Short selling and the rounding of analysts’ forecasts

Short selling and the rounding of analysts’ forecasts

Accepted Manuscript Short Selling and the Rounding of Analysts’ Forecasts Hae Mi Choi PII: DOI: Reference: S1544-6123(17)30305-7 10.1016/j.frl.2017...

941KB Sizes 1 Downloads 7 Views

Accepted Manuscript

Short Selling and the Rounding of Analysts’ Forecasts Hae Mi Choi PII: DOI: Reference:

S1544-6123(17)30305-7 10.1016/j.frl.2017.10.001 FRL 783

To appear in:

Finance Research Letters

Received date: Revised date: Accepted date:

2 June 2017 6 September 2017 3 October 2017

Please cite this article as: Hae Mi Choi , Short Selling and the Rounding of Analysts’ Forecasts, Finance Research Letters (2017), doi: 10.1016/j.frl.2017.10.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Highlights    

AC

CE

PT

ED

M

AN US

CR IP T



Regulation SHO led to an exogenous increase in short selling for pilot stocks. Analysts put less effort into collecting precise information of pilot stocks with downward price pressure. The likelihood of rounded forecasts significantly increase for pilot stocks during Reg SHO. Pilot stocks experience lower abnormal returns around analyst forecast announcements during Reg SHO. The effect of short selling on the rounding of forecasts is stronger for firms with more firmspecific information and firms with low levels of institutional holdings.

0

ACCEPTED MANUSCRIPT

Short Selling and the Rounding of Analysts’ Forecasts Hae Mi Choi1 1

Hae Mi Choi, Quinlan School of Business, Loyola University Chicago, email: [email protected], phone (office): (312) 915-6320.

CR IP T

ABSTRACT

This paper examines the causal effect of short selling on analyst forecast precision by exploiting a regulatory change in short-sale constraints (Regulation SHO) as a natural experiment. I find that short selling increases analysts’ rounding of forecasts, which indicates that analysts allocate less effort to gathering precise information on firms with downward price pressure. In the cross-

AN US

section, the effect of short selling on analyst forecast precision is stronger for firms with more firm-specific information and firms with low levels of institutional holdings. Keywords

Short selling; Regulation SHO; Rounding; Earnings Forecasts; Analysts

JEL classification codes

M

G10; G14; M40

ED

1. Introduction

Financial analysts are important information intermediaries whose earnings forecasts

PT

influence investors’ earnings expectations, investment decisions, and stock prices. However, prior research finds that analysts have limited resources and allocate their efforts based on their

CE

expectations of future firm performance. McNichols and O’ Brien (1997) find that analysts put less effort into firms that are expected to perform poorly, since they have stronger incentives to

AC

issue optimistic forecasts than pessimistic forecasts. A large stream of literature shows that analysts optimistically bias earnings forecasts due to their own incentives, such as maintaining favorable relationships with firm management, managing career concerns, and generating trading

1

ACCEPTED MANUSCRIPT

commissions (Francis and Philbrick, 1993; Jackson, 2005; Chen and Matsumoto, 2006; Mayew, 2008; Soltes, 2014; Brown, Call, Clement, and Sharp, 2015).1 Building on the previous studies showing that analysts’ information collection incentives

CR IP T

and effort allocation vary across firm characteristics, I examine whether short selling of the firm has a causal impact on analysts’ earnings forecast precision. More specifically, this study examines whether analysts’ tendency to issue rounded forecasts increases for firms with exogenous increases in short-selling activity. Short-sellers are informed investors who trade on

AN US

negative information about firms, which leads to a downward pressure on stock prices. Analysts are less likely to put effort into these firms since they have more incentives to issue favorable forecasts than negative forecasts on average (McNichols and O’ Brien, 1997; Hayes, 1998; Lim, 2001). The main prediction is that analysts, in response to increased downward short-selling

ED

expected to perform badly.2

M

pressure on stock prices, put less effort into collecting precise information about firms that are

Prior studies provide little directional evidence on whether firm characteristics and the

PT

behavior of stock prices have a causal effect on analysts’ strategic forecasting properties. The difficulty in establishing causality stems largely from the endogenous nature of firm

CE

characteristics and analysts’ forecasts. In this paper, I use a temporary regulatory experiment on

AC

short-sale constraints, Regulation SHO (hereafter Reg SHO), as a natural experiment to investigate the causal effect of short-selling pressure on analysts’ forecasting precision, namely their tendency to round forecasts. When the Securities and Exchange Commission (SEC)

1

There is vast evidence that analysts are on average optimistically biased (Stickel, 1990; Abarbanell, 1991; Dreman and Berry, 1995; Chopra, 1998; Lim, 2001, among others). 2

Ke et al. (2015) find that analysts issue less optimistic forecasts when short-selling activity increases.

2

ACCEPTED MANUSCRIPT

adopted Reg SHO from May 2005 to August 2007, this mandated a temporary suspension of short-sale price tests for a set of randomly selected pilot stocks. Stocks included in the Russell 3000 Index were ranked by average daily trading volume, and every third stock was suspended from the uptick test for the NYSE and the bid test for the NASDAQ. This led to an exogenous

CR IP T

decrease in short-sale constraints and an increase in short selling for the pilot stocks (SEC, 2007; Diether et al., 2009; Angelis et al., 2013; Grullon et al., 2015). Reg SHO provides a natural experimental setting in which to compare the effects of short selling before and after the regulation for pilot stocks, since stocks were chosen randomly (Fang et al., 2016; Ke et al., 2015;

AN US

Li and Zhang, 2015).

