Using Dollar-Weighted Returns to Measure Investment Success
Stock-market investors are getting poorer overall returns than historical buy-and-hold return results suggest.
ANN ARBOR, Mich. Stock-market investors are getting lower overall returns than they realize, says a University of Michigan business professor.
This is because stock returns historically have been measured on a buy-and-hold basis, which does not account for different returns that occur at different points in time.
A new study by Ilia Dichev, associate professor of accounting at Michigan's Stephen M. Ross School of Business, contends that traditional buy-and-hold returns fail to assess the return experiences of stock investors because they ignore the timing effects of capital flows in and out of securities.
Instead, Dichev suggests using a new, more accurate measure that involves the dollar-weighting of returns (i.e., the returns generated by bigger investment amounts receive greater weighting than the returns from smaller investment amounts) and, therefore, properly reflects the effect of investors' timing.
"The compounding of buy-and-hold returns reflects the investing experience of passive buy-and-hold investors and assumes an equal weighting of returns over time," Dichev said. "However, most investors are actively trading or otherwise engaging in stock-capital transactions.
"As a result, they are consciously or unconsciously timing particular stocks or the market as a whole, and varying their net investment exposure over time. Thus, while buy-and-hold returns provide a good benchmark to track the investment performance of stocks, they can be a poor measure of the actual return experience of investors, if capital-flow timing affects stock returns."
Capital flows can include straightforward items such as dividends, stock repurchases and stock issues, as well as the contribution and distribution of non-cash assets, exercise of stock options, spinoffs and equity carve-outs, issuing stock in mergers and acquisitions, and listing and delisting of new stocks. The computation of dollar-weighted returns across numerous reporting periods can be affected by capital flows, which may change the weighting of the returns for each period used in the overall calculation.
The fact that investors have varying stock exposure over time is insufficient to create differences between buy-and-hold and dollar-weighted returns, according to Dichev. However, the two measures will differ if there are significant correlations between the timing of capital contributions and distributions and past and future stock returns.
For example, if investors tend to pour capital into the stock market after superior past returns and before poor future returns, then dollar-weighted returns will be lower than buy-and-hold returns, because the new capital contributions will earn lower returns than the rest of the stock capital. Many investors, Dichev says, discovered this phenomenon when they bought up shares of high-flying technology stocks during the recent stock-market bubble, only to watch shares prices plummet and their investments turn sour when the bubble burst.
To support his conclusions, Dichev examined the long-run returns following all capital flows over the entire history of available stock returns and across the U.S. and many foreign countries. His findings indicate that aggregate dollar-weighted returns are systematically and considerably lower than buy-and-hold returns.
The return differential is 1.3 percent for the NYSE/AMEX market between 1926 and 2002, 5.3 percent for the more volatile Nasdaq between 1973 and 2002, and an average of 1.5 percent for 19 major stock markets around the world between 1973 and 2004.
"These results provide comprehensive evidence that stock investors' actual returns are considerably lower than those from passive holdings and very different from widely published and studied security returns," Dichev said.
Dichev's study, "What are Stock Investors' Actual Historical Returns," can be found at
For more information, contact:
Phone: (734) 936-1015 or 647-1847