Table 1: Gross Momentum Returns, Unadjusted for Costs (% per Month)
Portfolio Size SP500 1 2 3 4 5 6 7
Jan 1992-Dec 2010 0.63 1.15 2.68*** 2.94*** 2.95*** 3.04*** 3.04*** 3.07***
Max 5.23 135.09 77.43 55.05 41.10 33.62 28.57 25.60
Min -3.83 -7.14 -6.84 -6.55 -6.50 -6.41 -6.20 -6.33
Median 0.99 0.78 2.45 3.37 3.03 3.48 3.49 3.55
Monthly St Dev. 0.06 0.70 0.44 0.34 0.29 0.24 0.22 0.20
Correlation NA 0.20 0.26 0.29 0.33 0.38 0.40 0.40
Outperform SP500 NA 49% 64% 64% 68% 73% 73% 74%
Sub-periods
1992-2000 1.24 1.53 3.14 3.04 3.06 3.41 3.46 3.30
2001-2009 0.06 0.79 2.23 2.84 2.83 2.69 2.64 2.84
2007-2008 -1.58 -4.32 -4.31 -3.83 -3.60 -3.42 -3.22 -3.01
1992-2006; 09-10 0.95 2.23 4.16 4.31 4.24 4.30 4.24 4.20
Portfolio Size 8 9 10 15 20 30 40 50
Jan 1992-Dec 2010 3.06*** 2.91*** 2.82*** 2.54*** 2.36*** 2.28*** 2.20*** 2.13***
Max 23.50 23.72 21.27 16.80 13.94 11.13 10.41 9.70
Min -6.20 -5.94 -5.83 -5.73 -5.53 -5.07 -5.18 -4.85
Median 3.40 3.53 3.32 3.08 3.04 2.85 2.76 2.76
Monthly St Dev. 0.19 0.18 0.17 0.14 0.12 0.11 0.10 0.10
Correlation 0.43 0.45 0.46 0.51 0.55 0.62 0.64 0.65
Outperform SP500 74% 74% 75% 73% 76% 78% 78% 77%
Sub-periods
1992-2000 3.36 3.20 3.08 2.72 2.52 2.60 2.54 2.53
2001-2009 2.77 2.64 2.57 2.36 2.21 1.97 1.88 1.74
2007-2008 -2.81 -2.70 -2.71 -2.42 -2.17 -2.06 -2.12 -2.12
1992-2006; 09-10 4.14 3.93 3.81 3.40 3.12 3.00 2.92 2.83
Note: For the analysis, a six-month formation period (-5 to 0 months) is implemented, ranking each stock by its
formation period (six-month) performance, from best to worst. After ranking each stock, equally weighted
portfolios were formed that contained the best (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50) performing stocks
in the formation period. We established a 12-month holding period and the overall return of each time period
was calculated by averaging the performance of all stocks in each portfolio. Over multiple time periods, the
average annual total return (geometric mean) of each portfolio was calculated in order to reach the monthly gross
returns. Correlation is between each portfolio and the S&P 500. "Outperform S&P 500" means the percentage of
months where each portfolio outperforms the S&P 500 benchmark. Statistical significance of the overall returns
is given by two sample parametric t-tests comparing the returns of each portfolio with the S&P 500.
* Significant at the 10% level, ** significant at the 5% level, *** significant at the 1% level.
The results in Table 1 show that all portfolios, on average, outperform the S&P 500
benchmark by 0.52%-2.44% per month. 12 Consistent with Siganos (2007), larger momentum
profits were primarily seen in the smaller portfolios. However, the portfolio containing the
12
The S&P 500 was used as the benchmark in this analysis as it the most commonly used benchmark for U.S.
stocks. We also ran the risk analysis against the Willshire 5000 Index, arguably a more comparable benchmark,
and found similar results.
8
Electronic copy available at: https://ssrn.com/abstract=2602320
best performing stock performed the worst out of all portfolios (1.15%), which is inconsistent
with the findings of Ammann, Moellenback, and Schmid (2011) and with those of Rey and
Schmid (2007). Overall portfolio performance gradually increases until it reaches the highest
performance, 3.07% per month, in the top seven stock portfolio. The returns then decrease as
the portfolio holds more stocks. Regardless, the top 50 stock portfolio still outperformed the
S&P 500 by 1.50% per month in the overall sample period.
Gross returns were divided into two equal sub-periods, January 1992 to December 2000 and
January 2001 to December 2009. Each sub-period appears to consistently outperform the S&P
500 in both categories. However, momentum trading clearly struggled during the financial
crisis of 2007 and 2008, incurring heavy losses and faring much worse than the S&P 500. In
this period, all portfolios underperformed the S&P 500 by 0.54% in the top 50 portfolio, and
by as much as 2.74% per month in the one stock portfolio. These findings are consistent with
Andrikopoulos, Clunie, and Siganos (2013), who find no evidence of momentum returns
during a similar period, February 2007 to February 2010, in the U.K. market. In hindsight, it
would have been more profitable to either stay in cash or seek an alternative trading
strategy. 13 We hope that these findings inspire further research into whether it is possible to
capture reliable ex-ante cues from the formation period data that can inform investors as to
whether they should continue with the momentum strategy or opt for an alternative trading
strategy for those holding periods.
3.2 Transaction Costs
To more accurately analyze the true profitability of momentum trading, all applicable
transaction costs are applied to each portfolio. At the beginning and end of each holding
period, a flat $10 commission per trade was factored in for each buy and sell order.
13
Daniel and Moskowitz (2013) find that in extreme market environments, the loser portfolio provides a high
premium, while the winner portfolio returns are minimal following large market declines.
9
Electronic copy available at: https://ssrn.com/abstract=2602320