To my knowledge, this study is the first to examine the causal effect of short selling on analysts’ forecast precision, exploiting an exogenous shock to short-selling activity triggered by

M

Reg SHO. I take a difference-in-differences (DiD) estimation approach and compare the rounding of forecasts of pilot stocks (the treatment group)–stocks that were randomly chosen by

ED

Reg SHO–to the rounding of forecasts of non-pilot stocks (the control group). The testable

PT

hypothesis is that the increased trading activity of pessimistic investors, under weaker short-sale constraints, increased analysts’ rounding of forecasts for pilot stocks relative to non-pilot stocks.

CE

I find statistically and economically significant evidence that short selling increased

AC

analysts’ rounding of forecasts. The DiD estimates for pilot stocks show that the likelihood of rounded forecasts significantly increases during Reg SHO.3 This suggests that analysts put less effort into collecting precise information about firms with negative price pressure. When comparing the abnormal returns around analyst forecast announcements, I find that pilot stocks

3

The empirical findings are reported in Table 2, and a detailed discussion of the results appears in section 3.

3

ACCEPTED MANUSCRIPT

experience significantly lower stock returns during Reg SHO. Additionally, the returns are lower and the tendency to round forecasts is stronger for stocks with high trading volume levels around forecast announcements, where trading volume proxies for the level of short-selling activity.4

CR IP T

I also explore possible explanations for the changes in analysts’ rounding and effort allocation. In the cross-section, I find that the effect of Reg SHO on analyst rounding is stronger for firms with high levels of firm-specific information and firms with low levels of institutional holdings. Analysts issue optimistic forecasts to maintain a favorable relationship with firm

AN US

management, and this incentive to gather inside information from management increases with the level of firm-specific information. When the firm is expected to perform poorly, analysts put less effort into these firms and tend to refrain from precisely incorporating pessimistic information. Similarly, institutional investors are important in determining analysts’ career

M

outcomes. For example, institutional investors nominate analysts for the Institutional Investors All-American All Star analyst status. Accordingly, analysts have incentives to allocate more

ED

effort to firms with high institutional holdings, which is also consistent with the empirical

PT

findings. Overall, the findings suggest that increased short-selling pressure has a causal effect on analysts’ forecast precision (i.e., their tendency to round forecasts). There is cross-sectional

CE

variation across firms, where analysts’ rounding of forecasts varies with the analysts’ own

AC

incentives to allocate effort.

2. Data and Research Design

4

I focus on stock returns and trading volume around analyst forecast announcements, since returns and trading during longer horizons can be correlated with other firm characteristics.

4

ACCEPTED MANUSCRIPT

The sample of 986 pilot stocks and 1,966 non-pilot stocks is identified based on the SEC’s methodology as described in Fang et al. (2016). The list of 986 stocks that traded without being subject to any price tests during the term of the pilot program is based on the 2004 Russell 3000 index, excluding stocks not listed on NYSE, AMEX, or Nasdaq NM.5 Stocks are sorted by

CR IP T

their average daily dollar volume computed from June 2003 through May 2004 within each of the three listing markets, and every third stock within each listing market is designated as a pilot stock.

AN US

As the SEC announced the list of pilot stocks on July 28, 2004 and initiated the pilot program on May 2, 2005, I exclude quarters between these dates (Ke et al., 2015). I use eight quarters before the announcement date as the pre-period and eight quarters after the program start date as the during-period to test the effect of short selling on the informational efficiency of

M

stock prices. The pre-period is from 2003:Q3 to 2004:Q2, and the during-period is from 2005:Q3 to 2007:Q2. I also delete firms in the financial services (SIC 6000-6999) and utilities industries

ED

(SIC 4900-4949) because disclosure requirements, accounting rules, and processes by which

PT

accruals are generated are significantly different for these regulated industries (Fang et al., 2016). I then merge this sample with analyst forecast data from I/B/E/S. I use annual earnings

CE

(EPS) forecasts that are one-year-ahead forecasts and use the unadjusted file to mitigate the

AC

rounding problem in I/B/E/S (see, for instance, Diether, Malloy, and Scherbina, 2002). Using the split-adjustment factors from I/B/E/S, I adjust the unadjusted forecast so that it is on the same per-share basis as the unadjusted actual earnings and retain only the forecast revision closest to

5

I thank Fang, Huang, and Karpoff (2016) for sharing the list of pilot and non-pilot stocks.

5

ACCEPTED MANUSCRIPT

July (but not after July) in a particular year (see, e.g., Hong and Kubik, 2003). 6 Firm-level variables are obtained from Compustat Annual Updates, and institutional holdings data is from the Thomson Reuters Spectrum database. Spectrum collects quarterly data on stock holdings from the 13F reports that institutions are required to file if their holdings exceed $100 million.

