Introduction
.
The rise of passive management, index funds and ETFs, has been breathtaking. Is this just a fad, or is it here to last? I suggest it is here to stay, but there are ramifications. Certain market environments, such as those over much of the last decade, favor passive management. However, some of the conditions – low dispersion of stock returns and overall high returns – may be coming to a close.
.
Bubbles are often accompanied by new financial gimmicks. The quants are now driving trading. A 2017 article by CNBC quoted JP Morgan in estimating that just 10% of trading on the NYSE is by discretionary stock pickers. While some of the quants base their investing on traditional factors such as value and momentum, much of the transactions may be driven by short-term traders, volatility strategies that can unravel quickly, etc. Artemis Capital Management estimates that over $2 trillion is run by strategies that can influence, and be influenced by, volatility. They predict that this can unravel quickly and lead to a hyper-crash. Volatility was low until last fall when it rose quickly, but it has subsided again.
.
ETFs are also run by quants and are not without risks. They have known liquidity problems. By blindly buying all stocks in proportion to their market values regardless of whether those values are right, ETFs and index funds make the market less efficient. Of course, index products also have their virtues. They provide a cheap way to gain beta exposure to markets, and active fund managers have not done a good job at justifying their fees by outperforming.
.
This article explores some reasons why active management has performed poorly over the last decade and the possibility for that environment to change. Plus, it provides evidence of what areas of the market provide the most alpha potential and best risk-adjusted returns.
.
The Alpha Cycle
.
Figure 3 shows the percent of large cap fund managers that outperform the S&P 500 each year since 2000. The average is an abysmal 37%. However, notice (figures 3 and 4) that the probability of outperforming varies inversely with S&P 500 returns. Managers are better when returns are low, and vice versa. And, over the last decade, returns have been stellar, so this favored passive management.
.
.
The ability to outperform tends to diminish with higher returns perhaps because this is when market breadth narrows. That is, fewer stocks outperform the average stock during high return environments (figure 5) as perhaps trend following – momentum – takes hold. In When Does Value and Growth Outperform? I showed that growth investing historically performs better, relative to value, during late cycle as the economy slows. Most fund managers may give some weight to value considerations as they select stocks, and the high flyers often lack this characteristic. Thus, this could cause them to underperform during high return markets.
.
On the other hand, Savita Subramanian, Head of US Equity & Quantitative Strategy at Bank of America Merrill Lynch, noted in Barron’s that fund managers were generally overweight the high flying FAANG stocks and the technology sector as late as August 2018. Value has been out of favor for a number of years, so perhaps there are fewer value managers who survived long enough and could have benefited from the 2018 market turmoil. Last year, only 31% of fund managers outperformed. Plus, Antti Petajisto notes in the Financial Analyst Journal that more managers are hugging their indices (lower active share) than in the past, which means they are not taking full advantage of market disruptions.
.
Over the last 20 years, only 42.4% of stocks outperform the average stock on a rolling 12-month basis. On a monthly basis, the chance of outperforming by throwing a dart at a list of stocks is 48.5%. Thus, active management starts behind (less than 50% odds of getting the stock call right); although, this also means index funds have more stocks that underperform than outperform.
.
Declining market breadth as returns rise (figure 5) helps explain the underperformance of managers over the last decade. However, if you get the stock calls right, you have the potential to perform very well when returns are high. Figure 6 shows that the return spread between the top 10% and bottom 10% of stocks widens when average stock returns are higher. Momentum works better in high return environments. Yet, if active managers shy away from both positive and negative momentum, a growing spread hurts performance.
.
I expect more managers have been moving toward momentum strategies as value investing has been out of favor. This should have boosted 2018 alpha; however, last year momentum was a poor factor. For momentum to work, we need lasting trends, and in 2018, each time the trend was established, it reversed. 2018 started out strong, corrected fast, rose through the summer, and then was clobbered at the end of the year. Cyclical and defensive stocks both had their heydays and times of distress last year.
.
Figures 7 and 8 illustrate the general declining spread between winning and losing stocks and winning and losing sectors. The job of active managers is to identify these winners and losers, but getting the calls right have not paid as handsomely as in the past. The declining return spread may be due to a couple of conditions.
.
.
First, as more money is managed passively and more people buy indiscriminately, returns should move together.
.
