In 2021 Leapday successfully launched our flagship product, REACTION. Additionally, we continued adding to our library of research reports to uncover improvements to the signals and explored portfolio performance in large and small portfolio settings. The REACTION signals performed well in 2021 with metrics that met and exceeded average historical performance. A portfolio that traded both the t0 open + 60 min and t0 close REACTION signals with constant capital of $50M per entry point ($100M total) returned 12.9% with a Sharpe ratio of 1.66 in 2021 as compared to an average annual return of 12.1% and Sharpe ratio of 2.14 between 2016-20. If trading were restricted to only the 500 most liquid companies, the portfolio would have returned 11.2% with a Sharpe ratio of 1.62 in 2021, as compared to an average annual return of 7.5% with a Sharpe ratio of 1.61 between 2016-20. Additionally, when we apply the market direction filter (which was discovered in our 2021 research), the performance for 2021 outperformed the historical averages as well. All results are inclusive of conservative cost assumptions. Strong performance that is uncorrelated to the market and to other strategies has attracted top-tier clients in 2021 with limited licenses for REACTION remaining.
Read as PDF
2021 Year In Review
Early in 2021 Leapday had a successful launch of its first trading signal product, REACTION. Throughout the rest of the year, we continued to provide our audience with original research and insight to support the intuition behind the signals. To follow up our whitepaper that introduced REACTION, we published a series of reports that examined how the signals performed in a portfolio setting. First we showed that combining multiple entry points was additive for performance and reducing portfolio risk. Then, we showed that investors can expect strong results not only for less active stocks but also in the most liquid stocks as well. Next, we explored how a traditional PEAD strategy based on earnings surprise or momentum stacks up against REACTION and discovered that the REACTION signals strongly outperform PEAD signals and are uncorrelated to PEAD signals. Additionally, early in the year we presented our discovery of a market direction filter, which we found was useful to pre-screen REACTION signals based on recent market momentum. To wrap up the year, we’ve been working on a proprietary sentiment algorithm for predicting short-term market reactions to news. More to come on that in the new year.
Outlining This Report
The remainder of this report provides an analysis of 2021 performance for the Leapday REACTION model. This report will provide summary trade performance metrics as well as performance for portfolios constructed using the top one-third of 3d REACTION signals. We include the following details in this report:
Our whitepaper “Trading Events: Predicting Reactions After Earnings Announcements” was first published in January 2021 and introduced the REACTION signals. This paper provided summary metrics for the performance of the REACTION signals including average returns, hit rate, and avg. winner / avg. loser. To continue providing updates on these metrics, this report will provide summary metrics for 2021 performance as compared to 2016-20. Like the whitepaper, it will break down performance by liquidity groups. Additionally, we apply the market direction filter that was discovered in our research this year and provide the associated summary metrics.
For portfolios using $50M per entry point, we examine performance in 2021 as compared to 2016-20 (both without and with the market direction filter.) We also conduct the same analysis but restrict trading to only the most liquid 500 stocks. The full list of assumptions for constructing the portfolios is provided in the appendix.
For portfolios using $10M per entry point, we examine performance in 2021 as compared to 2016-20. We conduct this analysis without and with the market direction filter. These results show performance of a portfolio that is able to allocate a larger proportion of the portfolio to less active stocks as compared to the $50M portfolios.
Below we present summary metrics using the top one-third of 3d REACTION signals for the t0 open + 60 min and t0 close entry points. Metrics exclude trading costs. The breakdowns include results without and with the market direction filter and are further broken down by liquidity groups.
The liquidity groups were first introduced in the original REACTION whitepaper and are determined by a stock’s historical average trade value (calculated using the trailing 20 day average traded value). The 3 liquidity groups are:
Low Liquidity: <$3M historical average trade value
Medium Liquidity: $3M-$25M historical average trade value
High Liquidity: >$25M historical average trade value
Results for the t0 open + 60 min signals in 2021 are similar to the historical averages. Interestingly, in 2021 the low and high liquidity groups tended to outperform their historical averages while the medium liquidity group underperformed. Results with the market direction filter applied are marginally better for 2021 results as well.
In 2021 the t0 close signals slightly underperformed their historical averages. However, the high liquidity group had strong hit rates in 2021 well above the historical average.
Performance Update For Portfolios Using $50M Per Entry Point
Below we present performance for portfolios constructed using the top one-third of 3d REACTION signals for the t0 open + 60 min and t0 close entry points. We also construct a combined portfolio using both of these entry points. Combined portfolios in some of our previous reports also used the t1 open signals. However, since returns for the t0 close and t1 open entry points are fairly correlated, we do not include the t1 open signal in this study to avoid tilting the combined portfolio.
