Summary
In this report, we investigate whether moves in SPY prior to a REACTION signal’s release provide meaningful information in deciding whether to enter or avoid a trade. We find that large up moves in SPY prior to a signal are a good indication to avoid short signals and act on long signals. Likewise, we find that large down moves in SPY prior to a signal are a strong indication to avoid long signals and trade short signals.
Read as PDF
Introduction
In our whitepaper “Trading Events: Predicting Reactions After Earnings Announcements” (available upon request by emailing qr@leapday.io), we demonstrate the predictive power of the REACTION signals to forecast individual stock returns. One approach to use the REACTION signals is to incorporate them into an existing quantitative investment process. For example, systematic traders can use the signals in their existing models or create new models alongside other inputs. Also, discretionary traders can use the signals to manage risk around existing positions or trade opportunistically using the signals as a primary input into their decision making process.
This report is the first in a series of reports in which we demonstrate simple ways to extract additional value out of the REACTION signals. In this report, we focus on examining how moves in the aggregate stock market prior to a signal’s release can be used to filter REACTION signals and improve performance.
Creating a signal filter using moves in SPY
Leapday’s REACTION signals are available for 3 entry points after an event day, which is the first day the market is open subsequent to a company’s earnings announcement. These 3 entry points are: (1) 60 minutes after the open on the event day; (2) just before the market close on the event day; (3) before the market open on the next trading day after the event day.
For this study, we focus on the market close signal. This signal is available for 4 holding periods: 3d, 5d, 10d, and 15d. We use the 3d hold signal in the discussion below.
The purpose of this study is to examine if moves in the aggregate stock market prior to a signal’s release affect performance. The question we want to answer is: if the market has a large up move before a signal is released, would you still want to initiate a short trade? And vice versa for a large down move with a long trade. We use the SPDR S&P 500 ETF (SPY) as a proxy for moves in the aggregate market.
Since we are trading REACTION signals at the close, we are going to look at moves in the SPY from the close on the trading day prior to the event day to the close on the event day. This approach introduces a minor look ahead bias since the REACTION signals are delivered 10-15 minutes before the close. However, using prices of SPY 10 or 15 minutes before the close versus at the close will not change the outcomes by much. We examine signal performance when SPY is up or down above certain thresholds: 1.0%, and 0.5%. A 1.0% or larger up move represents 11.7% of days between 2016-2020 and 1.0% or larger down move represents 9.2% of days. A 0.5% or larger up move occurs on 26.5% of days and a 0.5% or larger down move occurs on 18.2% of days.
For the signal that we are studying (t0 close 3d signal), we will also examine the 1d, 2d, and 3d returns to see the daily breakdown that corresponds to the 3d signal. We calculate returns using the same methodology used in our whitepaper referenced above. Namely, all returns are market-adjusted returns. For the market adjustment, we subtract the event return from the return of an opposite side trade in the SPY with a beta-adjusted notional value that offsets the beta risk of the event trade.
To summarize, for this study:
We start with the REACTION signals generated just before the close on the event day. We focus on the 3d hold signal.
We track moves in the SPY from the close on the day prior to the event day to the close on the event day.
We look at performance of the signals on days in which moves in SPY moves are greater than: +/- 1.0%, and +/- 0.5%. We use market-adjusted returns when examining this performance.
Results – Up Move in SPY
We start by looking at the performance of the signals when there is an up move in the market (and include a comparison to the performance of all REACTION signals). We study two thresholds for up moves in SPY: 1.0% and 0.5%. We break down the analysis between long and short signals and look at 1d to 3d performance across various event thresholds.
As you can see from Table 1 above, short signals significantly underperform (and often have negative returns) on days in which the market is up prior to the signal’s release near the close. This result exists when the market is up 1.0% and up 0.5%, for holds 1-3 days, and for each event threshold.
How frequently would this market direction filter be triggered? Table 2 below shows how many short signals exist from 2016-2020 at different event thresholds, and how many of those occur on days when the SPY return is greater than the thresholds. About 12% of short signals occur on days when the market return is over 1.0% and 24% occur when the market is up 0.5%. As Table 1 above shows, short signals preceded by up moves in the market significantly underperform the average short signal. This result suggests that avoiding these trades would improve portfolio performance.
Next, we examine how long signals perform when preceded by a significant up move in SPY and how that compares to all long REACTION signals:
In Table 3 above we see that long signals preceded by a large positive return in SPY outperform other long signals over 2 and 3 day holds and this result exists across event thresholds.
This filter is applied fairly frequently given the last five years have been a bull market. As Table 4 below shows, about 15% of long signals are on days preceded by a 1.0% positive return and 27% are preceded by a 0.5% positive return. With long signals outperforming when this happens, it may make sense to give extra weight to signals on these days.
In summary, larger positive returns in SPY prior to a REACTION signal can be used to achieve higher returns with the REACTION signals. An investor may wish to avoid short trades and give greater consideration to long trades on such days.
Results – Down Move in SPY
We will next examine the performance of the REACTION signals when they are preceded by a down move in SPY. Below we present metrics of how long REACTION signals do when preceded by down moves and compare these results to all REACTION long signals.
