Module 4: Information Approach to Decision-Usefulness Information Perspective - approach to financial accounting theory which equates the extent of security price change with information content and hence with decisionusefulness. Accounting research shows that security market prices respond to the net income component of accounting information. (Ball & Brown 1968 study was first major evidence.) Accounting information is of use to investors in assessing expected values & risk of security returns -> gives accounting content to accounting information since investors change beliefs & actions. Outline of the Research Problem Investors have prior beliefs about expected return & risk of shares based on all publicly available information differ in amount of info obtained and ability to interpret (note: do not pay directly for information) could include expectations about current & future profitability Investors analyze income component of current year F/S & become more informed. Upon analysis they revise their beliefs: a) if they think profitability will increase they will buy at current market price b) if they think profitability will decrease they will sell at current market price Thus, volume of shares traded should increase when F/S info released. The greater the discrepancy between investor prior expectations and their current interpretation of earnings, the greater the volume of shares traded. Beaver (1968) found a dramatic increase in volume during the week of release of earnings announcements. Factors which complicate the actual finding of statistical evidence of securities’ price response to reported NI: (1) Efficient market theory says that market will react quickly to new information. If researcher is looking for effects a few days late the effects may have disappeared at this time. Researchers will generally use date firm's NI was reported in financial media ->ex. Financial Post. Consider impact of "information leaks"where market learns of the info early; may also take a few days for investors to digest the news…..typically would want to use a wider window in research study to address this problem. (2) Also need to obtain a proxy for what investors expected NI to be: use analysts’ forecasts of earnings for the quarter or year in question (could also use the firm’s forecast) project the time series of the firm’s past reported NI’s (ex. expect that the next year’s NI will be the same as this year’s. The change in NI is proxy for the unexpected component) (3) Other events occurring - need market model to segregate other events (firm specific) which affect returns. 1 Module 4: Information Approach to Decision-Usefulness Actual return Rjt = (j + ßjE(Rmt) + jt = Expected return expected return + abnormal return E(Rjt ) = Rf (1 – βj ) + βRMt Typically will be given estimated beta, risk free rate and return on market ->therefore can plug in the expected return calculation into the first formula to solve for abnormal return. Concerns: a) Other firm specific info occurs on the day that other firms in the sample release their earnings info -> thus, these firms can be removed from the sample; b) errors in estimating beta, especially if the firm’s beta changes over time -> thus, can get a 2nd opinion on beta by estimating it from F/S info instead of from market data or can estimate beta from a period after the earnings announcement and compare with the estimate from a period before the announcement. Ball and Brown study - studied 261 NYSE firms over 9 years (1957 to 1965) with primary purpose of studying information content of earnings. Reported earnings greater than market expectations -> GN Reported earnings less than market expectations -> BN GN earnings firms - positive abnormal return; BN firms - negative abnormal return Conclusion: market did respond to the good or bad news in earnings therefore net income had information content. Earnings Response Coefficients ERC research - identification and explanation of differential market response to earnings. Definition: ERC measures the extent of the security’s abnormal market return in response to the unexpected component of the firm’s reported earnings. Empirical evidence supports a differential market response depending on ERC's. Affect on ERC -> a lower demand for a firm's shares implies a lower increase in market price and return in response to GN, therefore a lower ERC. Reasons include: BETA The more risk related to the firm's expected returns the lower will be the investor's reactions to a given amount of unexpected earnings. Collins & Kothari, Easton & Zmijewski (1989) found evidence of a lower ERC for higher beta securities Beta is relevant risk measure of a security; high beta increases portfolio risk 2 Module 4: Information Approach to Decision-Usefulness CAPITAL STRUCTURE ERC for a highly levered firm is lower than for a firm with little or no debt (GN in earnings passes on to debt holders instead of shareholders)> supports expansion of disclosure of nature & magnitude of financial instruments including off-balance sheet) Dhaliwal, Lee & Fargher (1991) found evidence of a lower ERC for more highly levered firms Billings (1999) also finds lower ERC’s for firms with higher debt/equity ratios. PERSISTENCE Source of increase in current earnings