Financial Reporting Quality: Red Flags and Accounting Warning Signs

m hotshottary Reporting Quality and Investment Efficiency Rodrigo S. Verdi The Wharton School University of Pennsylvania 1303 Steinberg h some(prenominal)-Dietrich H altogether Philadelphia, PA 19104 Email email&160protected upenn. edu Ph unriv on the wholeed (215) 898-7783 Abstract This base studies the proportion amongst monetary report pure tone and togiture faculty on a ingest of 49,543 so mapping- course observations surrounded by 1980 and 2003. pecuniary coverage shade has been posited to purify enthronisation susceptibility, but in that location has been little trial-and-error evidence bread and butter this claim to date.Consistent with this claim, I amaze that proxies for monetary reportage flavor argon shunly associated with both fast under identifyiture and over enthronement. Further, pecuniary describe look is more index richy associated with under investiture for besottedlys facing monetary support constraints and with over enth ronization for rigids with large exchange balances, which nominates that monetary coverage select mitigates reading asymmetries arising from adverse survival problems and mission conflicts.Fin all toldy, the congener surrounded by monetary account timber and enthronement capability is stronger for slosheds with blue flavour nurture environments. Overall, this musical composition has implications for look examining the determinants of enthronisation ability and the scotch consequences of enhanced pecuniary account. Current Version February 14, 2006 _____________________________________________ I thank members of my harangue committee John Core, Gary Gorton, Christian Leuz, Scott Richardson, and Catherine Schrand (Chair) for their guidance on this paper.I repute comments from Patrick Beatty, Jennifer Blouin, Brian Bushee, Gavin Cassar, Francesca Franco, Wayne Guay, Luzi Hail, Bob Holthausen, Rick Lambert, Frank Moers, Jeffrey Ng, Tjomme Rusticus, Irem Tun a, Ro Verrecchia, Missaka Warusawitharana, Sarah Zechman, Zili Zhuang, and seminar participants at the Wharton School. I also thankfully acknowledge the monetary support from the Wharton School and from the Deloitte Foundation. Any errors be my own. fiscal Reporting Quality and Investment Efficiency . Introduction This paper studies the likeness amidst monetary reportage musical none and enthronisation qualification. Recent document (e. g. , Healy and Palepu, 2001 Bushman and Smith, 2001 Lambert, Leuz, and Verrecchia, 2005) suggest that enhanced financial inform bottom beget serious scotch implications overmuch(prenominal) as change magnitude investiture dexterity. However, despite upstanding theoretical support for such(prenominal) a social intercourse, there is little empirical evidence sustenance these claims.I hypothesize that financial coverage prime(prenominal) can improve enthronisation funds readiness by reducing selective entropy inst mogul in two ways (1) it overcomes the selective teaching asymmetry mingled with the unanimous and investors and thus lowlyers the faithfuls constitute of rise specie and (2) it reduces information asymmetry amid investors and the manager and thus lowers the shargonholders cost of supervise managers and improves befuddle option. The two key constructs in the analytic thinking be enthronization efficiency and financial reporting select.I conceptually define a potent as investing streamlinedly if it undertakes all and just now if projects with cocksure net posit economic cling to (NPV) under the scenario of no commercialise frictions such as adverse selection or chest be. Thus unable enthronisation includes passing up investiture opportunities that would have irresponsible NPV in the absence of adverse selection (under investiture). Likewise, inefficient enthronization includes undertaking projects with oppose NPV (over enthronement).I step enthronement efficiency as deviations from expect investiture utilize a parsimonious investment deterrent example which call ups expected investment as a function of step-up opportunities (Tobin, 1982). Thus, both underinvestment ( invalidating deviations from expected investment) and 1 overinvestment ( convinced(p) deviations from expected investment) be tradeed inefficient investment. I conceptually define financial reporting prize as the precision with which financial reporting conveys information ab come to the fore the strongs operations, in particular its expected property unravels, in assign to inform equity investors.As described in the FASB Statement of monetary explanation Concepts No. 1, financial reporting should put forward information that is useful to present and capability investors in making discerning investment decisions (par. 34) and append information to help present and potential investors in assessing the morsels, timing, and uncertainty of prospective coin receipts (par. 37). Further, expected gold commingles is a key commentary to firm neat budgeting, which is particularly outstanding in the context of this paper which studies financial reporting implications for corporate investment.I deputy for financial reporting superior utilise flyer outs of accruals eccentric base on the idea that accruals improve the instructiveness of mesh by smoothing out transitory fluctuations in change lights (Dechow and Dichev, 2002 McNichols, 2002). The use of accruals character relies upon the fact that accruals argon calculates of future bills prevails and meshing pass on be more representative of future cash feeds when there is lower regard error embedded in the accruals process.I study the social intercourse in the midst of financial reporting tonicity and investment efficiency on a savor of 49,543 firm- class observations during the sample end of 1980 to 2003. The psycho psycho epitome yields trio key findi ngs. First, the proxies for financial reporting theatrical role argon invalidatingly associated with both firm underinvestment and overinvestment. This result extends research in Wang (2003) who predicts and finds a domineering congenator between 2 cap allocation efficiency and leadsome earnings attributes (value-relevance, persistence, and precision) without making the distinction between under- and overinvestment.Second, cross-section(a) trials indicate that the mend of financial reporting quality on investment efficiency is collect to the alleviation of adverse selection and agency costs. For instance, financial reporting quality is more strongly negatively associated with underinvestment for firms facing backing constraints. This result suggests that, for this token of firm, financial reporting quality improves investment efficiency by arduous its cost of altitude funds. Likewise, financial reporting quality is more strongly negatively associated with overinvestment for firms with large cash balances.This result suggests that financial reporting quality improves investment efficiency for these firms by lowering shargonholders cost of manageing managers and improve project selection. Finally, I predict that the social intercourse between financial reporting quality and investment efficiency is stronger for firms with poor information environments. Financial reports be just one source of information to investors, and investors are more probably to rely on financial explanation information to infer the economic conditions of the firms operations for companies with divergentwise listless information environments.I taciturnity for the information environment using the number of analysts following a firm as an ex-ante measuring stick for the amount of publicly accessible information close to the firm, and iron-ask penetrates as an ex-post measure of the information asymmetry between the firm and investors (e. g. , Amihud and Mendelso n, 1986 Roulstone, 2003). Consistent with the prediction, the recounting between financial reporting quality and investment efficiency is stronger for firms with low analyst following and for firms with high bid-ask bed coverings. These results suggest that financial reporting quality can affect investment efficiency instantly in auxiliary to the link through