G. Lee Willinger

A TIME-SERIES APPROACH TO MEASURING THE DECLINE IN QUARTERLY EARNINGS PERSISTENCE

Abstract Although prior research documents an inter-temporal decline in earnings relevance for equity investors, precise evidence has not been collected on why the decline has occurred. We document a substantial decline in the persistence of quarterly accounting earnings over a 35-year period for a sample of New York Stock Exchange firms. Our findings hold regardless of whether firms are in industries with dramatic increases in spending on information technology through time or not. Further, neither ex ante measures of expected economic change (changes in barriers-to-entry and product type) nor an ex post measure of economic change (quarterly sales persistence) decline inter-temporally for our sample firms.

Economic determinants of quarterly earnings data

Abstract We conduct exploratory data analysis on the economic determinants of quarterly earnings data. After surveying the industrial organization literature, we selected firm-size, product-type and barriers-to-entry as independent variables useful in explaining autocorrelation in quarterly earnings and sales. We perform cross-sectional regression analysis on a sample of 364 calendar year-end. New York Stock Exchange firms. As our dependent variable, we employed both seasonal and non-seasonal lags of the levels and first differences of the sample autocorrelation functions (SACF) of quarterly earnings and sales data. Our results support a pervasive impact of firm-size on the levels of the SACF. This outcome is consistent with the notion that larger firms exhibit mare stable, higher pronounced levels of serial correlation in their quarterly earnings numbers than smaller firms. Results for the other economic variables were more contextual, depending m whether we used the full sample or a subgroup and on whether the data were differenced. Stronger results on the product-type and barriers-to-entry variables were documented for the seasonal firm subgroup at seasonal lags. This result is suggestive of an economic rationale for the seasonal behavior of quarterly earnings and sales data.

Differential Earnings Behavior and the Security Market Assessment of Variation in Seasonal Earnings Patterns

We provide new evidence regarding the anomalous security market behavior that pertains to the predictability of abnormal security returns at future quarterly earnings announcements (Bernard and Thomas [1990]) as well as the security markets ability to form quarterly earnings expectations that are consistent with the market factoring in the serial correlation in the seasonally differenced quarterly earnings series (Ball and Bartov [1996]; Soffer and Lys [1999]). Consistent with extant work, we find that just prior to the quarterly earnings announcement date, the security markets expectation of quarterly earnings is consistent with the market recognizing the serial correlation in seasonally differenced quarterly earnings but underestimating its magnitude. When we treat all firms as exhibiting the same quarterly earnings process, our results reinforce the findings of Soffer and Lys (1999) that security market earnings expectations at the beginning of the quarter are consistent with investors not recognizing the serial correlation in the seasonally differenced quarterly earnings series. By distinguishing between those firms whose quarterly earnings process is inconsistent with a seasonal random walk (SRW) process (i.e., “bad-fit” firms) vis-à-vis those firms whose quarterly earnings process is better described by the SRW process (i.e., “good-fit” firms), we provide additional insights on this relationship throughout the quarter leading up to the quarterly earnings announcement date. Specifically, we find that security market expectations of quarterly earnings are consistent with a market that recognizes the differential time-series properties of quarterly earnings. That is, at the beginning of the quarter, for the “bad-fit” (“good-fit”) firms, we find that the security market acts as if it is aware (unaware) of the correct sign of the serial correlation in the seasonally differenced quarterly earnings series. However, the security market still acts as if it underestimates the magnitude of the serial correlation.

The Security Market's Reaction to Firms' Quarterly Earnings Evidencing Varying Degrees of Autocorrelation

Abstract On a full sample basis, our results are consistent with a security market that significantly underestimates the magnitude of autocorrelation at the 1st and 4th lags where autocorrelation is high but estimates autocorrelation unbiasedly at lags 2 and 3 where autocorrelation is low. Reinforcing the full sample results, when we partition the sample firms into subsamples based upon the magnitude of first lag autocorrelation, we find results consistent with the security market significantly underestimating the level of autocorrelation at the 1st lag for the high autocorrelation subsample of firms, but not for the moderate and low autocorrelation subsamples.

An analysis of the accuracy of long-term earnings predictions

Abstract This paper provides information on the long-term predictive ability of annual earnings numbers. We obtained a sample of 486 calendar, year-end firms that had complete quarterly earnings-per-share (eps) before extraordinary items available from 1978 to 1998. Firm-specific, quarterly, autoregressive-integrated-moving-average (ARIMA) time-series models were used to generate one through five year-ahead annual eps predictions across the 1994–1998 holdout period. Analysis of mean absolute percentage errors indicates: (1) firm-specific ARIMA models outperform so-called, common-structure, “primier” ARIMA models, (2) forecast errors from the firm-specific ARIMA time-series models ranged from 0.358 to 0.547 for one through five year-ahead annual eps predictions, (3) long-term earnings forecast accuracy is linked to firm size and earnings persistence, and (4) further research is needed to develop more powerful, long-term earnings prediction models suitable for use in conjunction with the abnormal earnings valuation model.