82 Evidence of the Relationship Between Credit Ratings and Reporting Discontinued Operations Douglas K. Schneider, Ph.D. East Carolina University Denise Dickins, Ph.D. East Carolina University Mark G. McCarthy, Ph.D. East Carolina University Dennis O’Reilly, Ph.D. East Carolina University To test whether standard setters’ objective of improving the usefulness of the financial statements by enacting Accounting Standards Update (ASU) 2014-08 was achieved, this study compares the relationship between credit ratings and discontinued operations under Statement of Financial Accounting Standard (SFAS) No. 144 and ASU 2014-08. If discontinued operations are interpreted by credit ratings agencies as non-recurring, they should have no or low persistence and should be unrelated to credit ratings. The study finds that the relationship between reported discontinued operations and credit ratings under SFAS No. 144 is significant, the implication being that credit ratings agencies perceived discontinued operations as recurring. In contrast, the relationship is insignificant under ASU 2014-08, discontinued operations are now viewed as non-recurring and the objective of standard setters was achieved. The results of the study contribute to extant literature on discontinued operations and the relevance of separately stated or disclosed items. Keywords: Discontinued Operations; Credit Ratings; APB No. 30; SFAS No. 144; ASU 2014-08 Data Availability: All data are available from public sources. Introduction The Financial Accounting Standards Board’s (FASB) Conceptual Framework recognizes that one objective of financial reporting is to provide information that is useful for assessing the amount, timing, and uncertainty of an entity’s future cash flows (FASB, 2010). A guiding philosophy is that an income statement is more useful for this purpose if non-recurring items are separated from income from continuing operations. This study examines whether ASU 2014-08, Reporting Discontinued http://journals.sfu.ca/abr ADVANCES IN BUSINESS RESEARCH 2020, Volume 10, pages 82-94 83 Operations and Disclosures of Disposals of Components of an Entity, improved the usefulness of financial statements by testing whether the relationship between credit ratings and discontinued operations changed after its implementation compared to the earlier period when discontinued operations were defined more broadly under Statement of Financial Accounting Standard (SFAS) No. 144. Credit ratings are important due to their significant impact on financial markets and financing decisions (e.g., Blume et al., 1998; Hand et al., 1992; Kisgen, 2006), and financial information impacts credit ratings (e.g., Blume et al., 1998; Ziebart & Reiter, 1992). The study’s results inform academics, standard- setters, and financial statement users. For the most part, the results of extant literature support that gains and losses reported as special items are transitory (Jones & Smith, 2011), or at least more transitory than other components of income (Burgstahler et al., 2002). But, Dechow and Ge (2006) found that the persistence of core earnings can be impacted by special items if investors misunderstand their transitory nature. Separately stated items should have less persistence than recurring or core earnings and therefore should have little or no relationship with market valuations such as stock price and credit ratings. While the presentation of discontinued operations as a separately stated item in the income statement has been consistent over the last 45 years, the rules for what qualifies as a discontinued operation have changed. Beginning in 1973, under Accounting Principles Board (APB) Opinion No. 30, asset disposals qualifying for discontinued operations accounting were narrowly defined as a component of a business that represented a major line or class of customer. SFAS No. 144 replaced APB No. 30 in 2001 and expanded the type of asset disposals that qualified by redefining a component as comprising “operations and cash flows that can be clearly distinguished, operationally and for financial reporting purposes, from the rest of the entity” (SFAS No. 144 – codified under ASC 205- 20-20, 2). It also eliminated the requirement that the disposed component represent a major line or class of customer. In 2014, ASU 