Ranking Stocks by Market Capitalization: Size Matters 6° When you look at the results for investable microcap names, those with market cap- italizations between a deflated $50 million and a deflated $250 million, you see that most of the return for tiny stocks disappears. Using the Compustat dataset, between 1964 and 2009 investable microcap stocks earned an average annual compound return of 12.70 percent, whereas when using the CRSP dataset over the same period microcap stocks earned 11.82 percent per year. When you use the CRSP dataset to review the full period between 1926 and 2009, the investable microcap stocks compound at 10.92 per- cent per year, slightly better than our Small Stocks universe. A far better way to analyze the effects of market capitalization on a company’s returns is to evaluate all fully investable stocks by decile. When you look at the returns in the All Stocks universe by market capitalization decile, a fairly different picture emerges. Looking at Figure 5.2, we see that within the universe of investable stocks, there is an advantage to smaller-cap stocks, but it is not of the magnitude of other studies that allow noninvestable micro- caps. Here, the smallest decile by market capitalization had the highest compound return between December 31, 1926, and December 31, 2009, and the largest two deciles had the lowest compound returns, but the amounts are not huge: The tenth decile had the highest return at 10.95 percent per year, whereas the first decile (largest stocks) had the lowest return at 8.82 percent, a difference of 2.13 percent. Table 5.Decile Addendum 1 shows what $10,000 grows to for each decile, from decile one—those 10 percent of stocks from All Stocks with the highest market capitalization—to decile ten, which is the 10 percent of the All Stocks universe with the smallest market capitalization. The table also shows the Sharpe ratio for each decile. 109% 109% 106% 106% 106% 10.1% 193% 102% ae s00% }———_su_ gg _f 38% 80% 6.0% — - - —— 40% + ~ “J 1 2 3 ‘ s 6 7 8 9 10 All stocks (highest) (Lowest) FIGURE 52 Average annual compound return by decile, All Stocks universe, January 1, 1927, to December 31, 2009100 WHAT WORKS ON WALL STREET TABLE 618 ‘Average Annual Compound Rates of Return by Decat 1960s" 1970s 1980s 1990s 2000st Large Stocks earnings/price decile 1 8.48% 138% 18.48% 16.08% 11.60% Large Stocks earnings/price decile 10 15.07% 093% 13.52% 18.53% ~188% Large Stocks 8.16% 6.65% 17.34% 1638% 242% *Rtuns for January 1, 1964, to December 31,1968 + Returns for January 1, 2000, t December 31, 2008, ALL DECILES As Figure 6.5 and Table 6.19 show, for the All Stocks universe, all four of the cheapest deciles outperform the All Stocks universe, while having lower standard deviation of returns than the All Stocks universe itself. Deciles $ through 10 all underperform with deciles 7 through 10 having significantly higher standard deviations than the All Stocks universe. Decile 10—the most expensive stocks with the highest PE ratios—underperforms an investment in U.S. T-bills. 18.0% -—— : a 163% 16.0% 1 ggy 15.05. 239% 14.0% - = 123% 12.0% + = aki 10.0% | sox + ee 4.0% 20% 9.9% cos +a th 2 3 4 5 (Highest) FIGURE 65 8 9 10 Allstocks (Lowest) All Stocks averag December 31, 2008 mnnual compound return by decile, Ea 198 Pri verse, January 1, 1964, to‘Results for Earnings/Price Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2008 $10,000 grows to Average return Compound return Standard deviation _Sharpe ratio $10,202,345 18.23% 16.25% 18.45% 061 $1,256,873 16.89% 15.40% 16.10% 0.65, $3,835,747 15.19% 1381% 15.56% 087 $2,037,160 13.68% 12.25% 15.85% 046 ‘$1,208,080 12.50% 10.99% 16.40% 037 77740 11.58% 391% 17.21% 028 $530,331 11.08% 9.02% 19.28% 021 $313,515 10.43% 178% 21.95% 0.13 $166,562 9.35% 631% 359% 0.06 $118,820 937% 5.53% 2652% 0.02 $1,329513 13.26% 11.22% 18.99% 033 "For Large Stocks, as Figure 6.6 and Table 6.20 show, we see that the four cheapest deciles all do better than the Large Stocks universe, whereas the bottom six all underperform Stocks, with decile 10 doing just slightly berter than T-bills. 360% 140% 20% 7 8 9 10 Lame (Lowest) stocks Earnings/Price universe, January 1, 1964. toWHAT WORKS ON WALL 120 TABLE 717 ‘Average Annual Compound Rates of Return by Decade—All Stocks 19605" 1970s 1980s 1990s ‘All Stocks EBITDA/enterprise value decile 1 14.18% 13.82% 19.95% 18.59% ‘All Stocks EBITDA/enterprise value decile 10 18.72% 261% 1187% 13.28% All Stocks 13.36% 7.56% 16.78% 15.35% aturns for January 1, 1964, to December 31, 1968. +t Retums for January 1, 2000, to December 31,2008, TABLE 718 ‘Average Annual Compound Rates of Return by Decade—Large Stocks 1960s" 1970s 1980s 1990s Large Stocks EBITDA/enterprise value decile 1 §.35% 12.75% 19.31% 16.96% Large Stocks EBITDA/enterprise value decile 10 15.70% o% 10.03% 21.78% Large Stocks 8.16% 6.65% 17.34% 16.38% * Returns for January 1, 1964, o December 31, 1968. + Returns for January 1, 2000, to December 31,2008. DECILES As Figures 7.5 and 7.6 and Tables 7.19 and 7.20 show, deciles 1 through 5 of EBITD, from All Stocks all outperform the All Stocks universe, underperform the All Stocks universe. Deciles 9 and 10 hi underperforming 30-day U.S. T-bills. 806 ————_—___— — 160% + (highest) FIGURE 75 ‘Average annual compound return by EBITDA/enterprise December 31, 2009 whereas deciles 6 through 10 ave the dubious distinction of iue decile, All Stocks universe, January 1, 1964, to All StocksEBITDA to Enterprise Value 121 16.0% - = — = 139% 140% - - — : 125% 120% - 113% ____— 102% 100% 2 = 80% - — - 6.0% +s — sa" 4.0% ~ - — § ——_— 2.0% ~ —a — —e- — 0.0% : = = = = 1 2 3 4 5 6 7 8 9 10 large (Highest) (Lowest) Stocks FIGURE 76 Jpvorage annual compound return by EBITDAVenterprise Value decile, Large Stocks universe, January 1, 1964, to December 31, 2009 TABLE 7.19 Results for EBITDA/Enterprise Value De‘ f 31, 2008 “Analysis of All Stocks Universe, January 1, 1964, to ‘$10,000 grows to 1 (highest) uigi4yi7 $8,275 696 4972319 273427 $1,620973 $1,072,180 $775,885 $316,287 106512 10 lowest) ‘$109,001 All Stocks $1,329513 TABLE 720 “Jos EBITDATEnterprise Value Decile Analysis of Large Stocks Universe, January 1, 1964, to “napo0 grows to Average return __ Compound return ‘Standard deviation Sharpe ratio. 15.52% 13.91% 16.82% 053 13.88% 12.52% 15.58% 0.48 12.53% nT 15.09% aa {continued on next page)-to-Sales Ratios 163 TABLE 919 mary Results for Sales to Price Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009 $10,000 grows to Average return Compound return Standard deviation _ Sharpe ratio highest) $5,044,457 16.5% 14.49% 20.68% 0.46 $5,128,129 16.61% 145% 18.98% 050 $3,587,828 15.57% 13.64% 18.32% 047 $2,744,842 14.78% 12.98% 17.74% 045 $2,075,249 1387% 1230% 17.12% 0.43 $1,465,785 12.95% 11.45% 16.32% 0.40 $806,593 151% 10.01% 16.37% 031 $965,406 9.89% 8.14% 17.89% 018 $173,581 8.84% 6.40% 21.16% 007 1 lowest) $45,711 7.05% 3.36% 26.24% 0.06 $1,329513 12.26% 11.22% 18.99% 033 160% + ~~ | 14sm 145% 9 10 Allstocks (Highest) (Lowest) ‘annual compound return by Sales to Price decile, All Stocks universe, January 1, 1964 to December 31, 2009 20 percent of Large Stocks with the lowest PSRs. Base rates are also better for decile 2, that group beating Large Stocks in 81 percent of all rolling five-year periods and in 94 t of all rolling ten-year periods. Yet, as we saw with All Stocks, after the decile 1 and ion, each decile declines in returns as you climb the PSR mountain, with decile 10°s behind those from U.S. T-bills. Table 9.20 and Figure 9.6 summarize the findings for Stocks.164 WHAT WORKS ON WALL STREET TABLE 920 Summary Results for Sales to Price Decile Analysis of Large Stocks Universe, January 1, 1958, (0 December 31, 2009 Decile $10,000 grows to Average return __Compound return tion _ Sharpe ratio 1 thighest) $1,470,652 13.16% 11.48% 037 a $1,872,056 13.38% 17% 040 3 1324507 12:76% 11.21% 037 4 $1,188,394 12.19% 10.8% 038 5 $1,037,505 11.88% 