"Pacific Accounting Review Management earnings forecasts in a continuous disclosure environment H. Chan, R. Faff, Y.K. Ho, A. Ramsay, Article information: To cite this document: H. Chan, R. Faff, Y.K. Ho, A. Ramsay, (2007) "Management earnings forecasts in a continuous disclosure environment", Pacific Accounting Review, Vol. 19 Issue: 1, pp.5-30, https:// doi.org/10.1108/01140580710754629 Permanent link to this document: https://doi.org/10.1108/01140580710754629 Downloaded on: 12 July 2017, At: 17:07 (PT) References: this document contains references to 29 other documents. To copy this document: [email protected] The fulltext of this document has been downloaded 960 times since 2007* Users who downloaded this article also downloaded: (2014),"Earnings management and voluntary disclosure of management's responsibility for the financial reports", Asian Review of Accounting, Vol. 22 Iss 3 pp. 233-256 <a href="https://doi.org/10.1108/ ARA-11-2013-0075">https://doi.org/10.1108/ARA-11-2013-0075</a> (2010),"Corporate environmental disclosure, corporate governance and earnings management", Managerial Auditing Journal, Vol. 25 Iss 7 pp. 679-700 <a href="https:// doi.org/10.1108/02686901011061351">https://doi.org/10.1108/02686901011061351</a> Access to this document was granted through an Emerald subscription provided by emerald-srm:414810 [] For Authors If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service information about how to choose which publication to write for and submission guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information. About Emerald www.emeraldinsight.com Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online products and additional customer resources and services. Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation. *Related content and download information correct at time of download. Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)The current issue and full text archive of this journal is available at www.emeraldinsight.com/0114-0582.htm Management Management earnings forecasts earnings forecasts in a continuous disclosure environment 5 H. Chan Department of Finance, University of Melbourne, Melbourne, Australia R. Faff Department of Accounting and Finance, Monash University, Melbourne, Australia Y.K. Ho Department of Finance and Accounting, National University of Singapore, Singapore, and A. Ramsay Department of Accounting and Finance, Monash University, Melbourne, Australia Abstract Purpose – The purpose of this paper is to assess management earnings forecasts in a continuous disclosure environment. Design/methodology/approach – A large sample of hand checked Australian management earnings forecasts are examined. These data are analysed using a series of logistic regressions. Hypotheses are proposed and tested based on Skinner’s litigation cost hypothesis. Increases in non-routine management earnings forecasts post-2000; and increases in the proportion of such forecasts that contain bad news are predicted. The relationship between forecast speci?city and forecast news content is investigated. Findings – It was found that, post-2000, legislative changes and increased enforcement action by ASIC were followed by increased disclosure of non-routine management earnings forecasts. For routine forecasts, no signi?cant increase in forecast disclosure is observed. This result is consistent with Skinner as is the ?nding that the increased disclosure is only apparent for bad news non-routine forecasts. For the second objective, evidence was found that the larger the gap between market expectations and actual performance the more speci?c the forecast, but only for bad news forecasts. Originality/value – The study extends the small amount of research investigating the characteristics of management earnings forecasts. It also provides an assessment of the effectiveness of efforts by ASIC to ensure that management meet their continuous disclosure obligations. Keywords Disclosure, Earnings, Financial forecasting, Australia Paper type Research paper The authors would like to thank Peter Clarkson, Peter Easton, Abe Herzberg, Greg Pound and Philip Sinnadurai and workshop participants at the AFAANZ 2004 conference at Alice Springs, Monash University, University of Auckland, University of Adelaide and University of Melbourne for comments on earlier drafts. They also thank two anonymous referees for Paci?c Accounting Review constructive comments and Alex Featherstone for research assistance. Financial assistance Vol. 19 No. 1, 2007 pp. 5-30 provided by a Monash University Faculty of Business and Economics research grant is q Emerald Group Publishing Limited gratefully acknowledged. The ?rst author gratefully acknowledges ?nancial assistance provided 0114-0582 by a Monash University Research Fellowship. DOI 10.1108/01140580710754629 Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)1. Introduction PAR Investors rely on a broad set of information. Such information varies across many 19,1 dimensions or characteristics including its timeliness, quality, routine, and speci?city. Many types and sources of potentially relevant information exist: the past time-series of earnings and the forecasts made by ?nancial analysts are but two noteworthy examples. Another source of highly relevant earnings-related news for investors is voluntary earnings forecasts issued by ?rms’ management. 