Price Action Lab Blog

Premium Market Analysis

Quantitative trading

Simple Explanation For Stock Market Rally

There is a simple explanation for the rally in stocks this year. Explanatory hypotheses are valid if they generate testable predictions. This one does unlike a number of suggested alternative hypotheses that are not falsifiable.

I like simple explanations. The mainstream financial media and a number of analysts believe that the stock rally this year is driven by expectations of higher growth due to anticipated tax cuts and policies by the new administration. They may be right but these explanations are overly complicated and possibly a result of wishful thinking.

After watching the price action near the close of last Friday it occurred to me that the rally in stocks may be related to algos and retail traders executing an extreme micro mean-reversion strategy. What do I mean by that?

Quants who are studying mean-reversion know that last year the extreme micro (daily) mean-reversion strategy of buying a lower close and exiting at the following up close has generated superior risk-adjusted returns when compared to buy and hold.

Below is a backtest of this strategy in SPY with $100K initial fully invested equity and $0.01 commission per share.

The simulated return is about 19% with only -3.83% max drawdown. Sharpe ratio is 2.29. In comparison, buy and hold for 2016 returned 11.33% with -9.2% max drawdown (not counting dividends.) Of course, this is theoretical performance and the actual varies due to slippage, etc.

The point is that the spectacular performance of this simple strategy was noticed by traders and also identified by machine learning algos. The result was that the trade due to execution of this strategy got crowded this year as traders and algos (naively) expect to profit from it. Year-to-date the return is less than 1.5% as shown below:

The main subject of this article is how this strategy has contributed to the stock rally this year. This is the hypothesis: traders and algos rush to buy near the close if they believe there will be a down day hoping to sell at a higher price afterwards.

The result of this herd behavior is that a down day suddenly turns into an up day due to massive buying. This was also the case last Friday with SPY in negative territory throughout most of the day only to rally near the close and finish with a gain. This is shown in the intraday chart below.

Source: Yahoo! Finance

So this is what happened: traders and algos in anticipation that the market will close down bought SPY to execute a trivial mean-reversion strategy and ended up pushing the market higher, essentially invalidating their strategy.

This is the main mechanism of the market to invalidate, or said in a more fancy way, arbitrage out edges. Strategies lose their edge if used by many traders. This is one reason that the edge must be unique and executed by only a few. Sharing an edge is catastrophic for its future performance. This is one reason that social strategy development and trading are doomed.

Below is a chart of the count of days with a lower open and also lower close in the last 290 days (252 in 20016+ 38 year-to-date)

It may be seen that the number of days that would have affected extreme mean-reversion signals, as described above, has declined from 108 in early 2016 to 90 as of Friday. This is a significant downtrend and mostly caused by algos and retail traders that buy in anticipation of a down day, essentially causing an up day. Notice how during chop and corrections this count normally increases while it decreases during uptrending markets.

The prediction

Since traders and algos pushed the market up in anticipation of a lower close, in turn invalidating the signal they attempted to follow, normally there should be a lower open on Monday as they exit the trade.

I remind here that explanatory hypotheses in any domain, including physics, can never be proven; they can be falsified or corroborated. Due to incompleteness, mathematical and physical, the truth can never be known but that is a debatable philosophical problem that also suffers from incompleteness.

If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY

Subscribe via RSS or Email, or follow us on Twitter.

Charting and backtesting program: Amibroker


Technical and quantitative analysis of Dow-30 stocks and 30 popular ETFs is included in our Weekly Premium Report. Market signals for longer-term traders are offered by our premium Market Signals service. Mean-reversion signals for short-term SPY traders are provided in our Mean Reversion report.

Copyright Notice

Leave a Reply

Theme by Anders Norén