10.82% 037 6 $553 407 10.27% 8.12% 028 7 4865 962 10.08% 8arh 026 8 $369,957 9.55% 8.17% 020 8 $04 900 43% IN 015 10 (lowest) $82,579 738% 470% 001 Large Stocks $872,861 11.72% 10.20% 032 140% roe 15818 ag 10.6% aa 10.0% eerie, 82% say 8.0% : 60% : am 4.0% 2.0% 00% Pe (Highest) (Lowest) Stocks FIGURE 96 ‘Average annual compound return by Sales to Price decile, Large Stocks universe, January 1, 1964, to Decomber 31,2009 IMPLICATIONS While no longer the king of the hill, stocks with low price-to-sales ratios beat the market con- sistently over time. The only time high PSR stocks manage to beat their benchmarks is when investors are in an irrational and exuberant frenzy, throwing caution (and smart investment decisions) to the wind and dashing headlong into stocks with the sexiest stories. Yet as we've seen, those times are brief and there was hell to pay after the party, with both the All Stocks: and Large Stocks high PSR deciles doing worse than keeping your money in cash. Both suffered ‘extreme declines from which they have not yet recovered. We'll see later how incorporating PSR into a group of other value ratios provides even better protection for investors.Price-to-Book Value Ratios: A Long-Term Winner with Long Periods of Underperformance 187 of looking at a variety of factors, as we do in Chapter Fifteen. Finally, Figures 10.5 and 10.6 revisit the return by decile for the All Stocks universe and the Large Stocks universe. Now let’s move on to dividend yield to see what it tells us about a stock's prospects. 14.0% 2% aon 12.0% +2238 12% 10.6% 105% 10.0% - 32% am 85% oe 2.0% 6.0% 40% 4 2.0% 0.0% 2 a =F 22 8 SB 8 0 stoeks (Highes) (Lowest) FIGURE 105 ‘Average annual compound return by book-to-price decile, All Stocks universe, January 1, 1927, to December 31, 2009 120% - 210% 105% 104% 10.2%, 100% %*—s5—2 97% 83% 20% 1%, 60% 2.0% 1 20 3 8 5 6 7 8 9 0 Large (ighest (Lowest) Stocks FIGURE 106 ‘Average annual compound return by book-to-price decile, Large Stocks universe, January 1, 1927, to Decomber 31, 2009‘Dividend Yields: Buying an Income ABLE 11:13 inued) ry Results for Dividend Yield Decile Analysis of All Stocks Universe, January 1, 1927, to December 31, 2009, 10,000 grows to Average retun Compound return __—Standard deviation _Sharpe ratio 5 $86,874,622 13.45% 115% 1851% 035 6 ‘$65,007,160 13.14% 11.18% 1891% on 7 $50,206,616 12.85% 1081% 19.10% 030 8 41,851,221, 1273% 1057% 19.67% 028 . $32,361,872 1259% 10.23% 81% 025 $21,823,969 12.48% 37% 22.18% 02 $99,542,780 1306% 10.48% 2187% 025 Yield Decile Analysis of Large Stocks Universe, January 1, 1927, to December 31, 2009 $10,00grows to __Average return ‘Compound return Standard deviation Sharpe ratio $51,678,232 1290% 1085% 19.36% 030 $59523,837 1281% 11.04% 5% 033 60,461,472 12.87% 11.06% 18.05% 038 $50,153,882 12.59% 1081% 17.96% 032 $29,832,588 11.36% 10.11% 18.35% 028 $23,573,631, 182% 381% 18.10% oz 19314301 144% 954% 18.56% 024 19,021,660 11.0% 95% 18.39% 03 18,188,725 11.80% 9.46% 20.57% om ‘0,405,908, 11.18% 846% 213% 018 s21ei7g72 175% 8.69% 19.35% 02 ree re sa ~ igs) 13 10 (Lowest) Allstocks ‘compound return by dividend yield decile, All Stocks universe, January 1, 1927, to December 31, 2009200 WHAT WORKS ON WALL STRI (Wighest) (Lowest) Stocks FIGURE 114 Average annual compound return by dividend yield decile, Large Stocks universe, January 1, 1927, to December 31, IMPLICATIONS High-dividend yields alone do not provide a constant panacea for investments in broa indexes or a more specific strategy that includes other factors. In other words, you're nt going to want to rely just on dividend yield to make your investment choice. We see ii Chapter Thirteen that shareholder yield does a better job and offers a smoother ride thi dividend yield alone. Most investors’ maximum time horizon is five years, and after revie ing Figure 11.1 we see that there have been a significant number of five-year periods i which high-yielding stocks did substantially worse than the market. Indeed, between 19} and 2000, the highest-yielding stocks from All Stocks actually underperformed the Stocks universe, if only by a small margin of five basis points. Investors who want to dividend yields as a sole determinant of value should stick to large, better-known com nies, since those companies typically have stronger balance sheets and longer operating his tories. Indeed, in the second case study at the end of this chapter, we see that when y include other criteria such as strong cash flows, large sales, and large numbers of s! outstanding, large stocks with high-dividend yields offer excellent risk-adjusted returns. CHAPTER ELEVEN CASE STUDY 1 DO HIGH PAYOUT YIELD AND DIVIDEND CUTS EFFECT DIVIDEND STRATEGIES? People who favor stocks with high dividend yields often pay close attention to how mi of a company’s earnings are being paid out in dividends. The reasonable assumption is thi stocks that payout a great deal of their earnings in the form of dividends might not perfo as well as those that choose instead to reinvest in the company's business. While I do n222 WHAT WORKS ON WALL STREET 16.0% 140% 120% 100% + 1 2 3 4 5 6 7 8 9 10 Allstocks (tighest) (lowest) FIGURE 125 ‘Average annual compound return by buyback yield decil All Stocks universe, January 1, 1927, to December 31, 2008 For Large Stocks, deciles 1 through 7 all beat Large Stocks, but deciles 4 and 6 do only marginally most likely because of the problem mentioned above on placing st with zero values in the right decile. As with All Stocks, only deciles 9 and 10 should completely avoided (see Table 12.24 and Figure 12.6). TABLE 1228 ‘Summary Results for Buyback Yield Decile Analysis of Large Stocks Universe, January 1, 1927, to December 31, Decile ‘$10,000 grows to Average return Compound return Standard deviation Sharpe. (highest) $50,019,446 15.88% 12.98% BN% 035 2 145,516,544 15.15% 12.24% B2% 031 3 $113.823,003 14.68% 11.91% 276% 030 4 ‘$25,737,208 1251% 992% 220% 022 5 $38,376,139 12.46% 10.45% 19.09% 023 6 $22,742,972 11.90% 976% 19.72% 02 7 $53,593,671 13.19% 10.90% 20.32% 023 8 $19536,117 11.84% 9.56% 20.28% oz 9 $8,615,546 10.90% 848% 21.18% 0.16 10 (lowest) $1,397,168 867% 6.13% 21.93% 0.05 Large Stocks $21,617,372 11.75% 969% 19.35% 02410 Large (lowest) Stocks le, Large Stocks universe, January 1, 1927, to December 31, 2009, studying the most effective way to return cash to shareholders, repurchasing shares open market provides higher returns than paying cash in the form of dividends. with the highest buyback yields do significantly better than the market—over the years besting the All Stocks universe by 3.23 percent on an average annual com- basis. Stocks with the lowest buyback yields perform horribly, losing 4.52 percent to the All Stocks universe. The spread between the best buyback group and the is 7.75 percent per year and seems to be very consistent when we analyze the rolling generated over all five-year holding periods. In that analysis, we see that the group worst buyback yields consistently lose to the All Stock universe over 98 percent ling five-year periods. All in all, buyback yield is an excellent factor to look at mining a stock’s relative attractiveness.240 WHAT WORKS ON WALL SI 10 all stocks (Highest) (Lowest) FIGURE 135 ‘Average annual compound return by shareholder yield decile, All Stocks universe, January 1, 1927, to December 31, TABLE 1319 ‘Summary Results for Shareholder Vield Decile Analysis of All Stocks Universe, January 1, 1927, to December 31, Decile $10,000 grows to Average return Compound return _ Standard deviation _ Sharpe highest) ——$298,963,138 18.50% 1322% 20.19% ost 2 164,632,206 14.35% 12.41% 18.66% 0.40 3 138,927,917 14.08% 12.18% 18.43% 033 4 '$116,208,174 13.93% 11.94% 18.91% 037 5 ‘$71,061,645 13.42% 11.28% 19.61% 032 6 $47,745,958 13.15% 10.75% 20.71% 028 7 $21,990,224 12.55% an% 2.60% on 8 $9,775,556 11.91% 8.65% 24.65% 0.18 9 $14,358,062 13.01% 9.15% 25.68% 0.16, 10 (lowest) $1,334,762 