Blog - Investing Notes

June 12, 2008 - Rough Test of a Simple Sector ETF Momentum Strategy (Updated to Append Comments 6/17/08)

Does a simple momentum trading strategy applied to the major stock market sectors outperform the overall market? To investigate, we apply a simple six-month past total return momentum trading strategy to the nine sector exchange-traded funds (ETF) defined by the Select Sector Standard & Poor's Depository Receipts (SPDR), all of which have trading data back to December 1998:

Materials Select Sector SPDR (XLB)
Energy Select Sector SPDR (XLE)
Financial Select Sector SPDR (XLF)
Industrial Select Sector SPDR (XLI)
Technology Select Sector SPDR (XLK)
Consumer Staples Select Sector SPDR (XLP)
Utilities Select Sector SPDR (XLU)
Health Care Select Sector SPDR (XLV)
Consumer Discretionary Select SPDR (XLY)

The simple trading strategy starts with $10,000 and each month puts all funds into the one of the above ETFs that has the highest total return over the prior six months. Using monthly adjusted closing prices for these ETFs and for the S&P 500 index mimicking SPY (to represent the overall stock market) over the period 1/99-5/08 (113 months), we find that:

We make the following assumptions in testing this trading strategy:

• We trade at the close on the last trading day of each month (we assume total six-month past returns of the ETFs calculated just before the close are very similar to those calculated at the close).
• Trading friction is 0.25% of the balance whenever we switch ETFs.

The following chart compares the cumulative value of a $10,000 initial investment in this sector ETF momentum trading strategy to that of a $10,000 buy-and-hold investment in SPY. The momentum trading strategy easily outperforms buying and holding SPY. The average monthly return for the momentum strategy is 1.1%, compared to 0.2% for SPY. The standard deviation of monthly returns for the momentum strategy is 5.7%, compared to a less volatile 4.0% for SPY.

What drives the outperformance?

The next chart shows the distribution of ETFs selected by the momentum trading strategy over the entire test period. The energy sector comprises 39 of the 107 monthly selections (36%).

What would happen to the momentum strategy if there were no energy sector boom (if we exclude XLE as a choice)?

The final chart adds to our comparison the cumulative value of a $10,000 initial investment in XLE and the cumulative value of a $10,000 initial investment in the sector ETF momentum trading strategy excluding XLE as a monthly choice. The average monthly return for XLE is 1.3% with standard deviation 5.8%. The average monthly return for the restricted momentum trading strategy is 0.4% with standard deviation 5.1%. Results suggest that at least boom sector is important to the success of the momentum trading strategy.

Note that the sample is very small in term of number of independent six-month momentum calculation intervals (only about 19).

In summary, a simple momentum strategy applied to sector ETFs over the past decade outperforms the broad stock market, thanks mostly to the booming energy sector.

For related research, see Blog Synthesis: Momentum Investing/Trading.

A reader poses the following question:

"Have you tested other time parameters, such as ranking periods of one month, three months, year-to-date and one year and holding periods of one 1 month, three months, year-to-date and one year? Is a six-month ranking period optimal for the various holding periods?"

Response:

The sample is small (these sector ETFs have not been around long), and data mining bias is especially pernicious for small samples. The six-month ranking period is intuitively large enough to gauge sector momentum but small enough to react to changes in business conditions that might favor one sector over others. This ranking period may not be optimal, but the sample is not large enough to isolate an optimal approach with confidence.

Academic studies of momentum most often use 12 months as the ranking period, with six months probably in second place. Many such studies also impose a skipped month to avoid a known short-term reversion tendency. In other words, they rank on the past 12 months, skip a month and then buy the top entry from the 12-month ranking period. These studies generally have much longer sample periods than 1/99-5/08 and typically analyze individual stocks rather than funds. Many of these studies also use a holding period longer than one month. (Hence the notation 12-1-6 means a 12-month ranking period, one skipped month and a six-month holding period.) Browse the summaries of formal research at Blog Synthesis: Momentum Investing/Trading for more.

To sum up, trying to isolate an optimal ranking period from this small sample would involve so much luck that the results are as likely to mislead as illuminate.