A diversified portfolio of reasonably price quality growth stocks outperforms over time. Specifically, a portfolio that avoids high debt and high P/E stocks and focuses on high ROE companies earns above average returns with below average risk.
.
Figure 1 shows that annual variation in returns of the S&P 500 is primarily driven by earnings growth and changes in P/E.
.
Figure 1: Figure 1: S&P 500 Total Return Decomposition
Return = ∆P/E * ∆E + D/P
Source: Spellman, FactSet, S&P 500 total returns
.
All else equal, a stock should appreciate as its earnings rise. Growth companies often have unique products or services to offer. However, popular growth stocks are often overpriced, and cannot grow at above average rates forever as products mature and competition intensifies. Thus, while growth is important to returns, one must not overpay for it. As growth slows or becomes negative, P/E declines. Therefore, one can limit the risk of owning growth stocks by avoiding high P/E stocks and investing in low to moderate P/E companies.
.
High growth may also arise due to cyclical tailwinds, and people may miss signs of shifting wings. Thus, a better proxy of lasting growth than past earnings growth is high ROE (earnings/equity). Risk can further be reduced by owning quality firms with low to moderate levels of debt. Firms with high levels of debt may not have sufficient cash flows to finance their projects or could have aggressive management teams. Firms with high ROE, which is not leverage with substantial debt, may have some type of competitive advantage (leading to higher margins) and/or operating advantage (leading to higher asset turnover). High ROE companies can use their above-average earnings to plow cash back into the company to generate ongoing growth, pay down debt, pay dividends, buy-back shares, or engage in other shareholder friendly activities.
.
Data
.
Data is from the FactSet Fundamentals global database from 12/31/1994 through 12/31/2016. This period includes several economic and market cycles. All US equity securities greater than or equal to $200 million in market capitalization were considered. In total, there were 281,369 possible data items, with only 18,878 not available. For instance, in 12/31/2016, there were 3,263 securities.
.
Stocks were rebalanced every three months into three fractiles based three variables, total debt to total assets, ROE, and E/P. The variables were fractiled within their respective FactSet sectors to avoid sector biases impacting overall portfolio returns. E/P was utilized instead of P/E in order to classify negative earnings (negative P/E) stocks in the same fractile as high P/E stocks. Equal-weighted returns for each fractile and the universe were computed for three, six, and twenty-four months. Conclusions from each period are the same: quality growth stocks at reasonable prices outperform (“QGV” outperforms).
.
Is There a Free Lunch?
.
Superior ROE, low and moderate debt, and low and moderate P/E stocks generate superior performance.
.
Figure 2: Growth of Hypothetical $1 from 12/31/1994 through 12/31/2016
.
The best three fractiles based on total debt to total assets, ROE, and E/P is coded with 1-1-1. 1 for debt implies low debt to asset companies, whereas 3 implies high debt to assets firms. 1 for ROE implies high ROE companies, and 1 for E/P implies high E/P (or low P/E) stocks. The target QGV group includes the 1-1-1’s, 1-1-2’s, 2-1-1’s, and 2-1-2’s. Great companies are not necessarily cheap (1s for E/P), so the target was expanded to include X-X-2’s. Also, depending on the sector, moderate amounts of debt may be beneficial and does not impose undue risk, so 2-X-X’s are also included in the target. The “other” group includes the other 23 groupings (1-1-3, etc.). Over the entire period, there were 53,210 security-periods for the target, 209,281 security-periods for other, and 18,878 security-periods with at least one missing data item.
.
Figure 2 shows that the target group generates far superior returns than the universe of securities. $1 at the end of 1994 would have grown to $5.01 for the universe, but to $14.15 for the target. The weighted average annual alpha of the target group versus the universe is over 4% per year and just under 4% per year versus the other group of securities. Furthermore, the target produces higher return with a lower beta and standard deviation of returns, thus there is a “free lunch.” Perhaps the low debt ratio and reasonable valuation limits volatility, while high sustainably high ROE leads to above average growth and returns and low volatility.
.
QGV Outperforms Consistently and Predictably
.