9.58% 607% 25.78% 0.04 All Stocks $38,542,780 13.06% 10.46% 267% 025 “The full decile analysis for the Large Stocks universe is similar to that for All St Figure 13.6 and Table 13.20 show that the returns by decile again decline in order, the 10 percent of stocks in decile 1 with the highest shareholder yield strongly outperfor ing those in decile 10 with the lowest shareholder yield. The spread between decile 1 decile 10 is 6.87 percent a year over the 83 years studied, also creating a huge gulf in minal portfolio value.older Yield 24 14.0% — 128% 20% i 306% 104% 10.0% 3.0% 4 6.0% 40% 20% a 2 a 4 is 6 7 8 9 10 Large (Highest) (lowest) stocks RE 136 Large Stocks universe, January 1, 1927, to E 1320 “Results for Shareholder Vield Decile Analysis of Large Stocks Universe, January 1, 1927, t December 31, 2008 ‘$10,000 grows to Average return Compound return Standard deviation _Sharpe ratio 217,331,288 14.86% 12.79% 1931% 040 ‘$91,230,690 13.40% 11.61% 199% 037 $41,545,263 12.32% 10.58% 17.91% 031 $37,632,925 12.22% 10.43% 18.07% 030 $28,687,731 11.94% 10.07% 18.37% 028 $22,224,338 11.76% 9.73% 19.11% 025 $9,720,088 11.00% 8.64% 20:73% a8 $13,821,435 111% 8.10% 217% aig $8,115,198 11.42% BAN 23.57% 14 1,182,249 8.89% 5.92% 23.90% 0.04 $21,617,372 175% 9.60% 19:35% 024 ‘older yield and buyback yield are superior to dividend yield alone in selecting ile buyback yield alone beats dividend yield and shareholder yield, shareholder re muted declines than buyback yield and superior base rates to buyback yield.244 ‘WHAT WORKS ON WALL STREET even though the company knows that it will be unable to sell its products but it then accrues those “sales” to create a number that might be ultimately bogus. For this reason, many believe that stocks with the highest accruals-to-price ratios will ultimately be the most likely to have negative earnings surprises that will adversely affect a stock’s price. Let’s see if this is true. Table 14.1 shows the results for $10,000 invested between 1963 and 2009 in each decile from All Stocks ranked by accruals-to-price, with the tenth decile being the 10 percent of stocks from All Stocks with the lowest accruals-to-price. Stocks with the lowest accruals-to- price clearly do better than the market, with the three deciles with the lowest accruals-to-price all beating the All Stocks universe, and the two deciles with the lowest accruals-to-price doing | so by a significant margin. At the same time, the two deciles with the highest accruals-to-price do significantly worse than the All Stocks universe. Table 14.2 shows the detailed results for be under pressure to “pad” sales by doing such things as sending excess inventory to stores deciles 1 and 10. | TABLE 141 Summary Results for Accruals-to-Price Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009 Dex $10,000 grows to Average return Compound return ation Sharpe ratio (highest) 487,988 137% 879% 018 2 463,666 1.12% 870% a8 $585,626, 11.85% 9.25% 020 $708,311 12.13% 371% 023 $947,994 12.57% 10.40% oz $1,244,211 12.97% 11.08% 033 ‘$1,105,662 12.44% 107% 033 81,487,495 12.94% 11.45% 040 $2,874,786 14.66% 13.10% 04g 10 (lowest) $4,800,813 16.70% 14.36% oar All Stocks $1,329,513 13.26% 2% 033 TABLE 142 ‘Summary Annual Return and Risk Results Data: All Stocks Accruals-to-Price Decile 1, All Stocks Accruals-to-F Decile 10, and All Stocks, January 1, 1964, to December 31, 2009 All Stocks accruals-to-price All Stocks accrual decile 1 decile 10 Al Stocks Arithmetic average 1.37% 16.70% 13.26% Geometric average 879% 14.36% 122% Median return 13.62% 19.06% 17.16% Standard deviation 2AT% 20.13% 18.99% Upside deviation 12.57% 13.10% 10.98% Downside deviation 18.47% 14.94% 13.90% (continued on next pagel257 Results for Cash Flow to Debt (%) Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009 ‘$10,000 grows to _ Average return Compound retwm Standard deviation Sharpe ratio $1,708,196, 142% 118% 0.65% 03 $2,142515, 14.40% 1238% 19.09% 039 $2,286,232 14.30% 125% 17.99% oa 2715232, 1477% 1296% 17.80% 045, $2,498,614 1450% 1275% 17.88% 043 $3,123,702 18.14% 1330% 17.89% 048 $2,358,119 14.36% 1261% 11.52% 043, $1,122,798 12.47% 1081% 11.20% om ‘172921 8.53% 639% 19.70% 007 $29,894 6.20% 241% 6.84% 0.10 S1329513 13.26% 2% 18.9% 0x LE 1426 ‘Annual Return and Risk Results Data: All Stocks Cash Flow to Debt (%) Dec Decile 10, and All Stocks, January 1, 1964, to December 31, 2009 1,All Stocks Cash Flow to Al Stocks cash flow to Al Stocks cash flow to debt (%) decile 1 dobt(%} decile 10 All Stocks 42% 620% 13.28% 11.83% 241% 11.22% 17.07% 10.49% 17.16% 20.85% 25.0% 18.99% 13:17% 1738% 10.98% 14.29% 20.16% 13:90% 625 1228 0.00 positive periods 36 39 329 negative periods 216 213 23 jpeak-to-rough decline 54.86% 22.73% 55.50% 108 128 1.00 im = 0) 439 153 447 (at = 5%) 033 0.10 033 (MAR = 10%) 13 038 009 $1,708,196, $2998 $1,329,513 46 88% -1156% 48.49% 121.96% 17331% 28.19% -1781% 52.46% = 18.68% 220% 027% 31.49% 718% 20.23% ~a0% 33.26% 288% 186% 313% 20.10% 822% 26.8% 345% BIM% 309% 14.83% 1.01% 291% 052% 2.05% 208% 148% -073% 55.52% 53.99% 51.26% rithm ern minus 2 tines the standard deviation Jxpocted return is ethmete return plus 2tmes the standard devition.265 unting Ratios For this decile analysis, we have data only from September 30, 1971, s0 this study cers just over 38 years. Table 14.41 shows the 10 deciles from All Stocks based upon rs INGnancing. Decile 1 is made up of the 10 percark ‘of companies from All Stocks that ce the highest external financing, whereas “Jecile 10 is made up of the 10 percent of cony Ses with the lowest use of external financing. Clearly, as Table 14.41 shows, high levels ‘nancing are toxic to a company’s stock remamns. Decile 1—the 10 percent of A from All Srocks with the highest external fnancing-—actually loses money over the ead, whereas decile 10—the stocks with the lowest external financing— ter than the All Stocks universe. Over the full period of the study, hrank to $9,759, an average annual compound loss of -.06 7 And that’s before taking inflation into accoure, If you did consider inflation, the | value of the $10,000 would have been a mere $1,843. So much for buying stocks oat generate their own financing from cash flow. forms significantly ,000 invested in decile 1 sl Le 14a ‘to December 31, 2008 iverse, October 1, 1971, Results for External Financing Decile Analysis of Al ‘Stocks Uni ‘$10,000 grows to Average return Compound rotum — Standard deviation _ShavPe ratio $9,759 4.16% 0.06% 78.28% 0.18 $95,130 897% 6.07% 22.98% 0.05 $348,976 1202% 973% 20.17% 023 $588,344 13.46% 11.24% 19.82% 031 $853,099 1467% 12.33% 231% 036 181,940,145 1579% 13.68% 13.20% 045, $1,078,798 14.75% 13.02% 143% 0.48 $1,685,795 15.90% 14.27% 16.88% 058 $1,443,408 15.47% 13.88% 16.62% 053 ‘$1,535,774 15.80% 14.07% 11.34% 082 13.35% 11.28% 19.34% 0.32 ile one also has the dubious distinction of having a maser decline of 92 per- mm peak to trough; you can find all the other sume information in Table 14.42. n Pe@ifor decile 1—featured in Table 14.43—are negate, with the group beating in just 7 percent of all rolling five-year periods and not even one rolling ten-year “The only time that stocks with high external financing do well is during times of inflation. Their best five years were the five years ending November 1980, when ind return of 30.22 percent. fe 10-—featuring the stocks with the lowest exvernal financing—tells a completely story. Here, $10,000 invested grows t0 $1,535,774, an average annual compound ei 14.07 percent and nearly three times the $587,200 you'd have earned from an tia the All Stocks universe. The base rates_-fearured in Table 14.44—are over: positive, with the group beating All Stocks in 91 percent of all rolling five-year ae e100 percent of all rolling ten-year periods.279 1t to which operating/reporting outcomes provoke excessive investor optimism ... In words, a high level of net operating assets, scaled to control for firm size, indicates of sustainability of recent earnings performance. Looking at Table 14.65, we see that the data confirm their hypothesis. Decile 1, those with the highest change in net operating assets, earned an average annual compound of just 3.5 percent, significantly worse than the All Stocks universe and worse than tment in U.S. T-bills. Deciles with the lowest change in net operating assets—espe- deciles 9 and 8—significantly outperformed an investment in All Stocks, earning and 14,09 percent, respectively. The detailed information on deciles 1 and 10 is fea in Table 14.66. The base rates for decile 1—those stocks with the highest change in rating margins—are atrocious. They managed to beat the All Stocks universe in just of all rolling five-year periods and in none of the rolling ten-year periods and are in Table 14.67. The base rates for the best performing decile—decile 9—is fea- in Table 14.68 and reveals the opposite, beating All Stocks in 98 percent of all rolling. 