How does investor relations disclosure affect analysts' forecasts?
Abstract
We study the relationship between investor relations disclosure and analyst forecast properties in Australian firms, a setting dominated by small firms with limited analyst coverage and requiring continuous disclosure of price sensitive information. We find increasing disclosure in the time period investigated is associated with greater accuracy in firms disclosing fewer items. Disclosure was unrelated to forecast dispersion, possibly due to the low analyst following. In periods of uncertainty, the investor relations awards effectively discriminated quality from quantity of disclosure. These findings highlight the importance of active communication with analysts, particularly in firms providing less disclosure and during periods of uncertainty.
1. Introduction
Factors influencing corporate disclosure and the use of such disclosed information are closely related processes. In this article, we focus on one such interrelationship, the relationship between web-based corporate disclosures and analyst forecast properties. According to Alford and Berger (1999) and Hope (2003), the extent and quality of a firm's disclosure, analyst following and the properties of its analysts’ forecasts are largely simultaneously determined. These simultaneous relationships suggest the presence of complex information dynamics at the firm level where the level and quality of analyst coverage and the firm's disclosure policy affect each other (Aerts et al., 2007). Information asymmetry between the managers of the firm and external parties such as outside investors, analysts and other intermediaries creates a demand for information. Such information consists of both financial and nonfinancial disclosure where credible, relevant and timely disclosures reduce agency costs between the various internal and external parties. Financial economists agree that the desirable benefits of a rich disclosure environment and low information asymmetry are efficient resource allocation, increased liquidity, lower volatility, higher analyst forecast accuracy and overall reduction in the cost of capital (Kothari et al., 2009).
The disclosure channel we investigate is investor relations which highlights effective communication between the firm and the investment community with its two main objectives being to reduce information asymmetry and to increase the firm's visibility (Marston, 1996; Brennan and Tamarowski, 2000; Bushee and Miller, 2007). The Australasian Investor Relations Association (AIRA) maintains that ‘the need for a focused investor relations association is now greater than ever with the growth in investment markets, greater institutional and retail shareholder activism and an increased focus on corporate disclosure by regulators’ (Australasian Investor Relations Association, 2005).
In particular, we study the relationship between investor relations disclosure and two aspects of analyst forecasts: forecast error and dispersion. The effect of investor relations on analyst forecast error and forecast dispersion was initially examined by Farragher et al. (1994) where they reported no association between investor relations and forecast error and a negative relationship with dispersion. Lang and Lundholm (1996) investigated the relationship between firm disclosure practices, analyst following and analyst forecasts. Their measure of disclosure practices was the 1985–1989 FAF report (Report of the Financial Analysts Federation Corporate Information Committee) where analysts evaluate firm disclosure via annual published information, other published information and investor relations. They reported that disclosure provided by firms determines both analyst following and forecast characteristics, and analyst following increases with disclosure. Aerts et al. (2007) replicated Lang and Lundholm (1996) with web-based corporate disclosure and compared differences between North American and European firms in the disclosure and analyst relationship. While there was support for Lang and Lundholm's findings with their North American sample, the relationship was weaker in Europe.
Using the terminology in Kothari et al. (2009), we investigate disclosure quantity rather than content, and this is measured by disclosures through investor relations on corporate websites. We utilise analyst forecasts because financial analysts are an important link in the capital market, providing earnings forecasts, recommendations and other information. Leuz (2003) argued that analyst forecasts may proxy for managerial monitoring quality while Chang et al. (2000) associated disclosure transparency with accurate analyst forecasts and vigorous analyst activity. Gelb and Zarowin (2002) suggested that analyst forecasts mirror how firms manage their relationship with analysts. When disclosure policies are more informative, firms have a larger analyst following, more accurate forecasts, less dispersion and also fewer revisions (Lang and Lundholm, 1996). Therefore, analyst forecasts are affected by a firm's disclosure policy and analysts, as a group respond to a firm's disclosure.1
With a sample of 22 countries, Hope (2003) showed a positive relationship between firm level annual report disclosures and analyst forecast accuracy where strong enforcement of accounting rules contributed to forecast accuracy. Aerts et al. (2007) highlighted differences between corporate web disclosure and analyst behaviour in North American firms compared to those in Continental Europe. They indicated that in North American firms, analyst following directs corporate performance disclosure resulting in a reduction in forecast dispersion although such a relationship does not exist in Continental Europe.
This study contributes to the literature in several ways. First, we investigate the Australian market where investor relations is at its relative infancy, compared to other developed markets such as the United States and UK. Investor relations has also become more important with the current uncertainty in the global financial markets and the growing strength of retail and institutional shareholder activism in Australia. While this article updates previous studies conducted in the United States and Europe (Lang and Lundholm, 1996 and Aerts et al., 2007), it remains important to investigate this investor relations–analyst relationship for Australia because Aerts et al. (2007) showed that while, in the United States, analysts are a powerful group who direct corporate performance disclosure, they did not display such behaviour in Continental Europe. Information dynamics in capital markets are influenced by political and institutional settings, and investor relation disclosure may have a differential impact on analyst coverage and following, thus affecting their forecast properties. The Australian institutional setting differs from the United States and Europe in two significant ways. Regulation Fair Disclosure (Reg FD) in the United States has prevented privileged disclosure between firms and analysts, mandating for fair and accessible corporate disclosure to all analysts and investors. In the absence of special relationships between firms and analysts, the overall power of analysts to direct corporate disclosure increased. While Reg FD applies only to US firms, Australian firms are also discouraged from having special relationships with analysts, and in addition, they operate under the Continuous Disclosure Regime that requires the reporting of price sensitive information when the information becomes known to the firm. While this could suggest that investor relations may become less useful, Brown et al. (1999) found increased disclosure under this regime only in small and poorer performing firms. Another institutional difference between US and Australian firms compared to European firms is the strong use of bank financing in the latter where the power of influence is with the firm's creditors.
The Australian market also differs from the US and European markets because it includes many small, less-recognised firms with low analyst coverage. As of 27 July 2012, the ASX 200 market capitalisation was A$1012.24 billion, while the figures for the ASX 50 and ASX 100 were A$853.12 billion and A$950.60 billion, respectively. Even though the ASX 200 represents the investable index containing the largest and most liquid firms, there is still a prevalence of small illiquid firms. One of the aims of this study is to determine where Australian analysts lie in the power continuum with firms as discussed in Aerts et al. (2007). The reasons mentioned above lead us to the expectation that Australian analysts are more like their US counterparts, though perhaps not as persuasive.
One of the prevailing difficulties faced in the disclosure literature is disclosure measurement, particularly items believed to be disclosed voluntarily by firms. We further contribute to the disclosure literature by augmenting the quantity measure based on a checklist of items reported in the investor relations section of a firm's corporate website with a quality measure. This measure is AIRA's annual investor relations awards that celebrate the quality of a firm's investor relations. These awards are determined by domestic and internal equity analysts and fund managers and provide an alternative measure of disclosure quality. It can also be seen as an external validation of our investor relations checklist. Agarwal et al. (2008) and Peasnell et al. (2011) also used investor relations awards to proxy for quality. Agarwal et al. (2008) reported superior abnormal returns before and after the award nominations. On the other hand, Peasnell et al. (2011) found that US firms with established investor relations reputations could not avoid a perceived decline in management credibility and reporting integrity during a period of corporate scandals.