6 There are two main objectives of this study. First, drawing on changes in Australia’s continuous disclosure regulatory environment, we propose and test hypotheses based on Skinner’s (1994) litigation cost hypothesis. We predict: . increases in the number of non-routine management earnings forecasts following increased regulatory activity by the Australian Securities and Investments Commission (ASIC) post-2000; and . increases in the proportion of such forecasts that contain bad news. Second, we investigate the relationship between forecast speci?city (point, range, maximum and minimum) and forecast news content. Prior research on this issue (Kasznik and Lev, 1995; Bamber and Cheon, 1998) contains con?icting predictions and ?ndings. We extend prior research by considering a different market and regulatory environment, by using several performance metrics to assess news content and by investigating the possible effect of changes in the regulatory environment on this relationship. These issues are of interest because improved disclosure has implications for the ef?ciency of capital markets and subsequent allocation of resources in the economy. Further, more timely and effective disclosure reduces opportunities for insider trading by management. Our study contributes to the literature by extending the small amount of research investigating the characteristics of management earnings forecasts. Hirst et al. (2006), note that managers have considerable control over the characteristics associated with a management earnings forecast and so this “is a rich area in which to advance our knowledge of the earnings forecasting process” (Hirst et al., 2006, p. 26)). We also contribute to the prior academic literature by considering Skinner’s (1994) litigation cost explanation for management’s voluntary disclosure of bad earnings news in a market which is similar to the USA in terms of the obligation to disclose, but different in terms of its legal and regulatory environment. Listing rules on the Australian Stock Exchange (ASX) require listed companies to continuously disclose information likely to have a material effect on share price. The study also adds to the literature that provides an assessment of the effectiveness of efforts by ASIC to ensure that management meet their continuous disclosure obligations. Our investigation of the relationship between forecast speci?city and forecast news content provides further evidence that may help resolve the con?icting predictions and results in the prior literature. We ?nd evidence that the larger the gap between market expectations and actual performance the more speci?c the forecast, but only for bad news forecasts, this is consistent with Kasznik and Lev (1995). Our study also contributes to the literature by applying and reporting both analyst’s forecasts and prior period earnings as benchmarks for assessing the news content of forecasts. This allows comparison of results from the two news benchmarks allowing us to gauge Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)whether the method of assessing news may be driving the prior results. Our modelling Management of the forecast speci?city and forecast news relationship is innovative and allows a earnings more complete control for news content than has been seen in the prior literature. forecasts Finally, our study is a substantial extension of existing Australian research on management earnings forecasts, in that we cover more recent time periods, with a larger sample of hand checked announcements from the of?cial Signal G database. A feature of our paper is that, rather than relying on secondary media sources (as is 7 generally the case in USA management forecast disclosure research), we source our management earnings forecast data direct from the of?cial company announcement database maintained by the ASX. Our results in relation to our ?rst objective show that, post-2000 (de?ned as the period after 1 January 2000), legislative changes and increased enforcement action by ASIC were followed by increased disclosure of non-routine management earnings forecasts. For routine forecasts, no signi?cant increase in forecast disclosure is observed. This result is consistent with Skinner (1994) as is the ?nding that the increased disclosure is only apparent for bad news non-routine forecasts. For routine forecasts, no signi?cant increase in bad news forecast disclosure is observed. For our second objective, we investigate competing explanations of the relationship between the news content of forecasts and forecast speci?city. We ?nd evidence that the larger the gap between market expectations and actual performance the more speci?c the forecast, but only for bad news forecasts (Kasznik and Lev, 1995). We test whether changes in the regulatory environment have any effect on this relationship and ?nd weak evidence that the observed relationship is stronger post the regulatory changes. The remainder of the study is structured as follows: section 2 outlines prior research leading to the formulation of testable hypotheses. Section 3 outlines our research method, data and sample, Section 4 outlines and discusses our results and section 5 presents a summary and conclusions. 