GLD/SPY Rotation Strategy

Michael A. Gayed, CFA is Chief Investment Strategist at Pension Partners, where he structures portfolios. Prior to this role, Michael served as a Portfolio Manager for a large international investment group, trading long/short investment ideas in an effort to capture excess returns. In 2007, he launched his own long/short hedge fund, using various trading strategies focused on taking advantage of stock market anomalies. Michael earned his B.S. from New York University, and is a CFA Charterholder.


A hypothetical “buy and rotate” strategy which would go either 100% in SPY or 100% in GLD depending upon which was outperforming the other on a rolling 20 day (1 trading month) basis.

1. I downloaded daily price data for both GLD and SPY going back to inception of GLD (11/18/2004) using Yahoo Finance (Adj. Closing Prices).

2. Calculated End of Day returns for both

3. Created a price ratio (GLD divided by SPY)

4. Created a 20 day simple moving average of that ratio

5. Put an if/then statement such that if the GLD/SPY price ratio was greater than the 20 day MA of the ratio, to track the performance of GLD. If not, track performance of SPY. Assumption is that using prices very close to end of day closing prices creates the rotation right at that moment for holding to be there next day

Here is how the chart stacks up. Note I did not include any assumptions for slippage or commission costs:

12/21/04 data is because its 20 days after inception of GLD, when strategy starts running. Here is what the annual performance looks like:



With the exception of 2010, GLD/SPY rotation strategy beats S&P 500 every year. Perhaps more interesting is the correlations. Strategy correlation I calculated against Gold is about 0.68. Strategy correlation to S&P 500 is about 0.34. The strategy would have been all in Gold for 863 days (about 52% of the time), and in SPY 793 days (about 48%).

Couple of takeaways:

1. Gold does not have to be bought and held to provide value – it can be bought and rotated against equities depending on relative momentum.

2. Gold does not have to be treated as a static (ex. 5%) allocation at all times – if anything all-in asset-allocation decisions could yield stronger results than buy and hold of S&P 500 and Gold (slippage/commission excluded).

3. The biggest outperformance beat occurred in 2008, where the rotation strategy would have gotten you out of stocks and into Gold as the Lehman Crash was occurring.

4. Even though in 2009 Gold and the S&P 500 performed the same, the rotation strategy still would have outperformed.

Of course, past performance is not indicative of future results, but the simple binary decision of being either long SPY or long GLD depending on which is outperforming the others does seem to suggest alpha can be generated.

Michael A. Gayed

What's been said:

Discussions found on the web:

Read this next.

Posted Under