Dataset to determine the differential impact of discretionary and non-discretionary disclosure on firm value in emerging and developed markets
Description
The objective of this dataset is to investigate the role of discretionary and non-discretionary disclosures in financial reports across both emerging and developed markets. It seeks to assess the significance of these disclosures in varying reporting environments and the extent to which their importance holds up across these markets. Two key hypotheses are proposed: (1) discretionary and non-discretionary disclosures are individually and collectively significant in both emerging and developed economies, and (2) the level or magnitude of these disclosures is influenced by a firm's reporting environment. Discretionary disclosure is evaluated through the tone of language used in the Chairman’s letters in annual reports, employing Diction measures such as optimism, certainty, activity, realism, and commonality. Non-discretionary disclosure, on the other hand, is measured using book value per share and earnings per share. The dataset also includes variables representing the reporting environment, such as ownership structure, economic conditions, and regulatory frameworks. Control factors like firm size, age, and risk are also incorporated. The findings reveal that the use of optimistic language in Chairman’s letters ranks highest across all reporting environments, followed by language emphasizing realism. This suggests that discretionary disclosures are often geared towards projecting a positive outlook on the company by including tangible and easily recognizable facts. Moreover, the comparison of emerging and developed markets shows distinct differences in reporting environments: emerging markets tend to exhibit weaker regulatory frameworks, lower rule of law, and less economic openness, yet higher economic growth. In terms of ownership structures, institutional ownership is more prevalent in developed markets, while there is no significant difference between foreign and local ownership across emerging and developed markets. Additionally, the dataset was utilized to explore whether discretionary and non-discretionary disclosures can predict firm value, and whether this predictive ability holds in different reporting environments.
Files
Steps to reproduce
The companies listed on the main exchanges in South Africa (Johannesburg stock exchange); India (Mumbai Stock market), the UK (London Stock exchange), and the USA (New York Stock exchange) were sorted using the market capitalization at the end of 2019. The top 100 companies by market capitalization in 2019 were identified and their chairman’s letters were then collected for each year from 2000 up to the end of 2019. Any top 100 company that did not have a chairman’s letter was excluded from the sample. The final sample consisted of 62 companies from South Africa, 62 from India, 68 from the UK, and 59 from the USA. The Chairman’s letters of the sample companies were then downloaded from the firms’ websites and from Annuakreports.com. The AnnualReports.com was preferred as the Chairman’s letters were saved in text format which was required for subsequent Diction analysis. The graphs, tables, and pictures were manually removed from the Chairman’s letters so that only text information remains. The Chairman’s letters were then analysed by Diction which then provided the discretionary disclosure measures including optimism, certainty, activity, realism, and commonality. All these variables measure the tone of the language in the Chairman’s letter and are meant to be proxies for discretionary disclosure in the firm’s annual reports. The yearly non-discretionary disclosure data were then collected for companies that have Chairman’s letter and therefore discretionary disclosure data. In some cases, a company would have a chairman’s letter but have missing data for earnings per share (EPS) and book value per share (BVP) if these variables are not reported. The reporting environment data for each company in a sample country was collected from World Bank Governance Database.