2. Theory development We begin by brie?y reviewing the prior literature on voluntary disclosure by management. Analytical research on discretionary information disclosure concludes that, in the absence of costs, full disclosure will occur (Grossman, 1981; Milgrom, 1981). However, in the presence of costs and given uncertainty about management’s private information, the failure by management to disclose does not necessarily imply bad news and, hence, full disclosure does not necessarily result. In relation to voluntary management earnings forecasts, Penman (1980) concludes that in an unregulated market, on average, ?rms with good news are more likely to disclose earnings forecasts than those with bad news. Ajinkya and Gift (1984) hypothesize that managers issue voluntary earnings forecasts when they believe that the market’s earnings expectations are signi?cantly out of line with their own earnings expectations. In order to avoid large movements in share prices at earnings announcement, management is motivated to voluntarily release an earnings forecast to adjust the market’s expectations more into line with their own. Ajinkya and Gift argue that this motivation for forecast disclosure applies equally to both good and bad news earnings forecasts. Skinner (1994) investigates the voluntary disclosure of bad news earnings forecasts by smaller USA listed companies. He ?nds that voluntary management earnings Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)forecasts are more likely to occur when there are large negative earnings surprises. He PAR argues that, contrary to Ajinkya and Gift, managers behave as if they bear large costs 19,1 when investors are surprised by large negative earnings surprises. Skinner attributes the increased propensity to preannounce bad news to a reduction in litigation costs resulting from the voluntary release of bad earnings news. Investors who lose money may sue managers for failing to promptly disclose bad earnings news. Lawsuits may result in personal liability on behalf of managers. A further reason for voluntary 8 disclosure is that managers incur reputational costs if they fail to disclose bad news in a timely fashion. Firms that fail to keep the market informed may be shunned by fund managers and ?nancial analysts. Francisetal. (1994) cast doubt on the link drawn by Skinner between early forecast disclosure of bad earnings news and avoidance of litigation. Skinner (1997) acknowledges that timely bad news earnings forecast disclosure may not prevent litigation, but provides evidence that such disclosure reduces the cost of such litigation. Skinner (1997) also points out the potential endogeneity between forecast disclosures and subsequent litigation. This issue is addressed by Field et al. (2005), who use a simultaneous equations methodology to investigate the relationship between forecast disclosure and litigation. They ?nd no evidence to support the notion that forecast disclosure causes litigation, but ?nd some evidence that forecast disclosures may deter certain types of litigation. Johnson et al. (2001) investigate the effects of the passage of the Private Securities Litigation Reform Act of 1995 on forecast disclosures. They ?nd that bad news disclosures increase after the passage of the Act and indicate that such a result is inconsistent with the view that the primary motivation for disclosure of bad news is to protect against shareholder litigation. Skinner’s arguments are couched in terms of stockholder lawsuits and revolve around private legal actions under SEC Rule 10b-5. In Australia, class actions and private shareholder litigation against management occur less frequently[1]. However, enforcement actions by regulators may also impose costs. Since 1994, the Australian regulatory environment has been one of “continuous disclosure”. The continuous disclosure regime comprises two parts consisting of the ASX Listing Rules and the Corporations Act which outlines the continuous disclosure principle and the statutory enforcement of this principle, including penalties for parties that fail to comply with the disclosure regime[2]. In Australia the principle of continuous disclosure (imposed by the exchange listing rules) is enshrined in the Corporations Act which gives statutory backing to the ASX’s disclosure obligations[3]. ASIC then centrally regulate the application of the continuous disclosure principle. Empirical research into the effects of the introduction of the continuous disclosure regime in Australia is found in Brown et al. (1999). They examine the impact of legislative enforcement of the Australian continuous disclosure requirements on the quantity and timeliness of voluntary disclosures made by ASX listed ?rms[4]. Their sample period (August 1992 to March 1996) covers approximately two years either side of the introduction of legislative enforcement of the continuous disclosure requirements. Their results indicate that “although total disclosures increased post-sanctions, disclosures classi?ed as ‘price-sensitive’ by the ASX only became more frequent for ?rms without large analyst following and for ?rms which are more likely to reveal relatively bad news” (Brown et al., 1999, p. 161)[5]. Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)Further research by Gallery et al. (2002) investigates whether disclosure of Management Australian earnings forecasts is opportunistic or continuous. They study management earnings earnings forecasts in the period June 1994 to December 1996. They ?nd that earnings forecasts changes are pre-empted by forecasts in only 14 per cent of cases, with bad news being pre-empted more often than good news. Gallery etal. (2002) conclude that their results are more consistent with opportunistic rather than continuous disclosure. Galleryetal. 