2014-08 revised the criteria for discontinued operations treatment to again be more restrictive by adding a requirement that the disposal represent a “major strategic shift” for the company (ASU 2014-08, 69). In each instance, the standard setters’ goal was to improve the usefulness of the resulting financial statements (e.g., SFAS No. 144, 6). Given the more inclusive definition of discontinued items under SFAS No. 144, it may have been that the FASB believed under APB No. 30 companies were including at least some non-recurring items in continuing operations. There is evidence that supports the misclassification of non-recurring items. For example, comparing the persistence of continuing operations under the two regimes, Curtis et al. (2014) found the broader scope of discontinued operations under SFAS No. 144 produced more persistent income from continuing operations among firms reporting discontinued operations. By returning to a more restrictive definition of discontinued operations under ASU 2014-08, it appears standard setters believe SFAS No. 144 was too expansive, and the results of extant research suggest they may be correct. For example, Barua et al. (2010) found that companies shift operating expenses to discontinued operations to increase core earnings and to meet or beat analysts’ forecasts. SFAS No. 144’s definition of discontinued operations may have enabled managers to smooth or otherwise manipulate earnings (Barua et al., 2010; Dickins et al., 2017). The present study finds that the relationship between reported discontinued operations and credit ratings under SFAS No. 144 is significant; credit ratings reflect that at least some of discontinued operations were recurring. In contrast, the relationship is insignificant under ASU 2014-08; discontinued operations are now more likely to be viewed as non-recurring and the objective of standard setters was achieved. These results contribute to extant research about the goal of the Conceptual Framework to provide decision-useful financial reporting (e.g., Burgstahler et al., 2002; Dechow & Ge, 2006; Jones & Smith, 2011). The results also address a gap in the literature on discontinued operations (e.g., Barua et al., 2010; Curtis et al., 2014). 84 Background, Literature Review, and Hypothesis Background In 1973, the Accounting Principles Board (the Board) issued APB No. 30 to, among other things, provide criteria to assist preparers in determining how to account for the disposal of a segment of the business. The Board noted that many accountants believe the income statement is more useful if the results of continuing operations are reported separately from the operations of a segment of the business that has been or will be discontinued.1 APB No. 30 mandated such treatment. APB No. 30 narrowly defined a segment of a business as a component of an entity that represents a major line of business or class of customer. It specifically noted that companies should distinguish the disposal of a business segment from those asset disposals that occur in the normal course of operating a business. The Board stated that the results of discontinued operations were required to be reported after income from operations but not as an extraordinary item. The 2001 release of SFAS No. 144 expanded the definition of disposals to be accounted for as discontinued operations to include “component operations.” A component “comprises operations and cash flows that can clearly be distinguished, operationally and for financial reporting purposes, from the rest of the entity” (SFAS No. 144, 5). This standard also eliminated APB No. 30’s requirement that the disposed component represent a major line of business or class of customer. Instead, sales of individual buildings, or the closing of individual stores or plants could qualify as discontinued operations. Not surprisingly, and consistent with the expectations of the FASB (SFAS No. 144, 6), the number of U.S. publicly-traded companies reporting discontinued operations was significantly greater under SFAS No. 144 (mean of 315 per year from 2002 to 2014) than under APB No. 30 (mean of 88 per year from 1989 to 2001) (difference of 227, p < 0.001). Literature Review While there are several studies that analyze the impact of financial information or events on credit