Accounting Information and Cost of Capital

Transcription

1 Accounting Information and Cost of Capital SFB Jour Fixe 29 Nov 2006 Joachim Gassen Institute for Accounting and Auditing School of Business and Economics Humboldt-Universität zu Berlin gassen@wiwi.hu-berlin.de

2 2 Walk through Idea and motivation Background Role of accounting on capital markets Pricing of information Prior empirical findings The suggested research program Research questions Prior work Work plan Linkages to current SFB projects

3 3 Idea and motivation Financial accounting disclosures are assumed to be one of the corner stones of the infrastructure fostering efficient capital markets and as such are assumed to have an economically significant impact on disclosing firms cost of capital: Arthur Levitt (SEC chairman, 29 Sep 1997): High quality accounting standards result in greater investor confidence, which improves liquidity, reduces capital costs, and makes fair market prices possible. Nevertheless: Weak support by economic theory Weak support by empirical findings

4 4 The role of accounting on capital markets From a capital market perspective, financial accounting fulfills two distinct roles: Providing public information for valuation of asset claims Providing public information as a contracting device From a valuation perspective Financial accounting should convey information about future cash flows to investors From a contracting perspective Financial accounting should play an governing role, setting incentives for management action, thus influencing future cash flows to investors

5 5 Pricing of information on capital markets Starting point: Rational Expectations (e.g.: Grossman/Stiglitz, 1980; Hellwig, 1980; Diamond/Verrecchia, 1981; Merton, 1987): Investors rationally acquire information to value their claims Competing literature on the impact of information asymmetry Estimation risk (e.g. Klein/Bawa, 1976; Barry/Brown, 1985): Investors have limited information (compared to nature), causing an unsystematic risk component, which is diversifiable Covariance estimation risk (e.g. Lambert et al., 2006a): Investors have limited information (compared to nature) about the covariance of future cash flows to investors, causing a systematic risk component, which is (at least in part) not-diversifiable Market microstructure models (e.g. Admati, 1985; Easley/O Hara, 2004; Hughes et al., 2006; Lambert et al. 2006b): Investors have heterogeneous information, causing a market microstructure effect, whose diversifiability is currently under dispute. Heterogeneous agent models (overview: Hommes, 2006)

6 6 Selected prior empirical findings on the link between Disclosure quality (scores) and bid-ask-spread: - (Welker, 1995) implied cost of equity capital: - (Botosan, 1997) implied cost of equity capital: amount:-, timeliness:+ (Botosan/Plumlee, 2002) Financial accounting rules and liquidity proxies: - (Leuz/Verrecchia, 2000) implied cost of equity capital:?/+ (Daske, 2005) accounting quality and liquidity proxies: -/? (Gassen/Sellhorn, 2006) Private Information (PIN-score) and Fama/French 3-factor model errors: + (Easley et al., 2002) Analysts information and implied cost of equity capital public info: -, private info: + (Botosan et al., 2004)

7 7 Our research program: research questions 1. Using different robust proxies for public information and cost of capital, Is the level of public information priced? If yes, does information constitute an additional priced risk factor? Is there a crowding out effect between public and private information? 2. Building on the contracting role of accounting, Is there a competing explanation for a link between accounting information and cost of capital? Is there empirical evidence for this link? Is contractual efficiency a priced risk factor?

8 Competing theories accounting disclosure Fama/French type factors market risk value risk size risk cost of capital efficient contracting microstructure

9 9 What we have done so far? Ashbaugh-Skaife et al. (2006a): Developed and tested a proxy for information impounded into prices (percentage of zero return score, PZR-score), serves as important proxy for information level Gassen/Sellhorn (2006): Tested the impact of accounting change on several metrics of information asymmetry to explore the link between accounting disclosures and market liquidity/microstructure Gassen et al. (2006): Verified that level of conditional conservatism is reliably related to demand for the use of accounting information as a contractual device. Justifies the use of conditional conservatism as a proxy for the level of contractual importance of accounting. Developed and justified a firm-specific measure of conditional conservatism Ashbaugh-Skaife et al. (2006b): Showing that the level of conditional conservatism relates to the level of firm-specific information impounded into prices, indicating that there might be a trade-off between public information and the contractual role of accounting

10 10 Research plan RQ 1 1. Develop proxy for public information Prior used proxies capture only limited aspects and/or available for a small sample only Idea: Combine PIN and PZR scores 2. Develop proxies for cost of capital Proxies are very noisy Idea: Develop a qualitative factor based on alternative proxies (implied equity, implied debt, asset pricing, fundamentals) 3. Pick samples and conduct analyses Prior studies focus on small within regime/country samples Idea: Go cross-country to make use of greater variance in explanatory variable, use U.S-sample for additional finer tests 4. Deal with endogeneity issues The link between public information and cost of capital suffers from potential self selection and general omitted variable problems Idea: Use 2SLS and/or alternative robust estimation techniques

11 11 Research plan RQ 2 1. Model link between accounting, demand for accounting as a contracting device and cost of capital Prior research has focused on the link between cost of capital and public financial accounting disclosures explained by the valuation role of accounting Idea: Explicitly model the impact of accounting on efficient contracting and the impact of efficient contracting on cost of capital (for a large economy with diversification) 2. Develop proxies for the level of accounting being used as a contracting device Idea: Use conditional conservatism 3. Set up the empirical design, conduct the analyses and (again) deal with endogeneity issues Idea: Proxy for the contracting role of accounting will most probably be based on firm specific time-series, thus sample will be crosssectional. Also see RQ1

12 12 Linkages to other projects of the SFB A1 (Maug, management compensation) Explores the link between information signals and efficient contracting A4 (Demougin, risk sharing in organizations) Refines contracting theory to multi-tasking setting and to unverifiable reputation effects A6 (Kübler, strategic risk in experiments) Helps to understand the usage of information at the micro level. B1 (Härdle, dynamic semiparametric modeling) Dynamic semiparametric factor modeling might help to address endogeneity in estimated models B4 (Dittmann/Maug) Implied cost of equity capital methods can be used as input to our cost of capital factor B6 (Imkeller, models of financial markets with heterogeneous information) Models developed by this group can be used to derived refined hypotheses about the impact of information asymmetry on capital markets) Other (B5, Spokoiny, data analysis of multi-dimensional structures and structural changes; C1, Uhlig, macroeconomic risk factors; C4, Reiss, time lags) FEDC

13 13 Backup: Construction of the public information proxy PZR-Score overall info low due to lack of private info overall info high due to public info Increasing level of public information overall info low due to lack of public info overall info high due to private info PIN-Score

14 14 Backup: Conditional conservatism environment Market efficiency RETURNS stock market NEWS EARNINGS Timeliness (conservatism implies asymmetrical timeliness for good versus bad news) accounting In empirical accounting research, stock returns are frequently used to represent news. This assumes market efficiency. (In inefficient markets, ceteris paribus, earnings may appear more timely.)