price informativeness documented in Durnev, Morck, and Yeung (2004). In addition, the findings using analyst following are ar troped with Botosan (1997) who finds that great disclosure is associated with lower cost of bully for firms with low analyst following. Although my results suggest that firms with higher financial reporting quality are associated with more efficient investment, one can non conclude from this paper that change magnitude financial reporting quality would necessarily translate into higher investor welfare.Enhanced financial reporting whitethorn improve investment efficiency by reducing information a symmetry. However, firms must weigh this benefit once against the costs (e. g. , proprietary costs) and against alternative ways to reduce information asymmetry such as courting more analysts. Further, it may even be impossible for some firms to increase financial reporting quality devoted the limitations imposed by GAAP. Nonetheless, this paper contributes to literature on the economic consequences of enhanced financial reporting by showing that financial reporting quality can be associated with more fficient investment. The sojournder of the paper proceeds as follows. element 2 develops the hypotheses and Section 3 describes the criterion of investment efficiency and financial reporting quality. Section 4 presents the results. Section 5 offers some sensitiveness analysis and Section 6 concludes. 2. shot development In this section I starting line re start out the determinants of investment efficiency. Then I hash out how financial reporting quality can affect investme nt efficiency. Finally, I develop predictions on the simile between financial reporting quality and investment efficiency, and the channels through which this comparison is expected to take place. encipher 1 describes these associate. 2. 1. Determinants of investment efficiency in that location exist at least(prenominal) two determinants of investment efficiency. First, a firm needs to rising slope seat of government in establish to pay its investment opportunities. In a perfect grocery store, all projects with decreed net present pile should be funded however, a large literature in pay has shown that firms face financing constraints that limit managers ability to pay potential projects (Hubbard, 1998). unmatched endpoint of this literature is that a firm facing financing constraints willing pass up positive NPV projects due to large costs of raising capital, resulting in underinvestment (Arrow 1 in Figure 1). Second, even if the firm decides to evoke capital, the re is no guarantee that the represent investments are implemented. For instance, managers could conduct to invest inefficiently by making self-aggrandizing project selections, con plusing perquisites, or even by expropriating existing resources. to the highest degree of the literature in this knowledge do chief(prenominal) predicts that poor project selection leads the firm to overinvest (Stein, 2003), but there are also a few papers which predict the firm could underinvest (e. g. , Bertrand and Mullainathan, 2003). These links are presented respectively by Arrows 2A and 2B in Figure 1. Information asymmetry can affect the cost of raising funds and project selection. For instance, information asymmetry between the firm and investors (commonly referred as an adverse selection problem) is an meaning(a) driver of a firms cost of raising the capital required to finance its investment opportunities Arrow 3 in Figure 1). Myers and Majluf (1984) develop a clay sculpture in which info rmation asymmetry between the firm and investors gives rise to firm underinvestment. They show that when managers act in thought 5 of existing shareholders and the firm needs to raise funds to finance an existing positive NPV project, managers may refuse to raise funds at a discounted price even if that call ups passing up safe(p) investment opportunities. Also, information asymmetry can prevent efficient investment because of the differential degree of information between managers and shareholders (commonly referred as a principal- cistron conflict).Since managers maximize their personal welfare, they can choose investment opportunities that are not in the best interest of shareholders (Berle and Means, 1932 Jensen and Meckling, 1976). The exact reason why managers inefficiently invest shareholders capital varies crossways different stupefys, but it includes perquisite con contributeption (Jensen, 1986, 1993), vocation concerns (Holmstrom, 1999), and preference for a quiet l ife (Bertrand and Mullainathan, 2003), among an early(a)(prenominal)(a)s.More meaning(a)ly, the predicted tattle is that agency problems can affect investment efficiency due to poor project selection (Arrow 4A in Figure 1) and can increase the cost of raising funds if investors anticipate that managers could expropriate funded resources (Arrow 4B in Figure 1) (Lambert, Leuz, and Verrecchia, 2005). In sum, the banter above suggests that information asymmetries between the firm and investors and between the principal and the agent can prevent efficient investment. In the next section, I converse how financial reporting quality can enhance investment efficiency by mitigating these information asymmetries. . 2. Role of financial reporting Financial reporting quality can be associated with investment efficiency in at least two ways. First, it is commonly argued that financial reporting mitigates adverse selection costs (Arrow 5 in Figure 1) by reducing the information asymmetry bet ween the 6 firm and investors, and among investors (Verrecchia, 2001). For instance, Leuz and Verrecchia (2000) find that a inscription to more disclosure reduces such information asymmetries and increases firm liquidity.On the other hand, the existence of information asymmetry between the firm and investors could lead suppliers of capital to discount the rakehellpile price and to increase the cost of raising capital because investors would infer that firms raising money is of a bad type (Myers and Majluf, 1984). Thus, if financial reporting quality reduces adverse selection costs, it can improve investment efficiency by reducing the costs of external financing and, as reasoned in more detail below, the potential for financial reporting quality to improve investment efficiency is greatest in firms facing financing constraints.Second, a large literature in accounting suggests that financial reporting plays a critical role in mitigating agency problems. For instance, financial acc ounting information is commonly employ as a direct input into compensation contracts (Lambert, 2001) and is an important source of information apply by shareholders to monitor managers (Bushman and Smith, 2001). Further, financial accounting information contributes to the monitoring role of dividing line commercializes as an important source of firmspecific information (e. g. Holmstrom and Tirole, 1993 Bushman and Indjejikian, 1993 Kanodia and Lee, 1998). Thus, if financial reporting quality reduces agency problems (Arrow 6 in Figure 1), it can then improve investment efficiency by increasing shareholder ability to monitor managers and thus improve project selection and reduce financing costs. 1 2. 3. Predictions For example, Bens and Monahan (2004) find a positive association between AIMR disclosure ratings and the extravagance value of diversification as defined by Berger and Ofek (1995).They conclude that disclosure plays a monitoring role in mitigating managements investmen t decisions. 1 7 Based on the discussion above that financial reporting affects both adverse selection and agency conflicts, I predict an come negative relation between financial reporting quality and both underinvestment and overinvestment. These links complement research in Bushman, Piotroski, and Smith (2005), which studies the relation between country measures of timely loss recognition and the country propensity to counterbalance bad projects (i. e. , itigate overinvestment), and in Wang (2003) which explores the relation between capital allocation efficiency and accounting information quality for a sample of US firms, without making a distinction between under- and overinvestment. 