ratings, none address the current study’s contribution, the relevance of discontinued operations under varying accounting standards. For instance, Jorion et al. (2009) provided evidence that decreases in credit ratings are associated with certain changes in accounting quality. Ayres et al. (2010) investigated whether credit analysts incorporate information from tax reporting alternatives in addition to financial data prepared using Generally Accepted Accounting Principles (GAAP). They reported evidence that book–tax changes signal negative information to credit rating agencies. More closely related to this investigation, Chen et al. (2015) tested whether credit ratings are influenced by underfunded multiemployer pension plan obligations (MEPPs) under two different accounting standards. The newer standard, effective in 2011, required more disclosure related to underfunded MEPPS. They found that both before and after the passage of the accounting standard, the underfunding was negatively associated with a firm’s credit rating suggesting information conveyed in the expanded disclosure was already known to creditors. Other studies provided evidence consistent with the idea that rating agencies use both public and private information when assigning credit ratings (Ziebart & Rieter, 1992; Hand et al., 1992). Specific to the impacts of discontinued operations, Curtis et al. (2014) examined the influence of changes in accounting on the usefulness of disaggregated income components in predicting a company’s future income. They found the broader scope of discontinued operations under SFAS No. 144 produced more persistent income from continuing operations among firms reporting 1 Interestingly, APB 30, paragraph 5, references “accountants” vs. users of financial information. Also, while the FASB does not explicitly express agreement with this view, agreement is implicit in that the FASB adopted provisions in line with this view. 85 discontinued operations. Similarly, under SFAS No. 144, Barua et al. (2010) found that firms shift losses to discontinued operations to meet or beat analysts’ expectations. Hypothesis The semi-strong version of the efficient market hypothesis (EMH – Fama, 1965; Fama, 1970) suggests that market prices reflect all available public information, including accounting information (Beaver, 1968; Beaver, 1970; Jensen, 1978). Credit ratings are valuation measures which are influenced by accounting and other public and private information (Ziebart & Reiter, 1992; Hand et al., 1992). Companies perceived as having greater financial prospects (higher future cash flows) generally receive higher credit ratings. Chen et al. (2015) found underfunding of pension liabilities disclosed in footnotes to financial statements to be associated with credit ratings under two different accounting standard regimes. In addition, Jorion et al. (2009) found that changes in the quality of accounting information impact credit ratings. This evidence, which is consistent with the EMH, demonstrates credit ratings reflect economic information no matter how (or if) it is reflected in the financial statements. Therefore, to the extent SFAS No. 144’s expanded criteria resulted in gains and losses from recurring asset disposals being reported as discontinued operations, credit rating agencies will consider them in estimating companies’ future cash flows. In contrast, by narrowing the definition of discontinued operations, ASU 2014-08 likely decreased the possibility that gains and losses are separately stated, improving the transparency of transitory items, and enabling credit rating agencies to better predict future cash flows. Our hypothesis, which is examined under two periods to detect changes resulting from adoption of ASU 2014-08, is: H1: Discontinued operations are a significant predictor of credit ratings. Discontinued gains increase credit ratings and discontinued losses reduce credit ratings. Methodology Sample To evaluate the study’s hypothesis, six years of data are collected and segmented into two equal three-year periods representing the SFAS No. 144 years (2012 to 2014) and the ASU 2014-08 years (2015 to 2017). Sample firms are those located in the U.S. (Compustat fic = ’USA’), traded on a major stock exchange (Compustat data item stko = 0), reporting discontinued