Our third contribution is an examination of analyst behaviour during periods of uncertainty for firms with an established record of investor relations disclosure. Ang and Ma (2001) examined how analysts behaved during the Asian crisis and reported that while they did not panic, they failed to adjust their forecasts to deteriorating market conditions. In this article, we investigate whether analysts behaved differently during the global financial crisis for a group of firms with which they had an established relationship of disclosure. With an established relationship, these analysts are predicted to be more certain about future firm prospects and therefore not suffer the problems referred to in Ang and Ma (2001).
The corporate disclosure in focus is web-based investor relations disclosure because this channel of communication is known to be dynamic, cost-effective and timely (Hedlin, 1999; Cormier and Magnan, 2004; Ettredge et al., 2005). It also has the ability to reach a wide spectrum of users of varying financial sophistication and is not subject to the same regulatory constraints as other disclosures such as annual reports and earnings announcements (Cormier and Magnan, 2004). Without a doubt, corporate reporting on the Internet has burgeoned over the last decade. Ashbaugh et al. (1999) reported that 70 per cent of their sample was involved in Internet reporting, and there was considerable difference in the quality of practice. In a survey of 270 companies across six countries including Australia, Belgium, France, the Netherlands, South Africa and the UK, Bollen et al. (2006) found that the extent of investor relations activities on the Internet was a function of firm size, internalisation, proportion of shares held by individual investors and the disclosure environment. The importance of investor relations in small and mid-cap firms was highlighted by Bushee and Miller (2007) where the hiring of an external investor relations firm led to improved visibility in terms of increases in disclosure, media coverage and also analyst following. Corporate reporting via the Internet in Australia is continually developing, with Lodhia et al. (2004) reporting that in 2001, the top 50 Australian companies had not fully utilised the potential of the Internet to disclose information to stakeholders.
Voluntary disclosure is disclosure in excess of legal requirements and represents the management's discretion to provide accounting or other information which are believed to be relevant to the needs of investors in decision making (Meek et al., 1995). Previous studies have indicated that a firm's disclosure policy is perhaps the most significant aspect of investor relations management. Some incentives for voluntary disclosure via investor relations include improving liquidity (Brennan and Tamarowski, 2000), increasing analysts following (Lang and Lundholm, 1996) and reducing cost of capital (Brennan and Tamarowski, 2000). Others have shown that investor relations has a positive impact on share prices, overcomes adverse selection (Lang and Lundholm, 1993) and increases firm visibility (Bushee and Miller, 2007).
This article investigates whether investor relations disclosure affects the information asymmetry between the firm and its analysts because a firm's disclosure is an input into analyst forecasts. Two properties of forecasts that proxy for information asymmetry are examined, forecast accuracy and dispersion. Specifically, disclosure is predicted to affect the information set available to analysts and the way analysts interpret such information as the investor relations process controls and manages information flows. Several factors influence management's readiness to release information, and more disclosure is associated with lower cost of capital as a result of less risk, increased liquidity and higher institutional ownership (Diamond and Verrecchia, 1991). While incentives for releasing good news are apparent (i.e. an increase in share price), benefits also exist for withholding bad news. In addition to a decrease in share price, management may be trying to buy time to reverse poor performance. Even when disclosure is made, the market may not correctly interpret the information disclosed because of imperfect information, noise in the communication channel, the firm's disclosure credibility and disinterest in the information (Ang and Ciccone, 2001). Given these issues, we expect investor relations disclosure to improve the prevailing information set, hence increasing forecast accuracy and reducing forecast dispersion.
The results show increased investor relations disclosures over the time period covered in this article, signifying the Internet as an emerging communication channel within the Australian financial community. In contrast to prior studies, the general results showed no relationship between disclosure and forecast error or between disclosure and dispersion.2
In further analysis, we find for firms disclosing relatively fewer items, there was a negative relationship between disclosure and forecast error. In these firms, the AIRA awards played an additional role in identifying investor relations quality. As highlighted in Bushee and Miller (2007) and Hong and Huang (2005), investor relations is more important in smaller and less-recognised firms and for firms in our sample that disclosed fewer items.
The remainder of the article is structured as follows. Section 2. reviews the literature, and Section 3. describes the sample data and the research method employed. Section 4. presents the results, while Section 5. summarises the key findings.
2. Literature review
Corporate disclosure is crucial for efficient capital markets to function well and to allocate resources effectively (Healy and Palepu, 2001). Today's stock price reflects future cash flows expected by investors, so it is important to satisfy the demand for information from investors. It is also crucial for investors to be informed of a firm's future prospects on a timely basis so that investment decisions can be made. This suggests the importance of voluntary and continuous disclosure in providing information which is related to changes in market expectations.
A well-organised investor relations process increases the quality of voluntary disclosure, avoiding the information ‘lemons’ problem and agency costs resulting from information asymmetry between the firm and its investors. Hence, the negative impact on the value of the company is reduced (Healy and Palepu, 2001). Gelb and Zarowin (2002) show that enhancement in disclosure results in increased future earnings changes. This suggests that enhanced disclosure quality reduces the incentives of investors to search for private information, and therefore, there is less private information in the capital market. Bushee and Miller (2007) and Brennan and Tamarowski (2000) illustrate that an increase in the firm's visibility or a decrease in the private information search incentives through disclosure activities will positively change the perception of uninformed investors.
Traditionally, investor relations functions include activities such as road shows, conference calls, quarterly and annual results announcement, analysts briefing, annual general meeting, site visits and one-to-one meetings. One of the earliest studies on investor relations was Marston (1996) who looked at the organisation of this function in large UK public companies. She defined investor relations as the link between a company and the financial community providing information to help the financial community and the investing public evaluate a company. Investor relations is a multidisciplinary management function that is becoming increasingly important, especially in the area of disclosure theory (Armitage and Marston, 2008).
Investor relations has also been well-recognised as a major player in determining corporate image, and it has become part of a coordinated marketing communication strategy (Dolphin, 2004). Rao and Sivakumar (1999) view investor relations departments in the Fortune 500 Industrials as a boundary-spanning structure. This is because it sends information to the capital market's environment, presenting the firm in a favourable light. It also signals commitment to institutionalised beliefs and values that represent the organisation to valued constituencies. According to Tuominen (1997), disclosed information is the most important instrument in managing investor relationships because short-term episodes form the basis of long-term investor relations.
Marston and Straker (2001) suggested that the development of equity markets has urged management to improve its relationship with the investment public due to the need for equity funding from investors. Bushee and Miller (2007) found that an initiation of the investor relations programme in small firms significantly improved its management access, visibility and analyst following. The other benefit was increased liquidity in the capital market, mainly for small and new firms (Hong and Huang, 2005).