We begin with a chart that displays portfolio performance of the combined portfolio inclusive of costs. We show performance for a portfolio that trades all companies as well as a portfolio that restricts trading to only the most liquid 500 stocks. We provide performance without and with the market direction filter.
Below we provide a comparison of returns and Sharpe ratios for 2021 vs. historical results from 2016-2020. We examine these results for a portfolio that trades all companies as well as one that trades only the most liquid 500 stocks. These breakdowns include results without and with the market direction filter as well as without and with trading costs.
Combined portfolio returns for 2021 exceeded historical performance. Sharpe ratios in 2021 were slightly below average likely due to the increased volatility in the market relative to the historical period. Examining the individual entry point portfolios we see that the t0 open + 60min outperformed in 2021 while the t0 close underperformed. The liquid 500 stocks portfolio performance also outpaced its historical average. Additionally, results showed improvement when applying the market direction filter and outperformed historical results with the filter applied.
Finally, we examine correlation between the return streams of the various portfolios and SPY.
Correlation between the REACTION portfolios and SPY remains insignificant in 2021. Given the low correlations for all the different combinations of entry points and stock universe it is clear that REACTION provides a unique return stream that is not related to equity returns.
Performance Update For Portfolios Using $10M Per Entry Point
The results above show performance for portfolios constructed using $50M per entry point. Due to constraints on trade size based on stock liquidity, the majority of performance is driven by high liquidity names. To get a better look at the performance of smaller names, we lower the portfolio size to $10M per entry point which results in a larger proportional allocation to less active stocks. Below is a chart of the combined portfolio inclusive of costs, without and with the market direction filter.
Below are detailed metrics including performance by entry point, without and with the market direction filter, and without and with trading costs.
In 2021 less liquid stocks continued to show increased performance over a portfolio that allocated more capital to high liquidity names. Relative to historic results, performance for the combined portfolio was in line and showed an improvement when the market direction filter was applied. As with the $50M per entry point portfolio the t0 open + 60min $10M per entry point portfolio beat its historical average while its counterpart with a t0 close entry underperformed its historic average without the market direction filter applied.
2021 was a landmark year for Leapday with the successful launch of its REACTION signals product. The results above indicate that the REACTION model has built upon its continued strong out of sample performance. The t0 open + 60min signals had an especially good year, with performance exceeding the historical average by over 10% for a portfolio with $50M of starting capital. The combined portfolio using the t0 open + 60min signals along with the t0 close signals exceeded its historical performance. Finally, an analysis of the correlation between the returns of the REACTION portfolios and SPY shows that the signals continue to provide a truly uncorrelated return stream. Strong performance and uncorrelated returns have attracted new clients in 2021 for the REACTION product with limited licenses remaining.
Appendix - Assumptions Used In Portfolio Construction
We use the strongest one-third of the 3d REACTION signals for each entry point (t0 open + 60 min, t0 close) to build a simulated portfolio. Below we present all the assumptions used in the portfolio construction.
Enter and exit trades as follows:
t0 open + 60min signal: Enter a trade 60 minutes after the open on the first day the market is open subsequent to a company’s earnings announcement (t0 open + 60 minutes) and exit the trade on the close 3 days later (t2 close).
t0 close signal: Enter a trade at the market close on the first day the market is open subsequent to a company’s earnings announcement (t0 close) and exit the trade on the close 3 days later (t3 close).
Create a market-neutral portfolio by hedging trades using an opposite side trade in SPY with a beta-adjusted notional value that offsets the beta risk of the event trade.
Starting capital of $50M ($10M) per entry point and $100M ($20M) for the combined portfolio, assuming constant capital throughout the portfolio (i.e. - we do not reinvest profits).
Portfolio leverage used during periods of high activity
Trade no more than 6 percent of the starting capital per position.
Trade no more than 1.5 percent for the t0 open + 60 min signal. Trade no more than 2 percent of a company’s historical average trade value per position for t0 close signal. (The historical average trade value is calculated using the trailing 20 day average traded value.)
Beyond the two constraints on trade size stated above, trade equal weight.
Round-trip trading costs are based on breakpoints for liquidity group as follows:
Low (<$3M historical average trade value): 20 bps (40 bps for t0 open + 60 min signals)
Mid ($3M-$25M historical average trade value): 10 bps (20 bps for t0 open + 60 min signals)
High (>$25M historical average trade value): 5 bps (10 bps for t0 open + 60 min signals)
Costs are allocated half on entry and half on exit
In the metrics presented below, we assume $0 in hedge costs (for the SPY hedge) and the hedge does not use any equity or margin
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.