9 (2003) focus on management earnings forecasts released by analyst-followed ASX ?rms in the period September 1994 to June 1998. Gallery et al. conclude that their results “appear to be inconsistent with the intentions of statutory continuous disclosure rules” (Gallery et al., 2003, abstract). Thus the initial introduction of statutory penalties for failing to adhere to continuous disclosure requirements seemed to have relatively little impact on corporate voluntary disclosure. However, in a 1999 survey of directors of ASX listed companies, 80 per cent of respondents expressed the view that continuous disclosure played an important role in improving disclosure in Australian capital markets (Corlettetal., 2000). This ?nding is consistent with the view that from around the beginning of 2000, a number of changes occurred that, arguably, substantially increased the effectiveness of Australian continuous disclosure regulation. One change was in the attitude and activity of the ASX and ASIC. ASIC, in particular, became more active in drawing attention to failures to keep the market informed[6]. The ASX, in conjunction with ASIC, established the continuous disclosure monitoring program in 2000 (Hsu et al., 2006). Another important change was the Financial Services Reform Act 2001. Following this legislation, a failure to comply with ASX listing rule 3.1 is treated as a breach of a ?nancial services civil penalty provision (as well as a criminal offence) and is decided on the civil standard of proof – the balance of probabilities. Any person such as a director involved in the contravention may be subject to a civil penalty order such as disquali?cation, ?nes and compensation. This legislation made it easier for ASIC to prove contraventions. During this period, other factors were also working to strengthen managers, investors and regulators understanding of the need for greater disclosure and transparency. The series of corporate collapses (for example, Enron, WorldCom in the USA, HIH in Australia) placed the focus on the need for better corporate governance and disclosure. Principle 5 of the ASX Corporate Governance Best Practice Guidelines encouraged development of written policies and procedures designed to ensure compliance with continuous disclosure obligations. Finally, the end of the 1990s bull market and the period of “irrational exuberance” by investors further focused attention on the need for disclosure and transparency. Taken together, these factors point to a substantial increase in enforcement and regulatory activity that is consistent with increased regulatory costs being imposed on ASX listed companies. We date this change from January, 2000[7]. In Australia, management may release earnings forecasts as part of a routine information event such as the Chairman’s address at the Annual General Meeting or with the release of the half-yearly pro?t result. Many companies will issue management earnings forecasts as part of these routine information releases. Alternatively, management may release a non-routine earnings forecast at any time throughout the year (Galleryetal., 2002; Galleryetal., 2003). The continuous disclosure regime and mechanisms for its enforcement are arguably directed more towards Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)ensuring that management make appropriate non-routine disclosure to ensure that the PAR market is fully informed. Based on the foregoing discussion, we propose H1: 19,1 H1. Management earnings forecasts issued in the post-2000 period are more likely to be non-routine announcements than those issued in the pre-2000 period. The “litigation cost” hypothesis (Skinner, 1994) indicates that higher litigation costs 10 will prompt more voluntary “bad news” disclosures by management. We outlined above a series of factors that point to a substantial increase in enforcement and regulatory activity that is consistent with increased regulatory costs being imposed on ASX listed companies. Adapting Skinner’s arguments to the Australian continuous disclosure environment, we hypothesize H2: H2. Non-routine management earnings forecasts issued in the post-2000 period are more likely to contain disclosures of “bad news” rather than “good news” when compared to those issued in the pre-2000 period. Another dimension of management earnings forecasts is the speci?city of the forecast. Skinner (1994) divides management earnings forecasts into point (most speci?c), range, lower bound, upper bound, qualitative (least speci?c). Kasznik and Lev (1995) argue that the greater the divergence between managerial earnings expectations and market earnings expectations, the more speci?c, or “harder” the forecasts need to be to achieve this re-alignment. This suggests that the larger the discrepancy between management and market earnings expectations (as indicated by analyst’s earnings forecasts), the more speci?c the management earnings forecast needs to be (Kasznik and Lev, 1995, p. 122). However, where the management earnings forecast is good news, a speci?c (point) forecast may open management to loss of reputation or even litigation if the speci?c forecast is not attained. This suggests an asymmetric linkage wherein the relationship between the size of the earnings adjustment and forecast speci?city will be demonstrated only for bad news forecasts. Notably, Bamber and Cheon (1998) propose a different view. They argue that SEC rule 10b-5 (1), which makes it unlawful for corporations or their managers to issue misleading statements, decreases