CR IP T

The holdings are aggregated over all institutions to arrive at the total institutional holdings. In most tests, I require all firms and analysts to have data across the entire sample period. The resulting sample includes 60,612 analyst forecast announcements from 1,490 firms, of which 497 firms are pilot stocks and 993 firms are non-pilot stocks.

AN US

I create an indicator variable PILOT to denote firms that are pilot stocks. Specifically, PILOT equals one if a firm’s stock is designated as a pilot stock under Regulation SHO’s pilot program, and zero otherwise. Pilot firms constitute the treatment group, and non-pilot firms serve

M

as the control group. I also construct an indicator variable to indicate time periods during Reg SHO. DURING equals one if the forecast is made within the eight-quarter period after the

ED

adoption date of Reg SHO, and zero otherwise. The dependent variable ROUNDING is an

PT

indicator variable that equals one if the analyst i’s forecast for firm j in year k is an integer.

CE

The main regression specification is as follows:

AC

The main explanatory variable of interest is the interaction effect between PILOT*DURING, whose coefficient (

measures the DiD effect, which is the differences in rounded forecasts

(relative to non-pilot stocks) of pilot stocks during the adoption of Reg SHO. If the tendency to 6

I use the most recent forecasts before the cut-off date of July to use a common time frame to compare the information contained in analysts’ forecasts (see, e.g., Crichfield et al., 1978; Hong and Kubik, 2003). The results are robust to alternative cut-off dates. Following Hong and Kubik (2003), I retain only firms with fiscal year ending in December.

6

ACCEPTED MANUSCRIPT

issue rounded forecasts increases (decreases) during Reg SHO, I expect

to be positive

(negative). Here

is a vector of analyst characteristics measured in year k−1 and is a vector of stock-level control variables measured

CR IP T

serving as control variables;

in year k−1. Analyst characteristics include All Star status (an indicator variable that equals one if the analyst is identified as an All Star by the All American Institutional Investor magazine), brokerage size (the number of analysts employed by the brokerage firm, in natural logarithm),

AN US

experience (number of years since first forecast issuance reported in I/B/E/S, in natural logarithm), coverage (number of firms covered by the analyst, in natural logarithm), and horizon (the number of days from forecast issue date to forecast period end date, in natural logarithm). Firm characteristics include firm size, market-to-book ratio, institutional investor holdings, and

M

trading volume five days prior to analysts’ forecast announcements. Heteroskedasticity and

PT

3. Results

ED

autocorrelation consistent (HAC) standard errors are clustered by analyst-firm.

CE

Table 1, Panel A reports summary statistics on the key variables used in the study. The mean (median) ROUNDING in the sample is 0.054 (0), which indicates that rounded forecasts

AC

constitute around 5.4% of all forecasts in the data sample. Descriptive statistics for analyst and firm characteristic variables are also reported. The average portfolio size is 15 firms, and the average analyst in the sample has been providing forecasts in I/B/E/S for 7 years. Around 18% of analysts in the sample are recognized as an All Star analyst. Panel B compares the main variables

7

ACCEPTED MANUSCRIPT

between pilot and non-pilot stocks, before and during Reg SHO.7 A univariate comparison shows that the rounded forecasts increase for pilot stocks during Reg SHO, while they remain unchanged for non-pilot stocks. The abnormal returns around analyst forecast announcements are significantly lower for pilot stocks (a decrease of around 0.6%) than for non-pilot stocks (a 0.1%

CR IP T

decrease) during Reg SHO. Trading volume increases for both pilot and non-pilot stocks, although the economic magnitude is small. This suggests that trading volume might have increased from greater short-selling activity during Reg SHO. I do not find significant differences in the information environment, since the earnings surprise, which is the difference

AN US

between actual earnings and analysts’ median forecasts, remains similar.

The testable hypothesis is that short-selling activity triggered by Reg SHO increases analysts’ tendencies to round forecasts. Here, I estimate equation (1) using the linear probability

M

model.8 Table 2 reports the primary DiD tests that compare the likelihood of rounded forecasts of pilot stocks during Reg SHO. The dependent variable is ROUNDING in all specifications.

ED

Column (1) presents a parsimonious specification excluding firm and analyst characteristics. Column (2) includes only the firm characteristics that may affect analysts’ incentives to gather

PT

information. In column (3), I include analyst characteristics that may affect the informativeness

CE

of the forecast revision, and in columns (4) and (5), I include trading volume of the firm. The main coefficient of interest is PILOT*DURING, as this captures the changes in rounded forecasts

AC

of pilot stocks during Reg SHO. In all columns, the coefficient of PILOT*DURING is positive and significant at the 1% level, suggesting that analysts’ tendencies to round forecasts increased

7

I do not compare the post-Reg SHO period, since there was a permanent removal of the uptick rule for all stocks in June 13, 2007. 8 Findings are materially similar when using a logit model. I report the results from the linear probability model to include HAC standard errors that are robust to the autocorrelation in analysts’ forecasts.

