Preserving Amortized Costs within a Fair-Value-Accounting Framework: Reclassification of Gains and Losses on Available-for-Sale Securities upon Realization
Minyue Dong† University of Lausanne Stephen G. Ryan‡ New York University Xiao-Jun Zhang* University of California, Berkeley
First draft: December 2008 Current draft (revision in process): March 2011
†
University of Lausanne, Faculty of Business and Economics, UNIL-Dorigny, Building Internef, CH-1015 Lausanne, Switzerland, (41)216923367;
[email protected]. ‡ Stern School of Business, Kaufman Management Center, 44 West 4th Street, New York, NY 10012; (01)212998-0020;
[email protected]. * 545 Student Services Building #1900, Berkeley, CA 94720, USA, (01)510-642-4789;
[email protected]. Comments and suggestions from Anne Christine d’Arcy, Jonathan Glover, Pierre Liang, Jack Stecher, Danqing Yang and seminar participants at Carnegie Mellon University, Chinese University of Hong Kong, University of Lausanne, and Columbia Business School’s 2010 Burton Workshop are gratefully acknowledged. We thank Jialu Shan and Joseph Cadora for research and editorial assistance.
Preserving Amortized Costs within a Fair-Value-Accounting Framework: Reclassification of Gains and Losses on Available-for-Sale Securities upon Realization ABSTRACT: SFAS No. 115 requires firms to recognize available-for-sale (“AFS”) securities on the balance sheet at fair value with accumulated unrealized gains and losses (“AUGL”) recorded in accumulated other comprehensive income (“AOCI”), a component of owners’ equity. Firms reclassify AUGL to net income when they realize gains and losses. We refer to the amount of this reclassification each period as “RECLASS.” Beginning in 1998, SFAS No. 130 requires firms to present the components of other comprehensive income, including RECLASS, prominently in their financial statements. For a sample of 200 large U.S. commercial banks from 1998-2006, we examine the incremental value relevance of RECLASS beyond AUGL and other components of book value of equity and comprehensive income in a levels valuation model. We find the coefficient on RECLASS is significantly positive and closer to the coefficient on the relatively permanent net income before extraordinary items and discontinued operations than to the coefficients on AUGL and more transitory components of book value and comprehensive income. We find this high value relevance for RECLASS despite also finding that investors assess at least a normal amount of value relevance to AUGL. We interpret these findings as implying that investors assign value relevance to the realization of prior unrealized gains and losses. We find markedly higher value relevance for RECLASS than Barth (1994) and Ahmed and Takeda (1995) do for realized gains and losses for sample periods ending prior to the issuance of SFAS No. 130. This is consistent with Biddle and Choi (2006) and Chambers et al.’s (2007) findings that the value relevance of comprehensive income increased after the issuance of SFAS No. 130. We also find that Chambers et al.’s finding that AUGL is valued by the market on far above a dollar-for-dollar basis is eliminated when RECLASS is included in the valuation model. This implies Chambers et al.’s finding is driven by AUGL’s association with RECLASS. We conduct three analyses to investigate possible explanations for the value relevance of RECLASS. First, we find RECLASS is more value relevant for banks holding more liquid securities, inconsistent with investors focusing on RECLASS because of unreliably measured AUGL. Second, we find RECLASS is more value relevant for higher growth banks, consistent with realization of gains and losses interacting with or indicating future bank growth. Third, supporting the growth explanation, we find RECLASS is significantly positively associated with one-year-ahead comprehensive income controlling for other components of book value and comprehensive income, more so for banks holding more liquid securities and growing faster. We interpret our findings as implying that the FASB should continue to require prominent financial statement presentation of realized gains and losses, an amortized cost accounting construct, within the fair value accounting framework for AFS securities. The FASB acknowledges the importance of retaining amortized cost information within a fair value framework in a May 2010 exposure draft proposing balance sheet presentation of both the amortized costs and fair values of many financial instruments. Key words: Available-for-sale securities; Reclassification; Fair value accounting; Realization.