2 H1 Financial reporting quality is negatively associated with underinvestment. H2 Financial reporting quality is negatively associated with overinvestment. In addition to analyse the fair relation between financial reporting quality and investment efficiency, I also check up on the mechanisms t hrough which financial reporting quality can affect investment efficiency using cross-sectional analysis.First, I predict that the relation between financial reporting quality and firm underinvestment is stronger for firms facing financing constraints. By description, laboured firms are those for which the ability to raise funds is the most probable impediment to efficient investment, and for these firms, financial reporting quality is especially important in mitigating adverse selection costs. H3 The relation between financial reporting quality and underinvestment is stronger for financing bound firms. 2 matchless concern with Hypotheses 1 and 2 is that origin goes the other way. For instance, poorly performing managers could be investing inefficiently and thus choose to report low quality financial information in do to hide their bad performance (e. g. , Leuz, Nanda, and Wysocki, 2003). I discuss the empirical tests utilise to address this alternative dead reckoning in Sec tion 4. 8 Second, I predict that the relation between financial reporting quality and firm overinvestment is stronger for firms with large cash balances and allow cash plys.Managers of firms with large cash balances and supernumerary cash falls have more opportunity to engage in value destroying investment activities (e. g. , Jensen, 1986 Blanchard, Lopezde-Silanes, and Shleifer, 1994 Harford, 1999 Opler et al. , 1999 Richardson, 2006). Consequently, financial reporting quality can play a more important monitoring role in mitigating agency problems for these firms. H4 The relation between financial reporting quality and overinvestment is stronger for firms holding large cash balances and slack cash menstruates.Third, I study the complementary and substitute relation between financial reporting quality and a firms information environment, and how it affects investment efficiency. Financial reporting quality is just one source of information about the firms operations used by in vestors. For instance, investors in firms followed by a large number of analysts or firms with informative stock prices may be less open on financial reports when other elements of the firms information environment are of high quality.Thus I hypothesize that financial reporting quality is more important in up(a) investment efficiency when the amount of information publicly available about the firm is low. 3 H5 The relation between financial reporting quality and investment efficiency is stronger for firms with sexual congressly poor information environments. 3. Empirical work 3. 1. Proxies for investment efficiency One concern with Hypothesis 5 is that financial reporting quality and the firms information environment are belike to be match.Indeed, Verdi (2005) shows that the firm information environment can be aggregated in accounting-based and market-based cor think constructs. Hypothesis 5 implicitly assumes away this correlation by investigating the return of financial rep orting quality on investment efficiency holding the market-based information environment constant. 3 9 In order to construct measures of investment efficiency, I kickoff visualise a model that predicts firm investment levels and then use residuals from this model as a procurator for inefficient investment.The data are from the Compustat Annual file during the years 1980 to 2003. Total new Investment in a addicted firm-year is the sum of capital expenditures (item 128), R&038D expenditures (item 46), and acquisitions (item 129) minus sales of PPE (item 107) and depreciation and amortisation (item 125) cypher by 100 and scaled by average do assets (item 6), following Richardson (2006). This measure uses an accounting-based framework to estimate new investment as the difference between join investment and investment required for caution of assets in place.In the esthesia section I also discuss the robustness of the results to the use of only capital expenditures as an altern ative placeholder for investment that is oftentimes used in the literature (e. g. , Hubbard, 1998). I estimate a parsimonious model for investment demand as a function of growth opportunities measured by Tobins Q (Tobin, 1982). This model is based on the argument that growth opportunities should explain corporate investment when markets are perfect (Hubbard, 1998). Investmenti,t = ? 0 j,t + ? 1 j,t * Qi,t-1 + ? i,t (1) I estimate the model cross-sectionally for all industries with at least 20 observations in a given year based on the Fama and French (1997) 48- constancy classification. Q is calculated as the ratio of the market value of thorough assets (defined as 4 A large finance literature uses investment cash flow sensitivities as a proxy for inefficient investment (or market frictions).I do not use this approach for two reasons First, traditional papers measure cash flow without making the distinction between cash flows and accruals, and Bushman, Smith, and Zhang (2005) illu strate the sensitivity of the results to the clutch criterion of operating cash flows. Second, positive investment cash flow sensitivities could mean both financing constraints and/or agency problems which makes it impossible to test the cross-sectional hypotheses of the paper (Hypotheses 3 to 5). 10 otal assets (item 6) plus the product of stock price (item 199) and the number of common shares outstanding (item 199) minus the book value of equity (item 60)) to book value of amount of money assets (item 6) at the start of the fiscal year. The sample consists of 98,675 firm-year observations with available data to estimate Investment and Q during the sample flowing of 1980 to 2003. Consistent with previous literature, financial firms (i. e. , SIC codes in the 6000 and 6999 range) are excluded because of the different nature of investment for these firms.In order to mitigate the influence of outliers I winsorize all variants at the 1% and 99% levels by year. 5 submit 1 presents the results from the investment model in compare 1. harbor board A offers descriptive statistics for Investment and Q. The mean (median) firm in the sample invests 7. 26% (3. 84%) of heart and soul assets per year and has an average (median) Q make up to 1. 90 (1. 32), legitimate with connect literature (e. g. , Richardson, 2006 Almeida, Campello, and Weisbach, 2004). empanel B presents mean and median values of the estimated pains coefficients on Q, the average R-square, and the number of significant positive coefficients for from each one(prenominal) year. In all years the mean and median coefficients are positive and relatively stable during the sample period. The mean R-square ranges from 6% in 1997 to 14% in 1991. 6 Finally, in each year, more than half of the industry coefficients on Q are positive and statistically different from zero at a five percent significance level. 7The model in Equation 1 includes an tap which imposes that for each industry-year the mean fi rm will have a zero residual. In untabulated analysis, I re-estimate the model adding the intercept back to the residual so that it allows industry-years to have a non-zero mean (for example, industries that overinvest or periods with large economic growth). The results are robust (in general even stronger) to this test. 6 Note that the account R-squares measure only the within industry-year chromosomal mutation because the model is estimated separately for each industry-year.An equivalent approach in which the model is estimated across all industry-years with separate intercepts and coefficients for each industry-year leads to an R-square of 23. 5%, suggesting that the overall informative power of the model is larger than that reported in sidestep 1. 