8

ACCEPTED MANUSCRIPT

significantly for pilot stocks after the implementation of Reg SHO.9 In column (5), I include only stocks with above-median trading volume levels around analyst announcements, assuming that short-selling activity increased more for stocks with high trading volume during Reg SHO. Using trading volume as a noisy proxy, I find that analysts’ tendency to round forecasts is stronger for

CR IP T

firms with more short-selling activity.10 More rounded forecasts indicate that analysts gather less precise information for pilot firms with increased short-selling pressure. It could also be that analysts strategically decrease their forecast precision to prevent negative information leakage to short sellers. However, I do not find evidence of this alternative interpretation, since rounded

AN US

forecasts increase for both positive and negative forecast revisions (untabulated).

The findings in Table 2 indicate that analysts tend to put less effort into pilot stocks with more pessimistic investors. Accordingly, I explore whether pilot stocks do experience downward

M

stock price pressure when short-selling activity increases, by comparing the abnormal stock returns around analyst forecast announcements. Table 3 reports the changes in abnormal stock

ED

returns of pilot stocks during Reg SHO. The dependent variable is the cumulative abnormal

PT

return (CAR) around the analyst announcement window of [-5, +1] days. The coefficient of PILOT*DURING is significantly negative, which indicates that pilot stocks have lower returns

CE

from increased short-selling activity. In columns (4)-(6), I repeat the analysis including only stocks with above-median trading volume levels, and I find that the returns are significantly

AC

lower for these firms. This finding is consistent with the assumption that trading volume

9

The coefficient of PILOT is significantly negative, which indicates that pilot stocks have less rounded forecasts before Reg SHO. There can be coincidental differences between pilot and non-pilot firm characteristics related to forecast rounding, but the difference is not by design. However, the focus of the research design is to compare the changes in rounding of pilot stocks, before and after the implementation of Reg SHO. 10 Blau and Wade (2012) find that short-selling increases prior to analyst stock recommendations. In a similar vein, Baklaci, Suer, and Yelkenci (2016) find that short-selling activity increases stock price volatility.

9

ACCEPTED MANUSCRIPT

increases with short-selling activity. Overall, the findings show that Reg SHO led to increased short-selling trades in pilot stocks, which resulted in a downward price pressure. Next, I explore several explanations for why analysts tend to allocate less effort to firms

CR IP T

with more short-selling activity. Prior literature demonstrates that analysts’ effort allocation depends on their various incentives, and they have stronger incentives to issue optimistically biased forecasts. The first incentive is to maintain favorable relationships with firm management, since firm insiders are valuable sources of information. For example, Brown, Call, Clement, and

AN US

Sharp (2015) find that private communication with firm management is more useful than the analyst’s own primary research; therefore, maintaining a favorable relationship is a top priority. This incentive increases with the level of firm-specific information. When firms perform poorly, I expect analysts to put less effort into gathering precise negative information, especially when

M

analysts need to maintain a favorable relationship with management. To test this possible explanation, I compare the rounding of forecasts between firms with high and low levels of firm-

ED

specific information in Table 4. Following Frankel, Kothari, and Weber (2006), I measure the

PT

degree of firm-specific information by calculating the correlation between the firm and market return, and use it as a proxy for the importance of maintaining a favorable relationship with firm

CE

management. A firm has a high degree of firm-specific information if the 1- R2 from the firm’s market-model regression is high. I sort firms by their level of firm-specific information for each

AC

year. Columns (1)-(3) include firms with above-median levels of firm-specific information, and columns (4)-(6) include firms with below-median levels. The results show that the increase in rounded forecasts is significant only for firms with high levels of firm-specific information. Consistent with the existing literature, analysts’ incentives to issue optimistic views are stronger

10

ACCEPTED MANUSCRIPT

for these firms. In sum, the findings in Table 4 indicate that analysts put less effort into firms with more firm-specific information when they expect these firms to perform badly. Analysts’ second incentive is related to their own reputation building and career concerns.

CR IP T

Institutional investors provide important feedback to analysts that impacts their reputation and career outcomes. Although analysts tend to refrain from collecting information about pessimistic firms with high short-selling pressure, they still have an incentive to put effort into pilot stocks with more institutional holdings.11 Harford et al. (2016) find that analysts allocate more effort

AN US

into firms that are important for their careers. Accordingly, I expect that rounding decreases with the level of institutional holdings. To test this explanation, I next examine whether the effect of increased short-selling activity from Reg SHO is stronger for pilot stocks with low levels of institutional holdings. In Table 5, I sort firms into two groups based on their levels of

M

institutional holdings each year. Columns (1)-(3) include firms with institutional holdings below the median level, while columns (4)-(6) include firms above the median. The findings in columns

ED

(1)-(3) show that the increase in rounded forecasts is pronounced for pilot stocks with lower

PT

levels of institutional holdings, as the coefficient of PILOT*DURING is positive and significant. In contrast, the effect of Reg SHO is insignificant for pilot stocks with high levels of institutional

CE

holdings in columns (4)-(5). In column (6), rounded forecasts decrease for stocks with more institutional investors. The results show that analysts’ effort allocation among pilot stocks

AC

increases with institutional holdings and their reputational concerns. The findings are also consistent with those of Dechow and You (2012), who show that the rounding of forecasts

11

Analysts’ forecast accuracy is an important factor in performance assessment, and poor forecast accuracy could lead to negative career outcomes (Hong, Kubik, and Solomon, 2000; Hong and Kubik, 2003).

