Where’s the Alpha? Characteristics of Alpha-Driven Stocks
The end of the year is good a time for reflection and analysis, and many investors take the opportunity to use the typically less hectic market period between mid-December and mid-January to examine their process, figure out what’s worked over the past year, and think about improvements for the next year. A key candidate for annual adjustment is the definition of the investment universe, or if the investment universe is fixed (e.g. on an external benchmark), the process to prioritize securities for analyst research.
To address the coverage and/or prioritization process, one of the preliminary questions to answer is “What qualities are typically associated with high-alpha driven stocks?” To help answer this question we’ve put together an analysis on alpha-driven securities in the Russell 3000 over the past 2 years. Below, we’ll share some of our results, as well as provide our weekly market and factor update.
Alpha-Driven Stock Analysis
What is an “alpha-driven” stock? Our definition centers on breaking down the volatility of a security. In particular, securities that are alpha-driven move because of micro-reasons largely specific to the company, and not due to external forces such as macroeconomic reports, politics, and sector moves. We call these alpha-driven moves “idiosyncratic risks” or “specific risks".
For the purposes of our analysis below we define an alpha-driven security as one where 2/3rds (67%) or more of its volatility is idiosyncratic.
Alpha-Driven Names as a % of the Russell 3000 (1/1/18 - 12/1/19)
The graph below shows the number and percentage of securities in the Russell 3000 that meet our criteria above, measured at the beginning of every month over the past 2 years.
We can see that the total number of alpha-driven names in the Russell 3000 has declined over the past two years. In particular, the 4Q18 market drawdown precipitated a big dip in alpha-driven names sending the percentage of alpha-driven names below 40%, and we haven’t seen much of a rebound since then. It’s clear the market has become more factor-driven over time, meaning that there has been essentially less “good” stock picking in 2019.
Factor Characteristics of alpha-driven names
Although the total number of alpha-driven names have decreased over the past 12 months, we still see about 1000-1500 names in the Russell 3000 that qualify under our definition. Below we examine few key characteristics of the median alpha-driven stock versus the median Russell 3000 name.
Median Market Cap: Alpha-Driven R3K vs. Russell 3000 (1/1/18 - 12/1/19)
Unsurprisingly, the median dollar market cap of the alpha-driven names was lower than that of the broader Russell 3000 throughout this period. It has also decreased during the last 2 years whereas the median market capitalization of Russell 3000 name remained roughly constant. This is contemporaneous with the prior observation showing an overall lower count of higher alpha names and indicates fewer stock picking opportunities in the midcap and largecap segments of the Russell 3000.Median Liquidity Exposure: Alpha-Driven R3K vs. Russell 3000 (1/1/18 - 12/1/19)
Note: All Y-axis values for factor exposures are expressed as z-scoresLiquidity is defined as the ratio of the stock’s ADTV (Average Daily Trading Volume) to its market cap. Interestingly enough, this chart shows that the higher alpha names have actually been more liquid than the broader Russell 3000, with the lone exception of Feb 2019. This indicates that even though these names aren’t as factor driven, they’re being traded at a higher frequency relative to the median security in the Russell 3000 index. This could mean that institutional investors are crowding these stocks, or perhaps that highly liquid and idiosyncratic sectors such as Biotech are driving this pattern. Next week we will dive into a sector-specific and crowding analysis of the alpha-driven securities to determine whether our thesis is correct.Median Value Exposure: Alpha-Driven R3K vs. Russell 3000 (1/1/18 - 12/1/19)
Relative to the Russell 3000, the alpha-driven bucket has significantly less exposure to Value, although this exposure appears to have increased over time. Value exposure reached a nadir in July 2018, rose to a peak of 0.14 in June 2019, and then fell pretty rapidly until rebounding in October. It now appears to be back on the rise, but still substantially lower than the broader index. Median Growth Exposure: Alpha-Driven R3K vs. Russell 3000 (1/1/18 - 12/1/19)
Perhaps our most fascinating learning here is that median Growth exposure for the high-alpha names has largely been lower than that of the index. Many investors assume that Growth names tend to have higher alpha characteristics, and may decide that this is where they want to focus their research efforts. But, as the data shows, companies with higher growth exposure have been more associated with factors over the past year.
Summary of Observations
- The number of alpha-driven names has significantly declined over the past 2 years in the US market, making it more difficult to select mid/large-cap securities.
- Possibly due to the number of fewer investment opportunities available in the alpha-driven space, we see higher higher Liquidity exposure for the median alpha-driven security than the median security in the Russell 3000.
- The median alpha-driven name has substantially lower Value exposure than the median of the Russell 3000 index, but it has been on the rise over the past two years
- Perhaps most surprising, Growth exposure for high-alpha names has been mostly lower than the index.
Next week we’ll extend our analysis to the sector and industry levels in order to determine which industry groups have been the most alpha-driven during this period.
US & Global Market Summary
US Market: 12/6/19 - 12/12/19
- Stocks rallied on Thursday, following initial reports that a US / China deal was imminent, as results of the UK election took some uncertainty off of the table with regard to the Brexit deadlock.
- On Friday (not captured in above chart), the major indices closed flat to slightly up as investors parsed the limited details surrounding the Phase One trade deal. The market had already priced in the 12/15 tariffs being removed, and the rollback in tariffs on $120B in products to 7.5% was less substantial than many expected.
- The Commerce Department reported that US retail sales were +0.2% in November, lower than the consensus estimate of +0.5%. This datapoint signaled a slower start to the holiday season and suggests that consumer spending could be softening in 4Q.
Factor Update - US Model
- The heavy rotation that we had seen from Value into Growth over the past few weeks saw an abrupt reversal, as Value escaped Oversold territory with a +0.34 standard deviation increase.
- Profitability also bounced off of a low and started to dig its way out of Oversold space.
- The relationship between Volatility and Market Sensitivity continued to diverge, with Volatility seeing slight gains and approaching Overbought, while Market Sensitivity continues to fall away from being Overbought in late November.
- Momentum peaked at +1.31 SD above the mean and now appears to be reverting.
- Earnings Yield continued to fall, and has now crossed over into Oversold territory.
- US Total Risk (using the Russell 3000 as proxy) fell by 17bps.
Factor Update - Worldwide Model
- Global Profitability saw a massive rally on a normalized basis, up +0.71 standard deviations in the past week which pulled it out of Extremely Oversold territory.
- Similar to the US, we saw a reversal of the Value to Growth rotation, with Growth seeing some weakness on a normalized basis while Value was up +0.57 standard deviations (and is now approaching Overbought space).
- Momentum peaked at +1.10 SD above the mean on 12/11 and appears to be in the process of reverting.
- Volatility and Market Sensitivity both saw slight declines, as Market Sensitivitycrossed over into negative territory.
- Earnings Yield was again the biggest loser, falling towards an Oversold designation only a few weeks after being labeled as an Overbought factor.
- Global Risk (using the ACWI as proxy) decreased by 15bps.