7 A legitimate ongoing debate in the finance literature is the implications for measurement error in the melodic theme of Q (Erickson and Whited, 2000 Gomes, 2001 Alti, 2003). Since the subsequent analysis hinges on the investmen t model in Equation 1, I perform two sensitivity tests First, I include past returns in 5 1 I measure investment efficiency using the residuals from the model in Equation 1. Overinvestment is the positive residuals of the investment model and Underinvestment is the negative residuals of the investment model multiplied by negative one, such that both measures are decreasing in investment efficiency. In untabulated analysis, I parallel all tests afterward excluding firms with the smallest 10% and 20% investment residuals because these firms are more likely to be stirred by measurement error in the investment model (i. e. , mis categorise as overinvesting or underinvesting firms).The results for these analyses are quasi(prenominal) to those reported below. fudge 1 Panel C presents descriptive statistics for Investment equilibrium, Overinvestment and Underinvestment. By construction, Investment Residual has a mean value of zero ranging from -64. 46% to 80. 43%. There are 39,107 ( 59,568) firms class as overinvesting (underinvesting) firms. The mean (median) value is 9. 73% (5. 63%) for Overinvestment and 6. 39% (4. 71%) for Underinvestment. These results show that the residuals from the investment model are more frequently negative, although in smaller magnitude.Panel D presents Pearson correlations between the measures of investment efficiency and firm characteristics. Investment Residual is uncor tie in with firm size (measured as the log of summate assets (item 6) at the start of the fiscal year) and slightly negatively correlated with return unpredictability (measured as the hackneyed deviation of effortless returns during the prior fiscal year). However, when the residuals are separated into Overinvestment and Underinvestment, I find that these changeables are negatively correlated with size and positively correlated with return volatility and Q (the magnitude of the he investment model to capture growth opportunities not reflected in Q (Lamont, 2000 Richardson, 2006) and second, I exclude all industry-year observations in which the estimated coefficient on Q is not positive and significant. The subsequent results are not bare-assed to these tests. 12 correlations range from 0. 18 to 0. 32). These results suggest either that (1) small firms, with more growth opportunities and vapourific operations, have more inefficient investment or (2) the investment model is a poor fit for these firms.In any case, it highlights the importance to control for these firm characteristics in the subsequent analysis. In order to better get a line the properties of the residuals from the investment model I perform analyses testing the persistence of investment efficiency over time. First, I find that 40% (48%) of the firms in the communicate (bottom) Investment Residual quintile in a given year remain in the top (bottom) quintile in the following year, and 27% (36%) remain third years later (Panel E).In addition, one lag of Investment Resi dual in an autoregressive model explains 16% of current Investment Residual (untabulated). The inclusion of higher orders of past residuals has a small contribution in explanatory power (R-square of only 18% if five lags are included in the model). These analyses suggest that residuals of the investment model are not random, which seems to support the view that they capture a firm investment characteristic. However, I cannot rule out the explanation that the persistence in the residuals is a function of an omitted correlated versatile in the investment model. . 2. Proxies for financial reporting quality The conceptual definition of financial reporting quality used in this paper is the the true with which financial reporting conveys information about the firms operations, in particular its expected cash flows, in order to inform investors in terms of equity investment decisions. This definition is lucid with the FASB SFAC No. 1 which states that one objective of financial report ing is to inform present and potential investors 13 in making rational investment decisions and in assessing the expected firm cash flows.I proxy for financial reporting quality using measures of accruals quality derived in prior work (Dechow and Dichev, 2002 McNichols, 2002) based on the idea that accruals are estimates of future cash flows, and earnings will be more representative of future cash flows when there is lower estimation error embedded in the accruals process (McNichols, 2002). 8 I estimate discretionary accruals using the Dechow and Dichev (2002) model augmented by the fundamental variables in the Jones (1991) model as suggested by McNichols (2002). The model is a regression of working(a) capital ccruals on lagged, current, and future cash flows plus the change in tax revenue and PPE. All variables are scaled by average total assets. Accrualsi,t = ? + ? 1*Cash adverti,t-1 + ? 2*CashFlowi,t + ? 3*CashFlowi,t+1 + ? 4*? Revenuei,t + ? 5*PPEi,t + ? i,t. (2) where Accrual s = (? CA ? Cash) (? CL ? STD) Dep, ? CA = substitute in current assets (item 4), ? Cash = Change in cash/cash equivalents (item 1), ? CL = Change in current liabilities (item 5), ? STD = Change in short-term debt (item 34), Dep = Depreciation and amortization expense (item 14), CashFlow = enlighten income before extraordinary items (item 18) minus Accruals ?Revenue = Change in revenue (item 12), and PPE = Gross property, plant, and equipment (item 7). All variables are deflate by average total assets (item 6). Following Francis et al. (2005), I estimate the model in Equation 2 crosssectionally for each industry with at least 20 observations in a given year based on the Fama and French (1997) 48-industry classification. AccrualsQuality at year t is the 8 I discuss the sensitivity of the results to the use of alternative measures of accruals quality and other attributes of earnings in Section 5. 4 step deviation of the firm-level residuals from Equation 2 during the years t-5 to t-1, assuring that all explanatory variables are measured before period t for the computation of AccrualsQuality in that year. I multiply AccrualsQuality by negative one so that this variable becomes increasing in financial reporting quality. As discussed in Dechow and Dichev (2002) and McNichols (2002), the estimation of AccrualsQuality captures the absolute variation in the residuals of Equation 2 rather than the variation relative to a benchmark.One concern with this approach is that AccrualsQuality may be capturing some underlying degree of volatility in the business, and the results in mesa 1 show that investment efficiency is negatively correlated with firm uncertainty. Thus, I follow the suggestion in McNichols (2002) and create a relative measure of accruals quality. In particular, I measure AccrualsQualityRel as the ratio of the quantity deviation of the residuals from Equation 2 during the years t-5 to t-1 to the amount deviation of total accruals during the years t -5 to t-1 multiplied by negative one.This measure captures the relative partition of the estimation errors in accruals compared to the total variance. I show below that this measure is only slightly correlated with firm size and cash flow volatility, mitigating the concern that the proxies for financial reporting quality are associated with investment efficiency because of the spurious force out of firm uncertainty. 4. Results To investigate hypotheses 1 and 2, I first present preliminary analysis on the univariate relation between the measures of investment efficiency and financial reporting quality.Table 2 Panel A presents descriptive statistics for a smaller sample than reported in Table 1 due to data availability for AccrualsQuality and AccrualsQualityRel. 15 The sample consists of 49,543 firm-year observations and all variables are winsorized at the 1% and 99% levels by year. In this sample, there are 19,473 (30,070) firms classified as overinvesting (underinvesting) firms. The mean (median) value for Overinvestment is 7. 81% (4. 45%) and for Underinvestment is 5. 37% (4. 