Agarwal et al. (2008) investigated the impact of effective investor relations on the market value of firms by using firms nominated for Annual US Investor Relations Magazine – Investor Relations Awards. They found that firms winning awards were associated with positive abnormal returns. Although it might partly be due to their past performances, the effect of investor relations still produced positive abnormal returns. Another benefit from effective investor relations is reduction in cost of capital as Armitage and Marston (2008) found that greater disclosure to rating agencies and lenders reduced the cost of debt.
Hedlin (1999) suggested that the globalisation of financial markets has hastened the integration of the Internet into investor relations because it increased the coverage of investors and provided information on a timely basis. Companies also incurred lower costs by providing financial information on the corporate website (Ettredge et al., 2001), which provides equal access to information for all market participants (Hedlin, 1999). It also achieves the objective of investor relations, which is to provide information to help investors make capital allocation decisions.
Lang and Lundholm (1996) showed that company-provided disclosures are a significant determinant of analyst behaviour in predicting forecasts, depending on the credibility and the accuracy of the company's disclosure policy. Williams (1996) found that the accuracy of prior earnings forecast by management serves as an indicator to analysts of the credibility of a current management forecast. Therefore, this suggests that if management establishes a forecasting ‘reputation’ based on prior earnings forecasts and maintains good investor relations disclosure, it is likely that analysts will view the company as a primary source of information. In this case, the analysts’ forecasts will be more homogeneous, and there will be less variation among analysts.
Farragher et al. (1994) used analysts’ forecast properties to proxy for the information risk investors face, finding an inverse relationship between the quality of the corporate ratings and the dispersion of analyst forecasts, but not the accuracy of the forecasts. This is consistent with Bowen et al. (2002) that conference calls are associated with more accurate and less dispersed analysts’ forecasts. In a study of the behaviour of analysts around market crashes in Indonesia, Korea, Malaysia and Thailand during the Asian financial crisis, Ang and Ma (2001) reported that they did not anticipate issues with the firms they followed and failed to adjust their forecasts after the markets crashed. On the other hand, there was also no evidence of panic or herding among analysts. Given the discussions above, we predict that firms with more disclosure via investor relations will have more accurate and less dispersed analyst forecasts.
3. Data and research design
3.1. Sample selection
Our sample consists of 291 and 293 publicly listed firms in the S&P/ASX300 index for two sample years, 2005 and 2009.3
Analysis was conducted on these 2 years because of the availability of the hand collected IR disclosure data. Firm data were obtained from the Aspect FinAnalysis database and analyst forecasts from the Institutional Brokers’ Estimation System (I/B/E/S) database through the Wharton Research Data Service (WRDS) platform. Firms followed by fewer than two analysts were excluded from the sample, resulting in the final sample of 156 and 214 firms in 2005 and 2009, respectively.
3.2. Investor relations disclosure
With the items in the checklist shown in Table 1, we capture the efforts of Australian firms to enhance their visibility and recognition through the investor relations channel on the Internet by investigating the content and structure of the firm website. The emphasis is on voluntary disclosure to current and potential investors on the corporate performance of the firm, from the viewpoint of management and analysts. While the focus remains on such voluntary disclosure, the checklist also takes in account the mandatory disclosure items such as earnings and annual reports which are placed in the investor relations section. Therefore, structure is an important aspect as it allows investors to access relevant information with ease in a ‘one stop shop’.
| Items | Weight |
|---|---|
| Dedicated investor relations/investor recognition area on website | 4 |
|
Report the following information on the investor relations/investor recognition area on website: a. Annual reports b. Earnings releases c. Investments d. Share price information and analysis e. Financial history f. Speeches and events g. Press releases |
All 1 |
| Disclosure of analysts’ covering the firm on website | 2 |
| Provision of earnings/profit guidance to the market on website | 3 |
| Webcast (via the company website) analyst presentations, offering either audio or video (or both) | 4 |
| a. Open to the general investor public | 1 |
| Offer a conference call facility | 3 |
| a. Conference call facility open to the general public | 1 |
| Questions and answers from analysts’ presentations available on website? | 2 |
| Publish information relating to current and future projects on the website | 1 |
| Posting information such as Chairman's address given at the AGM on website | 1 |
| Posting information about the next AGM on website | 1 |
| Posting CEO presentations as a power point presentation on website | 2 |
| Broadcast CEO presentations on the corporate website, offering either audio or video (or both) | 2 |
| Make available letters to shareholders (for instance detailing dividend distributions, acquisitions, new directors) on website | 1 |
| Posting ASIC audited financial reports and financial accounts on website | 1 |
| Posting ASX announcements on website, in the same form as it would be seen on the ASX website | 1 |
|
Posting presentations on: a. Key business principles b. Financial strength c. Timelines and milestones of the company d. Market opportunities e. Company targets (financial and nonfinancial) f. Significant competitors g. Potential projects and opportunities |
All 1 except market opportunities and potential projects and opportunities, 1.5 |
|
Report information relating to: a. The substantial shareholders b. The top 20 fully paid shares c. The top 20 convertible notes |
All 1 |
| Posting information relating to the appointment of and dismissal of company directors’ on website | 1 |
| Reporting sale of projects or other business interests on website | 1 |
| Reporting changes in director interest on website | 1 |
| Posting analyst reports in power point format (or another format) on website | 3 |
| Posting presentations targeted at attracting potential shareholders on website | 2 |
| Making email details available on website where investors (either existing or potential) can request further information | 1 |
| Archiving financial reports of at least 2 years old and making them available on website | 1 |
| Have a ‘frequently asked questions’ (FAQ) hyperlink relating to IR on website | 1 |
| Posting a financial calendar detailing future milestones, corporate prospects, upcoming meetings, speeches and other events on website | 1 |
| Posting a chart illustrating the corporate organisational structure on website | 1 |
| Posting information related to corporate directors offering a brief synopsis of employment history, qualifications and other general information on website | 1 |
Prior research has defined explicit expectations about the content and structure of a firm's investor relations website (Geerings et al., 2003). Items from Geerings et al. (2003) and Argenti (2005) were included in the investor relations checklist. Bushee and Miller (2007) also suggested questions directed at companies’ disclosing the names and recommendations of analysts to be a differentiating factor of a company's investor relations policy. In addition to questions on earnings guidance, CEO presentations, CEO broadcasts and analyst presentations also distinguish well-developed investor relations policies from others. Brown and Hillegeist (2007) claimed that firms can improve their investor relations through the use of conference calls, live webcasts of CEO and analyst presentations as well as accommodating a frequently asked questions (FAQ) area on the firm's website. These assist the firm to enhance investor perceptions as well as minimise disagreement on its internal policies. Marston and Straker (2001) identified items that focused on stock exchange announcements, current and future projects, corporate presentations to shareholders and questions and answers from analyst presentations as offering investors a subjective view of the firm's investor relations policy as they allow the investing public to make informed decisions regarding the viability of the company. Other items that were included in the checklist relate to historic and current annual and interim reports, the appointment and dismissal of corporate directors and information relating to historic and current annual general meetings.