1 periods and in 100 percent of all rolling ten-year periods. Table 14.69 shows the rates for decile 10. LE 1465 Results for NOA Chant (2%) Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009 ‘$10,000 grows to Average return Compound return Standard deviation Sharpe ratio $48,817 6.86% 3.50% 2494% 0.08 $317,719 1053% 81% 22.16% 013 $619,971 11.60% 9.39% 19.90% 022 $1,101,751 12.59% 10.76% 17.98% 032 $1,561,767 13.17% 11.81% 16.51% 0.40 $1,804,138 13.38% 11.98% 15.91% 0.44 $2,741,682 14.38% 12.98% 1872% ost $4,303,364 15.74% 14.09% 16.99% 054 $4,915,954 16.43% 14.42% 18.60% 051 $1,859,794 14.76% 12.08% 21.99% 032 $1,329,513 13.26% 12% 18.99% 033 E1466 /Annual Return and Risk Results Data: All Stocks NOA Change (*%) Di ‘and All Stocks, January 1, 1964, to December 31, 2008 1, All Stocks NOA Change (%) All Stocks NOA change (%) All Stocks NOA change (%) decile 1 All Stocks 6.86% 14.76% 13.26% 3.50% 12.03% 11.22% 12.47% 20.88% 17.16% 74.94% 21.93% 18.93% 144T% 13.58% 10.98% 18.64% 15.91% 13.90% {continued on next page)288 WHAT WORKS ON WALL S} TABLE 1481 Base Rates for Large Stocks TATA Decile 1 and Large Stocks, January 1, 1964, to December 31, 2009 Large Stocks TATA decile tem 1 boat Large Stocks Percent Single-year return 159 out of 541 29% Rolling 3-year compound return ‘Wout of 517 14% Rolling 5-year compound return 32 out of 483 6% Rolling 7-year compound return 2out of 469 5% Rolling 10-year compound return 1 out of 433, 0% TABLE 1482 Base Rates for Large Stocks TATA Decile 10 and Large Stocks, January 1, 1964, to December 31, 2008 Large Stocks TATA decile Ave Item 10 beat Large Stocks Percent excess Single-year return 345 out of 541 64% 2.64% Rolling 3-year compound return 380 out of 517 18% 2.03% Rolling 5-year compound return 384 out of 493 78% 1.69% Rolling 7-year compound return 363 out of 469, ™% 12% Rolling 10-year compound return 353 out of 433, 82% 145% IMPLICATIONS Much as we saw with accruals to price, companies that are playing games with the end up disappointing investors with negative earnings surprises. Indeed, while we are focused on what attributes identify attractive stocks, several of these accounting vari might be very helpful for investors interested in shorting stocks, as several of them excellent job of identifying stocks that go on to crash and burn, TOTAL ACCRUALS TO AVERAGE ASSETS (TAAA) Another take on the total accruals phenomenon is to compare them with a company’s ‘age assets, using the same assumption that companies with lower total accruals to a assets will provide higher-quality earnings and that companies with higher total acer average assets will have lower-quality earnings and might therefore be bad investmer order to get the figure for average assets through the Compustat dataset, we must use terly data that start in 1971. So, for this test, we start on September 30, 1971, and i $10,000 in the various deciles by total accruals to average assets. As Table 14.83 makes clear, much as we saw with total accruals to total as lower, the better. The $10,000 invested in decile 10—those stocks from Alll Stocks wi lowest total accruals to average assets—grows to $1,188,797, an average annual pound return of 13.31 percent, That’s over two times the $587,200 you'd haveAccounting Ratios 289 the universe compounded tive, with the stocks with investing in the All Stocks universe over the same period, where se in 71 percent of ar 11.24 percent. All base rates, featured in Table 14.6%, are Post She Icwest total accruals to average assets beating the All Stocks univers Gil rolling five-year periods and in 73 percent of all rolling ten-year periods. ABLE 1483 Ronults for TAAA Decile Analysis of All Stocks Universe, October '- 1971, to December 31, 2009 Compound return Standard deviation _ Sharpe ratio ‘$10,000 grows to Average return $17,835 591% 152% 878% 2 $163,775 10.65% 7.58% 23.49% 3 $458,772 12.91% 10.52% 20.58% 4 s710,848 1375% 11.78% 18.54% 5 S773 498 13.81% 12.04% 17.63% 6 $808,397 13.81% 12.17% 11.01% 1 $1,082,228 14.87% 13.03% 17.01% 8 $1,270,308 15.23% 13.50% 17.40% a ‘$1,252,980 15.29% 13.46% 17.85% $1,188,797 18.65% 1331% 0.29% $587,200 13.35% 11.20% 19.34% 032 BLE 1484 aioe for All Stocks TAAA Decile 10 and Al Stocks, October 1, 1871, to De smber 31, 2008 Al Stocks TAAA deci ‘Average annual 10 beat All Stocks Percent excess return 258 out of 448 58% 2.19% 3-year compound return 269 out of 424 63% 1.78% 5-year compound return 284 out of 400 1% 1.80% 77-year compound return ‘216 out of 376 73% 1.86% 10-year compound return 247 out of 340 73% 1.98% The $10,000 invested in decile 1—those stocks with the highest total accruals to sspets—tells quite a different story. Here $10,000 grows to 4 met! $17,835, an ase | compound return of just 1.52 percent. That's significantly worse than your money in T-bills. And if you adjusted the rerum for inflation, you would find Yn the poor house. All base rates, featured in Table 14.85, are negative, with the hich the highest total accruals to average assets beating All Stocks i js 8 percent rolling five-year peri .ds. You definitely want to ‘ods and in no rolling ten-year perio ay from companies playing with the numbers. Table 14.86 shows the detailed for deciles 1 and 10.Accounting Ratios 293 TABLE 1490 Base Rates for Large Stocks TAAA Decile 1 end Large Stocks, October 1, 1971, to December 3, 2008 Large Stocks TAA decile ‘Average annual em ‘beat Large Stocks Percent excess return ‘Single-year return 179 out of 48 0% 440% alin 3-year compound return 114 out of 424 m% 536% “Bolling 5-year compound return 7B out of 400 19% 5.80% ‘compound return 53 out of 376, 1% 6.10% 33 out of 340 10% 8.19% to avoid stocks that have high total games with the books end up with prises than com- ‘we saw with total accruals to total assets, you want als to average assets. Companies that are playing ‘more opportunities to disappoint investors with negative earnings sur] i vativ h to bookkeeping. IMPOSITED ACCOUNTING RATIO Sspend the entire next chapter investigating whether we might get better, more consistent by looking at how stocks score on all value factors rather than just looking at the fac- “individually. It is reasonable to assume that a stock that scores very well on PE ratio and "A-to-enterprise value arid price-to-sales might do better than a stock that scores well but not another. Let’s start the investigation with our accounting variables and see if ts of combining several accounting variables do better than they do individually. cifically, we look at a combination of the followings Total accruals to total assets Percentage change in net operating assets (NOA) Total accruals to average assets Depreciation expense to capital expense we're looking for stocks with high earnings quality. For each combined group of ‘we assign a percentile ranking for each stock in the All Stocks and Large Stocks uni- ‘a scale of 1 to 100. Ifa stock has a total accruals to total assets ratio that is in the percent for the