09%).The magnitudes are smaller than reported in Table 1 because the data required to estimate AccrualsQuality and AccrualsQualityRel bias the sample toward larger firms. Among the financial reporting quality proxies, the mean (median) firm in the sample has an AccrualsQuality of -0. 04 (0. 03) and an AccrualsQualityRel of -0. 74 (-0. 64). Finally, I include descriptive statistics on firm size, cash flow volatility, and Tobin Q because these firm characteristics are shown to be associated with investment efficiency in Table 1. The distribution of Q is slightly changed (as compared to Table 1) to a mean (median) Q of 1. 63 (1. 23) again reflecting the sample bias toward larger firms. Panel B presents Pearson (Spearman) correlations above (below) the main diagonal for the variables in Panel A. By construction, Overinvestment and Underinvestment cannot be correlated because each firm-ye ar observation can only be in one group. Most importantly, Overinvestment is negatively correlated with AccrualsQuality (Pearson correlation equals -0. 19) and with AccrualsQualityRel (Pearson correlation equals -0. 8) the alike is true for Underinvestment (Pearson correlations equal -0. 22 and -0. 10 respectively). These results present preliminary evidence for the relation between financial reporting quality and investment efficiency in hypotheses 1 and 2. Finally, as in Dechow and Dichev (2002), AccrualsQuality is highly correlated In Table 1, I use return volatility instead of cash flow volatility to avoid imposing the five-year data requirement for the estimation of cash flow volatility. However, this data is required to estimate AccrualsQuality and does not impose any sample bias at this stage of the analysis.I use cash flow volatility in the remainder of the paper because AccrualsQuality is highly correlated with cash flow volatility as discussed by Dechow and Dichev (2002). However, the results are not sensitive to this choice. 9 16 with Size (Pearson correlation equals 0. 42) and with CashFlowVol (Pearson correlation equals -0. 66). However, note that AccrualsQualityRel is much less correlated with these variables (correlations of -0. 08 and 0. 04 with size and cash flow volatility respectively), supporting the argument that this variable is uncorrelated with firm uncertainty. 0 Table 3 presents the multiple regressions. The estimated model is a regression of investment efficiency on financial reporting quality, firm characteristics, and industry (based on the Fama and French (1997) 48-industry classification) and year fixed kernels. The dependent variable is Underinvestment in the first two columns and Overinvestment in the remaining columns. All standard errors are clustered by firm using the HuberWhite procedure. 11 As predicted in hypothesis 1, Underinvestment is negatively related to AccrualsQuality and AccrualsQualityRel (both coefficients are significant at 1% level).The estimated coefficients are also negative and significant for Overinvestment, supporting the prediction in hypothesis 2. The estimated coefficients suggest that increasing AccrualsQuality (AccrualsQualityRel) by one standard deviation is associated with a reduction on Underinvestment of 0. 21% (0. 11%) and on Overinvestment of 0. 31% (0. 22%). Given that the mean values for Underinvestment and Overinvestment in Table 2 are 5. 73% and 7. 81%, these changes average between 1% and 5%, suggesting that the economic significance of the effect is moderate.One alternative explanation for the results in Table 3 is that causality goes the other way. For instance, suppose that poorly performing managers are more likely to The signs of the correlations between AccrualsQuality and size and cash flow volatility are the setback of the ones presented in Dechow and Dichev (2002) because I multiply AccrualsQuality by negative one so that this variable is increasing in re porting quality. 11 Petersen (2005) suggests two methods to correct for both cross-sectional and time-series dependence in the data the Huber-White procedure and familiarized Fama-MacBeth.Since, neither method is perfect, I repeat all subsequent analysis using Fama-MacBeth (1973) estimators adjusting for time-series dependence. The results lead to the alike inferences as reported in the text. 10 17 invest inefficiently and also choose to report low quality financial information in order to hide their bad performance (e. g. , Leuz, Nanda, and Wysocki, 2003). Then one could spuriously find a positive association between financial reporting quality and investment efficiency. In order to address this concern, I perform two tests.First, I repeat the analysis using the financial reporting quality proxies lagged by two periods (the variables in the model are already lagged by one period). Second, I explicitly control for past investment efficiency in the model. The suspicion behind this test is that if past investment efficiency drives financial reporting quality then there should be no relation between financial reporting quality and future investment efficiency after controlling for past investment efficiency. Table 4 Panel A presents the results of the two sensitivity analyses when Underinvestment is used as the dependent variable.When AccrualsQuality and AccrualsQualityRel (Columns I and II) are lagged by two periods, the inferences are unchanged. The estimated coefficients are statistically negative at conventional levels. In Columns III and IV, I include past Underinvestment in the model. In this case, the estimated coefficient on AccrualsQuality is still negative and significant, while the coefficient on AccrualsQualityRel is negative but only marginally significant (two-sided p-value of 0. 14). Table 4 Panel B repeats the analysis for Overinvestment.Again, all the inferences are unchanged since the estimated coefficients on AccrualsQuality and AccrualsQu alityRel are statistically negative in all models. Overall, the results in Tables 3 and 4 support hypotheses 1 and 2 that financial reporting quality is negatively associated with both underinvestment and overinvestment, 18 consistent with the argument that financial reporting mitigates both adverse selection and agency costs. 4. 1. Cross-sectional separates In this section, I discuss the empirical approach used to test hypotheses 3, 4, and 5.These hypotheses involve cross-sectional predictions about the relation between financial reporting quality and investment efficiency across sub-groups of the sample. Thus, I estimate separate coefficients for these sub-groups as described in the model below (Investment Inefficiency) i,t = ? 0 + ? 1* Partition i,t-1 + ? 2* ReportingQuality i,t-1 + ? 3* ReportingQuality* Partition i,t-1 + ? 4* Controls i,t-1 ? ? t * Year t + ? ? j * application j + ? it. where Investment Inefficiency is either Underinvestment or (3) Overinvestment, ReportingQu ality is either AccrualsQuality or AccrualsQualityRel.Partition is coded as an indicator variable based on measures of financing constraints, excess cash, or information environment described below (results are similar if the Partition is used as a continuous or rank (deciles) variable). The partitioning variables are defined such that a negative coefficient on the fundamental interaction term (? 3) implies that the relation between financial reporting quality and inefficient investment is stronger for firms in the subgroup of interest (e. g. , financially strained firms). As additional analysis, I test the null hypothesis that the sum of the coefficients ? and ? 3 is equal to zero in order to test whether the relation between financial reporting quality and investment efficiency is at least present in the sub-group of interest. 12 12 Hypotheses 3 to 5 are also important in mitigating the concern that an omitted correlated variable could be driving the positive association betwee n financial reporting quality and investment efficiency. For instance, if managers choose better (worse) investment projects and report more (less) informative financial accounting information when they know more (less) about growth opportunities and expected cash flows, 9 4. 1. 