The 44 items in the checklist are a fair representation of a firm's commitment to investor relations with specific attention paid to analyst coverage and presentations, CEO webcasts and presentations to potential and current investors. These same items allow a distinction between firms that invest significantly in investor relations from others who place less importance on this channel. The use of the Internet also contributes to the novel approach we have employed in analysing a firm's investor relations policy. Argenti (2005) suggested that firms can enhance investor relations through the use of tactical relationships with existing and prospective clients where the use of the Internet can enhance the efficiency of information disclosures to the investment public and can lead to a diversified shareholder base.
Different weightings were assigned to items expected to distinguish the investor relations commitment between firms. Items were assigned higher weightings when they were anticipated to offer an insight into the structure of the corporate website and the advancement of the firm's policy. One such item, whether the firm had a dedicated investor relations section on the website received a weighting of four, the highest assigned to any item on the checklist. Navigation and structure were suggested by Geerings et al. (2003) to be an important aspect of the website because it reflects the firm's internal organisation and allows the investment public to easily access the relevant information in a timely manner. Items relating to whether the firm provides earnings guidance, conference call facilities and analyst reports received a weighting of three because they allow investors access to unbiased and impartial views regarding the future prosperity of the firms. Further, five items relating to the availability of the name of analysts following the firm, questions and answers from analyst presentations, CEO presentations and broadcast of these presentation and presentations attracting potential shareholders are assigned a weighting of two as these features differentiate established IR practices from less developed ones. The items receiving weightings of two and three reflect the firm's commitment to providing shareholders with information on firm prospects, both from the management's and external analysts’ (independent) perspectives.
A maximum weighted score of 62 can be awarded to any firm in the sample based on the information posted on the website.4
This disclosure score was used in Chang et al. (2008) who reported that firms with higher scores also had more analyst following, higher institutional holdings and higher market capitalisations and trading volume.
The disclosure checklist provides a measure of the quantity of disclosure. In addition, a measure of investor relations quality is proxied by the annual awards conferred by the Australian Investor Relations Association (AIRA) to firms that have excelled in their investor communications with the investment market over the previous 12 months. These award categories that take firm size into account are as follows: Overall Best IR by a company in the S&P/ASX 200 and Best IR by a company in the ASX 201–300.
3.3. Properties of analysts’ forecasts
A firm with a following of less than two analysts in any month is removed from the sample. We focus on the effect of disclosure generally and do not expect the relationship between disclosure and analyst forecasts to differ at different times of the year. Therefore, we examine both the forecast accuracy and dispersion during the 12 months leading up to the firm's financial year end. The measures, AFE and FD, used in the regression models are computed as the simple average of the measure across the 12 monthly reporting periods on the I/B/E/S tape during the firm's financial year.
(1)
(2)4. Results
4.1. Descriptive statistics
Table 2 Panel A provides the descriptive statistics for the disclosure scores and analyst forecasts for the 2 years examined. There is a significant increase in the mean scores, from 31.26 in 2005 to 40.80 in 2009.5
| n | Mean | Median | Max | Min | SD | 25th | 75th | |
|---|---|---|---|---|---|---|---|---|
| Panel A: All firm-year observations | ||||||||
| DISC | ||||||||
| 2005 | 156 | 31.26 | 31.00 | 54.00 | 6.00 | 7.61 | 27.00 | 37.00 |
| 2009 | 214 | 40.80 | 41.00 | 57.00 | 19.00 | 7.28 | 36.00 | 46.00 |
| t-test/U-test | −12.22*** | −10.70*** | ||||||
| AFE | ||||||||
| 2005 | 1500 | 1.202 | 0.549 | 8.736 | 0.000 | 1.700 | 0.178 | 1.453 |
| 2009 | 2060 | 5.525 | 1.669 | 58.554 | 0.000 | 10.779 | 0.500 | 4.988 |
| t-test/U-test | −17.90*** | −19.31*** | ||||||
| FD | ||||||||
| 2005 | 1500 | 0.537 | 0.368 | 2.978 | 0.000 | 0.537 | 0.204 | 0.656 |
| 2009 | 2060 | 2.313 | 0.976 | 25.483 | 0.000 | 3.911 | 0.476 | 2.204 |
| t-test/U-test | −20.34*** | −27.10*** | ||||||
| Panel B: Common sample | ||||||||
| DISC | ||||||||
| 2005 | 101 | 32.25 | 32.00 | 54.00 | 6.00 | 7.92 | 28.00 | 37.00 |
| 2009 | 101 | 42.55 | 44.00 | 57.00 | 19.00 | 7.49 | 37.00 | 47.00 |
| t-test/U-test | −30.05*** | −26.28*** | ||||||
| AFE | ||||||||
| 2005 | 1006 | 0.964 | 0.429 | 8.736 | 0.000 | 1.429 | 0.146 | 1.074 |
| 2009 | 1030 | 2.987 | 1.167 | 58.554 | 0.000 | 5.851 | 0.368 | 2.927 |
| t-test/U-test | −10.78*** | −13.50*** | ||||||
| FD | ||||||||
| 2005 | 1006 | 0.471 | 0.325 | 2.978 | 0.000 | 0.438 | 0.188 | 0.585 |
| 2009 | 1030 | 1.611 | 0.774 | 25.483 | 0.000 | 2.840 | 0.400 | 1.498 |
| t-test/U-test | −12.73*** | −19.80*** | ||||||
- This table presents the descriptive statistics of the IR disclosure scores (DISC) and analysts’ forecasts properties based on companies listed in S&P/ASX 300 in the sample years 2005 and 2009. Panel A presents the statistics for all firm-year observations. Panel B presents the statistics for the firms that were present in the sample in both 2005 and 2009. DISC is the IR disclosure scores, AFE is 100 × |(EPSt − median forecastt)|/ Share pricet, FD is 100 × standard deviation of analysts’ forecasts /Share pricet. *** denotes significance at the 1 per cent level.
This reflects the increased commitment to and investment in investor relations for firms in the S&P/ASX 300 index. The minimum scores (6 in 2005 and 19 in 2009) also indicate improving disclosure for these firms with the maximum (54 in 2005 and 57 in 2009) remaining stable over the 2 years.
Absolute forecast error (AFE) increased significantly from 1.202 per cent to 5.525 per cent. The standard deviation was also higher in 2009 at 58.554 compared to 8.736 in 2005, indicative of uncertainty with future earnings across the sample firms resulting from the global financial crisis (GFC). The uncertainty surrounding the GFC was also reflected in within-firm forecast dispersion (FD) where the mean dispersion was 0.537 per cent in 2005 and increased to 2.313 per cent in 2009. The 2009 forecast error and dispersion are consistent with Sidhu and Tan (2011) who also reported larger forecast error and dispersion during the GFC and Hope and Kang's (2005) finding of a reduction in forecast accuracy in periods of macroeconomic uncertainty. Viewing the disclosure score together with the forecast error and dispersion across the 2 years suggests a positive relation that is counter-intuitive. In 2009, the increase in disclosure was accompanied by increases in both forecast error and dispersion, possibly a result of the GFC. It appears that the expected negative relationship between disclosure and forecast error and dispersion does not hold. We examine and discuss these relationships further below.