universe, it receives a rank of 100; if a stock has a total accruals to ratio in the highest 1 percent for the universe it receives a rank of 1. We follow ‘convention for each of the factors. Thus, if a stock is in the lowest 1 percent of the based on its NOA change, it gets 100; if in the highest 1 percent, it gets a 1. For in which higher is better, we reverse the rankings, so if a stock ranks in the upper of depreciation expense to capital expense, it gets a 100, whereas if it is in the low- f, it gets a rank of 1. If a value is missing for a factor, we assign a neutral rank ‘ll factors are ranked, we add up all the factors’ ranks and assign the stocks to upon their overall cumulative ranking. Those with the highest scores are to decile 1, and those with the lowest scores are assigned to decile 10.294 WHAT WORKS ON WALL Lee’s call this composited group of accounting variables the Farnings Quali Composite. As Table 14.91 demonstrates, by combining several accounting variables a single master composite, you increase the robustness of the analysis and get bettes Consistent overall results. Starting on December 31, 1963, $10,000 invested in the st fn decile 1 with the best scores for each of the accounting variables grew to $8,992,076 the end of 2009, a robust average annual compound return of 15.93 percent. Its also trillion more than the $1,329,513 you'd have earned with an investment in the All St TIniverse, which compounded at 11.24 percent over the same period. Risk was sli higher for the stocks with the best accounting scores, with the standard deviation of r Of 19.11 percent, slightly higher than that of All Stocks’ 18.99 percent. Yer the much hi Sbsolute return from the stocks with the best accounting variables generated a much hi Sharpe ratio of .57 versus .33 for the All Stocks universe. TABLE 1491 ‘Summary Results for COMP Earnings Quality Composite Decile Analysis of All Stocks Universe, January 1, 1954 t0 December 31, 2003 Decile ‘510,000 grows to Average return Compound roturn Standard deviation _ Sharpe 1 (highest) $8,992,076 18.07% 15.99% 19.11% 0s? 2 ‘4,902,190 16.23% 14.42% 11.74% 053 3 $2,837,680 14.79% 13.06% 17.36% 046, 4 $1,532,421 13.15% 1.56% 16.79% 039 5 $1,313,455 1277% 11.19% 16.81% 037 6 $1,616,169 13.14% 11.69% 16.05% 042 1 $1,020,448 1230% 10.58% 1751% 032 8 $592,812 11.32% 9.28% 19.12% oz 8 ‘$407,634 1081% 839% 20.84% 016 10 (lowest) ‘$60,037 7.29% 397% 20 79% 008 All Stocks $1,329,513 13.26% N2% 18.99% 033 The base rates for the Earnings Quality Composite are all positive, with decile 1 ing All Stocks in 94 percent of all rolling five-year periods and in 100 percent of all r ten-year periods. Table 14,92 shows the base rates for all other rolling periods. As you see from Table 14.91, the stocks with the worst scores from the Earnings Quality Comy do horribly. The same $10,000 invested on December 31, 1963, grows to just $60,037, average annual compound return of a mere 3.97 percent. That’s worse than an invest T-bills. What’s more, it also has a high standard deviation of return of 24.79 percent, brit ing its Sharpe ratio to an anemic ~.04, Its maximum decline is also dreadful, with a ma peak-to-trough decline of 73.54 percent, compared to 54.83 percent for decile 1 and 5S. percent for the All Stocks universe. The base rates for decile 10—those stocks with the Collective scores on the four accounting variables and featured in Table 14.93—were ious, beating the All Stocks universe in just 6 percent of all rolling five-year periods a no rolling ten-year periods.302 WHAT WORKS ON WALL $1 In the article, authors Siva Nathan, Kumar Sivakumar, and Jayaraman Vijayakumar that combining PE and price-to-sales ratios provided better returns than either factor Indeed, they claimed that, “We show that an investment strategy using both P/E and ratios simultaneously, and going long or short stocks based on the magnitudes of ratios, yields an average annual excess return of 28.89%. This far exceeds the annual returns of 16.01% reported by O'Shaughnessy using a strategy of going long the 50 si with the lowest P/S ratios each year.” 1 was intrigued by this notion of merging value variables, even though I felt that time period of their study—1990 through 1996—wasn’t long enough to truly test the cacy or robustness of the strategy. The period they studied did not include the technol bubble between 1997 and February 2000, and it was also a period when value factors erally enjoyed excellent returns. For example, between January 1, 1990, and December 1996, the decile of All Stocks with the lowest PE ratios returned 17 percent per year, a the decile of stocks with the best EBITDAJEV returned 19.37 percent; whereas the di with the lowest PSRs returned 13.74 percent, and the Alll Stocks universe gained 13.80 cent, Of the three value factors, PSR actually underperformed the Alll Stocks universe the other two did significantly better. Nevertheless, I felt it was a compelling idea decided to run a test over the full period for which we have monthly data for all the tors, from 1963 to 2009. First, we analyze just balance sheet and cash flow factors for a “pure play” combi value factor that we will call Value Factor One. It is made up of the following factors: Price-to-book Price-to-earnings Price-to-sales EBITDA/EV. Price-to-cash flow eens Next, we see if the addition of shareholder yield can improve the results of the play Value Factor One. Thus our second factor combination tests the five listed above also includes shareholder yield (buyback yield plus dividend yield). For each combined group of factors, we assign a percentile ranking (from 1 to 1 for each stock in the Alll Stocks and Large Stocks universes. If a stock has a PE ratio is in the lowest 1 percent for the universe, it receives a rank of 100; if a stock has a PE ra in the highest 1 percent for the universe, it receives a rank of 1. We follow a similar vention for each of the factors. Thus, if a stock is in the lowest 1 percent of the univer based on its price-to-sales ratio, it gets 100; if it’s in the highest 1 percent, it gets a 1. value is missing for a factor, we assign it a neutral rank of 50. For sharcholder yield a EBITDA/EY, those stocks in the 1 percent of the universe with the highest values are ranl 100, whereas those within the lowest 1 percent are be ranked 1, Once all factors ranked, we add up all their rankings and assign the stocks to deciles based upon their ov all cumulative ranking. Those with the highest scores are assigned to decile 1, while th with the lowest scores are assigned to decile 10.318, WHAT WORKS ON WALL STREET 18.0% 16.0% 40x 120% 10.0% 80% 50%. 40% 1 2 3004S 6 7 8 8 10 AllStocks (Highest) (lowest) FIGURE 155 ‘Average annual compound return by VC1 decile, All Stocks universe, January 1, 1964, to December 31, 2008 (Highest) (Lowest) Stocks FIGURE 156 ‘Average annual compound return by VC1 decile, Large Stocks universe, January 1, 1964, to December 31, 2009Combining Value Factors into a Single Composite Factor 323 SUBSTITUTING BUYBACK YIELD FOR SHAREHOLDER YIELD IN VALUE FACTOR TWO ‘Whar if you are interested only in stocks that have the best value characteristics, but you What torent as to whether they pay a dividend yield? In this case, you could add just Puy: back yield to the equation and create a new composited value factor that we will call Value Ractor Three. In this instance, we combine the following factors: Price-to-book Price-to-earnings Price-to-sales EBITDAEV Price-to-cash flow Buyback yield Just as we did with the first two composited value factors, we assign @ percentile rank to ead of the six factors. Thus a stock with a PE ratio in the lowest 1 percent of the uni- to ach fl receive a score of 100, and a stock with a PE in the highest 1 percent of the vervetee will receive a rank of 1. We follow a similar convention for each of the six fac- ave {fa value is missing for a factor, we assign a neutral score of 50. Once all