affects the ERC: - if earnings are expected to persist into the future ERC will be higher - non-persistence (i.e. unusual, non recurring items) results in lower ERC Kormendi & Lipe (1987) have provided evidence that ERC's are higher when the persistence of unexpected current earnings changes is higher Ramakrishnen & Thomas (1991) - persistence of earnings depends on whether the firm uses historical cost accounting (higher ERC if historical cost used) Price irrelevant (i.e. no cash flow impact): persistence is 0 for paper changes Transitory (i.e. current year affect): persistence is 1 for an impact on NI that is not expected to persist Permanent (i.e. expected to continue indefinitely): persistence is higher than 1 for earnings changes which are expected to persist for several years. (1 + Rf )/ Rf EARNINGS QUALITY Higher quality of earnings (as defined by the higher main diagonals of the related info system) results in a higher ERC. Bandyopadhyay (1994) - studied ERC’s of SE and FC oil & gas firms (effect of capitalization and subsequent w/o of dry holes introduces price irrelevant components under FC, thus he predicted SE would have higher quality earnings) - found significantly greater ERC for SE firms in 2 day period around release of earnings announcements - predicted that the higher ERC’s for SE firms would be less pronounced during periods of low exploration activity Lev & Thiagarajan (1993) - identified 12 fundamentals used by financial analysts in evaluating earnings quality, assigned a score of 0 or 1 to specific fundamentals for each firm (and grouped firms by earnings quality) - found higher quality earnings groups had higher ERC’s - also found that there was a positive relationship between persistence and quality of earnings - market, perhaps with assistance of analysts, is quite sophisticated in use of info: uses B/S info to augment the info content of earnings announcement. GROWTH OPPORTUNITIES To the extent that current good news in earnings suggests growth opportunities the ERC will be high (e.g. an expanding new division with high profitability -> good to have segment disclosure in F/S) Collins & Kothari (1989) - studied ratio of MV of equity to BV of equity - found positive relationship between this measure & ERC of firm 3 Module 4: Information Approach to Decision-Usefulness SIMILARITY OF INVESTOR EXPECTATIONS Although different investors may have different expectations of firm’s future performance based on prior expectations - to the extent that they use a common info source such as analysts’ forecasts, they will put the same interpretation on firm’s announcement of earnings. More similar the investors’ expectations, greater the effect of a $ of abnormal earnings on share price . (More precise the analysts’ forecasts -> more similar are investors’ earnings expectations -> greater the ERC.) IMPLICATIONS OF ERC RESEARCH Important to know how accounting information is important to investors -> can then prepare more useful financial statements. Information content of reported net income can be measured by the extent of the share price change upon release of current income. Disclosure is important given specific research findings: Lower informativeness of price for smaller firms -> these firms should provide more overall disclosure Highly levered firms -> should expand disclosure of nature & magnitude of financial instruments, including off balance sheet instruments Growth opportunity firms -> should disclose segment info Earnings persistence & quality of earnings -> should disclose components of NI Unusual, Non-recurring and Extraordinary Items The reporting of unusual, non-recurring and extraordinary items is one of the most important issues in current financial accounting practice -> frequently used by management to increase future reported core earnings, operating earnings and management compensation. Earnings management through use of unusual, non-recurring and extraordinary items to increase future reported core and operating earnings (typically involve the creation of loss provisions, which can be used in future periods to reduce costs when the provision is reversed) Excessive use of unusual, non-recurring and extraordinary charges is sometimes referred to as putting earnings "in the bank" (since the costs recorded earlier are used to reduce costs in future periods) Concern is lack of visibility in future years since disclosure relates to current periods, not future periods Core earnings are net income before extraordinary, unusual and nonrecurring items Operating earnings are net income before extraordinary items) IASB standards - IAS 1 prohibits the use of the term “extraordinary item” completely. doesn’t use the terms “unusual and non-recurring items.” requires separate disclosure in the income statement of low-persistence items (such as writedowns of inventory and property, plant and equipment, provisions for restructuring, and gains and losses on disposals of investments and property, plant and equipment) contains provisions that require fair presentation, including provision of sufficient information that the effects of significant transactions can be understood. Full disclosure should cover sufficient information so that investors can evaluate earnings persistence -> helps investors revise their probabilities of future firm performance. 