In this paper, we suggest a technology process assuming that a stock mutual fund, as an usual firm, can be seen as a decision-making unit, so that we can measure the effect of its decisions regarding allocation of resources for administrative and non-administrative expenses on its own levels of gain and risk. Theoretically, this approach enables us to take into account some of the critiques reported in the fund's performance literature. In practice, we can test it by applying nonparametric linear mathematical programming methods. Here, we perform an empirical exercise based on bootstrapped Directional Distance Function in order to analyze the management efficiency of stock mutual funds in Brazil. We are able to evidence a high level of persistence in terms of efficiency over years. We can identify that the most efficient funds have higher values of net assets and that they allocate more weighted resources to administrative expenses than do the inefficient ones. Therefore, we get a high correlation of average weighted expenses on administrative expenses and average bootstrapped efficiency, about 0.7. We claim that investors should also perceive this kind of efficiency measure proposed here-This paper has greatly benefited from comments and suggestions by participants at 2013 First International Workshop in Financial Econometrics.