operations (Compustat data item do) other than zero, and reporting a stock price (Compustat data item prcc_p) greater than zero at fiscal year-end.2 The sample of discontinued operations firms was merged by GVKEY with the Standard & Poor’s domestic long-term issuer credit rating (Compustat data item splticrm). The sample was further refined to eliminate firms with negative common shareholders’ equity (Compustat data item ceq), and those with no reported total assets (Compustat data item at), long-term debt - total (Compustat data item dltt), earnings before income and taxes (Compustat data item ebit), net income (Compustat data item ni), or interest expense (Compustat data item xint), as these data are necessary to construct ratios used in the regression model. The resulting sample includes 1,322 firm year observations. All continuous variables are winsorized at the 1st and 99th percentiles to reduce the likelihood of influential outliers. As depicted in Table 1, of the 1,322 observations, 798 are in the SFAS No. 144 period and 524 are in the ASU 2014-08 period. This decrease is expected due to the more stringent requirements for reporting discontinued operations under ASU 2014-08. 2 More specifically, data is from Compustat-Capital IQ, Monthly Updates, North America-Fundamentals Annual data base and the Ratings data base. 86 Table 1 - Frequency Distribution of Firms SFAS No. 144 Fiscal Years 2012-2014: Number Year Fiscal Year of Firms Percentage 1 2012 250 31.3 2 2013 283 35.5 3 2014 265 33.2 Sub-total 798 100.0 ASU 2014-08 Fiscal Years 2015-2017: Number Year Fiscal Year of Firms Percentage 1 2015 197 37.6 2 2016 163 31.1 3 2017 164 31.3 Sub-total 524 100.0 Total 1,322 Table 2 presents the observations by industry representation within each period. As depicted, reporting of discontinued operations is not distributed evenly across industries. The industries comprising the largest percentages of observations in the SFAS No. 144 period are durable manufacturers (18.30 percent), services (10.27 percent), retail (9.52 percent) and utilities (9.52 percent). In comparison, in the ASU 2014-08 period, durable manufacturers (20.42 percent), computers (10.11 percent), retail (9.54 percent), and services (8.78 percent) comprise the largest percentages. 87 Table 2 - Industry Classification of Sample Firms Industry Primary SIC codes SFAS No. 144 Fiscal Years 2012-2014 ASU 2014-08 Fiscal Years 2015-2017 Observations % Observations % Agriculture 1-999 0.75 3 0.57 Chemicals 2800-2824, 2840- 2899 56 7.02 42 8.02 Computers 7370-7379, 3570- 3579 and 3670-3679 52 6.52 53 10.11 Durable Manufacturers 3000-3999, excluding 3570- 3579 and 3670- 3679 146 18.30 107 20.42 Extractive Industries 2900-2999, 1300- 1399 68 8.52 35 6.68 Financial Institutions 6000-6499 60 7.52 44 8.40 Food 2000-2111 23 2.88 11 2.10 Insurance & Real Estate 6500-6999 23 2.88 15 2.86 Mining and Construction 1000-1999, excluding 1300-1399 35 4.39 22 4.20 Other 9000 and above 3 0.38 3 0.57 Pharmaceuticals 2830-2836 9 1.13 10 1.91 Retail 5000-5999 76 9.52 50 9.54 Services 7000-8999, excluding 7370-7379 82 10.27 46 8.78 Textiles & Printing 2200-2790 42 5.26 27 5.15 Transportation 4000-4899 41 5.14 23 4.39 Utilities 4900-4999 76 9.52 33 6.30 Total 798 100.0 524 100.0 This classification uses Compustat data item Standard Industrial Classification – Historical (sich). The distribution of observations by credit rating is presented in Table 3. In both periods, most of the observations are clustered in the middle of the credit ratings falling in the B- to A range (numerical sequencing of 7 to 17). Over 95 percent of the SFAS No. 144 observations fall within this range, while just under 95 percent of the ASU 2014-08 observations are in the range. 