factors are tanked, we add the scores and assign them ro deciles, with the 10 pereent of stocks with fhe highest scores going into decile 1 and the ten percent of stocks with the lowest scores cing into decile 10. Remember that decile 1 features the stocks with the lowest PEs, Re and highest buyback yields, whereas decile 10 is made up of the stocks with the jighest PEs, PSRs, and lowest buyback yields. Table 15.24 shows the results for decile 1 of Value Factor Three, and Table 15.25 ‘ows the base rates comparing decile 1 of Value Factor Three to decile 1 of Value Factor ee. Much as with Value Factor Two, including buyback yield enhances the performance + One—now the average annual compound return goes up to 'd deviation decreases to 17.68 percent and the Sharpe ratio es are virtually identical to Value Factor One, with the excep- year rolling periods increases to 96 percent against All ee whereas the base rates against Value Composite One, featured in Table 15.25, prove a bit over Value Composite Two. av eepe ABLE 1526 ry Return and Risk Results for Annual Data; All Stocks VC3 Decile 1 (High) and Alt Stocks VCt Decile 1, 4, 1984, to December 31, 2008 All Stocks VC3 decile 1 (high) All Stocks VC1 decile 1 192% 1909% 1739% 17.18% 491% 352% 17.68% 18.09% (continued on next page)(Combining Value Factors into a Single Composite Factor 337 ‘TABLE 1508.18 "Summary Return and Risk Results for Annual Data All Stocks VC3 High 25, All Stocks VC3 High 60, and All Stocks C3 Decile 1 (High), January 1, 1964, to December 31, 2009, Al Stocks VC3 high 25 All Stocks VC3 high 50 _All Stocks VC3 decile 1 (high) ‘Arithmetic average 70.60% 20.63% 19.22% Geometric average 18.46% 18.68% 17.39% Median return 21.49% 24.01% 24.91% ‘Standard deviation 19.06% 1837% 17.68% Upside deviation 12.72% 12.10% 11.48% ‘Downside deviation 13.64% 13.56% 13.30% Tracking error 9.38 861 752 ‘Number of positive periods 361 365 368 ‘Number of negative periods 191 187 184 laximum peak-to-trough decline 56.92% 55.86% 58.04% eta 0.88 0.86 0. 672 898 6.13 arpe ratio (Rf = 5%) on om 0.70 Sorti 062 0.4 0.56 10,000 becomes $24,288,981 $26 017,798 15,940,452 imum 1-year return 86.24% 44.88% —47.88% imum 1-year return 88.53% 85.38% 83.28% imum 3-year return =15.12% 14.20% -1737% imum 3-year return 45.78% 4257% imum 5-year return 392% -394% cimum 5-year return 39.98% 36.18% imum 7-year return 049% ~0.11% imum 7-year return 32.20% 30.43% jimum 10-year return 581% 647% imum 10-year return 79.09% 29.54% imum expected return* ~1153% =16.14% imum expected returnt 58.72% 54.57% imum expected return is arithmetic return minus 2 times the standard deviation. xximum expected return is arithmetic return plus 2 times the standard deviation. BLE 15¢s19 Rates for All Stocks VC3 High 25 and All Stocks, January 1, 1964, to December 31, 2009 All Stocks VC3 high beat All Stocks Percent excess return le-year return 420 out of 541 78% 6.98% F compound return 456 out of 517 8% 6.98% compound return 471 out of 493 96% 707% 19 7-year compound return 467 out of 469 100% 7.08% 1g 10-year compound return 433 out of 433 100% 7.02%346 WHAT WORKS ON WALL You must always consider risk before investing in strategies that buy stocks that significantly different from the market. Remember that high risk does not always mé high reward. Figures 16.4 and 16.5 show the compound average annual return and st dard deviations for the All Stocks strategies. All the higher-risk strategies are event dashed on the rocks, as Figures 16.2, 16.3, and 16.6 make clear. High Dividend Vic) TTT 13.30% Low 58 jameson 5.28% High BBY one erninen enamine 15.81% lowSH_ 1.0 5.06% High SH_YLO eens 15.56% High PSR joe 3.36% NO, LT 144% High £V/S mmm 5.06% Low ev/s — 15.79% High €V/E017OA 5.33% WEEE TT 16 55% High PE ems 5.53% Low PE Sa 16.25% High P/OCF “pam 3.49% Low P/OCE eters 16.25% High EV/CF 727% low EV/CF 16.10% High P/FCF 6.50% | Len PRE nem ncennememnmenri 84h High P/B00k 8.23%. Low P/Book 14.53% E Quality Comp Low | E Quality Comp High 15.3% FinStr Comp Low 487% Fin Str Comp High ome ren pm sete caaercncamat senannminai 15.07% Value Comp 1 Low Samm 2.9795 Vie COP 1 Hh SS 17.18% Value Comp 2 Low mamma 2.63% Value Comp 2 High SS 17.30% Value Comp 3 Low jum 2.47% Value Comp 3 High RR 17.39% Mi TT 21.2256 - 0.00% 5.00% 10.00% 15.00% 20.00% FIGURE 164 Compound average annual rates of return for the 46 years ending December 31, 2009, results of applying strategies the All Stocks universe“The Value of Value Factors 347 High-Dividend Yield Low BBY High BBY Low SH_YLD. High SH_YLD PSR Low PSR High Ev/s Low Ev/s High EV/EBITDA Low EV/EBITDA High PE Low PE High P/OCE Low P/OCF High EV/CF Low EV/CF High P/FCF Low P/FCF High P/Book Low P/Book lity Comp Low lity Comp High Fin Str Comp Low Fin Str Comp High “Value Comp 1 Low All Stocks 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% GURE 165 deviation of return for strategies from the All Stocks universe, 1964-2008 (higher is riskier) RACE CONSISTENCY should also look for strategies that do well over time and have the highest base rates to their universes, Stocks that score well on the various composited value factors as those with the best EBITDAVEV, the highest buyback and shareholder yield, the PE ratios and price-to-sales and price-to-cash flow ratios all have excellent long- base rates, On the other hand, stocks with the worst scores on the composited value , as well as those with high PE ratios and price-to-sales ratios and low shareholder buyback yield all have horrible base rates and usually only do well in unsustainable“The Value of Value Factors 343 All Stocks universe. However, higher risk does not always lead to higher returns. As Figure 16.2 shows, the higher risk of the high price-to-earnings, price-to-book, price-to- ‘cash flow, and price-to-sales ratios went uncompensated. Indeed, each of the strategies significantly underperformed the All Stocks universe. Buying the stocks in decile 1 of out ‘Value Composite Three turns $10,000 into $15,940,452 with a standard deviation of in of 17.68 percent, but buying the stocks in Value Composite Three in decile 10— f securities with the richest valuations—turns $10,000 into just $30,733, yer with a igher standard deviation of return of 28.14 percent. High Dividend Yield Low BBY High BBY | Low SH_YLD High SH_YLO High PSR Low PSR High EV/S. Low EV/S High EV/EBITDA Low EV/EBITDA High PE Low PE High P/OCF Low P/OCF High EV/CF Lowevicr | High P/FCF Low P/FCF High P/Book ratios for the various strategies applied to the All Stocks universe, 1964-2009 (higher is better)356 WHAT WORKS ON WALL STREEP You'll also need to think about the long-term results of our strategies after bear mar+ kets like those of 2000-2003 and 2007-2009. The back-to-back severe bear markets of the last decade soured many people on ever investing in the stock market, and such an asset allocation mistake will most certainly radically reduce the returns you earn on your invest ments. According to the Investment Company Institute, between December 31, 2006, a August 11, 2010, investors yanked nearly $159 billion from equity funds and poured m than $709 billion into bond funds. Given that bond yields are close to historic lows, 1 is almost certainly going to be a losing investment as time goes on because, when yiel increase, bond prices decline. It’s hard to remember the long-term prospects for equiti after suffering such a serious rout, yet it is when equities are priced well with low val tions and high-dividend yields that they go on to offer great returns for investors who the courage to stay true to their long-term asset allocation. Tables 16.1 and 16.2 show returns for the various strategies from All Stocks and Large Stocks by decades. TABLE 164 ‘Average Annual Compound Rates of Return by Decade, All Stocks Universe 1960s" 1970s 19808 All Stocks 13:36% 1.56% 16.78% VC3 high 16.50% 15.62% 22.56% VCS low 15.58% 1.48% 6.76% VC2 high 15.49% 1535% 23.21% VC2 low 16.29% 