Due to the different sample composition in 2005 and 2009, we provide the descriptive statistics for firms belonging to both year subsamples.6
Table 2 Panel B shows these 101 firms, on average, have higher disclosure scores but lower forecast error and dispersion when compared to the main sample. Similar to the main sample, the mean disclosure scores, forecast error and dispersion are significantly higher in 2009 compared to 2005.
Table 3 presents the results of the correlation analysis. Investor relations disclosure (DISC) is negatively correlated with AFE and FD in 2009 and positively (albeit weakly) correlated with FD in 2005. The negative relationship in 2009 is consistent with expectations and previous findings and highlights the importance of investor relations during periods of uncertainty (Lang and Lundholm, 1993; Aerts et al., 2007). In addition, we find investor relations disclosure (DISC) is positively associated with firm size (LNSIZE) and analyst following (ANALYST). DISC is also positively correlated with earnings surprise (SURPRISE) and the proportion of forecasts made for the first time or revised (NEWFOR), suggesting firms with earnings that suffer from greater variability are more likely to have better disclosure quality. The negative correlation between the DISC and the historical correlation between annual returns and earnings (CORR) suggests firms are likely to make more disclosures if it is more difficult for the market to determine the relationship between the firm's earnings and share price returns. Most of the coefficients between the control variables are in the expected direction with the highest correlation being between LNSIZE and ANALYST.
| AFE | FD | DISC | LN (SIZE) | ROE_STD | ANALYST | SURPRISE | LEVERAGE | NEWFOR | |
|---|---|---|---|---|---|---|---|---|---|
| Panel A: 2005 | |||||||||
| FD | 0.593*** | ||||||||
| DISC | −0.020 | 0.049* | |||||||
| LNSIZE | −0.308*** | −0.289*** | 0.427*** | ||||||
| ROE_STD | 0.261*** | 0.254*** | −0.156*** | −0.248*** | |||||
| ANALYST | −0.351*** | −0.309*** | 0.303*** | 0.753*** | −0.276*** | ||||
| SURPRISE | 0.348*** | 0.358*** | 0.064** | −0.278*** | 0.120*** | −0.242*** | |||
| LEV | −0.065** | −0.136*** | 0.057** | 0.093*** | 0.010 | 0.113*** | 0.022 | ||
| NEWFOR | 0.091*** | 0.158*** | 0.170*** | 0.134*** | −0.007 | 0.035 | 0.109*** | −0.043* | |
| CORR | 0.186*** | 0.245*** | −0.035 | −0.261*** | 0.079*** | −0.255*** | 0.186*** | −0.080*** | 0.043* |
| Panel B: 2009 | |||||||||
| FD | 0.661*** | ||||||||
| DISC | −0.039* | −0.040* | |||||||
| LNSIZE | −0.223*** | −0.228*** | 0.475*** | ||||||
| ROE_STD | 0.123*** | 0.103*** | −0.079*** | −0.212*** | |||||
| ANALYST | −0.378*** | −0.337*** | 0.201*** | 0.565*** | −0.297*** | ||||
| SURPRISE | 0.326*** | 0.392*** | 0.013 | −0.141*** | −0.014 | −0.167*** | |||
| LEV | 0.148*** | 0.095*** | 0.023 | −0.011 | 0.040* | 0.013 | −0.010 | ||
| NEWFOR | 0.078*** | 0.092*** | 0.069*** | 0.082*** | 0.016 | −0.056** | 0.012 | −0.038* | |
| CORR | 0.105*** | 0.135*** | 0.065*** | −0.066*** | 0.131*** | −0.193*** | 0.088*** | 0.011 | 0.033 |
- This table presents the Spearman's correlation matrices among the independent variables, dependent variables and control variables. The coefficient of the Spearman's correlation is tested by using a one-tailed parametric t-test of the entire sample. DISC is the IR disclosure scores collected from the checklist, LNSIZE is the natural log transformed market capitalisation of the firm; ROE_STD is volatility of the firm's return on equity; ANALYST is the number of analysts following; SURPRISE is the earning surprise measured by |(EPSt − EPSt−1)|/ share pricet; NEWFOR is the proportion of forecast that is made for the first time or revised averaged over the 12 months prior to the financial year end; CORR is the historical correlation between annual returns and earnings over the preceding 10 years; LEV is the firm leverage. ***, **, * denote significance at the 1 per cent, 5 per cent and 10 per cent levels, respectively.
To further examine the influence of DISC during the different time periods, we analyse the effect of IR on the change in AFE and FD (between 2005 and 2009) by conducting additional tests on the sample of firms common across the two periods (see Table 2 Panel B). In untabulated results, we find both changes in AFE and FD are negatively associated with the DISC score in 2005 (−0.26 and −0.18, respectively). That is, firms with better disclosure are likely to see a smaller increase in their forecast error and dispersion during periods of uncertainty.7
This finding supports the effectiveness of establishing an investor relations reputation, particularly in times of market uncertainty.
4.2. AIRA award recipients versus nonrecipients
Table 3 compares the disclosure scores between AIRA award recipients and nonrecipients to determine the validity of the disclosure score. The mean score for award recipients is 42.83, while for nonrecipient, the mean is 36.25. This difference is significant at p <0.01 for both the t-test and the Mann–Whitney U-test. The difference in the disclosure score between the award recipients and nonrecipients provides assurance to the validity of the constructed disclosure score as award recipients are expected to disclose more investor relations items than nonrecipients. We also compare the firm characteristics of AIRA award recipients and nonrecipients. As expected, award recipients are larger in market capitalisation terms and have more analyst following. However, they have similar volatility of the firm's return on equity, earnings surprise, firm leverage and return–earnings relationship. The analysis highlights the need to control for the effects of size and analyst following when testing the effects of investor relations disclosure (Table 4).
| Nonaward recipients | Award recipients | t-test | Mann–Whitney U-test | |||
|---|---|---|---|---|---|---|
| Mean | Median | Mean | Median | |||
| DISC | 36.25 | 37.00 | 42.83 | 44.00 | −4.020*** | −3.692*** |
| SIZE ($ million) | 3322 | 927 | 17 400 | 2777 | −4.740*** | −2.782*** |
| ROE_STD | 0.194 | 0.070 | 0.096 | 0.070 | 1.360 | 1.199 |
| ANALYST | 6.859 | 7.000 | 9.333 | 8.500 | −3.440*** | −3.003*** |
| SURPRISE | 0.121 | 0.030 | 0.148 | 0.030 | −0.530 | −0.046 |
| LEV | 2.194 | 1.910 | 2.250 | 1.940 | −0.190 | −0.378 |
| NEWFOR | 0.338 | 0.250 | 0.386 | 0.350 | −0.710 | −1.078 |
| CORR | 0.021 | 0.020 | −0.022 | 0.050 | 0.450 | 0.516 |
- This table examines the difference in the IR disclosure scores (DISC) and firm characteristics between the Australian Investor Relations Association (AIRA) award recipients (N = 30) and the nonaward recipients (N = 340). See Table 3, for variable definitions. ***, **, * denote significance at the 1 per cent, 5 per cent and 10 per cent levels, respectively.