management incentives to issue speci?c disclosures. They argue that managers faced with greater exposure to legal liability (as indicated by poor earnings news relative to prior year earnings) or higher proprietary information costs, are likely to voluntarily issue earnings forecasts that are qualitative or in the form of minima or maxima and, hence, are less likely to be inaccurate. This is consistent with Skinner (1994) who ?nds that good news disclosures tend to be point or range estimates, while bad news forecasts tend to be qualitative. The USA literature has con?icting predictions and evidence about the relationship between forecast news content and forecast speci?city. In Australia, Coulton and Taylor (2003) investigate 240 stand-alone earnings forecasts issued by Australian listed companies[8]. Drawing on Bamber and Cheon (1998), they hypothesize and ?nd that Australian ?rms’ good news stand-alone earnings forecasts are signi?cantly more speci?c than their bad news earnings forecasts. Similar to Bamber and Cheon (1998), Coulton and Taylor (2003) use prior year earnings rather than analyst forecasts in determining whether management earnings forecasts convey good or bad news. In contrast, Gallery et al. (2002) ?nd that Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)Australian bad news earnings forecasts are more quantitative than good news Management earnings forecasts. earnings Further complicating the relationship between forecast news content and forecast forecasts speci?city is the well documented relationship between forecast horizon and forecast speci?city (Baginski et al., 1994; Coulton and Taylor, 2003, p. 10). As more of the ?nancial reporting period elapses and less time remains before the release of the annual earnings announcement, management become more con?dent, more sure of the 11 outcome. This greater certainty leads to more speci?c earnings forecasts. Hence any assessment of the relationship between forecast speci?city and good/bad news must control for the possible effects of differences in forecast horizon between good/bad news earnings forecasts. Given the con?icting nature of theory and prior results, we specify hypothesis H3 in null form: H3. After controlling for the effects of forecast horizon, there is no relationship between the speci?city of management earnings forecasts and their news content. 3. Research method 3.1 Data and sample Our sample comprises that subset of Australian listed companies which are followed by analysts, over the period September 1994 to December 2001. Our primary motivation for imposing the analyst coverage restriction is that a critical part of our analysis focuses on earnings news benchmarked against analysts’ forecasts. Analyst earnings forecast data are sourced from the Institutional Broker Estimates System (I/B/E/S) summary earnings ?le. Management earnings forecast data are sourced from the of?cial Signal G electronic records of all company announcements to the ASX. These data were accessed in two ways. First, for the period September 1994 to 31 December 1995 all Signal G announcements from the Securities Industry Research Centre for Asia Paci?c (SIRCA) for companies with analyst’s earnings forecasts available were manually read for any reference to earnings outlook, forecasts or upgrades. Second, for the period 1 January 1996 to 31 December 2001, the Signal G company announcements by all analyst-followed ASX companies on the I/B/E/S database were accessed on Integrated Real Time Equity System (IRESS). The following announcements were manually checked and read: . announcements surrounding the company’s AGM for the prior ?nancial year (especially the Chairman’s address and letters to shareholders, but not the preceding annual report); . announcements surrounding the release of the current half yearly earnings announcement; and . all announcements within three months prior to the current annual earnings announcement. In addition, the headline of all announcements made by analyst-followed companies between the AGM for the prior ?nancial year and three months prior to the current annual earnings announcement were checked and, if necessary, the full text of the Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)announcement was read. This thorough search for management forecast disclosures PAR identi?ed in excess of 2,400 such disclosures. 19,1 3.2 Variable de?nitions 3.2.1Managementearningsforecasts. Initially, we de?ned any Signal G announcement in which management referred to current period or future period pro?ts, revenues, 12 distributions/dividends or production as an earnings-related prior announcement. This process identi?ed 2,424 announcements. Subsequently, announcements that did not speci?cally refer to pro?t (totalling 781 in number) were jettisoned and not analysed further. This left a sample of 1,643 management earnings forecasts made by 464 different companies. All qualifying forecasts were included whether they were the ?rst or subsequent forecast for the period. To run the regression models reported later in the paper, we restricted the sample to those cases where: . forecast horizon (FHOR, see section 3.4.2).3 days; and . a valid news classi?cation (either bad, neutral or good) could be assigned from either or both the analysts and prior year earnings benchmarks or the observation related to a qualitative forecast. This resulted in a ?nal sample size of 1,472. Where multiple pro?t measures were referred to in a single announcement, we choose the measure closest to net pro?t after tax and included that in our analysis[9]. 