Swing Trading Signals


Since 2013

  • 100% Quantified, data-driven and Backtested
  • We always show our results!
  • Signals every day via our site or email
  • Cancel at any time!

Short-Term Factor Strategies And Transaction Costs

Last Updated on 10 February, 2024 by Abrahamtolle

Short-term alpha strategies can be controversial in the trading space; this piece goes over whether or not they can survive transaction costs.

In June 2022, David Blitz, Matthias Hanauer, Iman Honorvar, Rob Huisman, and Pirm van Pliet published a paper called Beyond Fama-French Factors: Alpha from Short-Term Signals. The writers aimed to stress-test short-term factor strategies. Short-term alpha signals are usually dismissed in most asset pricing models due to market friction-related issues, and the authors tried to falsify that.

But is this truly the case? Can investors attain considerable net alpha through a combination of signals? Do composite strategies generate economically highly significant alphas? This piece aims to explain the answers better based on the mentioned research paper.

Data

The research team considered month-end stock numbers from December 1985 to December 2021 in the MSCI World Standard Index.

Stock numbers averaged 1,750 in the 36-year sample period, with a low of 1,296 and a high of 2,065.

Additionally, researchers took total monthly returns in USD for all stocks into account during the study, and other characteristics were considered when constructing the six FamaFrench control factors.

Do Short-Term Factor Strategies Survive Transaction Costs?

Methodology

For the study’s first analysis, equally weighted quintile portfolios were created by ranking stocks on their respective signal scores at the end of each month in the study timeframe. After doing this, the researchers computed the returns of the quintile portfolios over the subsequent month.

Empirical Results

Let’s look at the empirical results:

Individual short-term signals

The researchers discovered the annualized mean return ranged between 5% and 8%. Associated t-statistics for all signals except idiosyncratic volatility (iVOL) range between 3% and 7%. IVOL, in this instance, has a highly negative market beta because it structurally goes long on low-risk stocks and short on high-risk stocks. 

You can find the top-minus-bottom quintile performance for individual short-term signals below.

Further empirical results were obtained from the following: composite strategy, out-of-sample and post-publication performance, other investment frictions, robustness tests, and the role of sentiment and limits to arbitrage.

Cumulative performance for the multi-signal strategy based on factor strategies

Cumulative performance for the multi-signal strategy

Research Conclusions

Throughout the research, high turnover’s relevance in asset pricing literature was questioned because they failed to remain steady after accounting for market frictions.

The results in our research paper showed a statistically and economically high net alpha is obtainable from short-term signals.

The research approach combined numerous short-term signals that could produce significant diversification benefits, mitigate transaction costs by exclusively trading liquid stocks, and utilize effective buy-sell rules.

Reasons why we shouldn’t discard short-term signals to produce alpha

The first argument is that the standard academic factor construction methodology introduced by Fama and Fench in 1993 gives a high weight of 50% to small-caps that comprise only around 10% of total stock market capitalization.

The second argument is gross and net performance can improve significantly by shifting the focus from single signals to a combination of carefully chosen signals. Additionally, incorporating signals with minimal correlations grants powerful diversification benefits that result in lower volatility and higher gross returns.

Thirdly, the authors argue that most previous studies considered a naive trading strategy built on simply constructing entirely fresh top and bottom trading portfolios monthly.

It was also discovered that trading costs were often overestimated by researchers who cited the outdated model from Keim and Mahavan (1997); this model focused on the 1991-1993 trading period.

Key Insights

The first insight is that traders can generate statistically and economically highly significant alphas when they follow practical trading guidelines through composite strategies of short-term reversals, momentum, analyst revisions, risk, and monthly seasonality.

Secondly, alphas remain profitable after we take costs in the post-publication and out-of-sample periods for numerous signals into account, and this is as the performance of short-term compositive strategies has contracted over time.

Thirdly, market frictions such as implementation lags and short-selling limitations fail to explain the results adequately. Performance is to be equally strong on the short and long sides, while the alpha remains robust in incorporating implementation lags of one or two days.

This insight provides evidence consistent with the hypothesis that sentiment-related mispricing influences return predictability, and higher returns were recorded after highly sentimental periods.

FAQ:

Why are short-term alpha signals often dismissed in asset pricing models?

The research paper explores the viability of short-term alpha strategies in trading and aims to stress-test them against market friction-related issues. Short-term alpha signals are typically dismissed due to concerns related to market friction, as highlighted in the paper. The FAQ further delves into the challenges associated with these signals.

How were quintile portfolios created in the study’s methodology?

The research team considered month-end stock numbers from December 1985 to December 2021 in the MSCI World Standard Index. Details about the data and sample period are provided. Quintile portfolios were created by ranking stocks on their respective signal scores at the end of each month in the study timeframe. The returns of these portfolios were then computed over the subsequent month.

What were the empirical results for individual short-term signals?

The empirical results showed annualized mean returns ranging between 5% and 8% for individual short-term signals, with associated t-statistics for most signals falling between 3% and 7%. Detailed information about each signal’s performance is provided. The research concluded that a statistically and economically high net alpha is obtainable from short-term signals. It questioned the relevance of high turnover in asset pricing literature and emphasized the benefits of combining multiple short-term signals.

Leave a Reply

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}

Monthly Trading Strategy Club

$42 Per Strategy

>

Login to Your Account



Signup Here
Lost Password