"We find that investors in value mutual funds have produced an average IRR that is meaningfully lower than the average returns reported by the corresponding value mutual funds. On average, investors that invest in value mutual funds do not benefit from the excess returns reported by those funds because of the timing of their allocations. Thus, caution needs to be used when interpreting the documented value premium in mutual funds since it does not reflect what the average investor actually received. In fact, over time periods with a documented high value premium, the average value investor in mutual funds has actually done worse than a buy-and-hold investor in an S&P500 index fund! One can conclude from this finding that, on average, value mutual fund investors time poorly, directing money to value mutual funds when value stocks are “expensive” and offer a lower premium and redeem money from value mutual funds when value stocks offer a high premium."

"This finding could support the following interpretations. First, it suggests that value mutual fund investors have not been extracting profits from investors on the other side of their trades and thus we would not expect the anomaly to be arbitraged away. In fact, we could argue that the activities of value mutual fund investors might contribute to the persistency of the value anomaly rather than arbitrage it away."

"We find that to the average investor, the value premium has largely been an illusion. Although the buy-and-hold average returns for value mutual funds have outperformed the market portfolio, the dollar-weighted average returns of these same funds meaningfully underperformed the market portfolio. Put in other words, while “value” managers have largely been successful in exploiting the value premium to outperform the market, the “value” investors, who ultimately allocate capital to these value managers, have managed to reverse the sign on the premium earned through their timing decisions. Specifically, they seem to allocate to value mutual funds when value stocks are “relatively” expensive and offer a low premium and vice versa."

"We examine the difference between dollar-weighted returns, or the internal rate of return (IRR) of a mutual fund that considers the flows into and out of each fund, and the time-weighted returns or the buy-and-hold returns that are usually reported for each fund. We find that the average investor does not time his allocation well and actually underperforms the buy-and-hold benchmark by almost 2% per year. This has significant implications. This loss in return through poor timing must be captured by someone else and addresses the question of how the value premium persists and who is on the losing end of the trade. Indeed, for fund categories where the average manager has outperformed the broad market benchmark net of fees as measured by buy-and-hold returns (like value and small cap style category), the investor’s negative timing skill actually results in an underperformance as measured by IRR."