for investment are the following: iShares 20+ Year Treasury (NYSEARCA:TLT), Power Shares QQQ Trust, Series 1 (NASDAQ:QQQ), SPDR S&P500 (NYSEARCA:SPY), SPDR S&P400 (NYSEARCA:MDY), iShares Russell 2000 (NYSEARCA:IWM), and iShares MSCI Emerging Markets (NYSEARCA:EEM).

This experiment has been ongoing since July 2013. It extends over a period of 16 months. Within this time interval, the system had been invested 12 months in QQQ, 2 months in EEM, and 1 month each in IWM and TLT. The equities markets have been in a relentless uptrend, and the system adapted to this environment by being invested 15 months in equities, and only one month in bonds. The performance was quite satisfactory: 38.71% total return, 27.82% average annualized return (CAGR%), and -3.17% maximum drawdown (DD%).

For comparison, we simulated the performance of a portfolio of the same assets with a fixed allocation, 50% in TLT, and 10% each equity ETF. It also delivered good performance. These results are shown in table 1.

Table 1. Portfolios performance July 2013 to November 2014

Portfolio

Total return%

CAGR%

DD%

Adaptive Allocation

38.71

27.82

-3.17

Fixed Allocation

20.12

14.74

-3.45

Simulation of Adaptive Asset Allocation Portfolio

These two portfolios were simulated with historical data, downloaded from Yahoo Finance.

We use the monthly price data from April 2003 to November 2014, adjusted for splits and dividends. The time selection is restricted by the availability of data; EEM was created in April 2003.

From July 2003 to November 2014 there were 57 reallocations, about 5 trades per year. The system was invested 46 months in EEM, 33 months in TLT, 27 months in QQQ, 18 months in IWM, 10 months in MDY, and 3 months in SPY, for a total of 104 months in equities, and 33 months in bonds.

For comparison, we also simulated the performance of a portfolio of the same assets with a fixed allocation, 50% in TLT, and 10% in each equity ETF. As can be seen in table 2, the adaptive allocation strategy delivered significantly better performance. Not only is the return much higher at 24.16% vs. 8.95%, it also had a much lower drawdown at 15.62% vs. 29.26%.

The adaptive strategy was able to avoid the huge losses of 2008-09, while the fixed allocation strategy simply averaged the losses of stocks with the gains from bonds.

Table 2. Portfolios performance July 2003 to November 2014

Portfolio

Total return%

CAGR%

DD%

Adaptive Allocation

1,083.24

24.16

-15.62

Fixed Allocation

165.88

8.95

-29.26

A nice illustration of the simulation results is presented in the figure below, where the equities of the two portfolios are plotted in a linear scale. Notice that the adaptive allocation equity increase accelerates, due to compounding. That effect is not apparent in the fixed allocation equity.

Source: This chart is based on EXCEL calculations using the adjusted monthly closing share prices of TLT, QQQ, SPY, MDY, IWM, and EEM.

To get a more realistic comparison, in the next figure we plot the equities in a logarithmic scale. In this plot, both equities grow linearly. The slope of the adaptive strategy is much larger and it is also very stable vs. time.

Source: This chart is based on EXCEL calculations using the adjusted monthly closing share prices of TLT, QQQ, SPY, MDY, IWM, and EEM.

Note: This article is written for educational purposes. It contains no specific investment advice.

Disclosure: The author is long QQQ, SPY, IWM.

The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.