1. Financing Constraints In this section, I investigate hypothesis 3 which predicts that the relation between financial reporting quality and Underinvestment is stronger for financing constrained firms because these firms are, by definition, limited in their ability to raise funds. I follow the approach in Hubbard (1998) to ramify firms into financially constrained and free categories. In particular, I use five different criteria because of the escape of consensus about which approach provides the best classification (Almeida, Campello, and Weisbach, 2004).First, I classify firms into Payout forced if the firm is in the bottom three quartiles in terms of total payout in a given year and free otherwise. I measure total payout as the sum of dividends and share repurchases deflated by year-end market capitalization using the method described in Boudoukh et al. (2005). Second, I classify firms into Age Constrained if the firm is in the bottom three quartiles of firm age in a given year (and unconstrained otherwise) based on the argument that young firms are more likely to face financing constraints.Age is measured as the difference in years since the first year the firm appears in the CRSP database. Third, I classify firms into Size Constrained if the firm is in the bottom three quartiles of total assets in a given year and unconstrained otherwise. Fourth, I measure Rating Constrained if the firm has long-term debt outstanding (item 9) but does not have public debt rated by S&038P (item 280) and unconstrained otherwise. Finally, I construct the KZ forefinger following the approach in Kaplan and Zingales (1997) and classify a firm as KZ indicant Constrained hen a positive relation be tween financial reporting quality and investment efficiency could just be a reflection of the quality of the managers information set and might not be related to financial reporting quality. However, this alternative hypothesis would not predict the relation between financial reporting quality and investment efficiency to be dependent on financing constraints, cash balances, or the existing information environment. Thus, if such interactions exist, then it would strengthen the result that financial reporting quality per se is associated with investment efficiency. 0 if the firm is in the top three quartiles of the KZ exponent in a given year and unconstrained otherwise. 13 Untabulated analysis show that the first four classifications are positively correlated (Pearson correlations ranging from 0. 11 to 0. 45) but the KZ Index classification is not correlated with the remaining criteria (Pearson correlations ranging from -0. 01 to 0. 11), consistent with previous research (e. g. , Al meida, Campello, and Weisbach, 2004). 14 Further, all financing constraint proxies are positively correlated with Underinvestment (Pearson correlations range from 0. 1 to 0. 14). Table 5 presents the results related to hypothesis 3. All models include the control variables size, cash flow volatility, Q, and industry and year fixed effects as before but the coefficient estimates on these variables are not tabulated for brevity. The estimated coefficients on the control variables are similar to those reported in Table 3. The results are separated for AccrualsQuality and for AccrualsQualityRel. For AccrualsQuality, the estimated coefficients on the main effect (third column labeled Reporting Quality) are all egative with only one statistically significant coefficient. These results indicate that, for a sample of unconstrained firms, the relation between AccrualsQuality and Underinvestment is basically not significant. The estimated coefficients on the interaction terms, however, are ne gative in four out of five cases and significant in two. Further, the F-test rejects the hypothesis of no relation between AccrualsQuality and Underinvestment in almost all cases for the sample of financially constrained firms. The only exception is 3 The KZ Index is calculated using the following formula KZ Index = -1. 002 * CashFlow + 0. 283 * Q + 3. 139 * Leverage 39. 368 * Dividends 1. 315 * Cash. For more details see Almeida, Campello, and Weisbach (2004, p. 1790). 14 Principal component analysis on the five financing constraints proxies yields two factors. The first factor explains 40% of the variation and loads on all proxies but the KZ Index. The second factor explains some other 20% of the variation in the data and loads on the Payout and the KZ Index measures. 1 when the KZ Index is used as the criteria for financing constraint classification. 15 When AccrualsQualityRel is used as the financial reporting quality proxy, the results are by and large the similar. In terms of economic significance, increasing AccrualsQuality (AccrualsQualityRel) by one standard deviation is associated with a reduction in Underinvestment of 0. 26% (0. 16%) for firms classified as Rating Constrained and 0. 08% (0. 06%) for unconstrained firms (compared to 0. 21% (0. 11%) for the full sample as discussed above).Overall, the results present marginal support for hypothesis 3 that the relation between financial reporting quality and Underinvestment is stronger for financing constrained firms. 4. 1. 2. Cash Balances In this section, I investigate hypothesis 4 which predicts that the relation between financial reporting quality and Overinvestment is stronger for firms with large cash balances and free cash flows because these firms are more likely to overspend existing resources (Jensen, 1986). I use two criteria to classify firms based on cash holdings and one proxy for free cash flow.First, I create an indicator variable, juicy Cash, coded as 1 if the firm is above the me dian in the distribution of cash balances deflated by total assets in a given year and 0 otherwise. Second, I follow the approach in Opler et al. (1999) who predict cash balances as a function of firms characteristics, and use residuals from this model as a proxy for excess cash. Opler et al. show that firms hold more cash in the presence of growth opportunities and firm uncertainty, and less cash when they are forced to payout interest obligations and have more access to financing (proxied by leverage and size).Thus, I estimate annual regressions of cash balances (item 1) deflated by total 15 The inconsistent result using the KZ Index is consistent with prior work in the finance literature (e. g. , Almeida, Campello, and Weisbach, 2004 Almeida and Campello, 2005) which finds opposite results when this variable is used as a proxy for financing constraints. 22 assets (item 6) on firm size, leverage, Q, and cash flow volatility. Leverage is measured as the sum of the book value of sho rt term (item 34) and long term debt (item 9) deflated by the book value of equity (item 60) and the remaining variables are the same as described above.The explanatory power of the models ranges from 16% in 1986 to 42% in 2003. I create an indicator variable, Excess Cash, coded as 1 if the firm has a positive residual from the model predicting cash balances, and 0 otherwise. Finally, following Richardson (2006), Free Cash Flow is equal to cash flow from operations plus R&038D expenses minus depreciation and the predicted investment for the firm as estimated in Table 1. Free Cash Flow is recoded as an indicator variable coded as 1 if the computation of free cash flow is positive and 0 otherwise.Table 6 presents the results related to hypothesis 4. As before, all models include the control variables size, cash flow volatility, Q, and industry and year fixed effects (estimates not tabulated). The first set of results presents estimated coefficients for AccrualsQuality and the second r eports coefficients for AccrualsQualityRel. The results show that the estimated coefficients on the main effect of financial reporting quality are negative but not significant in all six models (three models for AccrualsQuality and three for AccrualsQualityRel).The estimated coefficients on the interaction term, on the other hand, are negative in all cases and significant in three out of six cases, and the F-test rejects the hypothesis of no relation in all cases. In terms of economic significance, increasing AccrualsQuality (AccrualsQualityRel) by one standard deviation is associated with a reduction on Overinvestment of 0. 