3.2.2Forecastspeci?city. We follow Skinner (1994) in classifying management pro?t forecasts into point, range, maximum, minimum and qualitative categories. In a point forecast, a precise single numerical ?gure is given ($X) or can be readily quanti?ed (for example, pro?t is expected to equal that of the previous year). A range forecast contains a precise numeric range of pro?t (for example, “pro?t will be between $X and $Y”) or the range can be readily quanti?ed (for example, “pro?t will be within 10 per cent of last year”). A maximum forecast sets a maximum or upper bound to pro?t, whereas in a minimum forecast, a minimum or lower bound to pro?t is speci?ed. Qualitative forecasts provide a general statement that is not capable of any precise numeric interpretation. For the purposes of hypothesis testing, we regard maximum and minimum forecasts as being equally “speci?c” (or quantitative). Hence, for testing H3 we treated forecast speci?city (SPEC) as a variable with four ordered categories: qualitative (assigned a value ¼ 1), maximum/minimum (assigned a value ¼ 2), range (assigned a value ¼ 3), and point (assigned a value ¼ 4). As such, a higher value of SPEC indicates a greater degree of speci?city. 3.2.3 Classi?cation into good/neutral/bad news. Two methods of determining good/neutral/bad news were employed in the study; news benchmarked: (A) against analysts forecasts; and (B) against prior years earnings. 3.3 The analyst’s forecast method (A) As our sample included only analyst followed Australian listed companies, our preferred method was to compare the management earnings forecast to the mean analyst earnings forecast at the date of the announcement (Gallery et al., 2003). The following procedure was employed. Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)First, using I/B/E/S data, the number of shares on issue was obtained. This was Management multiplied by the mean analyst EPS forecast immediately prior to the management earnings forecast obtained from the I/B/E/S detail ?le to give the implied analyst net pro?t after forecasts tax – this is our analysts benchmark (AB). Next, the implied analyst net pro?t after tax was compared to the management earnings forecast. If the point earnings forecast or the mid-point of a range earnings forecast was more than 10 per cent above (below) the implied analyst net pro?t after tax (AB), this was classed as good (bad) news, while all 13 remaining valid cases are classed as neutral news. For management earnings forecasts in the form a maximum or minimum[10], we ?rst determine the most speci?c $ amount we could attach to the forecast (S and min S , respectively). S and S were then compared to the implied analyst net pro?t max min max after tax as follows: If S is greater than 10 per cent above (40 per cent below) the min implied analyst pro?t, it is classi?ed as good (bad) news, and otherwise it is deemed neutral news. If S is greater than 40 per cent above (10 per cent below) the implied max analyst pro?t, it is classi?ed as good (bad) news, and otherwise it is deemed neutral news.[11][12] For the hypothesis testing, the news variables based on analyst forecasts are a series of dummy variables labelled D ;D ; and D and are scored 1 when GoodA NeutA BadA the decision rules above are applied and zero otherwise. In some analysis (regarding H3) we also utilise an aggregate analysts’ news variable, D , which takes a value NewsA of 2 1 for bad news; a value of zero for neutral news and a value of 1 for good news. 3.4 The prior year earnings method (B) In this method good/bad news is assessed relative to prior year net pro?t after tax. The following procedure was employed. For point and mid-point of range forecasts the management forecast was compared to the previous year net pro?t after tax as follows. If the management earnings forecast was more than 10 per cent above (below) the previous year net pro?t after tax (NPAT ), this was classi?ed as good (bad) news, while all remaining cases are t21 classi?ed as neutral news. For minimum and maximum management earnings forecasts, S and S were min max determined as outlined above and compared to NPAT as follows. If S is greater t21 min than or equal to NPAT (20 per cent below NPAT ), it is classi?ed as good (bad) t21 t21 news, otherwise it is deemed neutral news. If S is greater than 20 per cent above max NPAT (equal to or below NPAT ), it is classi?ed as good (bad) news, and t21 t21 otherwise it is deemed neutral news[13]. For the hypothesis testing, the news variables based on the prior year earnings benchmark are a series of dummy variables labelled D ;D ; and D and are GoodP NeutP BadP scored 1 when the decision rules above are applied and zero otherwise. In some analysis (regardingH3) we also utilise an aggregate prior year earnings news variable, D , which takes a value of 2 1 for bad news; a value of zero for neutral news and a NewsP value of 1 for good news. 3.4.1 Routine/non-routine announcements. We classify management earnings forecasts into routine and non-routine forecasts (Galleryetal., 2002; Galleryetal., 2003). Routine announcements containing earnings forecasts are those made at or accompanying the Annual General Meeting (AGM) or the half-yearly pro?t announcement. All other announcements containing earnings forecasts are deemed Downloaded by VICTORIA UNIVERSITY (Australia) At 17:07 12 July 2017 (PT)"