88 Table 3 - Numerical Equivalents of Standard & Poor’s Credit Rating Numerical SFAS 144 2012-2014 ASU 2014-08 2015-2017 Credit Rating Equivalent Firms Percentage Firms Percentage AAA 22 1 0.13 0 0.00 AA+ 21 3 0.38 1 0.19 AA 20 6 0.75 5 0.96 AA- 19 3 0.38 7 1.34 A+ 18 13 1.63 4 0.76 A 17 44 5.51 28 5.35 A- 16 52 6.52 25 4.78 BBB+ 15 68 8.52 56 10.71 BBB 14 126 15.79 76 14.53 BBB- 13 90 11.28 60 11.47 BB+ 12 69 8.65 58 11.09 BB 11 81 10.15 59 11.28 BB- 10 76 9.52 61 11.66 B+ 9 58 7.27 34 6.50 B 8 59 7.39 23 4.40 B- 7 40 5.01 15 2.87 CCC+ 6 3 0.38 4 0.76 CCC 5 4 0.50 2 0.38 CCC- 4 0 0.00 1 0.19 CC 3 1 0.13 1 0.19 C 2 0 0.00 0 0.00 D 1 3 0.38 3 0.57 Total 798 100.00 *523 100.00 *For the ordered Probit model one observation was deleted due to missing values, so only 523 observations were used. Research Design The following regression equation, which is consistent with extant related research (e.g., Alalia et al., 2012; Ayres, 2016; Blume et al., 1998), is used to test the study’s hypothesis using an ordered Probit3 model: CREDITRATING i,t+4 = α0 + δ1DOATi,t + δ2INTCOVi,t + δ3LEVi,t + δ4RETURNi,t + δ5DEBTCOVi,t + δ6LOGATi,t + δ7BETAi,t + δ8INDUSTRYi,t + δ9YEARi,t + ei,t. 3 The model was also estimated using a generalized linear model (GLM) and the results were consistent with the ordered Probit findings. 89 Variables used in the equation are summarized in Table 4. The dependent variable, CREDITRATING, is the Standard and Poor’s long-term domestic credit rating (Compustat data item splticrm), converted from letter grades to numerical equivalents (Ayres, 2016). Credit ratings are provided monthly. To allow sufficient time for a firm’s credit rating to reflect the financial reporting of the most recent fiscal year-end, CREDITRATING is measured four months after the date of each observation’s fiscal year-end (Ayres, 2016). DOAT, the variable of interest, is constructed as discontinued operations (Compustat data item do) divided by total assets (Compustat data item at).4 A significant DOAT coefficient suggests credit rating agencies perceive discontinued operations as being relevant to assigning a credit rating – they are expected to have some persistence. Variables that prior research has shown to be highly correlated with credit ratings are included in the regression equation. These are, leverage (LEV), which is calculated as long-term debt (Compustat data item dltt) divided by long-term debt plus total assets. This variable is expected to have a negative coefficient since higher leverage is expected to be negatively related to a firm’s credit rating. RETURN, calculated as operating income (Compustat data item ebit) divided by sales (Compustat data item sale), is expected to be positive since favorable operating performance positively impacts credit ratings. The interest coverage ratio, INTCOV, is earnings before interest and taxes (Compustat data item ebit) divided by interest expense (Compustat data item xint). The coefficient on INTCOV is expected to be positive since a higher coverage ratio represents a better ability to meet interest payments. LOGAT is the log of total assets. A positive coefficient is expected since larger firms tend to be less risky. Debt coverage, DEBTCOV, is measured as earnings before interest (Compustat data item ebitda) less interest expense (Compustat data item xint) and taxes (Compustat data item txt), divided by total liabilities (Compustat data item lt). DEBTCOV is expected to have a positive coefficient as greater liquidity suggests lower risk. Previous studies provide evidence that as a firm’s equity risk increases (BETA), it is less likely to service its debt. Consistent with prior studies investigating credit ratings, (Blume et al.,1998; Jorion et al., 2009; Ziebart & Reiter, 1992), the study includes BETA in the equation and estimated it using monthly returns over the sixty-month period prior to the fiscal year-end. It is expected to have a negative coefficient as more risk, higher BETA, should result in lower credit ratings. To control for industry-specific fixed effects, the study includes INDUSTRY, which is one of 15 (16 industries detailed in Table 2, minus 1) indicator variables to control for industry-specific fixed effects (Barth et al., 1998; Easton & Pae, 2004). In addition, to control for the year specific fixed effects, the variable YEAR is included. Table 4 - Definition of Regression Dependent and Independent Variables Compustat Data Items in (italics) (addition (+) and subtraction (-) assumed to occur before division (/)) Variable Definition 4 Because discontinued operations are typically comprised of two components, gains or losses on the sale of assets and net operating costs expected to be incurred prior to disposition of the discontinued unit, and the recorded value of total assets is more stable than operating income or sales, DOAT is deflated by total assets. Deflating by total assets is also consistent with the methodology used by Guragai et al. (2020). 