Second, since the Great Financial Crisis, people have been more focused on macro, as opposed to company-specific, conditions. Rightly or wrongly, the economy, the Fed, and political news seem to be front and center on investors’ minds. Figures 10 and 11 show that stock correlation has generally risen over the last 10 years, which means that stock picking efficacy has waned. Correlation was 0.36 from 1/00 through 9/07 and 0.50 from 10/07 through 1/19. Also, it appears (there are exceptions) correlation rises when the ISM Manufacturing PMI (a gauge of economic conditions) declines, and vice versa (figure 10). The ISM is correlated with market returns (figure 9). So, an improving economy drives up returns, and higher returns are associated with lower correlation (and lower volatility since correlation directly impacts volatility) (figure 11). You’d think that lower correlation with higher returns would be better for stock picking, and it may be if other factors, such as the degree momentum drives market returns, are unchanged.
.
Last year, large-core managers navigated, not so well, a perfect storm. Again, only 31% of active large cap core managers outperformed in 2018. This poor showing is not surprising since active managers favored FAANGs and technology (possibly in part because of fear of underperformance that drove index hugging) that were pummeled at year-end, correlation spiked last year (figures 10-11), momentum did not work as trends constantly reversed and value investing was not too effective during much of the year, and breadth declined as the year moved toward a close (figures 7-8).
.
Long-term, high correlation creates opportunity. If markets are more correlated, and stock fundamentals are not, this implies that there is more mispricing. Eventually, that mispricing should be corrected.
.
Small-Cap Value Alpha
.
Speaking of opportunity, alpha appears to be easier to come by in small-cap value. On average, 46% of small value managers outperformed the Russell 2000 Value Index per year (figure 12) since 2000 (compared to 37% for large cap core managers outperforming the S&P 500). Plus, unlike large-cap, outperformance does not appear to be correlated with index returns (figure 13). Small-cap managers did a better job at outperforming their index than their large-cap peers in every year except four (figure 14).
.
There are many reasons that small-cap managers perform better than their large-cap brethren.
.
Small companies are underfollowed, so there is more chance of mispricing and opportunity. As of the end of 2018, there were an average of 19.8 FactSet collected FY1 EPS estimates for S&P 500 companies and only an average of 5.9 for the 444 S&P 600 Value (small-cap value) stocks.
.
By nature, this may mean (I need to ponder this one more) that there is less trend following in small-caps than large-caps, since fewer people are following the stocks.
.
Finally, and perhaps most importantly, the reason it may be easier to outperform in small-caps is that smaller companies are a very diverse group of stocks based on returns, growth rates, valuation, profitability, and financial forecasts. Given this, there is greater opportunity to gain by picking winners and losers in this space.
.
Figure 15 shows the median, 10th percentile, 90th percentile, and difference between the 10th and 90th percentile for returns and other financial characteristics of stocks sorted from high (1) to low (5) market cap each quarter since 12/31/2002 through 12/31/2018. The minimum cut-off for market cap was $100 million. Stocks were sorted within FactSet sectors to reduce the influence of sector biases influencing the results. Figures 16-27 provide visual depictions of the distributions for returns and various financial statistics for large-caps (1) and small-caps (5).
.
.
While small caps underperformed large caps within their FactSet sectors (average three-year return of 14.0% versus 32.6%), the ability to generate alpha is much higher in the small-cap space. The 90th percentile 36-month return of small-caps was 127%, versus large-caps of 109%, while the bottom 10% return for small-caps was almost a complete loss (down 94%) versus down -44% for large caps. More alpha can be made by getting the stock call right in small company investing.
.
The distribution of returns may be wider the smaller the company because there is much more disagreement for smaller firms, maybe due to lack of coverage or the influence of minor factors having a greater impact. Panel D of figure 15 shows that the range of P/S multiples for small-caps is a staggering 50, whereas it is 7 for large-caps. Valuation reflects expectations for growth, risk, and profitability (see box titled The Math of the P/S Ratio). Thus, there is much more disagreement about the future, or actual dispersion of future fundamentals, for small versus large stocks.
.
.
To standardized estimate dispersion, I divided standard deviation of the FY2 sales estimate by the average FY2 sales estimate. The last column of panel D shows that disagreement over future sales is much higher for small (25%) than large (10%) firms. Plus, as mentioned earlier, there are fewer estimates for small companies (5) than large stocks (19).
.
Fundamentals also vary much more for small stocks than large stocks. The difference in net profit margin is a whopping 122% for small-caps (90th vs 10th percentile) versus 25% for large-caps. While margins are lower for small-caps (median of 2% for small versus 9% for large), the top small-cap companies have about the same margins as large-caps (median of 20% versus 25%). While the median three-year sales growth for small firms (15%) lags large-caps (24%), the top small-caps have higher growth than large companies (140% versus 116%). The difference in three-year sales growth is huge in both small- and large-caps, but the 90th-10th percentile difference for small-caps is higher (174% versus 128%).