4.3. Multiple regression results
Regression analysis of the effect of investor relations disclosure on analyst forecast error was conducted and the results presented in Panel A of Table 5. Model 1 includes the quantity measure (DISC) only; Model 2 includes the quality measure (AWARD) only, and Model 3 includes both measures.
| Expected sign | Model 1 | Model 2 | Model 3 | ||||
|---|---|---|---|---|---|---|---|
| Coeff. | t-stat. | Coeff. | t-stat. | Coeff. | t-stat. | ||
| Panel A: AFE as dependent variable | |||||||
| DISC | − | −0.020 | −0.506 | −0.019 | −0.462 | ||
| AWARD | − | −0.419 | −0.417 | −0.367 | −0.362 | ||
| FD | + | 2.123 | 16.333*** | 2.121 | 16.312*** | 2.122 | 16.300*** |
| LNSIZE | − | 0.133 | 0.476 | 0.100 | 0.378 | 0.142 | 0.506 |
| ROE_STD | + | 0.331 | 0.439 | 0.316 | 0.419 | 0.321 | 0.425 |
| ANALYST | − | −0.221 | −1.985** | −0.220 | −1.980** | −0.221 | −1.981** |
| SURPRISE | + | −0.768 | −0.639 | −0.776 | −0.646 | −0.749 | −0.622 |
| LEV | + | 0.429 | 2.468*** | 0.426 | 2.456*** | 0.429 | 2.469*** |
| NEWFOR | + | 2.396 | 1.684** | 2.287 | 1.619* | 2.373 | 1.664** |
| CORR | − | −0.502 | −0.869 | −0.512 | −0.888 | −0.495 | −0.856 |
| NEG | + | 2.057 | 2.829*** | 2.030 | 2.787*** | 2.040 | 2.796*** |
| Constant | −2.555 | −0.513 | −2.528 | −0.507 | −2.758 | −0.550 | |
| Year fixed effects | Yes | Yes | Yes | ||||
| N | 370 | 370 | 370 | ||||
| Adjusted R2 | 0.627 | 0.627 | 0.626 | ||||
| F statistic (p-value) | 57.419 (0.000) | 57.398 (0.000) | 52.517 (0.000) | ||||
| Model 4 | Model 5 | Model 6 | |||||
| Panel B: FD as dependent variable | |||||||
| DISC | − | 0.005 | 0.327 | 0.006 | 0.373 | ||
| AWARD | − | −0.163 | −0.398 | −0.180 | −0.437 | ||
| LNSIZE | − | −0.039 | −0.343 | −0.021 | −0.193 | −0.035 | −0.303 |
| ROE_STD | + | 0.047 | 0.153 | 0.043 | 0.142 | 0.042 | 0.137 |
| ANALYST | − | −0.209 | −4.778*** | −0.209 | −4.780*** | −0.209 | −4.769*** |
| SURPRISE | + | 3.740 | 8.381*** | 3.757 | 8.430*** | 3.747 | 8.382*** |
| LEV | + | 0.178 | 2.542*** | 0.179 | 2.562*** | 0.178 | 2.542*** |
| NEWFOR | + | 1.231 | 2.144** | 1.248 | 2.190** | 1.219 | 2.118** |
| CORR | − | 0.270 | 1.154 | 0.279 | 1.194 | 0.273 | 1.166 |
| NEG | + | 0.828 | 2.836*** | 0.823 | 2.812*** | 0.819 | 2.796*** |
| Constant | 2.167 | 1.074 | 1.992 | 0.986 | 2.066 | 1.016 | |
| Year fixed effects | Yes | Yes | Yes | ||||
| N | 370 | 370 | 370 | ||||
| Adjusted R2 | 0.435 | 0.436 | 0.434 | ||||
| F statistic (p-value) | 29.465 (0.000) | 29.475 (0.000) | 26.744 (0.000) | ||||
- This table reports the regression results of analysts’ forecast error on the amount of IR disclosure, IR disclosure quality proxy by the AIRA Awards and the control variables. See Tables 2 and 3, for variable definitions. ***, **, * denote significance at the 1 per cent, 5 per cent and 10 per cent levels, respectively, based on one-tailed tests.
In Models 1, 2 and 3, the coefficients on DISC and AWARD are not significant singularly and together, respectively. When FD is the dependent variable, the coefficients on DISC and AWARD in Models 4, 5 and 6 are not significant. This indicates that the quantity and quality of investor relations disclosure does not affect forecast error and dispersion generally. The coefficients on the control variables such as number of analysts, earning surprise, financial leverage, proportion of new forecasts, and loss making firms are found to be consistent with expectations.
Due to the counter-intuitive results in Table 5 and Bushee and Miller's (2007) proposition that the incremental benefits of investor relations are likely to be larger for small or mid-cap firms that are lacking in visibility, we divide firms in each year into four subsamples based on disclosure score and estimate Models 1 and 2 for these subsamples.
Table 6 Panel A presents the summary statistics for the four quartiles where Quartile 1 contains firms with the largest quantity of disclosure and Quartile 4, the lowest. When examining the mean values of DISC, AFE and FD for the first three quartiles, we find negative relationships between DISC and AFE, and between DISC and FD. We also find more variability in DISC in the last quartile compared to Quartiles 1–3. We do not find significant differences in AFE and FD for firms in Quartile 4 compared to firms in the other quartiles.