11

ACCEPTED MANUSCRIPT

decreases with the level of institutional ownership, since institutional investors are important factors in analysts’ career outcomes. Additional alternative explanations for the changes in the rounding of forecasts include

CR IP T

changes in expected trading commissions and the information environment. Jackson (2005) shows that trading commissions are lower for firms with pessimistic investors than for firms with optimistic investors.12 Moreover, Dechow and You (2012) show that analysts gather less precise information and are more likely to issue rounded forecasts for firms that generate less brokerage

AN US

or investment banking business. If analysts expected pilot stocks to have lower trading commissions due to poor performance, they would have put less effort into pilot stocks during Reg SHO. On the other hand, increased short-selling activity of pilot stocks leads to more trading activity, which increases analysts’ trading commissions. Consistent with the latter conjecture, I

M

find that trading volume increases for pilot stocks during Reg SHO in Panel B of Table 1.

ED

Lastly, the information environment of the pilot stocks may change due to more shortselling activity. Fang et al. (2016) show that earnings management decreased for pilot stocks,

PT

which makes earnings less predictable. To the extent that poor performance is associated with more earnings uncertainty and poorer disclosures, the changes in the information environment

CE

would affect analysts’ rounding of forecasts. I use the firm’s earnings surprise as a proxy for the

AC

information environment to examine any changes in the information environment of pilot stocks.13 From Table 1, Panel B, I find that the earnings surprises of pilot stocks do not change during Reg SHO, which indicates that the information environment remains similar. Therefore,

12

Irvine (2004) and Jackson (2005) show that analysts issue optimistically biased forecasts to generate more trading activity. 13 Li and Chen (2016) use analyst forecast dispersion as a proxy for the difference in opinions among investors.

12

ACCEPTED MANUSCRIPT

the alternative explanations related to trading commissions and the information environment do not explain the changes in analysts’ rounding of forecasts. Overall, I find that short selling has a significant causal effect on analysts’ incentives to

CR IP T

gather precise information and on their forecasting properties. In response to pilot stocks’ lower returns, analysts allocate less effort to pilot stocks due to their own opportunistic incentives. In the cross-section, analysts’ rounding of forecasts is more pronounced for firms with lower levels

AN US

of institutional holdings and for firms with more firm-specific information.

4. Conclusion

This paper examines the effect of short selling on analyst forecast precision using a

M

natural experiment–Reg SHO. I find that short selling increases analysts’ tendency to round forecasts. The findings suggest that analysts put less effort into collecting precise information

ED

about pilot stocks with increased short selling, due to the downward pressure on their stock

PT

prices. In the cross-section, rounded forecasts increase for firms where it is important for analysts to maintain favorable relationships, and for firms that are less important to analysts’ career

CE

outcomes and reputation. The results indicate that analysts exhibit opportunistic behavior when allocating their effort, in response to the changes in short-selling activity from Reg SHO. Since

AC

analysts are important information intermediaries in financial markets, investors can benefit from understanding analysts’ incentives and effort allocation across firms when using their forecasts. The current findings add to the literature on the interaction between financial analysts and short sellers, and to the literature that examines the effect of regulatory changes in financial markets.

13

ACCEPTED MANUSCRIPT

References Abarbanell, J., 1991, Do analysts’ earnings forecasts incorporate information in prior stock price changes? Journal of Accounting and Economics 14, 147-165. Angelis, D., G. Grullon, and S. Michenaud, 2013, Downside risk and the design of CEO incentives: Evidence from a natural experiment, Working paper, Rice University.

CR IP T

Baklaci, H. F., O. Suer, and T. Yelkenci, 2016, A closer insight into the causality between short selling trades and volatility. Finance Research Letters 17, 48-54. Blau, B. M., and C. Wade, 2012, Informed or speculative: Short selling analyst recommendations. Journal of Banking & Finance 36(1), 14-25. Brown, L. D., A. C. Call, M. B. Clement, and N. Y. Sharp, 2015, Inside the “black box” of sell‐side financial analysts, Journal of Accounting Research 53(1), 1-47.

AN US

Chen, S., and D. A. Matsumoto, 2006, Favorable versus unfavorable recommendations: The impact on analyst access to management‐provided information, Journal of Accounting Research 44(4), 657-689. Chopra, V. K., 1998, Why so much error in analysts’ earnings forecasts? Financial Analysts Journal 54, 30-37. Crichfield, T., T. Dyckman, and J. Lakonishok, 1978, An evaluation of security analysts’ forecasts, Accounting Review 651-668. Dechow, P. M., and H. You, 2012, Analysts’ motives for rounding EPS forecasts. The Accounting Review 87(6), 1939-1966.

M

Diether, K., K. Lee, and I. Werner, 2009, It’s SHO time! Short-sale price tests and market quality, Journal of Finance 64, 37-73.

ED

Diether, K., C. Malloy, and A. Scherbina, 2002, Differences of opinion and the cross-section of stock returns, Journal of Finance 57, 2113-2141. Dreman, D., and M. Berry, 1995, Analyst forecasting errors and their implications for security analysis, Financial Analysts Journal 51, 30-42.

PT

Fang, V., A. Huang, and J. Karpoff, 2016, Short selling and earnings management: A controlled experiment, Journal of Finance 71, 1251-1294.