I. INTRODUCTION In this paper, we examine the incremental value-relevance of periodic realized gains and losses beyond accumulated unrealized gains and losses (“AUGL”) and other components of book value and comprehensive income for commercial banks’ available for sale (“AFS”) securities. We focus on AFS securities because they are reported at fair value on the balance sheet, but amortized cost information about realized gains and losses is preserved and reported on the income statement through the use of “dirty surplus” accounting described below. The financial reporting for AFS securities contrasts with the typical accounting for financial instruments under current U.S. GAAP, in which one of fair value and amortized cost information is reported only in footnote disclosures or not at all. Prior research generally shows that investors react more strongly to amounts recognized in financial statements rather than simply disclosed in financial reports, either because they do not have the ability or inclination to evaluate the many disclosures in those reports or because they deem recognized amounts more reliable (Schipper 2007). Hence, AFS securities constitute a relatively unambiguous setting in which to test for the incremental value relevance of information about the fair values and amortized costs of financial instruments. Statement of Financial Accounting Standards (“SFAS”) No. 115, Accounting for Certain Investments in Debt and Equity Securities, requires a hybrid fair-value-on-the-balance-sheet and amortized-cost-on-the-income-statement approach to accounting for AFS securities. Specifically, firms record AFS securities on the balance sheet at fair value, with AUGL recorded in accumulated other comprehensive income (“AOCI”), a component of owners’ equity distinct from retained earnings. This is dirty surplus accounting because changes in owners’ equity occur without corresponding changes on net income. Subsequently, firms reclassify AUGL to net income when gains and losses are realized economically through sale of AFS securities or for
accounting purposes through transfer of the securities to trading or other-than-temporary (“OTT”) impairment write-downs. We refer to the amount of this reclassification each period as “RECLASS.” Advocates of fair value accounting often criticize this accounting for AFS securities as politically motivated and convoluted, particularly for liquid securities for which fair value is the most relevant, comprehensive, and timely measure of the value of the securities.1 However, this accounting has the desirable feature of preserving certain aspects of amortized cost accounting for AFS securities—particularly the realization of gains and losses reported on the income statement via RECLASS—within a primarily fair value accounting framework. Even if fair values and unrealized gains and losses are well measured, amortized costs and realized gains and losses may be incrementally value relevant beyond fair values for various reasons. The FASB acknowledges this fact in its May 2010 Exposure Draft, Accounting for Financial Instruments and Revisions to the Accounting for Derivative Instruments and Hedging Activities (“the ED”), stating that amortized costs may be relevant due to their association with contractual cash flows or correspondence with the firm’s business strategy.2 Consistent with this statement, in the ED
1
Schultz and Hollister (1993) and Johnson and Swieringa (1996) describe the political process and resulting compromises involved in SFAS No. 115. Interestingly, these compromises occurred despite considerable support for fair value accounting for all investment securities from securities and bank regulators at the time, apparently because of the limitations of amortized cost accounting revealed by thrift crisis that culminated during the period leading up to the issuance of SFAS No. 115. Notably, Richard Breeden, the SEC Chairman during this period, expressed strong support for market-based accounting for all investment securities, regardless of the intent of the holder. In a September 14, 1990 speech, “The Proper Role of Financial Reporting: Market Based Accounting,” at Smith Barney's Fourth Annual Financial Services Conference in Washington, D.C., Mr. Breeden stated “Marketbased information can permit regulators and investors alike to make a much more meaningful assessment of the real economic value and risk exposures of a financial institution. Knowing current market value also allows regulators to take appropriate actions before a situation deteriorates irretrievably…Because it is inherently difficult to distinguish portfolio categories based on intent and ability, particularly considering the dynamic market environment in which investment decisions are made, serious consideration must be given to reporting all investment securities at market value.” 2 See the “What are the Main Aspects of the Proposed Guidance?” section on pp. 3-5 and paragraphs BC57, BC58, and BC79 of the ED. The FASB
-2-
the FASB proposes dual presentation and reconciliation of the amortized costs and fair values of many financial instruments on the face of the balance sheet. The FASB’s proposal is consistent with at least two positions expressed by critics of fair value accounting. First, almost all of these critics question the reliability of the measurement of unrealized gains and losses—at least for some financial instruments or in some circumstances— either in absolute terms (e.g., Wallison 2008 and Forbes 2009) or relative to the certain measurement of realized gains and losses (Abdel-Khalik 2008). In our view, reliability of measurement is a relatively minor concern for most of banks’ AFS securities, which are dominated by federal governmental, government-sponsored agency (e.g., Fannie Mae), and other liquid securities,3 although it is a significant concern for illiquid (e.g., some structured assetbacked) securities. More interestingly, some critics of fair value accounting point out that realization of gains and losses is important for various purposes such as contracting and stewardship assessment (Watts 1993 and Holthausen and Watts 2001), capital regulation (Khan 2010), portraying firms’ business strategies (Nissim and Penman 2008), evaluating managers’ past performance and past predictions of current performance, and predicting future firm performance (Ronen 2008). In this study, we provide evidence regarding the benefit of preserving amortized costs within a fair value accounting framework in the specific context of the value relevance of RECLASS. We acknowledge that RECLASS constitutes a more limited and less visible preservation of amortized costs than the FASB’s proposal in the ED to require dual presentation and reconciliation of the amortized costs and fair values of many financial instruments on the
3
Trevor Harris points out to us that securities typically viewed as highly liquid may experience significant price pressures due to financial institutions’ concentrated and correlated trading behavior at certain points in time. This behavior may occur at the end of accounting periods due to performance evaluation of employees, end-of-period financial statement window dressing, or other reasons.