1.48% 6.719% VCt high 16.84% 14.78% 221% Vcr low 15.88% 220% 742% Financial Strength Comp high 15.13% 10.41% 1967% Financial Strength Comp low un 478% 10.34% Earnings Quality Comp high 11.30% 14.36% 20.46% Earnings Quality Comp low 14.95% 0.69% 5.89% Low P/Book 14.69% 13.29% 21.25% High P/Book 14.99% 280% 12.31% Low P/Free Cash Flow 14.40% 13.59% 172% High P/Free Cash Flow 151% 111% 132% Low Enterprise Value/Cash Flow 21.81% 11.89% 15.92% High EnterpriseValue/Cash Flow 15.00% TA2% 11.90% Low Price/Operating Cash Flow 15.86% 13.64% 19.27% High Price/Operating Cash Flow 16.70% 0.96% 719% Low PE 17.88% 13.03% 2038% High PE 14.74% 181% 9.21% Low Enterprise Value/EBITOA 14.14% 13.82% 19.95% High Enterprise Velue/EBITDA 1872% 261% 11.97% Low Enterprise Value/Sales 16.39% 12.44% 22.86% High Enterprise Vlue/Sales 16.41% 574% 12.95% Low PSR 16.90% 11.44% 20.03% High PSR 11.58% 5.60% 7.23% High SH_YLD 14.74% 13.09% 23.15%402 WHAT WORKS ON WALL TABLE 1918 ‘Average Annual Compound Rates of Return by Decade 19605" 19708 1980s 19908 Large Stocks ROE, decile 1 7.43% 206% 14.00% 18.77% Large Stocks ROE, decile 10 4.55% 181% 18.42% 16.99% Large Stocks 8.16% 6.65% 17.34% 16.38% “Returns for January 1, 1964, to December 31,1968. 1 Returns for January 1, 2000, to December 31, 2009, DECILE The decile analysis of the ROE stocks from All Stocks sees the top four deciles ou form the universe, but not by huge margins. Decile 5 matches the return for All Si yet decile 6 does somewhat better. It is not until deciles 9 and 10 that we see signif underperformance. The oddness of this distribution pattern seems to imply that inv should simply avoid stocks with the lowest ROE. Tables 19.19 and 19.20, as w Figures 19.5 and 19.6 summarize the results. 1 2 a 4 5 6 7 8 9 10 Allstocks (highest) (Lowest) FIGURE 195 ‘Average annual compound return by ROE decile, All Stocks unive , January 1, 1964, to December 31, 2009438 returns for both All Stocks and Large Stocks by 6- and 12-month price appreciation Fini Tables 20.43 and 20.44 show the returns for both groups from All Stocks and Large S by decade. 16.0% ————____—— 14a 140% — 17% 12.0% + — TEOK "10.6% 10.0% —_s- 80% — 6.0% + 4.0% 2.0% oo + 1 2 3 4 5 ‘ (Highest) FIGURE 209 ‘Average annual compound return by deci December 31, 2009 14.0% 123% 12.0% + — 21% 4, TR 08% 10.0% 6.0% 2.0% + + oon + 1 2 3 4 5 6 (Highest) FIGURE 2010 10.1% 6-month momentum, All Stocks universe, January 1, 1927, to 10.4% WHAT WORKS ON WALL 105% 80% 42% 8 9 10 AllStocks (lowest) 96% 39% 5.6% 8 9 10 AStocke | (Lowest) Average annual compound return by decile, 12-month momentum, All Stocks universe, January 1, 1927, to December 31, 2009Multifactor Models to Improve Performance 451 ABLE 218 inal Value of $10,000 Invested for Best and Worst Average Annual Compound Returns for Monthly Data, ry 1, 1928, to December 31, 2008 3-year Tyear any period period ll Stocks, 3 and 6 mo.smed, top 25 ‘SY minimum $10,000 value sani $2,499 $4837 10.144 $8,660 ll Stocks, 3 and 6 mo.>med, top 25 SY maximum $10,000 value $26,002 $35,153 $53,745 $78,745 $160,589 Small Stocks, 3 and 6 mo.>med, top 50 SY minimum $10,000 value 4512 $2556 $4880 $9561 $8,255 ll Stocks, 3 and 6 mo.>med, top 50 ‘SY maximum $10,000 value $28,319 $33,355 $50,088 S734) $128,530 3 Stocks minimum $10,000 value $3,308 $1,465 92.940 $5,733 $5,872 I Stocks maximum $10,000 value $33,348 $36,775 $62,313 $54,327 $89,249 iverse. The worst five years on a relative basis versus the Small Stocks universe were ending June 1937, when the strategy suffered a 99.1 percent absolute loss against Stocks, earning an average annual compound return of 37.57 percent versus 42.70 ent for the Small Stocks universe. Figure 21.1 shows the rolling excess (deficient) 1m for the strategy versus the Small Stocks universe. i average annual compound excess (deficient) return Small Stocks, 3 and 6 mo. >med, top 25 SY minus Small : January 1, 1928, to Decomber 31, 2009Multifactor Models to Improve Performance 461 BLE 2120 “Results for Strategies Benchmarked against the Returns of Small Stocks (Jan-28 to Doc-09}; Strategies ‘by Compound Return (Continued) Geometric Standard Excess Sharpe Tracking Largest ‘mean deviation (%) returns ratio__—error_—_ drawdown Bete 'Stocks>med by SY, top 25 8 15.99% 02% «= 5a SABE TBD UBT Stocks, 3 and 6 mo.>med, BS 15.98% 190% 54% OSB 785 -T88HH_ OB ‘Stocks >med by BP top 25 BBY 1567% 16% = 512 OMB TB ARTI 080 ‘Stocks >med by 3m, 6m 50 BP 15.61% 209% 500% =O 707 -73.00% © 086 Stocks, PB in top 3 deciles, }6 mo, mom>med! 25 by SY 1851% NI 490% = 0B -88STH OT ‘Stocks>med by SY, ‘0p 50 BP 1549% 1939% 49% = S421 787% =O Stocks >med by 3 and mo. top 25 BP 1535% 23% 40% = OMS 778.8% OT Stocks, PB in top 3d {6 mo, mom>median, B by 12m mom 1530% 242% = ATS 8H 88TH OR Stocks, 3 and 6 mo.>med, 50 SY 152% 1eavk = ATO 1a) 1838% = 078 ‘Stocks >med by BP, top 25 6 mo. 15.21% 72.10% «486% = 04S BB OKETH ORT Stocks, PB in top 3 deciles, and 6 mo. mom>median, ‘top 50by SY 15.15% 3.0% 480% OH 263 -8637% 088 ‘Stocks>med by BP, 6 mo, 109 5SY 15.14% 750% «= A50% =O NB 81.608 Stocks, PB in top 3 deciles, ‘Gand 6 mo. mom>median, _top SOby 12m mom 15.06% 786% = 450% OM 82 -8627% 098 i! Stocks>med by BP, ‘§rmo, top 50 BBY 1479% 0% = AMON TT RBH 080 I Stocks>med by BP, 6 mo, 1459% D2% AON OMS 795 -8352% «088 440% 211% «385% = OK} OSH 878% 080 Band6 mo, top 25 12mo. 14.28% ame = a0 161 -4282% = 097 all Stocks>med by SY, 3and6mo, top $0 12mo. 14.24% 21% = «369% = OM Tk TT O88 II Stocks>med by BP, nd 6 mo, top 50 12mo. 14.15% a a) all Stocks>med by BF, BBY, top 256 mo 139% 796% = 3am «= 001463 -8180%H IA (continued on next pagel462. WHAT WORKS ON WALL TABLE 2.0 ‘Summary Results for Strategies Benchmarked against the Returns of Small Stocks (Jan-28 to Dec-09). Strategi Sorted by Compound Return (Continued) Geometric Standard Excess. Sharpe Tracking —_Largest Strategy moan deviation (%) returns ratio error drawdown Small Stocks>med by SY, ‘and 6 mo, top 25 12 mo. 13.90% 260% 334% 039 923 16.45% Small Stocks>med by BP, BBY, top 50.6 mo. 13.59% 23.38% 303% 029 1444 -91.60% ‘Small Stocks>mad by BP, BBY, 3mo,, top 25 6 mo. 13.52% 28.47% 297% = 032, 1376 = 907% Small Stocks, PB in top 3 deciles, 3 and 6 mo, mom>0, top 25 by 12m mom 13.26% 26.45% 2n% = oat 1286 94.82% Small Stocks, PB in top 3 deciles, 3.and 6 mo, mom>0, top 25 by SY 13.24% 24.86% 260% 0331314 94.82% ‘Small Stocks, PB in top 3 deciles, 3 and 6 mo. mom>0, top 50 by 2m mom 13.14% 25.80% 259% 0321258 94.82% ‘Small Stocks>med by BP, BBY, 3mo., top 50 6 mo, 13.09% 26.04% 258% O31 1433 -9097% ‘Small Stocks, PB in top 3 deciles, and 6 mo. mom>0, top S0by SY 12.96% 25.02% 241% = 032 1290 -94.82% ‘Small Stocks>med by BBY, 3mo, top 5012 mo. 12.21% 25.07% 165% 028 1178 82.88% ‘Small Stocks>med by BBY, 3mo, top 25 12 mo. 12.06% 26.45% 150% 027 1227 82.88% Small Stocks 10.56% 23.19% NA 028 NA 86.12% TABLE 2121 ‘Summary Base Rate Information for all Ro id Ten-Year Periods for Strategies jenchmarked against the Returns of Small Stocks (Jan-28 to Dec-08); Strategies Sorted by Compound Return Percent of time Strategy 3years 5 years Tyears 10 Microcap BP top 3 deciles, 3 and 6 mo.>0, top 25 by 12m mom 736% 843% 907% 4.3% 949 Microcap BP top 3 deciles, 3 end 6 mo.>median top 25 by 12m mom 71.1% 828% 91.0% 33.2% Microcap BP top 3 deciles, 3 and 6 mo.>median top 25 by 12m mom 69.7% 820% 902% 926% Ex Microcap BP top 3 deciles, 3 and 6 mo.>0, top 50 by 12m mom 737% 85.1% 90.2% 93.1% saa ‘Small Stocks>med by SY, 6 mo, top 25 BP 740% 882% 944% 4.8% 30% Small Stocks, 3 and 6 mo.