Performance Persistence

Abstract

In this chapter, new empirical evidence on the performance persistence of active U.S. equity funds is provided for the period from 1992 to 2007, which immediately follows the sample period from 1962 to 1993 used in the seminal study of Carhart (1997). Existing studies on performance persistence are extended in several dimensions. First, all share classes that belong to the same fund are aggregated to one observation. This avoids potential biases due to double counting of funds with several share classes. Nanda, Wang, and Zheng (2009) report that the number of single-class funds in their sample decreased from 313 in 1993 to 125 in 2002. The number of multiple-class funds increased, instead, from 40 to 838 in the same period. In 2005, a total of 18,444 share classes existed in the U.S. which belonged to only 6,791 different funds (Table 1.5). However, Carhart (1997) and Huij and Verbeek (2007), for example, treat each share class of a fund as a separate entity. This might be misleading as the underlying portfolio and the portfolio manager are identical across all share classes.