This study uses the tools of backtesting and data visualization to examine the challenge posed by short selling the SPDR S&P 500 ETF Trust.

The last few weeks in the S&P 500 have been a bit of a roller-coaster ride, with the SPY declining just over 5% in an 8-day period and subsequently making a new all time high over the following 15 days (on a closing basis). The recent volatility may have caused you to question holding long exposure to the S&P 500, and even tempted you to go short.

This study presents an analysis, gathered from historical data, that may be relevant to anyone considering testing the short side on SPY.  

Backtest Analysis

So let’s take a look at the data. As the summary table below indicates, the historical probability of SPY closing higher ten days following any market open is 62%. So, within this timeframe, the S&P 500 has a natural upside drift.

‍

But, curiously, at the time of this study, the probability that SPY closes higher at least one of the days between day 0 and day ten is 88.82%. Let’s dig deeper.

The forward return chart below shows every single 10-day return period for the S&P 500 ETF, the SPY, since its inception.  

The blue lines show every single 10-day return period where the SPY closed higher at least once.  The red lines show every single 10-day return period where the SPY failed to close higher at least once, ie. T[0] was the highest price the SPY traded over the following 10 days.  

If it’s not already obvious, the density of the blue lines is a lot higher than the density of the red lines, simply because there are a lot more blue lines than there are red lines.  

The figure below highlights the occurrence of red line price paths by date and shows the 10-day returns of each date on the y-axis.

Medium-Term Timeframes

The same probabilities not only persist but intensify when we observe in 15-day and 20-day timeframes:

• In each 15-day period the SPY historical probability of closing higher on day 15 is 64% and the probability that it closes higher from time 0 in at least one of those days is 91.53%.  This historical tendency played out yet again in the most recent correction, which started on November 18th but saw SPY close higher on day 15. You can see that in the first chart – the candlestick chart at the start of the post.
• In each 20-day period, the historical probability that SPY closes higher on day 20 is 65%, and the probability that it closes higher from time 0 in at least one of those days is 93%.

Longer Timeframes

And do these tendencies persist over a 60-day period? Incredibly, when we look at the historical data, SPY has closed higher at least once over the next 60 trading days 97% of the time.

Takeaways

So, what are the takeaways from this?

1. The S&P 500 exhibits a natural upside drift. Historically, SPY has closed higher than its t[0] value well over 50% of the time, over virtually all n-Day forward timeframes.
   

2. More importantly, the price path the SPY takes to its final destination often registers many new highs from its starting point. In each of the 10, 15, 20 and 60 day timeframes the SPY has exhibited a historical tendency of closing higher at least once 89%, 92%, 93%, and 97% of the time respectively. This underscores how averaging down can actually prove itself a worthy strategy for index traders with long time horizons.

I hope you find this research helpful in your investment and trading process. Thank you for reading.