| DISC Quartile 1 | DISC Quartile 2 | DISC Quartile 3 | DISC Quartile 4 | F-test/ Kruskal-Wallis (between all quartiles) | t-test/ U-test (Q1 vs Q4) | t-test/U-test (Q1, 2, 3 vs Q4) | ||
|---|---|---|---|---|---|---|---|---|
| Panel A: Summary statistics | ||||||||
| DISC | Mean | 45.622 | 38.711 | 35.054 | 26.899 | 180.03*** | −21.34*** | −15.73*** |
| Median | 47.000 | 42.000 | 36.000 | 27.000 | 222.33*** | −11.80*** | −11.92*** | |
| Max | 57.000 | 45.000 | 41.000 | 35.000 | ||||
| Min | 37.000 | 31.000 | 28.000 | 6.000 | ||||
| AFE | Mean | 2.575 | 5.021 | 5.170 | 3.445 | 2.77** | 1.11 | 0.89 |
| Median | 1.309 | 1.283 | 1.481 | 1.191 | 0.43 | 0.12 | 0.43 | |
| Max | 27.661 | 54.569 | 58.554 | 40.125 | ||||
| Min | 0.037 | 0.034 | 0.035 | 0.006 | ||||
| FD | Mean | 1.181 | 1.903 | 2.273 | 1.474 | 2.87** | 1.15 | 1.06 |
| Median | 0.793 | 0.803 | 0.744 | 0.678 | 1.65 | 0.81 | 1.21 | |
| Max | 7.035 | 14.840 | 25.483 | 10.191 | ||||
| Min | 0.108 | 0.066 | 0.124 | 0.042 | ||||
| Expected sign | DISC Quartile 1 Model 7 | DISC Quartile 2 Model 8 | DISC Quartile 3 Model 9 | DISC Quartile 4 Model 10 | |||||
|---|---|---|---|---|---|---|---|---|---|
| Coeff. | t-stat. | Coeff. | t-stat. | Coeff. | t-stat. | Coeff. | t-stat. | ||
| Panel B: AFE as dependent variable | |||||||||
| DISC | − | −0.034 | −0.307 | 0.517 | 1.024 | −0.806 | −1.923** | −0.071 | −0.862 |
| AWARD | − | −0.098 | −0.097 | 1.095 | 0.439 | −0.467 | −0.150 | −3.524 | −1.800** |
| FD | + | 1.221 | 3.152*** | 2.558 | 7.677*** | 2.431 | 10.785*** | 2.003 | 7.420*** |
| LNSIZE | − | −0.381 | −1.223 | 0.869 | 1.073 | 0.480 | 0.690 | −0.605 | −1.229 |
| ROE_STD | + | 0.480 | 0.252 | 0.349 | 0.203 | −2.130 | −1.470 | −1.502 | −1.305 |
| ANALYST | − | 0.098 | 0.722 | −0.710 | −2.304** | −0.335 | −1.365* | 0.016 | 0.098 |
| SURPRISE | + | 5.183 | 2.198** | 2.480 | 1.014 | −6.733 | −2.310 | 1.649 | 0.959 |
| LEV | + | −0.005 | −0.024 | 1.983 | 5.117*** | −0.334 | −0.548 | 0.027 | 0.075 |
| NEWFOR | + | 2.537 | 1.169 | −3.444 | −1.135 | 4.530 | 1.383* | 6.035 | 2.500*** |
| CORR | − | 0.475 | 0.607 | 0.655 | 0.490 | −1.466 | −1.063 | −1.841 | −2.100** |
| NEG | + | 0.801 | 0.797 | 0.998 | 0.602 | 2.565 | 1.515* | 2.729 | 2.293** |
| Constant | 8.048 | 1.143 | −36.693 | −1.640 | 19.640 | 1.067 | 11.607 | 1.201 | |
| Year fixed effects | Yes | Yes | Yes | Yes | |||||
| N | 98 | 90 | 86 | 89 | |||||
| Adjusted R2 | 0.320 | 0.677 | 0.722 | 0.704 | |||||
| F statistic (p-value) | 4.801 (0.000) | 16.565 (0.000) | 20.878 (0.000) | 18.413 (0.000) | |||||
| Expected sign | DISC Quartile 1 Model 11 | DISC Quartile 2 Model 12 | DISC Quartile 3 Model 13 | DISC Quartile 4 Model 14 | |||||
|---|---|---|---|---|---|---|---|---|---|
| Coeff. | t-stat. | Coeff. | t-stat. | Coeff. | t-stat. | Coeff. | t-stat. | ||
| Panel C: FD as dependent variable | |||||||||
| DISC | − | 0.014 | 0.441 | 0.023 | 0.133 | 0.498 | 2.501 | −0.025 | −0.718 |
| AWARD | − | 0.519 | 1.882 | −1.280 | −1.532* | −0.280 | −0.182 | 0.021 | 0.025 |
| LNSIZE | − | −0.185 | −2.189** | 0.182 | 0.663 | −0.208 | −0.608 | −0.132 | −0.638 |
| ROE_STD | + | 0.874 | 1.672** | 0.118 | 0.203 | −0.270 | −0.378 | 0.025 | 0.052 |
| ANALYST | − | −0.050 | −1.341* | −0.278 | −2.786*** | −0.251 | −2.136** | −0.142 | −2.118** |
| SURPRISE | + | 0.962 | 1.485* | 3.281 | 4.411*** | 5.212 | 3.964*** | 2.092 | 3.051*** |
| LEV | + | 0.202 | 3.808*** | −0.028 | −0.216 | 0.604 | 2.061** | 0.132 | 0.884 |
| NEWFOR | + | 1.872 | 3.290*** | 0.878 | 0.856 | 0.781 | 0.485 | 3.246 | 3.418*** |
| CORR | − | −0.015 | −0.071 | 0.255 | 0.563 | 1.207 | 1.812 | −0.285 | −0.772 |
| NEG | + | 0.339 | 1.223 | 0.698 | 1.251 | 0.823 | 0.992 | 1.107 | 2.277** |
| Constant | 3.236 | 1.679* | −1.914 | −0.252 | −11.657 | −1.298 | 3.637 | 0.896 | |
| Year fixed effects | Yes | Yes | Yes | Yes | |||||
| N | 98 | 183 | 82 | 82 | |||||
| Adjusted R2 | 0.426 | 0.478 | 0.532 | 0.488 | |||||
| F statistic (p-value) | 7.531 (0.000) | 8.411 (0.000) | 10.505 (0.000) | 8.639 (0.000) | |||||
- This table reports the regression results of analysts’ forecast error on the amount of IR disclosure, IR disclosure quality proxy by the AIRA Awards and the control variables. The data are partitioned by the quantity of disclosure – Quartile 1 comprises firms with the highest quantity of disclosure and Quartile 4 comprises with the lowest quantity of disclosure. See Tables 2 and 3, for variable definitions. ***, **, * denote significance at the 1 per cent, 5 per cent and 10 per cent levels, respectively, based on one-tailed tests.
In further analysis, we find the coefficients on DISC and AWARD are negative and significant only in Quartile 3 (Model 9) and Quartile 4 (Model 10), respectively. While the results are weak, they show that in firms disclosing fewer items, additional disclosure mitigates the forecast error. Award recipients also had significantly lower forecast errors than nonrecipients. This result indicates that the quantity disclosed is not as effective as specific items being disclosed, and the awards provide recognition of the quality of the disclosure. Another indirect implication from this result is that investor relations appears to be most effective in smaller and less-recognised firms, where it is most needed (Bushee and Miller, 2007).
Separate analysis was conducted for 2005 and 2009 for Quartile 4 to examine the effects of the uncertainty surrounding the GFC. In 2009, the AWARD coefficient is negative and significant at p <0.10, while the DISC coefficient is not significant. However, with the 2005 observations, both the coefficients on AWARD and DISC are not significant. These results suggest that during periods of unprecedented uncertainty with analyst forecasts reflecting this uncertainty, the award recipients are able to distinguish themselves from the crowd with their effective investor relations disclosure.
