# SMARTBETA-1: before vs after launch (DESCRIPTIVE ONLY)

Preregistered in PREREG.md §6.4 (sealed 2026-09-26T16:00:46Z). **No inference is claimed.** The MSCI and S&P index backtests are licensed and were not used; the "before" evidence is the academic factor premium on the Kenneth R. French Data Library (derived statistics only). The mapping is a proxy that differs from the ETFs in universe (all US stocks vs large and mid caps) and construction (long-short or quintile sorts vs index tilts), and it is never evidence about the ETFs.

Windows (monthly): pre-launch full = first month available to the month before the inception month; pre-launch 1990 on = 1990-01 to the month before the inception month (roughly the providers' back-test era); post-launch = first full month after inception to 2026-08 (last French month). Means are annualised (× 12), t-statistics i.i.d.

## Summary

| ETF | Long-short factor | Pre-launch, full | Pre-launch, 1990 on | Post-launch | Long-only proxy minus market: pre (1990 on) → post | ETF minus SPY, post |
|---|---|---|---|---|---|---|
| MTUM | Mom | +8.1 pts (t +4.5) | +6.9 pts (t +1.9) | +2.6 pts (t +0.7) | +5.5 pts → +4.5 pts | +1.2 pts (t +0.5) |
| QUAL | RMW | +3.2 pts (t +2.9) | +4.6 pts (t +2.3) | +2.4 pts (t +1.0) | +1.1 pts → +1.0 pts | -0.4 pts (t -0.5) |
| VLUE | HML | +5.0 pts (t +3.7) | +3.3 pts (t +1.5) | -0.6 pts (t -0.2) | +2.9 pts → +2.6 pts | -0.1 pts (t -0.0) |
| SIZE | SMB | +2.6 pts (t +2.1) | +1.8 pts (t +0.8) | -1.5 pts (t -0.6) | none (no honest long-only match) | -2.2 pts (t -1.6) |
| USMV | Lo 20 minus Hi 20 (VW, variance) | +3.0 pts (t +0.9) | +1.8 pts (t +0.3) | -3.0 pts (t -0.5) | +0.2 pts → -0.7 pts | -3.3 pts (t -1.8) |
| SPLV | Lo 20 minus Hi 20 (VW, variance) | +2.5 pts (t +0.8) | +0.6 pts (t +0.1) | -2.0 pts (t -0.3) | -0.3 pts → -0.4 pts | -4.1 pts (t -1.7) |

## Detail per ETF

### MTUM

Long-short: Mom. Long-only proxy: Hi PRIOR (top decile, VW, prior 12-2).

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1927-01 to 2013-03) | 1035 | +8.1 pts (+4.51) | 16.7% | +7.1 pts (+5.44) | +7.0 pts | – | – | – | – |
| pre_backtest_era (1990-01 to 2013-03) | 279 | +6.9 pts (+1.90) | 17.6% | +5.5 pts (+2.07) | +4.7 pts | – | – | – | – |
| post (2013-05 to 2026-08) | 160 | +2.6 pts (+0.70) | 13.5% | +4.5 pts (+1.15) | +3.6 pts | – | – | +1.2 pts (+0.51) | +1.0 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -5.5 pts (t -1.34); vs 1990-on pre-launch -4.3 pts (t -0.84).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes; long-only proxy did at least as well as the ETF after launch: yes.

### QUAL

Long-short: RMW. Long-only proxy: Hi 20 (highest quintile, VW, operating profitability).

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1963-07 to 2013-06) | 600 | +3.2 pts (+2.92) | 7.8% | +1.1 pts (+1.84) | +1.2 pts | – | – | – | – |
| pre_backtest_era (1990-01 to 2013-06) | 282 | +4.6 pts (+2.28) | 9.8% | +1.1 pts (+1.18) | +1.4 pts | – | – | – | – |
| post (2013-08 to 2026-08) | 157 | +2.4 pts (+1.02) | 8.3% | +1.0 pts (+0.91) | +1.2 pts | – | – | -0.4 pts (-0.51) | -0.5 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -0.9 pts (t -0.34); vs 1990-on pre-launch -2.3 pts (t -0.74).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes; long-only proxy did at least as well as the ETF after launch: yes.

