A Comprehensive Look at the Empirical Performance of Equity Premium Prediction II
(With Amit Goyal and Athanasse Zafirov)
Our paper reexamines whether 29 variables from 26 papers published after Goyal and Welch (2008), as well as the original 17 variables, were useful in predicting the equity premium in-sample and out-of-sample. Our samples include the original periods in which these variables were identified, but ends later (in 2020). Most variables have already lost their empirical support, but a handful still perform reasonably well. Overall, the predictive performance remains disappointingIn the following description, you want to see a statistically significant in-sample (IS) statistic, two half (H1 and H2) coefficients that are not very different or sign changing, and an out-of-sample (OOS) R² that is statistically significant and positive. Variables without IS significance are not worth looking at. Beyond this, you want variables that did not just work in one year, say 1950, and then never again. (The full information about variables with better descriptions is in the paper).
| # | F | Variable | Paper | IS Coef | IS T | H1 | H2 | IS R² | OOS R² |
|---|---|---|---|---|---|---|---|---|---|
| 1 | M | BPS | Bakshi, Panayotov, and Skoulakis (JFE 2011) | 0.14 | 0.35 | 0.83 | −0.07 | 0.10 | −3.44 |
| 2 | M | BTZ | Bollerslev, Tauchen, and Zhou (RFS 2009) | 0.07 | 0.12 | 1.03 | −0.08 | 0.03 | −8.88 |
| 3 | M | CEP | Chen, Eaton, and Paye (JFE 2018) | 0.26 | 0.96 | 0.30 | 0.23 | 0.23 | 0.11 * |
| 4 | M | CP | Cooper and Priestley (RFS 2009) | −0.21 | −0.62 | −0.08 | −1.33 | 0.15 | −0.24 |
| 5 | M | DJM | Driesprong, Jacobsen, and Maat (JFE 2008) | −0.29 | −1.47 | −0.60 | −0.25 | 0.29 | −0.12 |
| 6 | M | HJTZ | Huang, Jiang, Tu, and Zhou (RFS 2015) | 0.45*** | 2.66*** | 0.48 | 0.30 | 1.04 | −1.24 |
| 7 | M | JT | Jones and Tuzel (RFS 2013) | −0.33* | −1.73* | −0.50 | −0.22 | 0.60 | −0.48 |
| 8 | M | JZZ | Jondeau, Zhang, and Zhu (JFE 2019) | 0.01 | 0.04 | 0.18 | −0.33 | 0.00 | −0.57 |
| 9 | M | KJ | Kelly and Jiang (RFS 2014) | 0.03 | 0.21 | −0.08 | 0.80 | 0.00 | −0.37 |
| 10 | M | KP | Kelly and Pruitt (JF 2013) | 1.05*** | 3.44*** | 1.08 | 0.48 | 3.79 | −2.21 |
| 11 | M | LYdtoy | Li and Yu (JFE 2012), dtoy | 0.17 | 0.40 | 0.34 | −0.23 | 0.10 | −0.57 |
| 12 | M | LYdtoat | Li and Yu (JFE 2012), dtoat | −0.09 | −0.32 | −0.02 | −0.69 | 0.03 | −0.07 |
| 13 | M | Mgap | Maio (JFM 2016) | 0.13 | 0.67 | 0.18 | 0.29 | 0.10 | −1.24 |
| 14 | M | Mdsp | Maio (RF 2013) | 0.28 | 0.74 | 0.39 | −0.22 | 0.27 | −1.39 |
| 15 | M | NRTZ | Neely, Rapach, Tu, and Zhou (MS 2014) | 0.29 | 1.61 | 0.33 | 0.22 | 0.47 | 0.20 * |
| 16 | Q | PW | Pollett and Wilson (JFE 2010) | 0.20 | 0.84 | 0.00 | 0.44 | 0.13 | −0.17 |
| 17 | M | RRZ | Rapach, Ringgenberg, and Zhou (JFE 2016) | −0.29 | −1.28 | 0.06 | −0.39 | 0.41 | 0.74 ** |
| 18 | M | Y | Yu (JFE 2011) | 0.01 | 0.06 | −0.31 | 0.36 | 0.00 | −1.07 |
| 19 | M | CSdp | Campbell-Shiller (JF 1988), d/p | 0.20 | 0.84 | 0.54 | 0.14 | 0.14 | −0.12 |
| 20 | M | CSdy | Campbell-Shiller (JF 1988), d/y | 0.26 | 1.03 | 0.74 | 0.15 | 0.22 | −0.32 |
| 21 | M | CSep | Campbell-Shiller (JF 1988), e/p | 0.30* | 1.72* | 0.97 | 0.04 | 0.31 | −1.24 |
| 22 | M | CSde | Campbell-Shiller (JF 1988), d/e | −0.10 | −0.34 | −0.64 | 0.11 | 0.03 | −1.19 |