The results for FD as the dependent variable are given in Table 5 Panel B, and the only significant coefficient is AWARD in Quartile 2.
We further segment the sample based on firm size and disclosure due to their positive correlation where large firms disclose more items. The segmentation attempts to partially account for the endogenous relationship between firm size and disclosure and also the analyst following in large firms. These results are presented in Table 6. In Model 16, where the subsample consists of firms below the median in terms of market capitalisation and disclosure, the AWARD and DISC coefficients are negative although only the former is significant. This result supports the previous result in Table 6, again highlighting the importance of the award in distinguishing among firms with low levels of disclosure.
Another aim of this study was to determine the role analysts play in disclosure in Australia where in North America unlike in Continental Europe, and they appear to drive corporate performance disclosure and are more powerful or persuasive (Aerts et al., 2010). The results generally show the expected negative relationship between analyst following and forecast error and dispersion. When the sample was segmented to control for the endogeneity between firm size and disclosure in Table 7, the same negative relationship was observed except for smaller firms with lower than median levels of disclosure. A comparison of analyst following between award and nonaward recipients also shows significantly higher analyst following for the former. Taken together, these results indicate that for the largest Australian firms, analysts are a persuasive group that directs corporate disclosure. For these firms, the analyst–management relationship is closer to the North American relationship. However, this is not the case for the smaller firms on the ASX/S&P300 who will benefit the most from analyst following and exposure.
| Expected sign | Model 15 (LNSIZE > Median, DISC < Median) | Model 16 (LNSIZE < Median, DISC < Median) | |||
|---|---|---|---|---|---|
| Coeff. | t-stat. | Coeff. | t-stat. | ||
| Panel A: AFE as dependent variable | |||||
| DISC | − | −0.096 | −0.648 | −0.070 | −0.809 |
| AWARD | − | −1.077 | −0.306 | −2.981 | −1.395* |
| FD | + | 3.012 | 4.965*** | 2.169 | 13.028*** |
| LNSIZE | − | −0.393 | −0.377 | −1.034 | −1.205 |
| ROE_STD | + | −5.199 | −1.526 | −1.474 | −1.568 |
| ANALYST | − | −0.395 | −1.340* | 0.012 | 0.063 |
| SURPRISE | + | −1.749 | −0.587 | −2.677 | −1.242 |
| LEV | + | −0.122 | −0.213 | −0.194 | −0.375 |
| NEWFOR | + | 7.095 | 1.410* | 4.918 | 2.251** |
| CORR | − | −1.460 | −0.837 | −1.270 | −1.400* |
| NEG | + | −0.129 | −0.053 | 3.027 | 2.557*** |
| Constant | 12.562 | 0.575 | 20.950 | 1.277 | |
| Year fixed effects | Yes | Yes | |||
| N | 70 | 112 | |||
| Adjusted R2 | 0.423 | 0.791 | |||
| F statistic (p-value) | 5.222 (0.000) | 36.078 (0.000) | |||
| Model 17 (LNSIZE > Median, DISC < Median) | Model 18 (LNSIZE > Median, DISC < Median) | ||||
| Panel B: FD as dependent variable | |||||
| DISC | − | 0.041 | 1.292 | 0.042 | 0.809 |
| AWARD | − | 0.176 | 0.231 | 0.445 | 0.347 |
| LNSIZE | − | −0.099 | −0.440 | 0.558 | 1.090 |
| ROE_STD | + | 1.263 | 1.758** | −0.137 | −0.242 |
| ANALYST | − | −0.097 | −1.556* | −0.291 | −2.726*** |
| SURPRISE | + | 0.113 | 0.175 | 5.504 | 4.697*** |
| LEV | + | −0.136 | −1.110 | 0.787 | 2.615*** |
| NEWFOR | + | 0.772 | 0.712 | 1.782 | 1.370* |
| CORR | − | −0.213 | −0.565 | 0.852 | 1.584 |
| NEG | + | 1.431 | 2.922*** | 1.204 | 1.717** |
| Constant | 2.210 | 0.468 | −11.805 | −1.207 | |
| Year fixed effects | Yes | Yes | |||
| N | 70 | 112 | |||
| Adjusted R2 | 0.396 | 0.532 | |||
| F statistic (p-value) | 5.119 (0.000) | 12.461 (0.000) | |||
- This table reports the regression results of analysts’ forecast error on the amount of IR disclosure, IR disclosure quality proxy by the AIRA Awards and the control variables. The data are partitioned by the quantity of disclosure. See Tables 2 and 3, for variable definitions. ***, **, * denote significance at the 1 per cent, 5 per cent and 10 per cent levels, respectively, based on one-tailed tests.
5. Summary and conclusions
Investor relations is well-recognised as an effective multidisciplinary management tool in the firm. It increases the credibility of voluntary disclosure and improves the information flow to a wider audience. By examining the properties of analysts’ forecasts, we provided insight into the effect of the quantity and quality of investor relations disclosure because analysts’ forecasts proxy for market expectations and analysts utilise the information disclosed by the firm. We use both a quantity (disclosure score) and quality (AIRA award) measure of investor relations.
We report an increase in the quantity of items disclosed when comparing disclosure in 2005 with 2009. This reflects the growing investment in investor relations and the recognition by Australian firms of its importance as a communication channel with the investing public.
We use a sample of 370 observations from two separate years to investigate the relationship between investor relations disclosure and the properties of analyst forecasts, and the results indicate that an increase in disclosure reduces the forecast error. This result is only evident in firms with lower levels of disclosure and therefore highlights the importance of such disclosure among smaller and less-recognised firms. The awards given by AIRA are also important in highlighting quality investor relations practices. Contrary to our prediction, we found no relationship between forecast dispersion and disclosure. Lang and Lundholm (1996) reported a similar result where they argued that the effect of an increased disclosure on the dispersion of analysts’ forecasts depended on the reasons behind the disclosure. Analysts place different weights on components based on their own preferences so there may be differences in information or differences in forecasting models. If analysts relied more on their private information of forecasting models, an increase in company's disclosure would then increase the forecast dispersion. The inclusion of 2009 in the analysis also shows that while firms increased their disclosure during times of greater uncertainty, analysts did not ‘use’ all the information provided. This could be because they were themselves uncertain about the information provided or adopted a ‘wait and see’ attitude, as evidence by the higher forecast error and wider dispersion in 2009.
Our findings indicate analysts use investor relations disclosures on the firm's website as an important source of information in assisting them in producing forecasts. Australian analysts, particularly those following larger firms, are not unlike their US counterparts in the way they direct firm disclosure to keep the market informed of the firm's activities and prospects. Almost every firm listed on S&P/ASX 300 has an investor relations section on their website to keep in close contact with investors. Investor relations disclosure appears to be an effective corporate governance tool to reduce agency costs and information asymmetry.