CE

Francis, J., and D. Philbrick, 1993, Analysts’ decisions as products of a multi-task environment, Journal of Accounting Research 31, 216-230.

AC

Frankel, R., S. P. Kothari, and J. Weber, 2006, Determinants of the informativeness of analyst research, Journal of Accounting and Economics 41, 29-54. Grullon, G., S. Michenaud, and J. Weston, 2015, The real effects of short-selling constraints, Review of Financial Studies 28, 1737-1767. Harford, J., F. Jiang, R. Wang, and F. Xie, 2016, Career concerns and strategic effort allocation by analysts, University of Washington Working paper. Hayes, R. M., 1998, The impact of trading commission incentives on analysts’ stock coverage decisions and earnings forecasts, Journal of Accounting Research 36(2), 299-320. Hong, H., and J. Kubik, 2003, Analyzing the analysts: Career concerns and biased earnings forecasts, Journal of Finance 58, 313-351.

14

ACCEPTED MANUSCRIPT

Hong, H., J. D. Kubik, and A. Solomon, 2000, Security analysts’ career concerns and the herding of earnings forecasts, RAND Journal of Economics 31, 121-144. Irvine, P., 2004, Analysts’ forecasts and brokerage-firm trading, The Accounting Review 79, 125-149. Jackson, A., 2005, Trade generation, reputation and sell-side analysts, Journal of Finance 60, 673-717. Ke, Y., L. Kin, J. Sheng, and J. Zhang, 2015, Does short selling mitigate optimism in financial analyst forecast? Evidence from a randomized experiment, Working paper, University of British Columbia.

CR IP T

Li, L., and C. Chen, 2016, Analysts’ forecast dispersion and stock returns: A panel threshold regression analysis based on conditional limited market participation hypothesis, Finance Research Letters 18, 100-107. Li, Y., and L. Zhang, 2015, Short selling pressure, stock price behavior, and management forecast precision: Evidence from a natural experiment, Journal of Accounting Research 53, 79-117. Lim, T., 2001, Rationality and analysts’ forecast bias, Journal of Finance 56, 369-385.

AN US

Mayew, W. J., 2008, Evidence of management discrimination among analysts during earnings conference calls, Journal of Accounting Research 46(3), 627-659. McNichols, M., and P. O’Brien, 1997, Self-selection and analyst coverage, Journal of Accounting Research 35, 167199. Securities and Exchange Commission (SEC), 2007, Economic analysis of the short sale price restrictions under the Regulation SHO pilot, Office of Economic Analysis.

M

Soltes, E., 2014, Private interaction between firm management and sell‐side analysts, Journal of Accounting Research 52(1), 245-272.

AC

CE

PT

ED

Stickel, S. E., 1990, Predicting individual analyst earnings forecasts, Journal of Accounting Research 28, 409-417.

15

ACCEPTED MANUSCRIPT

TABLE 1 Descriptive Statistics

CR IP T

Panel A reports the descriptive statistics for the main variables. The sample period is from 2003:Q2 to 2007:Q2. ROUNDING is an indicator variable that equals one if analyst i’s forecast for firm j in year k is an integer. All Star is an indicator variable that equals one if the analyst is included in the All Star analyst list by the Institutional Investors magazine in year k. Brokerage Size is measured by the number of analysts in a given brokerage firm in year k. Experience is the number of years analyst i issues a forecast for a firm, averaged across the firms the analyst covers in year k. Coverage is the number of firms covered by analyst i in year k. Horizon is the number of days from the analyst’s forecast to the actual earnings announcement date (in natural logarithm). Institutional Holdings is the fraction of institutional investor holdings in firm j at year k. Size is the market value of equity for firm j in year k (in natural logarithm). Market/Book is the market-to-book ratio of firm j in year k calculated as the market value of the firm’s equity at the end of year k plus the difference between the book value of the firm’s assets and the book value of the firm’s equity in year k, divided by the book value of firm j’s assets in year k. Trading Volume is the sum of daily trading volume during the five-day window prior to the analyst’s forecast announcement of firm j (in natural logarithm).

AN US

Panel B compares the main variables between pilot and non-pilot stocks, before and during Reg SHO. Stock returns is the cumulative abnormal return (CAR) around the analyst forecast announcement date [-5, 1]. Earnings surprise is the difference between actual earnings and the analyst median forecast, normalized by the prior year ending stock price. Other variables are defined in Panel A. Panel A

Mean

Analyst Forecast Variable

0.054

0

0.182 61.438 6.815 16.144 5.438

0.386 54.227 5.257 8.076 0.137

0 45 5 15 5.488

0.750 8.041 2.077 15.377

0.208 1.610 1.378 1.526

0.775 7.911 1.608 15.396

ED

PT

Analyst Characteristics All Star Brokerage Size Experience Coverage Horizon

Median

0.166

M

Rounding

Stddev

CE

Firm Characteristics

AC

Institutional Holdings Size Market/Book Trading Volume

16

ACCEPTED MANUSCRIPT

Panel B Before Reg SHO

During Reg SHO

T-stat

P-value

Rounding 0.049 0.051

Stock Returns Pilot Stocks Non-pilot Stocks

0.639 0.423

0.007 0.301

Trading Volume Pilot Stocks Non-pilot Stocks

15.295 15.202

15.541 15.491

Earnings Surprise Pilot Stocks Non-pilot Stocks

0.001 0.001

AC

CE

PT

ED

M

0.001 0.001

-2.322 -0.007

17

0.020 0.976

CR IP T

0.043 0.051

AN US

Pilot Stocks Non-pilot Stocks

8.011 2.227

<0.0001 0.026

-9.037 -15.118

<0.0001 <0.0001

0.412 0.908

0.680 0.327

ACCEPTED MANUSCRIPT

TABLE 2 Rounding and Short-Selling Activity

Dependent Variable

ROUNDING

PILOT DURING

(1)

