14 هفته چهارم
کد مقاله | سال انتشار | تعداد صفحات مقاله انگلیسی |
---|---|---|
13214 | 2012 | 19 صفحه PDF |
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Accounting and Economics, Volume 53, Issues 1–2, February–April 2012, Pages 271–289
چکیده انگلیسی
Many firms define their fiscal quarters as 13-week periods so that each fiscal year contains 52 weeks, which leaves out one or two day(s) a year. To compensate, one extra week is added every fifth or sixth year and, consequently, one quarter therein comprises 14 weeks. We find evidence of predictable forecast errors and stock returns in 14-week quarters, suggesting that analysts and investors do not, on average, adjust their expectations for the extra week. The ease with which 14-week quarters can be predicted, and expectations adjusted, suggests a surprising lack of effort on the part of analysts and investors.
مقدمه انگلیسی
Many companies define their fiscal quarters as 13-week periods rather than three calendar months.2 A 52-week fiscal year omits one day a year (two days in a leap year) requiring an extra week to be added to one fiscal quarter (the “14-week” quarter) every five or six years. Though a 52/53-week reporting format enhances comparability in most quarters, earnings and revenues will be approximately 7.7% (1/13) higher in the 14-week quarters. Fourteen-week quarters are predictable since the fiscal year convention is clearly described in all financial statements. Given the impact that an extra week has on revenues and earnings, as well as the simplicity of the adjustment required, we would expect expert financial intermediaries such as analysts to adjust their forecasts for the extra week.3 However, prior research on limited attention, analyst incentives and strategic disclosure suggests that analysts may overlook the presence of a longer reporting period and underestimate performance in 14-week quarters. Like other economic agents, analysts face cognitive processing constraints that prevent them from processing all available information simultaneously (Hirshleifer and Teoh, 2003). Faced with multiple tasks and limited attention, analysts are forced to allocate their effort to various tasks based on their incentives. Recent research suggests sell-side analysts have relatively modest incentives to allocate their time to making accurate forecasts compared to other job responsibilities (e.g., stock picking, attracting investment banking business, increase their own visibility).4 For example, Groysberg et al. (2010) report that analysts' relative forecast accuracy has no significant direct effect on their compensation. Similarly, Emery and Li (2009) find that forecast accuracy has little bearing on the selection of All-star analysts. Consequently, faced with multiple tasks and constrained by limited attention, analysts might rationally choose not to exert the necessary effort to identify 14-week quarters and adjust for the extra week. In this study, we consider this hypothesis by testing whether analysts systematically underestimate revenues and earnings in 14-week quarters relative to 13-week quarters for the same firms. In addition to financial analysts, we also consider the impact of 14-week quarters on investors. In spite of their obvious incentives to identify 14-week quarters, past research suggests limited attention may also cause investors to ignore the extra week. Hirshleifer and Teoh (2003) argue that investors are more likely to ignore footnote or other less prominent disclosure. Recognizing the presence of the extra week is important for investors since the impact of the extra week is transitory and contains no new information about firm performance. However, if investors do not account for the extra week in 14-week quarters, two empirical regularities should emerge. First, the relation between returns and earnings (and revenues) should be the same in 14-week quarters as it is in other quarters because investors naively treat unexpected earnings and revenues consistently. That is, investors will (mis) price “unexpected” revenues and earnings caused by the extra week. Second, since prices will drift up as news about higher revenues and earnings is revealed to the market, a trading strategy of buying and holding stocks of firms over their 14-week quarters will earn positive abnormal returns. Our sample consists of 658 firms with 933 14-week fiscal quarters over the years 1994 to 2006. Consistent with the additional week contributing to revenue and earnings, we find that both seasonally adjusted unexpected revenues (SUR) and earnings (SUE) are higher in 14-week quarters than in 13-week quarters.5 However, analysts seem to either ignore or not fully account for the extra week's revenues and earnings since both revenue and earnings forecast errors are significantly positive in 14-week quarters.6 Further, for the same fiscal quarter in the year following a 14-week quarter, analysts' revenue forecasts are optimistic suggesting that analysts expect the higher revenues that occurred in the 14-week quarter one year ago to persist even though the current year's quarter contains one less week. In additional tests, we find that pessimistic revenue forecasts in 14-week quarters are associated with analysts who do not mention the presence of an extra week in their reports, suggesting that analysts' lack of awareness of the extra week is contributing to the forecast errors. These findings alleviate potential concerns that our results are driven by something