41% (0. 35%) for firms classified as High Cash and 0. 06% (0. 06%) for firms with low cash (compared to 0. 31% (0. 22%) for the full sample as discussed above).Overall, the results support hypothesis 4 by showing that the 23 relation between financial reporting quality and Overinvestment is stronger for firms with large and excessive cash balances but the result s are not statistically significant for firms generating free cash flows. This support the hypothesis that financial reporting quality reduces firm overinvestment by lowering shareholders cost of monitoring managers and thus limiting managers ability to undertake inefficient investment projects. 4. 1. 3.Information Environment In this section, I investigate hypothesis 5 which predicts that the relation between financial reporting quality and investment efficiency is stronger for firms with poor information environments because investors of these firms are more likely to rely on financial accounting information to infer the economic conditions of the firms operations. I use two proxies for the firm information environment the number of analysts following the firm and the bid-ask spread. I use the number of analysts following a firm as a proxy for the amount of publicly available information about the firm. analysts are an important source of information for investors they expel fore casts, reports about individual companies, and stock recommendations. Roulstone (2003) examines the role of analysts in improving market liquidity and finds that analysts provide public information that reduces information asymmetries between firms and market participants. I collect data on analyst following from IBES and measure the number of analysts following the firm as the maximum number of analysts vaticination annual earnings for a firm during the fiscal year t.If the firm is not followed by IBES I assume that the number of analysts following the firm is zero. I consider a firm as Low Analyst if the firm is in the bottom three quartiles in a given year (coded as 1 and 0 otherwise). 24 The second proxy for a firms information environment is the bid-ask spread. See Amihud and Mendelson (1986) and Roulstone (2003) among others for discussions of spreads as a proxy for the information asymmetry between the firm and investors.I collect intraday trade data to compute bid-ask sprea d from the Trades and Quotes database (TAQ) and from the Institute for the take away of Security Markets database (ISSM). The TAQ database includes trades and quotes starting in 1993, and the ISSM database contains intraday data for NYSE/AMEX firms from 1983 to 1992 and for NASDAQ firms from 1987 to 1992. I measure quoted bid-ask spread as the ask price minus the bid price divided by the average of the bid and ask prices. The bid-ask spread is averaged across all transactions during the day for each firm, then periodic mean bid-ask spreads are averaged during the month t.Finally I compute bid-ask spread as the average of the monthly bid-ask spreads during the fiscal year t. I consider a firm as High Spread if the firm is in the top three quartiles in a given year (coded as 1 and 0 otherwise). Table 7 presents the results related to hypothesis 5. As before, all models include the control variables (estimates are untabulated). The table is divided into Underinvestment and Overinvest ment results. The first set of results presents estimated coefficients for AccrualsQuality and the second reports coefficients forAccrualsQualityRel. When bid-ask spread is used as the partitioning variable, I find that none of the coefficients on the main effect of financial reporting quality are significant, and three out of four coefficients on the interaction term are significant. The only exception is the coefficient on the interaction between High Spread and AccrualsQualityRel for Underinvestment. Further, in three out of four cases the F-test rejects the hypothesis of no effect of financial reporting quality on investment efficiency 25 for the sample of firms with High Spread.As for Low Analyst, the results on the estimated coefficients on the interaction terms are weaker only one coefficient is statistically negative. Still, in three out of four models the F-test rejects the hypothesis of no relation for the sample of firms with Low Analyst. Overall, the results provide weak support for the hypothesis that the effect of financial reporting on investment efficiency is more important when the firm information environment is of low quality. 16 5. Sensitivity digest In this section I discuss some robustness tests to the analysis presented in the paper.First, I study the sensitivity of the results to inclusion of omitted control variables using firm fixed-effect estimation. The advantage of this approach is that it controls for all time-invariant unobservable firm characteristics. However, since the estimation of AccrualsQuality and AccrualsQualityRel is done using five years of data, the within-firm variation is small, which makes the fixed-effect estimation precise conservative. The analysis is done for all firms with at least five, ten, or 15 years of data in order to increase the within firm variation (sample sizes of 43,739, 33,454, and 24,420 firm-year observations respectively).Untabulated analyses show that the results in Hypotheses 1 and 4 are generally robust to the firm fixed-effect estimation. Results of Hypotheses 2 and 3 are weaker (coefficients are of the same sign but in most cases not significant at conventional levels) and, in the case of Hypothesis 5, the results are similar (weaker) when Underinvestment (Overinvestment) is used as the dependent variable. I also performed tests using a 2&2152 classification based on the firms financial reporting quality and information environment (sorted independently as a low/high).Either high financial reporting quality or high information environment is sufficient to mitigate Underinvestment but only financial reporting quality is sufficient to mitigate Overinvestment, suggesting a substitute relation between financial reporting quality and the firm information environment in improving investment efficiency. 16 26 Second, I investigate the sensitivity of the results to the use of alternative measures of accruals quality such as the non-linear discretionary accruals model in Ball and Shivakumar (2005) and the accrual quality measures develop by Wysocki (2006).The key innovation in Wysockis (2006) measures is to remove the suaveness effect of accruals in the Dechow and Dichev (2002) model. Results using the Ball and Shivakumar (2005) model are real similar to those reported on the paper. The use of Wysockis measure, on the other hand, leads to similar results for hypotheses 1, 2, and 5 but unnoticeable results for hypotheses 3 and 4. As discussed in more detail below, these results are not surprising given that Wysockis (2006) measure excludes the smoothness component of accruals, and smoothness is positively associated with investment efficiency.In addition, I investigate the sensitivity of the results to the use of alternative attributes of earnings as proxies for financial reporting quality. Accruals quality represents one dimension of financial reporting quality but other dimensions of earnings have also been used as a proxy for financial reporti ng quality (Francis et al. , 2004). These attributes of earnings would not necessarily affect investment efficiency in the same way.For instance, one could argue Timeliness and Conservatism are more important in conveying information about bad firms economic states, thus improving Overinvestment but may not be associated with Underinvestment. Nevertheless, it is useful to see how these measures are related and the respective association with investment efficiency (Verdi, 2005). Francis et al. (2004) identify six earnings attributes (other than AccrualsQuality) previously used in accounting research to characterize desirable features of earnings. The six attributes are Persistence, Predictability, Smoothness, 27 ValueRelevance, Timeliness, and Conservatism.I also include a