90 DEPENDENT VARIABLE: CREDITRATING Standard & Poor’s Domestic Long-term Issuer Credit Rating (splticrm), converted from letter grades to numerical equivalents. To provide sufficient time for a firm’s credit rating to reflect the financial reporting of the most recent fiscal year-end, the rating is measured four months after the date of each observation’s fiscal year-end. INDEPENDENT VARIABLES: DOAT Discontinued Operations (do) divided by total assets (at). INTCOV Interest Coverage is earnings before interest and taxes (ebit) / interest expense (xint). If Interest Coverage Ratio is < -1 then it is set to = -1 and if > 25 then set to 25. An average of the three most recent years is used. LEV Leverage Ratio is long-term debt (dltt) / long-term debt (dltt) + total assets (at). An average of the three most recent years is used. RETURN Return is operating income (ebit) / sales (sale). An average of the three most recent years is used. DEBTCOV Debt Coverage is earnings before interest (ebitda) – interest expense (xint) – taxes (txt) / total liabilities (lt). If Debt Coverage is < -1 then it is set to = -1 and if > 5 then set to 5. An average of the three most recent years is used. LOGAT Natural log of total assets (at). BETA Market model betas were estimated using monthly returns over a 60 month period prior to the fiscal year-end of each observation. Descriptive statistics of variables included in the regression equations are presented in Table 5. Panel A shows the mean and median of the SFAS No. 144 variables reported for discontinued operations and Panel B presents the mean and median for the variables during the ASU 2014-08 period. Also presented are the 25th and 75th percentiles for each variable. Table 5 - Descriptive Statistics of Dependent and Independent Variables PANEL A SFAS No. 144 Fiscal Years 2012-2014 (n = 798): Variable Mean 25th Median 75th 91 Percentile Percentile CREDITRATING 12.312 10.000 13.000 14.000 DOAT 0.002 -0.001 0.000 0.002 INTCOV 5.887 2.181 3.892 7.938 LEV 0.466 0.339 0.453 0.570 RETURN 0.113 0.062 0.103 0.169 DEBTCOV 0.125 0.071 0.115 0.172 LOGAT 8.808 7.793 8.666 9.668 BETA 1.422 0.891 1.328 1.814 PANEL B ASU 2014-08 Fiscal Years 2015-2017 (n = 524): Variable Mean 25th Percentile Median 75th Percentile CREDITRATING 12.390 10.000 13.000 14.000 DOAT 0.002 -0.001 0.000 0.001 INTCOV 6.328 2.609 4.327 8.895 LEV 0.487 0.358 0.473 0.609 RETURN 0.106 0.059 0.106 0.164 DEBTCOV 0.125 0.072 0.121 0.175 LOGAT 9.008 7.900 8.846 9.928 BETA 1.317 0.905 1.234 1.644 The dependent variable CREDITRATING is not significantly different between the periods (SFAS No. 144 = 12.312 and ASU 2014-08 = 12.390). However, sample firms in the ASU 2014-08 period are more leveraged (LEV, p < 0.05), larger (LOGAT, p < 0.05), and less risky (BETA, p < 0.01) than sample firms in the SFAS No. 144 period (not tabulated). Differences may, in part, be due to period specific economics, the impact of which are accounted for in the YEAR fixed effects variables included in the regression equation. Discussion Results Correlation analysis presented in Table 6 depicts that some of the independent variables used in the regression equation are highly correlated. Statistical diagnostics reveal variance inflation factors (VIFs) associated with all of the independent variables are less than three. Therefore, collinearity is not adversely impacting the power of the study’s tests to detect significance of the relationship between DOAT and CREDITRATING. 92 Table 6 - Correlations of Independent Variables Used in the Regression Equation (n = 1,322) VARIABLE INTCOV LEV RETURN DEBTCOV LOGAT BETA DOAT -0.027 -0.010 -0.065* -0.011 -0.034 -0.012 INTCOV -0.578*** 0.258*** 0.636*** 0.158*** -0.253*** LEV -0.028 -0.417*** -0.118*** 0.172*** RETURN 0.345*** 0.328*** -0.225*** DEBTCOV -0.094*** -0.176*** LOGAT -0.190*** *, **, *** Significant at p < 0.05, p < 0.01, p < 0.001, respectively. Variable definitions are shown in Table 4. Two regressions are estimated, one for each of the time periods, SFAS No. 144 (years 2012 to 2014) and ASU 2014-08 (years 2015 to 2017). The results of the regressions are presented in Table 7. For ease of interpretation, the coefficients on the fixed effects