inherently different about 14-week quarters, other than the duration of the quarter. Buying and holding stocks of firms in their 14-week quarters produces positive abnormal returns of approximately 3.15% over the quarter (12.6% annualized). The (estimated) extra week's earnings are positively related to the abnormal returns for the quarter, suggesting that investors price the predictable earnings “innovations” caused by the extra week. Additionally, we find evidence that the level of disclosure (about the extra week) in 14-week earnings announcements of firms affects returns around earnings announcement dates. Specifically, returns are more positive for firms that do not explicitly disclose the presence of the extra week, providing further evidence that lack of awareness possibly contributes to investor (mis) reaction. These findings are consistent with investor inattention when disclosure is less prominent (Hirshleifer and Teoh, 2003). Our evidence is consistent with both analysts and investors failing to properly account for the extra week in their expectations and pricing models. The simplicity of adjusting for the extra week suggests a lack of effort rather than ability. Faced with limited attention and inadequate incentives, analysts do not appear to exert sufficient effort to account for 14-week quarters suggesting that the benefits associated with improved accuracy must be smaller than previously believed, consistent with Groysberg et al. (2010) and Emery and Li (2009). For investors, the evidence is also consistent with limited attention, where investors ignore less salient disclosure. The investor findings are interesting because it should only take a few attentive investors (perhaps one) to arbitrage away the abnormal returns associated with 14-week quarters. However, alternative explanations are hard to identify. For example, impediments to arbitrage seem unlikely because the trading profits are derived from taking a long (not short) position and the firms in our sample are generally large. The remainder of the paper is organized as follows. Section 2 further discusses 14-week quarters and introduces hypotheses. Section 3 describes the sample selection procedure and presents descriptive statistics. Section 4 reports results and Section 5 discusses additional analyses. Section 6 concludes.
نتیجه گیری انگلیسی
In this study we identify an attractive setting to test whether investors and information intermediaries (analysts) fixate on past earnings when forecasting future earnings. In the case of 14-week quarters, we know that a seasonal random walk model is downward biased and we know the magnitude (roughly 1/13) of the bias. Further, the fact that the timing of 14-week quarters is pre-determined gives us confidence that our variable of interest is uncorrelated with other firm performance-based characteristics that might drive our results. Finally, because the forecast correction needed in 14-week quarters is relatively straightforward and predictable (we know when these quarters will occur) any observed fixation is likely due to a lack of effort on the part of the average analyst, or due to limited attention of the marginal investor, as opposed to a lack of ability. Our evidence is consistent with analysts, faced with limited attention and insufficient incentives, ignoring the extra week in 14-week quarters and systematically underestimating both revenues and earnings in those quarters. For the same quarter in the year following a 14-week quarter, analysts again appear to overestimate revenues and, to a lesser extent, earnings, as if they “forget” there was an extra week in the same fiscal quarter of the previous year. In more detailed tests, we find that analysts who disclose their awareness of the extra week in the quarter (approximately one out of four disclose) generally have smaller forecast errors than those who are not aware of the extra week. Stock return tests lead us to similar conclusions about the marginal investor. First, we find that a trading strategy of buying stocks two days after the prior earnings announcement and holding until one day after the earnings announcement for the 14-week quarter, earns abnormal returns of 3.15% per quarter (12.6% annualized). Results of a regression of returns on unexpected revenues and earnings suggest that investors seem to price the “unexpected” earnings related to the extra week, which one would not expect in an efficient market, since such innovations are predictable and will not persist. Our findings add to the discussion on market efficiency. Using methods similar to past research but in a unique research setting, we find evidence consistent with analysts and investors failing to incorporate available information about future revenues and earnings into forecasts and prices. These results are quite surprising in light of the ease with which this information can be obtained and incorporated into expectations. Our evidence suggests opportunities for future research on potential impediments to the rational pricing of past accounting information. For analysts, perhaps there are not sufficient incentives to provide accurate forecasts. For investors, perhaps there are other impediments to arbitrage than those that researchers typically point to (e.g., firm size and inability to short). What is peculiar about our setting is that the returns are in the long-position, which eliminates many of the traditional concerns about implementability, but perhaps other impediments exist.