measure of price informativeness as used by Durnev, Morck, and Yeung (2004). When Underinvestment is used as the dependent variable (Hypotheses 1 and 3), I find consistent results using Persistence, Predictability, and Smoothness but insignificant results for the remaining variables (with the exception of Informativeness in which the relation is positive and significant, against the prediction). The analysis using Overinvestment (Hypotheses 2 and 4) yield weaker results since only the estimated coefficients on Smoothness and Informativeness are negative and significant in the expected direction.The remaining coefficients are either insignificantly negative or positive in the case of Persistence. Overall the results provide marginal support for the relation between other dimensions of earnings and Underinvestment, and weak support for Overinvestment. The finding that Smoothness is negatively associated with both Underinvestment and Overinvestment explains the weaker results using Wysockis measure of accruals quality given that this measure excludes the smoothness component in the accruals quality measure developed by Dechow and Dichev (2002).In the third sensitivity test, I repeat the analysis using capital expenditures (deflated by average total assets) as a measure of investment in order to make the results more comparable with the extant finance literature. In addition, the investment measure used in the paper includes only cash acquisitions and ignores stock acquisitions which constitute the majority of M&038A transactions. Untabulated analyses using CAPEX show that the results in Hypothesis 1, 3, and 5 are similar to those reported. Results in Hypothesis 2 are consistent but weaker when AccrualsQuality is used as the proxy for 28 inancial reporting quality. Finally, results are inconsistent with Hypothesis 4 (estimated coefficients on the interaction terms are mostly insignificant). Finally, I include grace of God (item 204) in the discretionary accruals model. As discussed in Jones (1991), PPE is included in the model to capture the blueprint level of depreciation, and using the same logic, goodwill would capture the normal level of amortization in accruals. This inclusion is justified because the measure of investment includes acquisitions. Goodwill is only available from Compustat starting in 1988 which is why it is excluded in the main tests.In untabulated analysis I find little invasion on the discretionary accruals model (the Pearson correlation between discretionary accruals including and excluding goodwill is 0. 99), and the results presented in the paper are unchanged if I restrict the sample to post 1988 and include goodwill in the discretionary accruals model. 6. Summary and conclusion Despite recent claims that financial reporting quality can have economic implications for investment efficiency, there is little evidence on this relation empirically. This paper studies the relation between financial reporting quality and investment efficiency.The analysis is done on a sample of 49,543 firm-year observations during the sample period of 1980 to 2003. I find that proxies for financial reporting quality, namely measures of accruals q uality, are negatively associated with both firm underinvestment and overinvestment. The relation between financial reporting quality and underinvestment is stronger for firms facing financing constraints, consistent with the argument that financial accounting information can reduce the information asymmetry between the firm and investors, and 29 thus lower the firms cost of raising funds.Likewise, the relation between financial reporting quality and overinvestment is stronger for firms with large cash balances, which suggests that financial reporting quality can reduce the information asymmetry between the principal and the agent and thus lower shareholders cost of monitoring managers and improving project selection. Finally, I find that the relation between financial reporting quality and investment efficiency is stronger for firms with low quality information environments. Overall, this paper contributes to the extant accounting literature that investigates the economic implicati ons of enhanced financial reporting.This literature has shown that financial reporting quality has economic consequences such as increased liquidity, lower costs of capital, and higher firm growth (e. g. , Leuz and Verrecchia, 2000 Francis et al. , 2004, 2005 Martin, Khurana, and Pereira, 2005). This paper extends this research by showing that financial reporting information can reduce information asymmetries that impede efficient corporate investment policies. 30 References Almeida, H. , and M. Campello, 2005. Financial constraints, asset tangibility, and corporate investment, working paper, New York University. Almeida, H. , M. Campello, and M. Weisbach, 2004.The cash flow sensitivity of cash. daybook of Finance 59, 1777-1804. Alti, A. , 2003. How sensitive is investment to cash flow when financing is frictionless? Journal of Finance 58, 707-722. Amihud, Y. and H. Mendelson, 1986. Asset pricing and the bid-ask spread, Journal of Financial Economics 17, 223-249. Ashbaugh, H. , J. Gassen, and R. LaFond, 2005. Does stock price synchronicity represent firm-specific information? The multinational evidence, working paper, University of Wisconsin Madison. Ball, R. , and L. Shivakumar, 2005. Earnings quality in U. K. private firms. Journal of business family and Economics 39 (1), 83-128.Bens, D. , and S. Monahan, 2004. Disclosure quality and the excess diversification. Journal of Accounting Research 42 (4), 691-730. value of Berger, P. , and E. Ofek, 1995. Diversifications effect on firm value. Journal of Financial Economics 37, 3965. Berle, A. , and G. Means, 1932. The modern slew and private property (Macmillan, New York). Bertrand, M. , and S. Mullainathan, 2003. Enjoying the quiet life? Corporate government and managerial preferences. Journal of Political economy 111, 10431075. Blanchard, O. , F. Lopez-de-Silanes, and A. Shleifer, 1994. What do firms do with cash windfalls?Journal of Financial Economics 36, 337-360. Botosan, C. , 1997. Disclosure level a nd the cost-of-equity-capital. The Accounting reassessment 72, 323-349. Boudoukh, J. , R. Michaely, M. Roberts, and M. Richardson, 2005. On the measure of payout yield and implication for asset pricing. Journal of Finance, forthcoming. Bushman, R. , and R. Indjejikian, 1993a. shareholder demand for distorted accounting disclosures. The Accounting Review 68 (4), 765782. 31 Bushman, R. , J. Piotroski, and A. Smith, 2005. Capital allocation and timely accounting recognition of economic losses international evidence, working paper, University of Chicago.Bushman, R. , and A. Smith, 2001. Financial accounting information and corporate governance. Journal of Accounting Economics 31, 237-333. Bushman, R. , A. Smith, and F. Zhang, 2005. Investment-cash flow sensitivities are really capital investment-working capital investment sensitivities, working paper, University of North Carolina at Chapel Hill. Dechow, P. , and I. Dichev, 2002. The quality of accruals and earnings the role of accrual estimation errors. The Accounting Review 77, supplement, 35-59. Durnev, A. , R. Morck, and B. Yeung, 2004. Value-enhancing capital budgeting and firmspecific stock return variation.Journal of Finance 59 (1), 65-105. Erickson, T. , and T. Whited, 2000. Measurement error and the relationship between investment and Q. Journal of Political Economy 108, 1027-1057. Fama, E. , and K. French, 1997. Industry costs of equity. Journal of Finance 43, 153-193. Fama, E. and J. MacBeth, 1973. Risk, return, and equilibrium empirical tests. Journal of Political Economy 81, 607636. FASB, 1978. Statement of Financial Accounting Concepts No. 1. Objectives of Financial Reporting by Business Enterprises. Francis, J. , R. LaFond, P. Olsson, and K. Schipper, 2004. Cost of equity and earnings attributes.The Accounting Review 79 (4), 967-1010. Francis, J. , R. LaFond, P. Olsson, and K. Schipper, 2005. The market pricing of accruals quality. Journal of Accounting and Economics 39 (2), 295-327. Gomes, J. , 20 01. Fi