(2)

0.005*** (0.001) -0.007*** (0.001) 0.000 (0.000)

Institutional Holdings Size

(5)

0.005***

0.004***

0.004***

0.007***

(0.001) -0.007*** (0.001) 0.002*** (0.001) -0.024*** (0.006) 0.003*** (0.000) -0.004*** (0.001)

(0.001) -0.007*** (0.001) 0.002** (0.001) -0.025*** (0.006) 0.003*** (0.000) -0.004*** (0.001) -0.000 (0.003) 0.000 (0.000) 0.001*** (0.000) 0.003 (0.002) 0.011*** (0.001)

0.055*** (0.007) 60610 0.003

-0.037*** (0.007) 53085 0.004

(0.001) -0.006*** (0.001) 0.000 (0.001) -0.032*** (0.006) 0.001* (0.001) -0.004*** (0.001) 0.000 (0.004) 0.000 (0.000) 0.001*** (0.000) 0.003 (0.002) 0.003** (0.001) 0.003*** (0.001) -0.014 (0.009) 51292 0.005

(0.002) -0.008*** (0.001) 0.004* (0.002) -0.017** (0.007) 0.001*** (0.001) -0.006*** (0.001) -0.001 (0.002) 0.000 (0.000) 0.001*** (0.000) 0.004* (0.002) 0.011** (0.004) -0.000 (0.001) -0.027 (0.020) 30608 0.004

ED

All Star Brokerage Size

PT

Experience

CE

Horizon

(4)

M

Market/Book

(3)

AN US

PILOT*DURING

Coverage

CR IP T

Table 2 compares the changes in rounded forecasts for pilot stocks during Reg SHO. The dependent variable is the tendency to round forecasts, ROUNDING, which equals one if analyst i’s forecast for firm j in year k is an integer. PILOT is an indicator variable that equals one if the stock was selected into the pilot program by the SEC, and zero otherwise. DURING is an indicator variable that equals one for the Reg SHO periods, and zero for the pre-Reg SHO periods. All other variables follow from Table 1. Columns (1)-(4) include all stocks, while column (5) includes stocks with above-median trading volume levels. HAC standard errors are clustered by firm-analyst and reported in parentheses. ***, **, and * denote significance levels of 1%, 5%, and 10%, respectively.

AC

Trading Volume Constant N R-sq

0.052*** (0.000) 60612 0.000

18

ACCEPTED MANUSCRIPT

TABLE 3 Stock Returns around Analyst Forecast Revisions

Dependent Variable

Stock Returns (CAR)

Sample

All Stocks

PILOT DURING REV

(1)

(2)

(3)

-0.153**

-0.169**

-0.143*

(0.076) 0.119 (0.077) -0.321*** (0.103) 16.795*** (0.895)

(0.081) 0.135* (0.080) -0.359*** (0.093) 16.839*** (0.797) 0.935*** (0.182) -0.048 (0.069) -0.078*** (0.009)

(0.082) 0.093 (0.075) -0.377*** (0.097) 16.955*** (0.675) 0.881*** (0.252) -0.079 (0.120) -0.081*** (0.015) -0.254*** (0.035) 0.001*** (0.000) 0.097*** (0.012) -0.124*** (0.016) 0.747* (0.422) 0.054 (0.102) -4.153 (2.886) 54322 0.117

Institutional Holdings

M

Size Market/Book

ED

All Star

Horizon

CE

Coverage

PT

Brokerage Size

AC

Trade Volume Constant N R-sq

High Trading Volume Stocks

0.421*** (0.102) 57888 0.113

(4)

(5)

(6)

-0.518***

-0.536***

-0.433***

(0.065) 0.254*** (0.032) -0.042 (0.115) 15.508*** (0.781)

(0.064) 0.277*** (0.032) -0.090 (0.084) 15.607*** (0.727) 0.982*** (0.365) -0.005 (0.076) -0.095*** (0.022)

0.298*** (0.114) 33941 0.099

-0.192 (1.001) 33941 0.101

(0.062) 0.192*** (0.027) -0.152 (0.092) 15.700*** (0.544) 1.075*** (0.272) -0.016 (0.103) -0.106*** (0.026) -0.361*** (0.047) 0.001** (0.000) 0.166*** (0.021) -0.071** (0.033) 1.526*** (0.578) 0.040 (0.135) -9.128** (3.809) 32047 0.105

AN US

PILOT*DURING

Experience

CR IP T

Table 3 compares the stock returns of pilot stocks around analyst forecast revisions. The dependent variable is the cumulative abnormal return (CAR) around the analyst forecast announcement date [-5, 1]. Columns (1)-(3) include all stocks, while columns (4)-(6) include stocks with above-median trading volume. All variables follow the definitions in Table 1. HAC standard errors are clustered by firm-analyst and reported in parentheses. ***, **, and * denote significance levels of 1%, 5%, and 10%, respectively.

