Internet investor relations, information asymmetry and the cost of capital: evidence from JSE listed companies

Thesis (PhD)--Stellenbosch University, 2016.ENGLISH SUMMARY : In South Africa, King III compliance is a JSE listing requirement. Chapter 8 of the King III code states that transparent and effective communication with stakeholders is essential for building and maintaining their trust and confidence. Investors and creditors, as important sources of finance, are two important stakeholders to companies. Investor relations is the field that is concerned with the management of relationships between companies and investors. Financial communication is an important component of investor relations, and entails much more than mere financial statements. While South Africa has recently been ranked number one by the World Economic Forum 2015–2016 Global Competitiveness Index for the strength of its auditing and reporting standards for the sixth consecutive year, the quality of investor relations, as a wider concept, is largely un-researched in South Africa. As opposed to financial statements which content is regulated by various standards, acts and codes, the investor relations activity is not regulated and companies have a wide variety of investor relations communication channels from which to choose. This dissertation is concerned with one of these channels, namely the corporate website. The purpose of this study was to investigate the quality of the corporate website for investor relations purposes in South Africa, to establish the determinants thereof, and to establish whether the use thereof has any effect on the level of information asymmetry and the cost of capital. Theoretically, a well-developed Internet investor relations strategy will increase company visibility, according to the investor recognition hypothesis of Merton (1987). An increased visibility may increase liquidity and, according to economic theory, an increased liquidity is linked to the cost of capital through information asymmetry. Considerable research has been done on the relationship between disclosure and both information asymmetry and the cost of capital. The vast majority of empirical research to date has relied on either the use of an indirect proxy for disclosure (e.g. analyst ratings) or a measurement of annual report voluntary disclosures. Prior research on Internet investor relations is limited. Although the weight of empirical evidence points to a negative association between disclosure/investor relations and both information asymmetry and the cost of capital, literature is far from reaching a consensus; and numerous studies have found no or even positive associations. Empirical research in these areas has further favoured the use of developed country data and this is the first study that endeavours to examine the determinants of Internet investor relations, as well as the effect thereof on information asymmetry and the cost of capital in the South African context. In the absence of a readily available and comprehensive measurement instrument, the first objective of this study was to develop a measurement instrument that could be used to measure the extent of Internet investor relations. The measurement instrument was developed using: (1) the best practice corporate website guidelines as published by the Investor Relations Society, (2) an extensive literature review to mitigate the risk of omitting important variables and to improve comparability with previous studies, and (3) a pilot study to evaluate the practicality of measuring the attributes as selected. The result was a measurement instrument that consisted of 346 attributes. Stratified random sampling with proportional allocation (using JSE industry membership) was used to select a sample of 85 JSE-listed companies. The corporate websites of these companies were assessed from March to September 2015. No research assistants were used and all assessments were done by one researcher (the writer of this dissertation). The scores of individual attributes were added together to calculate a disclosure score per company. Although the majority of attributes were measured as either available (1) or absent (0), 50 attributes were measured as partially available (0.5), based on the breadth and depth of content available. Non-functional and un-useful links were assessed as absent (0). Outdated information was assessed as either partially available (0.5) or absent (0). Where information was available, but as a result of factors such as poor layout, inconsistencies and incompleteness was not fully useful, the attributes were assessed as partially available (0.5). Although there was some subjectivity involved in such a methodology, the dissertation promotes the argument that such an approach was important to ensure that the quality of Internet investor relations would be measured, and not merely the quantity. No other study in the literature reviewed for this dissertation has assessed attributes as being only partially available based on the amount of information, and timeliness and usability concerns. Overall, the results showed that the majority of companies did not use corporate websites optimally to communicate with investors. Suggestions have been made on how companies can improve their Internet investor relations. To establish the determinants of Internet investor relations, numerous company characteristics that could explain variations in Internet investor relations levels were identified from the literature review. Using stepwise regression, it was found that company size, leverage, the audit firm used, industry membership, free float, and dual-listing status explained 69% of the total variation in Internet investor relations. As information asymmetry is not directly observable, this dissertation used five alternative proxies to estimate information asymmetry: the bid-ask spread, price impact, share price volatility, share turnover, and analyst following. In theory, the first three proxies are positively related to information asymmetry and the last two, negatively. The level of Internet investor relations was found to be statistically significantly and negatively related to the bid-ask spread and price impact, and positively related to share turnover and analyst following. Based on theory, the observed relationships – two negative; two positive – therefore all points towards a negative relationship between Internet investor relations and information asymmetry. By identifying additional variables that were used in literature to explain variations in information asymmetry and applying stepwise regression, this study constructed regression models that explained 51%, 54%, 90% and 91% of the variations in the share turnover, analyst following, price impact and bid-ask spread, respectively. The activity of Internet investor relations was found to be non-significant in explaining the share price volatility information asymmetry proxy. When examining the association between Internet investor relations and the cost of debt, it was found that Internet investor relations was statistically significantly and negatively related to the cost of debt. The cost of debt was measured as the interest expense for the year, scaled by the average interest-bearing liabilities. Although the explanatory power of this regression model was very low (adjusted R² of 14%), the adjusted R² compared favourably with previous disclosure–cost of debt studies. Guidara, Khlif and Jarboui (2014), for example, examined the relationship between annual report voluntary disclosure and the cost of debt using South African data and reported an adjusted R² of 8%. Owing to the non-availability of analyst forecast data for the study sample, ex ante cost of equity estimate methods could not be used. To estimate the cost of equity, this study thus used the capital asset pricing model. PwC (2015) valuation surveys have shown that the capital asset pricing model is the most often used method in cost of equity calculations in southern Africa. Criticism against the use of the capital asset pricing model was carefully considered in this dissertation and an adjustment was made to the cost of equity of smaller companies (i.e. companies with a market capitalization of less than R2 000 million). These adjustments were based on current valuation practice in South Africa. This study found that the level of Internet investor relations was statistically significantly and negatively related to the cost of equity. Together with share price, leverage, the market-to-book ratio and industry membership, the level of Internet investor relations was found to explain 59% of variations in the cost of equity. In a separate analysis of the cost of equity, before any adjustments to the smaller companies, the level of Internet investor relations was, however, found to be non-significantly related. Cost of capital – also named the weighted average cost of capital (WACC) – is the weighted average cost of equity and cost of debt. Irrespective of how the weightings were calculated (i.e. by means of book value or market value) or whether the cost of equity adjustments discussed in the two paragraphs above were made or not, the level of Internet investor relations was found to be statistically significantly and negatively related to the cost of capital. Overall, the results of this study suggested that companies may potentially benefit from a well-developed Internet investor relations strategy through decreased information asymmetry and cost of capital. Since disclosure studies are often criticized for not testing or controlling for endogeneity, the Wu-Hausman test statistic was applied, and duly confirmed the absence of endogeneity in all regression models.AFRIKAANSE OPSOMMING : Geen opsomming beskikbaar