Objective research to aid investing decisions
Menu
Value Allocations for July 2019 (Final)
Cash TLT LQD SPY
Momentum Allocations for July 2019 (Final)
1st ETF 2nd ETF 3rd ETF

“Best” Indicator Consistency Across Samples

Posted in Technical Trading

A subscriber inquired whether “The Only Indicator You Will Ever Need” really works. This technical indicator, applied to the Dow Jones Industrial Average by Jay Kaeppel, is a multi-parameter composite based on monthly closes as follows:

  1. Calculate the asset’s return over the past 11 months.
  2. Calculate the asset’s return over the past 14 months.
  3. Average these two past returns.
  4. Each month, calculate the 10-month front-weighted moving average of this average (multiply the most recent value by 10, the next most recent by 9, the value for the month before that by 8, etc. Then sum the products and divide by 55.)
  5. Hold the asset (cash) if this weighted moving average is above (below) its value three months ago.

We designate this indicator 11-14WMA3. To test 11-14WMA3 in realistic scenarios, we apply it to the entire available histories for three exchange-traded funds (ETF): SPDR S&P 500 (SPY)SPDR Dow Jones Industrial Average (DIA) and iShares Russell 2000 (IWM). We consider both buy-and-hold and a conventional 10-month simple moving average timing strategy (SMA10) as benchmarks. SMA10 holds the ETF (cash) when the ETF’s most recent monthly close is above (below) its 10-month SMA. Using monthly dividend-adjusted and unadjusted closes for the ETFs from their respective inceptions through October 2016 and the contemporaneous yield on 13-week U.S. Treasury bills (T-bills), we find that:

Testing assumes the following:

  • We use unadjusted ETF closes to calculate SMA10 and 11-14WMA3 signals, and dividend-adjusted closes to calculate returns. (Results are similar when using adjusted closes to calculate signals.)
  • We can slightly anticipate signals such that associated ETF-cash switches occur at the same monthly close.
  • One-way switching frictions (broker fee plus bid-ask spread) for these liquid ETFs are 0.1% of the balance.
  • When in cash, active strategies earn the T-bill yield.
  • Ignore tax implications of trading.

For SPY, the first available 11-14WMA3 signal is at the end of March 1995. The following chart compares the net cumulative performance of $10,000 initial investments in buy-and-hold, SMA10 and 11-14WMA3 as applied to SPY from the end March 1995 through October 2016. Some net performance statistics for these three strategies are, respectively:

  • Average monthly returns: 0.80%, 0.89% and 0.50%.
  • Standard deviations of monthly returns: 4.30%, 2.95% and 2.33%.
  • Compound annual growth rates (CAGR): 11.7%, 14% and 7.7%.
  • Maximum drawdowns (MaxDD): -51%, -15% and -20%.
  • Numbers of SPY-cash switches: 0, 25 and 31.

Applied to SPY, 11-14WMA3 suppresses volatility but substantially underperforms both buy-and-hold and SMA10 based on CAGR. 11-14WMA3 generates lower monthly volatility than SMA10 but a higher MaxDD. Results are not very sensitive to the assumed level of switching friction.

Next, we consider the DIA sample.

kaeppel-coppock-guide-strategy-spy

For DIA, the first available 11-14WMA3 signal is at the end of March 2000. The following chart compares the net cumulative performance of $10,000 initial investments in buy-and-hold, SMA10 and 11-14WMA3 as applied to DIA from the end March 2000 through October 2016. Some net performance statistics for these three strategies are, respectively:

  • Average monthly returns: 0.52%, 0.46% and 0.38%.
  • Standard deviations of monthly returns: 4.05%, 2.36% and 2.07%.
  • CAGR: 5.4%, 5.3% and 4.4%.
  • MaxDD: -47%, -18% and -16%.
  • Numbers of DIA-cash switches: 0, 35 and 15.

Applied to DIA, 11-14WMA3 modestly underperforms both buy-and-hold and SMA10 based on CAGR. However, 11-14WMA3 outperforms SMA10 on risk metrics and trades less frequently.

Finally, we consider the IWM sample.

kaeppel-coppock-guide-strategy-dia

For IWM, the first available 11-14WMA3 signal is at the end of July 2002. The following chart compares the net cumulative performance of $10,000 initial investments in buy-and-hold, SMA10 and 11-14WMA3 as applied to IWM from the end July 2002 through October 2016. Some net performance statistics for these three strategies are, respectively:

  • Average monthly returns: 0.91%, 0.61% and 0.45%.
  • Standard deviations of monthly returns: 5.40%, 3.52% and 2.93%.
  • CAGR: 9.5%, 6.8% and 5.0%.
  • MaxDD: -52%, -21% and -25%.
  • Numbers of DIA-cash switches: 0, 29 and 13.

Applied to IWM, 11-14WMA3 substantially underperforms both buy-and-hold and SMA10 based on CAGR. 11-14WMA3 generates lower monthly volatility than SMA10 but a higher MaxDD.

kaeppel-coppock-guide-strategy-iwm

In summary, evidence from robustness tests on tradable index proxies over different recent sample periods is mixed but mostly does not support belief that 11-14WMA3 is a “best” indicator.

Multi-parameter indicators such as 11-14WMA3 have considerable latitude for data snooping. Findings above suggest that 11-14WMA3 may perform best when the sample is rich in bear markets (the DIA sample commences with a long bear market). Hence, inclusion of the 1930s may be crucial to any reputation it has as a good indicator. 

Cautions regarding findings include:

  • Samples are not long in terms of number of independent 14-month measurement intervals and, especially, in terms of number of bear markets included.
  • The benchmark SMA10 performs differently for different assets and sample periods.
  • Some investors may bear switching frictions different from the assumed level.
Why not subscribe to our premium content?
It costs less than a single trading commission. Learn more here.
Daily Email Updates
Login
Research Categories
Recent Research
Popular Posts