variables are not tabulated. Table 7 - Results of Ordered Probit for SFAS 144 (2012-2014 Fiscal Years) and ASU 2014-08 (2015-2017 Fiscal Years) CREDITRATING i,t+4 = α0 + δ1DOATi,t + δ2INTCOVi,t + δ3LEVi,t + δ4RETURNi,t + δ5DEBTCOVi,t + δ6LOGATi,t + δ7BETAi,t + δ8INDUSTRYi,t + δ9YEARi,t + ei,t . SFAS 144, 2012-2014 ASU 2014-08, 2015-2017 Predicted Independent Variables Sign Coefficient p-value Coefficient p-value DOAT Positive 4.4447 0.013 0.2858 0.887 INTCOV Positive 0.0625 <0.001 0.1064 <0.001 LEV Negative -1.6250 <0.001 -1.2635 <0.001 RETURN Positive 0.4048 0.427 0.9983 0.128 DEBTCOV 14.02299 14.02299 Positive 1.84631 0.008 1.7843 0.058 LOGAT Positive 0.7602 <0.001 0.6397 <0.001 BETA Negative -0.4704 <0.001 -0.5221 <0.001 Total observations 798 *523 Model F-value 79.63 356.3425 p < 0.001 41.97 p < 0.001 R2 0.71 0.67 Variable definitions are shown in Table 4. The coefficients on INDUSTRY and YEAR are not presented for ease of interpretation. *For the ordered Probit model one observation was deleted due to missing values, so only 523 observations were used. 93 In the SFAS No. 144 period the coefficient on DOAT is positive and significant (p = 0.013) suggesting that discontinued operations in this period are viewed as persistent and considered when assigning credit ratings. In comparison, in the ASU 2014-08 period, the coefficient on DOAT is not significant (p = 0.887). These results demonstrate that in this period discontinued operations are viewed as having little or no persistence. In both periods, the coefficients on four of the six control variables are significant and the signs are in the expected direction. The variables LEV and BETA have significant negative coefficients while the variables INTCOV and LOGAT, have significant positive coefficients. The DEBTCOV coefficient is significant in the SFAS 144 model and moderately significant (p = 0.058) in the ASU 2014-08 model. The coefficient on RETURN is not significant in either time period. As a robustness test, RETURN is replaced with return on assets (ROA), measured as net income divided by total assets. Results are substantially unchanged. In the SFAS No. 144 period, the coefficient on DOAT remains significant (p < 0.01) and the coefficient on ROA is significant (p < 0.001), but the coefficient on DEBTCOV is no longer significant. In the ASU 2014-08 period, the coefficient on DOAT remains insignificant and the coefficient on ROA is significant (p < 0.01), and the coefficients on the remaining independent variables are significant in the direction predicted. Contributions This study found that during the SFAS No. 144 period, the relationship between reported discontinued operations and credit ratings was positive and significant. This suggests that the amounts reported as gains and losses from discontinued operations influence the reporting company’s credit rating. This result is likely due to the relatively broad definition of discontinued operations in use under SFAS 144. Although the FASB’s Conceptual Framework supports the notion that an income statement is more useful for decision making if non-recurring items are separated from income from continuing operations, it appears the provisions of SFAS No. 144 did not achieve this goal. It may be that this was an unintended consequence of expanding the definition of qualifying items from that of APB No. 30, or it may be that the more inclusive definition allowed managers greater flexibility to present recurring losses outside of continuing operations in the income statement. In contrast, no relationship was found between reported discontinued operations and companies’ credit ratings during the subsequent ASU 2014-08 period. This suggests that under the narrower definition, included gains and losses are more likely to be non-recurring. It appears standard setters recognized that the definition of discontinued items under SFAS No. 144 was not consistent with the goal of the Conceptual Framework, and ASU 2014-08’s more restrictive definition appears to improve the usefulness of the income statements. These results contribute to extant research about the goal of the Conceptual Framework to provide decision-useful financial reporting (e.g., Burgstahler et al., 2002; Dechow & Ge, 2006; Jones & Smith, 2011) and address a gap in the literature on discontinued operations (e.g., Barua et al., 2010; Curtis et al., 2014). Importantly, the results suggest that users of financial information are not misled when recurring gains and losses are reflected as separately stated discontinued operations. Consistent with the EMH, credit ratings reflect relevant economic information no matter how it is reported in the financial statements. Managers need not expend effort attempting to recharacterize gains and losses. This result informs standard-setters and supports the idea of a conceptual framework for accounting versus more prescriptive accounting standards. 