### VLUE

Long-short: HML. Long-only proxy: Hi 20 (highest quintile, VW, book-to-market).

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1926-07 to 2013-03) | 1041 | +5.0 pts (+3.75) | 12.3% | +5.1 pts (+3.27) | +3.5 pts | – | – | – | – |
| pre_backtest_era (1990-01 to 2013-03) | 279 | +3.3 pts (+1.52) | 10.6% | +2.9 pts (+1.51) | +2.6 pts | – | – | – | – |
| post (2013-05 to 2026-08) | 160 | -0.6 pts (-0.17) | 12.1% | +2.6 pts (+0.73) | +1.4 pts | – | – | -0.1 pts (-0.04) | -0.7 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -5.5 pts (t -1.55); vs 1990-on pre-launch -3.9 pts (t -0.98).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes; long-only proxy did at least as well as the ETF after launch: yes.

### SIZE

Long-short: SMB. Long-only proxy: none.

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1926-07 to 2013-03) | 1041 | +2.6 pts (+2.14) | 11.1% | – | – | – | – | – | – |
| pre_backtest_era (1990-01 to 2013-03) | 279 | +1.8 pts (+0.78) | 11.5% | – | – | – | – | – | – |
| post (2013-05 to 2026-08) | 160 | -1.5 pts (-0.60) | 9.2% | – | – | – | – | -2.2 pts (-1.64) | -2.6 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -4.1 pts (t -1.46); vs 1990-on pre-launch -3.4 pts (t -0.97).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes.

### USMV (non-blind)

Long-short: Lo 20 minus Hi 20 (VW, variance). Long-only proxy: Lo 20 (lowest-variance quintile, VW).

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1963-07 to 2011-09) | 579 | +3.0 pts (+0.94) | 22.4% | +0.1 pts (+0.10) | +0.7 pts | 0.78 | +8.0 pts | – | – |
| pre_backtest_era (1990-01 to 2011-09) | 261 | +1.8 pts (+0.32) | 26.7% | +0.2 pts (+0.09) | +0.8 pts | 0.73 | +13.8 pts | – | – |
| post (2011-11 to 2026-08) | 178 | -3.0 pts (-0.47) | 24.8% | -0.7 pts (-0.37) | -0.4 pts | 0.83 | +7.2 pts | -3.3 pts (-1.76) | -3.3 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -6.1 pts (t -0.84); vs 1990-on pre-launch -4.9 pts (t -0.57).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes; long-only proxy did at least as well as the ETF after launch: yes.

### SPLV (non-blind)

Long-short: Lo 20 minus Hi 20 (VW, variance). Long-only proxy: Lo 20 (lowest-variance quintile, VW).

| Window | Months | Long-short mean, a year (t) | Long-short vol | Long-only minus market, a year (t) | Long-only compound excess | Vol ratio to market | Drawdown cut vs market | ETF minus SPY, a year (t) | ETF compound excess |
|---|---|---|---|---|---|---|---|---|---|
| pre_full (1963-07 to 2011-04) | 574 | +2.5 pts (+0.76) | 22.4% | -0.1 pts (-0.09) | +0.4 pts | 0.78 | +8.0 pts | – | – |
| pre_backtest_era (1990-01 to 2011-04) | 256 | +0.6 pts (+0.10) | 26.7% | -0.3 pts (-0.13) | +0.3 pts | 0.74 | +13.8 pts | – | – |
| post (2011-06 to 2026-08) | 183 | -2.0 pts (-0.31) | 25.0% | -0.4 pts (-0.21) | +0.0 pts | 0.81 | +7.2 pts | -4.1 pts (-1.68) | -4.2 pts |

Welch t (post minus pre, long-short mean; context only): vs full pre-launch -4.4 pts (t -0.62); vs 1990-on pre-launch -2.5 pts (t -0.29).
Readings: factor premium lower after launch than before (full / 1990 on): yes / yes; long-only proxy did at least as well as the ETF after launch: yes.

Source: Kenneth R. French Data Library (files and SHA-256 in `MANIFEST.json`); ETF and SPY from Yahoo adjusted close. Derived statistics only; no raw series is published.