.
It makes sense that small companies have a wider dispersion of returns, growth, profitability, and estimates. While many of the small companies could be mature, firms start out as small. So, this small-cap universe probably includes more young, less established companies than the large-cap space. Young companies may start out as unprofitable, but as they grow, operating leverage may lead to substantial improvements in profitability. Also, smaller stocks tend to be more domestically based, which means that geographic diversification does not help them overcome gyrations in the US economy. Small companies may also be focused on fewer product lines, thus the loss of one or decline in competitive position may hurt them more than their large counterparts. On the other hand, if a product becomes a winner then there may be a longer runway for sales growth and for it to impact small firms’ financial successes.
.
Given the lack of coverage, higher dispersion of valuation and estimates, and greater range of growth and profitability, small-cap managers should have a better chance of outperforming than their large-cap peers, if they get the stock calls right. Also, notice in figure 15 that the advantage of going smaller grows gradually as one marches down the market cap spectrum from the 1s to the 5s. Thus, mid-cap managers have an advantage over large-cap managers as well.
.
Large-Cap Premium
.
Large-cap stocks have outperformed small-cap stocks since 1993. Remember, my sorting of size is within FactSet sectors, as I will later show that overall small-caps have outperformed large-caps since 1999. Figure 28 shows that the return gap for small minus large is about 16% to 20% depending on the time period. Thus, small-caps have really struggled. This underperformance could be due to (1) sales growth, (2) valuation, and (3) share growth.
.
Due to outliers, missing data, and perhaps bad data, the best we can do to determine the cause of the underperformance is to analyze the median numbers for P/S, sales growth, and share growth for each year from 2002 through 2018 for the median return stocks and try to reconstruct prior three-year price returns. Please consider the numbers in figure 29 with a grain of salt, as these are truly just best guess estimates of the drivers of the small-cap return short-fall. The figures in the columns represent characteristics of the median 36-month price return stock for the large (1) and small (5) quintiles of stocks at a point of time at the end of each year from 2002 through 2018. If the data was accurate, one would expect the calculated returns (third column) to equal the actual returns (second column), and while they are often close, they are not the same. Also, keep in mind that the median return company by market cap quintile may be far from representative of the distribution of stocks in the quintile.
.
Figure 29 shows that three-year sales growth rates for the median return small- and large-cap stocks are relatively similar (17.3% for small versus 16.4% for large). The three-year change in P/S multiples are also similar. P/S rose 6.3% (0.83 (P/S)-3 / 0.78 (P/S)0 – 1) for small versus 7.8% (1.58 (P/S)-3 / 1.47 (P/S)0 – 1) for large. The big difference in return appears to come down to share growth. While the annual average three-year growth of the median large company is -0.3% for share count (-0.3%), the smallest companies grew shares by 13.2%. More shares dilute current shareholders and drives down returns if the additional capital is not used for endeavors that commensurately boost corporate value.
.
Drivers of Returns
.
While one must be careful drawing too many conclusions from figure 29, it is still interesting. If small-caps underperformed due to share growth, then perhaps share growth is a good way to screen out losers smaller companies. Figure 30 shows that this is true. Lower share growth (see panel A) tends to boost return in each market cap quintile. The biggest drop off in 1-year returns comes when one moves from quintile 4 (moderately high share growth) to quintile 5 (highest share growth). Thus, watch out for firms that dilute their ownership base. I expect the most common large share growth scenarios result from acquisitions, which many studies show is a destructive use of capital.
.
Figure 30 also shows that moderate sales growth is best (panel C), while low expectations (high S/P or low P/S) is beneficial (panel B). Keep your expectations low and invest in steady growers – not too fast or slow – and earn higher returns. Note that moderate sales growth small-caps have the same returns as large companies! Thus, pushing for too high, or failing to deliver, growth is what hurts small firms the most.
.
Moderate price momentum (12-month prior returns) is also best (panel E). Again, temper those expectations.
.
Finally, more certainty in sales estimates is rewarded (panel D). High FY2 sales estimate standard deviation / mean estimate (the 1s) have lower returns. This implies that risk, as measured by estimate disagreement, is not being rewarded. However, figure 31 shows that low expectations (high S/P or low P/S) combined with a more certain future (lower FY2 sales estimate standard deviation/mean) generates solid returns.
.
.
The Most Efficient Mixes
.