A good investment approach should perform consistently and predictably over time. QGV outperforms the universe 72% of 3-month periods (63 of 88 periods), and this grows to 80% for 12-month and 84% for 24-month periods. This strategy even performs better in down markets. Figure 3 shows that the target QGV outperforms in nearly 96% of the down markets while participating in 2/3 of the rising markets which occur about 2/3 of the time (56 of 85 12-month periods). Down markets are risk-off markets when quality should outperform since safety is sought.
.
QGV outperforms in almost all down markets.
.
Figure 3: 12-Month Returns versus the Universe and Other
.
Further, figure 4 shows that QGV works well in most economic environments. Only during early stages of economic revivals, such as during the economic revival in the early 2000s and after the financial crisis, does the approach falter. While it still outperformed, it was also weaker in 2016 as economic conditions improved (based on measures such as the ISM Manufacturing Index). During these periods, the economy lifts all companies and investors seek risky securities instead of safety. Also, cyclical companies with low earnings, and therefore high P/Es, often lead the market. Still, QVG still participates in these rising markets, and normally leads during other periods.
.
QGV is least effective when growth is abundant and during “risk on” periods during early stages of economic recoveries.
.
Figure 4: Excess Returns from QGV Investing Over Time
.
QGV has a Winning Batting Average
.
In any given 12-month period, only 46.7% of the stocks outperform the average of the universe. The batting average can be improved by picking the right stocks (figure 5). While the improvement for QSV is modest, every little bit helps. This also implies that fundamental security analysis can add additional value to a pure quantitative process to QGV security selection.
.
Figure 5: QGV Improves the Batting Average
.
QSV Investing Works in Most Sectors
.
Investing in reasonably priced high ROE and low debt securities works in most areas of the market. Figure 6 shows the excess 12-month returns of the target group versus the universe of securities for each area of the market except one (miscellaneous). It is most effective in areas such as health care and other growth areas; however, on average, QGV also works well in cyclical and defensive areas of the market.
.
QGV works in nearly all areas of the market.
.
Figure 6: 12-Month Excess Returns by Area of Market
.
Historical Comparisons
.
Figure 7 shows the historic ranges for E/P, ROE, and debt to assets ratios. Overall, moderately priced stocks (2s for E/P) are a better deal than normal, and high ROE (1s) and low debt companies (1s) have a greater advantage than normal.
.
Moderate E/P companies are a good deal, and high ROE and low debt companies have a greater advantage than normal.
.
Figure 7: Characteristics of Fractiles (Median)
.
The difference between the first and third fractiles of E/P became most stretched during the financial crises as many firms reported losses. The historic median E/P of the lowest fractile is -1.0% versus -2.3% today. The median of highest E/P companies over the entire period is 7.4%, versus 5.9% today. Thus, the difference between cheap (high E/P) and expensive (low E/P) firms is currently 8.2% versus a historical median of 8.5%. This implies that cheap companies are more expensive than high-priced firms right now. On the other hand, average priced stocks (the 2s for E/P) are a better deal than normal. The difference between the second fractile of E/P and the third is 5.8% versus a historic average of 5.1%. Additionally, the difference between the first and second fractiles of E/P is 2.5% versus a historic median of 3.3%.
.
The gap between high and low ROE companies has expanded over time (currently 27.8% versus a median of 22.4%), as has the difference between high and average ROE firms (currently 13.1% versus a median of 10.6%). Thus, high ROE companies have a greater advantage than normal. Notice that the difference normally spikes during recessions, which implies that these high ROE companies are less cyclical, and therefore of better quality, than the low ROE stocks.
.
Low debt companies are 1s and high debt companies are 3s. Corporations have been raising leverage over the last three years; however, the safe low debt companies are relatively safer now versus the other fractiles of companies than in the past. The 1s have risen debt to assets from a median of 0.5% to 2.7%, while the high debt companies have increased from a median of 43.3% to 51.1%. Thus, the difference between low and high leveraged companies is now 48.3% versus a historic median of 42.7%. The average (2s for debt to asset) companies have also increased debt levels (currently 24.0% versus a historic 19.4% median) more than the low debt companies.