Table 5 above indicates that long signals preceded by down moves in the SPY strongly underperform the average long signal and are often losing trades. This applies across holding periods and event thresholds.
Since down moves in the stock market have been less common in the recent bull market, this translates to 8% of long signals being preceded by a 1.0% down move and 18% being preceded by a 0.5% down move as indicated in Table 6 below. However, when this does happen, it may be prudent to avoid the long signals.
Next, let’s look at how short signals do when preceded by a significant down move in SPY and how that compares to all short REACTION signals:
As shown in Table 7 above, short signals preceded by down moves do incredibly well, making 2-3x more profits than the average short signals with metrics that hold across holding periods and event thresholds.
About 8% of short signals are preceded by a negative return over 1.0% and 18% are preceded by a negative return over 0.5%, as shown in Table 8 below. Given the strength of the short signals on these days, an investor may use these days to overweight trades based on the REACTION signals.
In summary, negative returns in SPY preceding a REACTION signal are a valuable filter for extracting additional value from the REACTION signals. When a down move precedes a long signal, it is a good risk signal to avoid such trades and when it precedes a short signal, it is a very strong indication that those signals may outperform.
Results – Using Market Direction to Screen Signals
Thus far, we examined the performance of the signals when preceded by large moves in SPY and discovered that large up moves prior to a short signal are a good warning to avoid short signals and large down moves prior to a long signal are a good warning to avoid long signals. Next, we will present the metrics of all REACTION signals if you remove long signals on large down days and remove short signals on large up days.
Table 9 below shows the performance of all long REACTION signals compared to long REACTION signals that do not trade on days when SPY is down more than 1.0% and down more than 0.5% prior to the signal.
While long REACTION signals perform well on their own as shown above under All Long Signals, performance is boosted when avoiding trades in which there was a moderate down move in the market prior to the signal. It is worthwhile to note that excluding signals preceded by SPY moves greater than -0.5% boosts returns more than excluding signals preceded by SPY moves greater than -1.0%. This outcome arises because long signals preceded by SPY moves greater than -0.5% and -1.0% both underperform the average long signal (as shown in Table 5). Using the -0.5% threshold excludes a larger number of poor performing signals thereby boosting returns more.
Next, we take a look at the performance of all short REACTION signals compared to short REACTION signals that do not trade on days when SPY is up more than 1.0% and up more than 0.5% prior to the signal.
Again, while short REACTION signals perform well on their own as indicated above on Table 10 under All Short Signals, performance is boosted when avoiding trades in which there was a moderate up move in the market prior to the signal.
In the study above, we only examined using the moves in SPY as a risk warning to avoid unprofitable trades. However, our research in this report indicated that signal performance is also boosted by trading long signals on days in which the market is up substantially and trading short signals on days in which the market is down substantially. By overweighting these long and short trades, an investor can expect to see stronger performance than indicated above.
Discussion
We have shown that relatively large moves in the aggregate stock market (as measured by SPY) that precede a REACTION signal are useful indicators in determining if one should avoid such trades or overweight such trades. Specifically, signals perform poorly when preceded by large SPY moves in the opposite direction of the signal but perform strongly when preceded by large SPY moves in the same direction as the signal. We believe these results exist because investors tend to underreact to company specific news that coincides with the prevailing market sentiment on that day. Meanwhile investors overreact to company specific news that is contrary to the prevailing market sentiment on that day.
Above, we present metrics for REACTION t0 close 3d hold signals. The analysis above is robust and extends to other entry points and signals for other holding periods. To see expanded metrics and learn more about the various ways to use the REACTION signals, contact qr@leapday.io.
Conclusion
In this report, we present an approach to improve returns for the Leapday REACTION signals by filtering on large moves in SPY preceding the signals. The average long t0 close 3d signal returns 1.5% in 3 days using a restrictive trade threshold of 5% of events. When this signal is preceded by a positive daily return in SPY of 1.0% or more, the REACTION signal return increases to 2.4% in 3 days. When the signal is preceded by a negative return in SPY of 1.0% or more, the trade returns -0.5% on average. The average short t0 close 3d signal returns 1.5% in 3 days using a restrictive trade threshold of 5% of events. When this signal is preceded by an up move in SPY of 1.0% or more, the trade returns -0.3% on average. When this short signal is preceded by a down move in SPY of 1.0% or more, it returns 3.7% on average. This research indicates an investor may wish to avoid trades in the opposite direction of large market moves immediately preceding a REACTION signal and overweight trades in the same direction of large market moves.
Disclaimer
This material is solely for informational purposes and is not an offer or solicitation for the purchase or sale of any security, nor is it to be construed as legal or tax advice. References to securities and strategies are for illustrative purposes only and do not constitute buy or sell recommendations. The information in this report should not be used as the basis for any investment decisions. We make no representation or warranty as to the accuracy or completeness of the information contained in this report, including third-party data sources. The views expressed are as of the publication date and subject to change at any time. Hypothetical performance has many significant limitations and no representation is being made that such performance is achievable in the future. Past performance is no guarantee of future performance.
Comments