>m top 25 SY 749% 934% 946% 96.1% 93% (continued on next,470 WHAT WORKS ON WALL: DOUBLE VALUE WITH MOMENTUM. As Table 21.23 shows, between 1928 and 2009, there are two strategies from the All universe that have perfect ten-year base rates. The first strategy requires that stocks book-to-price greater than the median (i.e. eliminate the half of the stocks where inv are paying the most for every dollar of book value); have six-month price appreci greater than the median; and finally, buy the 25 stocks with the highest shareholder yi Not only does this strategy beat the All Stocks universe in 100 percent of all 1 ten-year periods, but it also beats the universe in seven out of every ten one-year hi periods. There is a price for the consistency however, as the strategy earned an a annual compound return of 15.32 percent, some 61 basis points behind the best- ing strategy. It also has a larger maximum decline of 82 percent and a lower Sharpe Nevertheless, a 100 percent base rate over all rolling ten-year periods between 1928 2009 is a rarity. The second strategy requires that stocks have both a book-to-price and sharehi yield greater than the median stock in All Stocks, and then buys the 25 stocks with the 12-month price appreciation. This strategy beat the All Stocks universe in eight out of rolling one-year periods and in 100 percent of all rolling ten-year periods. As with the egy above, you would pay a price for the strategy’s greater consistency—this str: earned an average annual compound return of 15.26 percent, some 67 basis points the best-performing All Stocks strategy requiring three- and six-month price appreci greater than the universe median and buying the 25 stocks with the best shareholder See Tables 21.20 through 21.25 for the summary results for other strategies. LONG-TERM SUCCESS OF MULTIFACTOR STRATEGIES For all three of our universes, history demonstrates that you can do vastly better ¢ passive investment in any given universe itself by using more than one factor to sel portfolio of stocks. Buying stocks with great price momentum or high shareholder alone works quite well, but we've seen that by combining these factors—first requ good price appreciation and then focusing on the stocks with the best shareholder yi generates significantly higher returns at lower levels of risk. This combination of value growth characteristics is one we revisit when we look at even more multifactor m using research from the Compustat dataset. For now, the long-term data show that uni several factors provides higher returns and lower risk than using single factor criteria, IMPLICATIONS Investors are best served by buying stocks that have jumped a series of hurdles rather just one. It is often best to marry both value and growth characteristics when looking stocks that will go on to offer investors the best absolute and risk-adjusted returns. seen that by combining single factors like price momentum and shareholder yieldDissecting the Market Leaders Universe: The Ratios That Add the Most Value 477 SUMMARY RESULTS FOR MARKET LEADERS STRATEGIES Here I confine myself to reviewing the summary data for each strategy, as my goal is to show you that what holds true for the broad All Stocks and Large Stocks universes is equally as compelling within the Market Leaders universe. In order to accommodate all the factors tested, the data here start on September 1965. Table 22.2 lists the various strategies and is sorted by average annual compound return. Notice that the high valuations for the factors are not as devastating to Market Leaders as they are for All Stocks and Large Stocks, primarily because of the sheer size of the market leading companies. For example, the three worst-performing Market Leaders strategies are buying the stocks from Market Leaders with the worst scores on Value Composite Three, which returned 7.71 percent per year; buying the decile of Market Leaders with the lowest shareholder yield, which returned 7.25 percent per year and, in last place, buying the 10 percent of stocks from Market Leaders with the lowest buyback yield (essentially those Market Leaders that were net issuers of shares), which returned 6.76 percent. TABLE 22 ‘Summary Results for Strategies Benchmarked against the Returns of Market Leaders Universe USA (Sep-65 to Dec-09}; ‘Strategies Sorted by Compound Return Geometric Standard Excess Sharpe Tracking —_Largest Strategy mean returns ratio error drawdown Beta Market Leaders, top two deciles Value Comp 2,top 25 by6mo.PA 1534% 17.17% 470% 0.60 603 52.74% = 0.98 Market Leaders, top two deciles Value Comp 2,top 50 by6mo.PA 15.00% 17.16% 3.73% «= 0.58 527 53.43% 1.00 Market Leaders top 25 by shareholder yid 1494% 17.00% 366% 058 652 -57.57% = 0.96 Market Leaders Composite Value 2 decile 1 (high) 14.84% —1852% «= 357% 053 710 62.16% 1.05 Market Leaders shareholder yield (%) decile 1 1A7M% = 16.77% 3.44% 058 638 58MM 085 Market Leaders, 3/6 mo. PA>median, top 25 by shareholder yid 1465% 16.32% = 3.38% = 058 509 56.33% = 0.95, Cornerstone Value improved 50 455% «16.19%» 3.28% = 0.58 511 55.01% 0.94 Market Leaders Composite Value 3 decile 1 (high) 1449% 18.29% 3.21% == 0.52 690 80.13% 1.03 Market Leaders EBITDA/EV decile 1 1439% 18.15% 3.12% = 052 753 -5221% = 1.01 ‘Market Leaders buyback Yield (%) decile 1 14.28% «= 16.52% © 3.01% = 0.56 524 83.60% 0.98 Market Leaders Composite Value decile 1 (high) 1427% 18.69% == 3.00% 0.50 675 -80.56% = 1.07 Market Leaders free CF to enterprise value decile 1 1375% 18.60% 248% 0.47 15 ~6279% 1.05 (continued on next page)Dissecting the Small Stocks Universe: The Ratios That Add the Most Value 489 Table 23.2 compares the universes’ performance for the Compustat dataset, which covers the 1963 to 2009 time period. Here we see that Small Stocks beats Alll Stocks by just 38 basis points, but actually had a lower Sharpe ratio, since Small Stocks were more ~ volatile than Alll Stocks. Over the same 46-year period, Small Stocks beat the Large Stocks universe by 1.40 percent. And while the Small Stocks universe beat the All Stocks universe ‘on an absolute basis, its base rates were worse than for those from the longer-term CRSP - data, beating All Stocks in 53 percent of all rolling five- and in 55 percent of all rolling ten- year periods. Nevertheless, as we are about to see, if you are searching for strategies that deliver the highest absolute total returns, the Small Stocks universe is a good place to start. MONTHLY DATA REVIEWED; SUMMARY DATA ACCESSED "The start date that allows a full test of all Small Stocks strategies is August 31, 1965—the nearly 44 years between August 31, 1965, and December 31, 2009. Table 23.3 shows the returns for all Small Stocks strategies. I must offer a disclaimer here—the five best-performing strategies do not actually come from our Small Stocks uni- verse, but from a universe of microcap stocks that I created to see what types of returns individual investors might achieve investing in the tiniest stocks. We've already discussed several of these strategies in Chapter Twenty-One, so you can refer back to the broader analysis in that chapter. TABLE 233 ‘Summary Results for Strategies Benchmarked against the Returns of Sm: Sorted by Compound Return ‘Stocks (Sep-65 to Dec-09); Strategies Standard Geometric deviation Excess Sharpe Tracking Largest Strategy mei (%)__returns error__drawdown Bota icrocap BP top 3 deciles, 3 and 6 ma.>0, top 25 by 12m mom 233% 2038% 10.97% «085 «873-82 89% © O88 Microcap PSR<1, pos 3 and 6 mo. rettop 10 12 mo. ret 2% «15% «109% «063-7 1.13 Microcap BP top 3 deciles, 9 and 6 mo >median top 25 by 12m mom 21.78% 201% 10.42% +084 = 1005 -55.64% «085 Microcap BP top 3 deciles, and 6 ‘mo.>0, top 50 by 12m mom 21.43% 19.17% 1007 086888 5.26%4 = Bd Microcap PSR<1 top 25 12 mo. ret 2033% «77.18% «BST 058155 -9.20% 1.08 Small Stocks Value Comp 3, decile 1 1937% —1892% © 8.01% ©=—«0767.95 88.68% 0.85 ‘Small Stocks Value Comp, decile 1 19.03% 18.14% 7.65% 0778.29 80.05% 0.81 ‘Small Stocks EBITDAVEV deci 18.96% —1870% 7.80% +075 «= 6878.9 086 ‘Small Stocks Value Comp 1, decile 1 18.85% © 1937% = 7.40% «0.727.868 80.29% 087 Small Stocks, PB in top 3 deciles, 3 and {6 mo, mom>median, top 25 by SY 18.84% © 1635% = 748% 0859.68 —49.20% © 070 Small Stocks psr<1 3 and 6 mo.>0, top 50, 12 mo. rat 18.80% — 2329% «7.44% «0599.98 -8557% = 1.02 (continued on next page)