0.282 (0.691) 57888 0.115

19

ACCEPTED MANUSCRIPT

TABLE 4 Firm-Specific Information and Rounded Forecasts

Dependent Variable

ROUNDING

Level of Firm-specific Information

HIGH (2)

(3)

0.007***

0.006***

0.005***

(0.001) -0.009*** (0.001) 0.000 (0.001)

(0.001) -0.009*** (0.001) 0.001 (0.001) -0.017** (0.008) 0.001 (0.000) -0.004*** (0.001)

(0.001) -0.007*** (0.001) 0.001* (0.001) -0.026*** (0.008) -0.002** (0.001) -0.004*** (0.000) 0.001 (0.004) 0.000*** (0.000) 0.001*** (0.000) 0.002* (0.001) 0.011* (0.007) 0.004*** (0.001) -0.056 (0.036) 35709 0.004

DURING Institutional Holdings Size

M

Market/Book

ED

All Star Brokerage Size

PT

Experience

CE

Horizon

AC

Trade Volume Constant N R-sq

0.030*** (0.001) 35751 0.001

(4)

(5)

(6)

0.001

0.001

0.003

(0.002) -0.004** (0.001) 0.012*** (0.003)

(0.002) -0.003*** (0.001) 0.010*** (0.003) -0.008 (0.007) 0.004*** (0.001) -0.005*** (0.001)

0.030*** (0.003) 25747 0.000

0.011*** (0.004) 25747 0.002

(0.004) -0.004** (0.002) 0.005 (0.004) -0.020*** (0.006) 0.006*** (0.001) -0.005*** (0.002) 0.001 (0.003) -0.000 (0.000) 0.001*** (0.000) 0.003** (0.001) 0.002 (0.007) -0.001 (0.001) -0.005 (0.034) 25506 0.005

AN US

PILOT

LOW

(1) PILOT*DURING

Coverage

CR IP T

Table 4 reports the multivariate DiD regression results of two groups of stocks sorted by the level of firm-specific information. Firm-specific information is computed as the 1- R2 from the firm’s market-model regression. Firms are sorted into two groups (above or below the median) based on the level of firm-specific information in their returns each year. The HIGH (LOW) group includes stocks above (below) the median level of firm-specific information. All variables follow the definitions in Table 1. HAC standard errors are clustered by firm-analyst and reported in parentheses. ***, **, and * denote significance levels of 1%, 5%, and 10%, respectively.

0.045*** (0.005) 35749 0.002

20

ACCEPTED MANUSCRIPT

TABLE 5 Institutional Holdings and Rounding of Forecasts

Dependent Variable

ROUNDING

Level of Institutional Holdings

PILOT DURING

LOW (1)

(2)

(3)

0.010**

0.009**

0.014***

(0.004) -0.012*** (0.004) 0.004 (0.003)

(0.004) -0.011*** (0.004) 0.004 (0.003) -0.017* (0.010) 0.003*** (0.000) -0.004*** (0.001)

(0.004) -0.012*** (0.004) -0.000 (0.003) -0.036*** (0.011) 0.001*** (0.001) -0.004*** (0.001) 0.000 (0.007) 0.000 (0.000) 0.001*** (0.000) 0.003* (0.002) -0.011 (0.008) 0.002** (0.001) 0.067* (0.039) 26458 0.004

Institutional Holdings Size

M

Market/Book

ED

All Star Brokerage Size

PT

Experience

CE

Horizon

Trade Volume

AC

Constant N R-sq

HIGH

0.036*** (0.003) 30315 0.001

(4)

(5)

(6)

-0.001

-0.001

-0.006***

(0.002) -0.002 (0.002) -0.000 (0.002)

(0.002) -0.003 (0.002) 0.000 (0.002) -0.019 (0.012) 0.003*** (0.000) -0.004*** (0.001)

0.026*** (0.002) 30297 0.000

0.030*** (0.011) 30295 0.002

(0.002) 0.001 (0.002) 0.001 (0.002) -0.021* (0.013) 0.001 (0.001) -0.004*** (0.001) -0.000 (0.001) 0.000*** (0.000) 0.001*** (0.000) 0.002 (0.002) 0.017*** (0.005) 0.004*** (0.001) -0.114*** (0.039) 27834 0.004

AN US

PILOT*DURING

Coverage

CR IP T

Table 5 reports the multivariate DiD regression results of two groups of stocks sorted by the level of institutional holdings. Firms are sorted into two groups (above or below the median) based on their level of institutional holdings each year. The HIGH (LOW) group includes stocks above (below) the median level of institutional holdings. All variables follow the definitions in Table 1. HAC standard errors are clustered by firm-analyst and reported in parentheses. ***, **, and * denote significance levels of 1%, 5%, and 10%, respectively.

0.033*** (0.008) 30315 0.003

21

AC

CE

PT

ED

M

AN US

CR IP T

ACCEPTED MANUSCRIPT