4 Module 4: Information Approach to Decision-Usefulness A caveat Is the best accounting policy that one which produces the greatest market price response? Increased information may be useful to investors but society may not be better off. Information has the characteristics of a public good: DEF'N - a public good is a good which can be consumed without loss in utility. Due to the public good nature of information the supply & demand forces don't work to produce the right amount of information. Annual reports are provided without charge to investors, however cost will eventually be recouped through higher products prices. Since no direct cost to investors they will want more! Concern - investors may consider increased information as being useful while society would consider the costs (i.e. higher product prices) outweigh the benefits to investors and capital markets. {Regulation of information production covered further in Lessons 9 & 10 - reread this section at that time} However, investor response can still guide accountants as to which accounting policies are most useful to investors. Example: If net income based on declining balance depreciation produces a greater ERC than net income based on straight line, then the main diagonal probabilities of the information system must be higher under declining balance depreciation than under straight line. This leads to a stronger effect on share price. Thus, the argument that accountants should produce information that is of greatest help to investors is still valid. Information Content of Other Financial Statement Information Does RRA information have additional explanatory power over historical cost NI & BV? Since historical cost and market values are usually significantly different in the oil & gas industry we would expect RRA info to have value. However, research is not finding RRA to have strong explanatory power. Following research uses wide window studies -> appears that RRA info competes with info from other sources): Magliolo sample of firms (1978 - 1983) - looked at undiscounted value of net reserves reported in RRA compared to the market valuation of these reserves. >> Reserve info provided by an investment service outperformed the RRA info. Doran, Collins, Dhaliwal 173 oil & gas firms (1979 - 1984) - calculated abnormal returns for 12 months up to year end; used regression analysis 1st 3 years - both historical & RRA had significant explanatory power (likely due to high volatility of oil & gas market prices) Next 3 years - neither had explanatory power (could be due to effect of expectation of oil glut) >> weak evidence to support RRA usefulness Harris & Ohlson 273 observations (1979 - 1983) - BV had significant explanatory power for market value of these assets - RRA had some explanatory power (less than BV) 5 Module 4: Information Approach to Decision-Usefulness Reasons for weak results for explanatory power of RRA: 1. Lack of reliability 2. Methodological problems - difficult to determine when market first becomes aware of RRA info 3. Historical cost based info may be more value relevant than implied in the previous discussion. (i.e. availability of alternative info sources) (2nd reason is most likely) Lev and Thiagarajan research (refer to ERC topic above) provides some evidence of the usefulness of balance sheet information (as opposed to supplemental note information, as in RRA). This indicates that the market response to balance sheet and supplementary information is routed through the ERC. OVERALL: 1. Empirical results support the rational investor and efficient securities market theories and indicate that at least net income information is found to be useful by investors Financial accounting information assists in: - revising expected security returns - estimating firm risk (Since beta is the primary risk measure for a diversified investor and is not directly related to F/S variables >> above finding is surprising) 2. Market does not respond to non-earnings information as strongly as it does to earnings information More difficult to find a market reaction to non-earnings accounting information. Methodological difficulties of finding a direct reaction are a possible explanation. o researchers may be unable to find just when the market first becomes aware of non-earnings information, or o the market may learn much of the other financial statement information from more timely sources 3. By routing the effect of balance sheet information on share price through the ERC rather than looking for it directly, it appears that the market does find other financial statement information useful. Investors: - attempt to predict future returns from their investments - seek all relevant info in this regard (not just accounting info) Accountants: - interested in using the extent of security market response to accounting info as a gauge of usefulness thus motivating their interest in empirical research on decision usefulness Caveat - refer to public good nature of accounting info before concluding that a/c policies & disclosures producing greatest market response are the best for society. this concern limits ability of decision usefulness research to aid accounting standard setters 6