| 23 | M | G | Guo (JB 2006), svar | −0.08 | −0.18 | 0.00 | −0.18 | 0.02 | −0.32 |
| 24 | M | KS | Kothari-Shanken (JFE 1997), b/m | 0.31 | 0.95 | 1.02 | −0.02 | 0.33 | −1.35 |
| 25 | M | BMRR | Boudoukh et al. (JF 2007), ntis | −0.36 | −1.44 | −0.64 | −0.09 | 0.44 | −0.48 |
| 26 | M | Ca | Campbell (JFE 1987), tbl? | −0.28* | −1.71* | −0.50 | −0.26 | 0.26 | 0.14 * |
| 27 | M | FFlty | Fama-French (JFE 1989), lty | −0.22 | −1.43 | −0.77 | −0.21 | 0.16 | −0.63 |
| 28 | M | FFltr | Fama-French (JFE 1989), ltr | 0.24 | 1.53 | 0.16 | 0.27 | 0.20 | −0.94 |
| 29 | M | FFtms | Fama-French (JFE 1989), tms | 0.18 | 1.11 | 0.19 | 0.21 | 0.11 | 0.02 |
| 30 | M | FFdfy | Fama-French (JFE 1989), dfy | 0.07 | 0.15 | 0.02 | 0.24 | 0.02 | −0.11 |
| 31 | M | FFdfr | Fama-French (JFE 1989), dfr | 0.23 | 0.90 | −0.02 | 0.39 | 0.18 | −0.34 |
| 32 | M | FS | Fama-Schwert (JFE 1977), infl | −0.23 | −1.01 | −0.15 | −0.43 | 0.17 | 0.11 * |
| 33 | Q | AMP | Atanasov, Moller, and Priestley (JF 2020) | −1.68*** | −3.57*** | −1.84 | −1.57 | 4.30 | −1.17 |
| 34 | Q | AMS | Adrian, Mönch, and Shin (FRBNY 2010) | 0.52 | 0.87 | 1.05 | −1.80 | 0.43 | −1.91 |
| 35 | Q | BY | Belo and Yu (JME 2013) | 0.67* | 1.67* | 1.36 | −2.86 | 0.72 | −1.25 |
| 36 | Q | CGP | Chava, Gallmeyer, and Park (JME 2015) | −1.73 | −1.65 | −2.20 | −1.24 | 4.52 | 2.61 ** |
| 37 | M | MA | Martin (QJE 2017), [rs]vix | 2.00* | 1.69* | 2.51 | 1.74 | 5.66 | −2.90 |
| 38 | Q | PW | Pollett and Wilson (JFE 2010) | 1.00 | 1.37 | 0.21 | 1.96 | 0.90 | 1.09 ** |
| 39 | Q | Co | Cochrane (JF 1991), i/k | −1.56*** | −3.42*** | −2.17 | −0.93 | 3.90 | 2.26 *** |
| 40 | Q | LL | Lettau-Ludvigson (JF 2001), cay | 0.34 | 0.51 | 2.64 | −0.54 | 0.17 | −11.57 |
| 41 | A | CGMS | Colacito, Ghysels, Meng, and Siwasarit (RFS 2016) | 1.76 | 0.83 | 1.20 | 2.93 | 1.34 | −0.26 |
| 42 | A | HHTac | Hirshleifer, Hou, and Teoh (JFE 2008), acc | 5.46*** | 2.98*** | 4.70 | 7.16 | 12.82 | 16.68 *** |
| 43 | A | HHTcf | Hirshleifer, Hou, and Teoh (JFE 2008), cf | −3.32 | −1.42 | −4.67 | −2.22 | 4.86 | −7.34 |
| 44 | M | MRgpce | Moller and Rangvid (JFE 2015), gpce | −3.86* | −1.77* | −3.46 | −4.78 | 6.37 | 6.85 ** |
| 45 | M | MRgip | Moller and Rangvid (JFE 2015), gip | −0.42 | −0.13 | −0.10 | −4.92 | 0.03 | −0.54 |
| 46 | A | PST | Piazzesi, Schneider, and Tuzel (JFE 2007) | 1.34 | 0.63 | 2.06 | −4.95 | 0.34 | 0.04 |
| 47 | A | BW | Baker-Wurgler (JF 2000), eqis | −4.48 | −1.65 | −8.05 | 2.03 | 3.70 | −15.68 |
| 48 | A | CGMS | Colacito, Ghysels, Meng, and Siwasarit (RFS 2016) | 0.65 | 0.28 | −2.66 | 7.85 | 0.16 | −3.73 |
| 49 | A | HHTac | Hirshleifer, Hou, and Teoh (JFE 2008), acc | 5.15*** | 2.73*** | 2.69 | 8.54 | 9.61 | 12.49 *** |
| 50 | A | HHTcf | Hirshleifer, Hou, and Teoh (JFE 2008), cf | −5.35*** | −3.08*** | −6.15 | −4.58 | 10.53 | 5.16 ** |
| 51 | M | MRgpce | Moller and Rangvid (JFE 2015), gpce | −5.64*** | −3.40*** | −6.02 | −4.48 | 11.91 | 4.86 ** |
| 52 | M | MRgip | Moller and Rangvid (JFE 2015), gip | −0.29 | −0.10 | 0.42 | −10.26 | 0.02 | −1.98 |
| 53 | A | PST | Piazzesi, Schneider, and Tuzel (JFE 2007) | 1.94 | 0.99 | 2.81 | −1.48 | 1.00 | 0.59 * |
| 54 | A | BW | Baker-Wurgler (JF 2000), eqis | −5.48*** | −2.67*** | −8.10 | −0.15 | 7.96 | −0.32 |