‘Smart’ copycat mutual funds: on the performance of partial imitation strategies

Financial Innovation

Table 2 Copycat Score and mutual fund performance

Quantile	CopyScore	Return	StDev	Alpha	MktRf	SMB	HML	MOM
5	9.81	0.78	4.84	0.136*	0.981***	0.273***	0.177***	− 0.020
4	3.17	0.74	4.84	0.099	1.000***	0.284***	0.121***	0.003
3	1.69	0.65	4.77	0.053	1.022***	0.162***	0.101***	0.001
2	0.86	0.62	4.66	0.044	1.017***	0.127***	0.066***	0.013
1	0.27	0.61	4.61	0.061	1.015***	0.088***	0.044***	0.008
0	0.00	0.52	4.53	− 0.005	0.999***	0.087***	− 0.014	0.016**
− 1	− 0.33	0.57	4.57	0.019	1.007***	0.103***	0.025**	0.016**
− 2	− 0.99	0.59	4.69	0.017	1.016***	0.154***	0.039***	0.005
− 3	− 1.88	0.59	4.80	0.009	1.021***	0.194***	0.039**	− 0.001
− 4	− 3.38	0.60	4.86	− 0.007	1.024***	0.219***	0.081***	− 0.005
− 5	− 8.42	0.68	5.01	0.036	1.018***	0.313***	0.106***	− 0.025**
5–0		0.27***	0.31	0.141**	− 0.018	0.186***	0.190***	− 0.037***
(− 5)− 0		0.16**	0.48	0.042	0.020*	0.226***	0.119***	− 0.042***

Mutual funds are sorted every quarter into uneven quantiles by their Copycat Score, and the subsequent quarter’s funds’ returns are recorded. Funds with a score of zero are placed in quantile 0, and those with negative and positive scores are separately ranked into quintiles. For each fund, gross returns are estimated by adding reported expenses to net returns. For funds with more than one share class, data is aggregated using lagged total asset size as weights. Monthly portfolio returns are calculated as the equally-weighted return of all funds within each quantile. The data spans the years 2000 to 2016. Average monthly return for each quantile is presented, as well as standard deviation of returns, and the output from a Carhart (1997) four-factor model, which includes the factor model alpha, as well loadings in the market, SMB, HML and MOM factors. Statistical significance of the regression results is denoted by ∗∗∗, ∗∗ and ∗ for significance at the 1%, 5% and 10% levels, respectively