Hi everyone,

I am carrying out a research on mutual fund performance, but I am a little bit stuck at the moment.

The goal of my analysis is: demonstrate that 'value' mutual funds outperform 'growth' mutual funds

Up to now, the workflow has been as follows:

- Select US equity funds in the CRSP Mutual Fund Holdings database, with complete quarterly portfolio holdings reports in the period from 2008 to 2018. Resulting sample size = 50

- Classify funds on the basis of their investment orientation. To this end, I have applied an adapted methodology from Morningstar (2002), thus carrying out a holdings-based analysis. Therefore I have identified the portfolio centroids for all the 50 funds on a quarterly basis, using a value-growth score as the x variable and a market cap score as the y variable.

- Retrieve quarterly returns for all the funds. After this, the funds have been classified, for every quarter, into 10 deciles. The deciles have been taken out of the x variable, so the value/gorwth score. Then, this exercise has been repeated along three market cap dimensions: small, medium, large. Therefore, for every quarter, there are 30 deciles, with at least 1 observation.

- Graph the scatter plot (see file) of the average return for each decile, for the entire sample period. The time variable has been treated both as a discrete and as a continous variable.

Now I would like to know how can I analyze these returns. I was thinking about using a CAPM, but I don't know which kind of parameters to specify, if any. Does someone have a suggestion/reference paper that can assiste me?