Table I

Correlations Between Prior Six-Month Return and Past Earnings

Surprises

Correlation coefficients are calculated over all months and over all stocks for the following variables. R6 is a stocks compound return over the prior six months. SUE is unexpected earnings (the change in the most recent past quarterly earnings per share from its value four quarters ago), scaled by the standard deviation of unexpected earnings over the past eight quarters. REV6 is a moving average of the past six months revisions in Institutional Brokers Estimate System (I/B/E/S) median analyst earnings forecasts relative to beginning-of-month stock price. ABR is the abnormal return relative to the equally-weighted market index cumulated from two days before to one day after the most recent past announcement date of quarterly earnings. The sample includes all domestic primary firms on New York Stock Exchange (NYSE), American Stock Exchange (AMEX), and Nasdaq with coverage on the Center for Research in Security Prices (CRSP) and COMPUSTAT. The data extend from January 1977 to December 1993.

ent measures of earnings surprises are not strongly associated with each other. The highest correlation (0.440) is between standardized unexpected earnings and revisions in analyst forecasts, while the correlation between analyst revisions and abnormal returns around earnings announcements is 0.115. The low correlations suggest that the different momentum variables are not entirely based on the same information. Rather, they capture different aspects of improvement or deterioration in a companys performance.

Panel A of Table II documents the stock price performance of portfolios formed on the basis of prior six-month returns, where portfolio 1 comprises past losers and portfolio 10 comprises past winners. Subsequent to the portfolio formation date, winners outperform losers, so that by the end of twelve months there is a large difference of 15.4 percent between the returns of the winner and loser portfolios. This difference is driven by the extreme decile portfolios, however. Comparing the returns on decile portfolios 9 and 2 reveals a smaller difference of 6.3 percent.

While there is prior evidence on the profitability of price momentum strategies, we go further and provide additional characteristics of the different portfolios. In Panel B, there is a fairly close association between past return performance and the portfolios book-to-market ratios (measured as of the portfolio formation date). The portfolio of past winners tends to include glamour stocks with low book-to-market ratios. Conversely, the portfolio of past losers tends to include value stocks with high book-to-market ratios. This is not necessarily surprising, however. Even if the different portfolios had similar book-to-market ratios at the beginning of the period, book values change very slowly over time but one portfolio rose in market value by 70 percent while the other fell by 31 percent. However, the ten portfolios display smaller differences

Table II

Mean Returns and Characteristics for Portfolios Classified by Prior

Six-Month Return

At the beginning of every month from January 1977 to January 1993, all stocks are ranked by their compound return over the prior six months and assigned to one of ten portfolios. The assignment uses breakpoints based on New York Stock Exchange (NYSE) issues only. All stocks are equally-weighted in a portfolio. The sample includes all NYSE, American Stock Exchange (AMEX), and Nasdaq domestic primary issues with coverage on the Center for Research in Security Prices (CRSP) and COMPUSTAT. Panel A reports the average past six-month return for each portfolio, and buy-and-hold returns over periods following portfolio formation (in the following six months and in the first, second, and third subsequent years). Panel В reports accounting characteristics for each portfolio: book value of common equity relative to market value, and cash flow (earnings plus depreciation) relative to market value. Panel С reports each portfolios most recent past and subsequent values of quarterly standardized unexpected earnings (the change in quarterly earnings per share from its value four quarters ago, divided by the standard deviation of unexpected earnings over the last eight quarters). Panel D reports abnormal returns around earnings announcement dates. Abnormal returns are relative to the equally-weighted market index and are cumulated from two days before to one day after the date of earnings announcement. In Panel E, averages of percentage revisions relative to the beginning-of-month stock price in monthly mean I/B/E/S estimates of current fiscal-year earnings per share are reported.

Panel D: Abnormal Return Around Earnings Announcements

after portfolio formation

Table II-Continued

with respect to their ratios of cash flow to price. The extreme portfolios feature low ratios of cash flow to price, but for different reasons. The portfolio of past losers contains stocks with relatively depressed past earnings and cash flow, while the portfolio of past winners contains glamour stocks that have done well in the past.

The last three panels of Table II provide clues as to what may be driving price momentum. Perhaps not surprisingly, the past price performance of the portfolios is closely aligned with their past earnings performance. There is a large difference between the past winners and past losers in terms of the innovation in their past quarterly earnings (Panel C). Past abnormal announcement returns (Panel D) also rise across the momentum portfolios, with a large difference (6.2 percent) between portfolios ten and one. Stocks that have experienced high (low) past returns are associated with large upward (downward) past revisions in analysts estimates (Panel E).4

More remarkably, the differences across the portfolios in their past earnings performance continue over the periods following portfolio formation. The spread between the SUEs of the winner and loser portfolios is actually wider in the following quarter. This may simply be a symptom of a misspecified model of expected earnings,5 so examining the behavior of returns around earnings announcement dates provides a more direct piece of evidence. We find that the market continues to be caught by surprise at the two quarterly earnings announcements following portfolio formation, particularly for the extreme decile portfolios.6 In particular, the abnormal return around the first subsequent announcement is higher by 2.6 percent for winner stocks compared

4 Note that in Panel E we report statistics for monthly percentage revisions in the consensus estimates (while portfolios are formed on the basis of a six-month moving average of revisions). The presence of reporting delays in the individual estimates underlying the consensus may induce apparent persistence on a month-by-month basis, so we report average percent changes over the first and second six-month periods following portfolio formation.

5 Fama and French (1993, 1995) argue that the statistical process for earnings changed during the 1980s. It might be suggested that this prolonged period of continuous rational surprises could account for part of the earnings surprise effects in returns. On the other hand, numerous studies document the existence of earnings surprise effects before the start of our sample period. See, for example, Givoly and Lakonishok (1979), Jones and Litzenberger (1970), and Latane and Jones (1979).

6 Note that the average abnormal return around announcement dates is positive. This is consistent with the findings of Chari, Jagannathan, and Ofer (1988).