As shown earlier, while large-cap may not be the easiest place to outperform, this does not mean it doesn’t have a place in a portfolio. Also, just because small-cap value is perhaps the best place to find alpha, does not mean one should put all his/her money in this space. Plus, I just showed that there are better areas to put money to work (e.g., value, low share growth, and more certainty in estimates).
.
Figure 32 shows the annual rolling returns (monthly) of the nine Standard & Poor’s style boxes (S&P 500 is large, S&P 400 is mid, and S&P 600 is small), and figure 33 shows the correlation of monthly returns. Keep in mind that these returns are not sector neutral like my studies discussed above for small to large, etc. Overall, small-cap tends to have higher cyclical sector exposure and more domestic exposure than large companies.
.
(See figure 32.) It is clear, small-caps stocks generated the highest return, and mid-cap also performed well. This is despite the fact that 90.5% of the market cap, as of 2/27/2019, of the S&P 1500 all cap index is large-cap. This means that 90.5% of the stocks (the S&P 1500 represents most stocks) are subject to low returns, and low risk-adjusted returns. While large-cap underperforms, it is still reasonably volatile. Small-cap is the most volatile, so mid-cap has the best, and slightly better than small, Sharpe ratios.
.
If you own large caps and index your money – the S&P 500 SPY is the largest ETF – then the best bet to diversify is in the small-cap space as its correlation is the lowest to the S&P 500 (figure 33). However, keep in mind that the correlation between all styles is pretty high (above 0.70).
.
We can also optimize to solve for the best risk-adjusted combination of portfolios. Doing so results in the efficient frontier as shown in figure 34. The optimization allowed up to 100% shorting in the nine style buckets. Also, keep in mind that the core boxes (S&P 500, S&P 400, and S&P 600) include both the value, core, and growth stocks; therefore, doubling up on value and growth was allowed.
.
The combination with the highest Sharpe ratio was completely out of the S&P 500, S&P 500 Value, S&P 400 Growth, S&P 600, and S&P 600 Value with 100% shorted to these indices. The S&P 500 Growth index (large-cap growth) had a 10% shorting weight. The highest Sharpe ratio combination loaded up on S&P 400 (mid-cap) with 139% weight, S&P 400 Value (mid-cap value) with 289% weight, and S&P 600 Growth (small-cap growth) with 182% weight.
.
.
The lowest risk combination was a composite of mostly the S&P 500 (100%), S&P 400 (96%), and S&P 600 (156%). The S&P 400 Growth and S&P 600 Value were 100% shorted. The single index with the lowest risk (grey dot in figure 34) was the S&P 500.
.
To achieve the highest return, one would have needed to be completely out of six styles and have a 98% weight to the S&P 400, 377% to the S&P 400 Value Index, and 226% to the S&P 600 Growth Index.
.
Isn’t it interesting and perhaps a sad state of affairs that no combination gives the S&P 500 the highest weight, yet it is the most common benchmark!
.
We can improve the S&P 500, though. The equal-weighted S&P 500 outperforms the market-weighted S&P 500 index (figure 36). Since the winners gain weight in market-weighted indices, we can consider the S&P 500 EW a value-like index and the S&P 500 somewhat designed like a momentum index (the weight of winners rises). The S&P 500 EW index also has more exposure to smaller stocks. Given earlier comments in this section about small stocks outperforming and earlier comments about value stocks outperforming, it is not surprising that the S&P 500 EW index performs better than the market-weight version (S&P 500).
.
The market-weight S&P 500 is not all bad. It shines during down markets. The S&P 500 EW index outperforms the S&P 500 index in 52% of the months, but when the market (i.e., S&P 500) is down, it only outperforms 40% of the time.
.
Turning to value versus growth, figure 37 shows that value is the better of the two since 1999. However, as with the S&P 500, down markets tend to favor the Russell 3000 Growth Index over the Russell 3000 Value Index. In the 86 months when the Russell 3000 (a combination of value and growth) was down (36% of the months), value outperformed growth 36 times, or a hit rate of just 42%.
.
Important note: These results – value underperforming growth during market corrections – appears to be at odds with my prior research in When Do Value and Growth Outperform? My prior research focused on sector-neutral strategies to avoid macro biases influencing results, while these indices have more or less exposures to certain sectors. Some of these sectors are historically more tied to the economy than others, with the value indices generally having more cyclical exposure (if we consider technology growth). If a market correction occurs due to fears of an economic recession, one may expect cyclical sectors, and value indices that are overweight them, to underperform. Although, my sector-neutral research shows that value does better than growth during down markets. This reason may be that value stocks are low expectations stocks; they already discount deteriorating conditions or a poor environment and they are already cheap (what is down cannot fall as far).
.