94 References Alalia, F., Anandarajanb, A., & Jiang, W. (2012). The effect of corporate governance on firm’s credit ratings: Further evidence using governance score in the United States. Accounting and Finance, 52, 291-312. Ayers, B. C., LaPlante, S. K., & McGuire, S. T. (2010). Credit ratings and taxes: The effect of book- tax differences on ratings changes. Contemporary Accounting Research, 27(2), 359-402 Ayres, D. 2016. Fair value disclosures of level three assets and credit ratings. Journal of Accounting & Public Policy, 35(6), 635-653. Barua, A., Lin, S., & Sbaraglia, A. (2010). Earnings management using discontinued operations. The Accounting Review, 85(5), 1485-1500. Barth, M., Beaver, W., & Landsman, W. (1998). Relative valuation roles of equity book value and net income as a function of financial health. Journal of Accounting and Economics, 25(1), 1-34. Beaver, W. (1968). The information content of annual earnings announcements. Journal of Accounting Research, 6, 67-94. Beaver, W. (1970). The time series behavior of earnings. Journal of Accounting Research, 8, 62-99. Blume, M. E., Lim, F., & Mackinlay, C. (1998). The declining credit quality of U.S. corporate debt: Myth or reality? The Journal of Finance, 53(4), 1389-1413. Burgstahler, D., Jiambalvo, J., & Shevlin, T. (2002). Do stock prices fully reflect the implications of special items for future earnings? Journal of Accounting Research, 40, 585-612. Chen, T., Martin X., Mashruwala, C., & Mashruwala, S. (2015). The value and credit relevance of multiemployer pension plan obligations. The Accounting Review, 90(5), 1907-1938. Curtis, A., McVay, S., & Wolfe, M. (2014). An analysis of the implications of discontinued operations for continuing income. Journal of Accounting and Public Policy, 33, 190-201. Dechow, P. M., & Ge, W. (2006). The persistence of earnings and cash flows and the role of special items: Implications for accrual anomaly. Review of Accounting Studies, 11, 253-296. Dickins, D., McCarthy, M., O’Reilly, D., & Schneider, D. (2017). Reporting of discontinued operations – where have we been and where are we going? CPA Journal, (February), 52-55. Easton, P., & Pae, J. (2004). Accounting conservatism and the relation between returns and accounting data. Review of Accounting Studies, 9(4), 495-521. Fama, E. (1965). The behavior of stock-market prices. The Journal of Business, 38(1), 34-105. Fama, E. (1970). Efficient capital markets: a review of theory and empirical work. Journal of Finance, 25, 383-417. Financial Accounting Standards Board (FASB). (2010). Statement of Financial Accounting Concepts No. 8, Conceptual Framework for Financial Reporting (FASB: Norwalk, CT). Guragai, B., Henke, T. S., & Young, G. (2020). Discontinued operations and dividend payout policy. Working paper presented at the 2020 Annual Meeting of the American Accounting Association. Hand, J., Holthausen, R., & Leftwich, R. (1992). The effect of bond rating agency announcements on bond and stock prices. The Journal of Finance, 47(2), 733-752. Jensen, M. (1978). Some anomalous evidence regarding market efficiency. Journal of Financial Economics, 6(2/3), 95-101. Jones, D., & Smith, K. (2011). Comparing the value relevance, predictive value, and persistence of other comprehensive income and special items. The Accounting Review, 86(6), 2047-2073. Jorion, P., Shi, C., & Zhang, S. (2009). Tightening credit standards: the role of accounting quality. Review of Accounting Studies, 14(1), 123-160. Kisgen, D.J. (2006). Credit ratings and capital structure. Journal of Finance, 61(3), 1035–1072. Ziebart, D., & Reiter, S. (1992). Bond ratings, bond yields and financial information. Contemporary Accounting Research, 9(1), 252-282.