-3-
face of the balance sheet. However, our examination of RECLASS is meaningful because this variable embodies the realization principle that is central to amortized cost accounting. We hand collected RECLASS for the 200 largest publicly traded U.S. commercial banks (based on total assets in 1998) for the period 1998-2006. We limit our sample to banks because total (realized) gains and losses on AFS securities often constitute significant portions of their owners’ equity (net income). Our sample period begins in 1998, when FAS No. 130, Reporting Comprehensive Income, first required firms to disclose RECLASS in a visible fashion.4 Our primary findings are as follows. First, we find that banks’ market values are significantly positively associated with AUGL in a pure balance sheet valuation model and higher than the coefficient on the remainder of book value in a combined balance sheet and comprehensive income valuation model, consistent with investors assessing at least a normal amount of value relevance to AUGL. We interpret this finding as implying that the incremental value relevance of RECLASS that we document in this paper is not solely attributable to RECLASS remedying the unreliability of the fair value accounting-based AUGL. Second, we find that investors ascribe considerable incremental value relevance to RECLASS, controlling for AUGL and other components of book value and comprehensive income. Specifically, we find that the market value of equity is significantly positively associated with RECLASS, with the coefficient on RECLASS being closer to the coefficient on the relatively permanent net income before extraordinary items and discontinued operations (“NIBEX”) than to the much lower coefficients on the more transitory AUGL and other components of book value and comprehensive income. These findings obtain even though
4 From 1993-1997, users of financial reports generally could have, with some effort, determined RECLASS from SFAS No. 115-required AFS securities footnote disclosures of AUGL, realized gains and losses, transfers of securities between the standard’s three categories (trading, AFS, and held-to-maturity), and OTT impairment writedowns.
-4-
RECLASS effectively just reclassifies one component of owners’ equity, AOCI, to another, retained earnings. Our remaining primary findings pertain to analyses that we conducted to investigate possible explanations for the significance of the realization of gains and losses recorded in RECLASS. Third, we consider the possibility that unrealized gains and losses are unreliable, as critics of fair value accounting often allege. This possibility is inconsistent with our aforementioned findings for AUGL and with the high liquidity of most AFS securities. Directly contradicting this possibility, we find that the incremental value relevance of RECLASS is stronger for banks that hold more liquid securities. We again interpret this finding as implying that the incremental value relevance of RECLASS is not primarily attributable to the unreliability of fair values. Fourth, we consider the possibility that RECLASS interacts with or indicates of future bank growth, as suggested by Ronen (2008).5 Consistent with this possibility, we find greater incremental value relevance of RECLASS for higher growth banks. Fifth, we find that RECLASS is significantly positively associated with year-ahead comprehensive income, controlling for AUGL and the other components of book value and comprehensive income. Consistent with the aforementioned value relevance results, this association is stronger for banks that hold more liquid securities and grow faster. We interpret the last two findings as implying that the incremental value relevance of RECLASS is primarily attributable to it informing about bank growth.6 5
Specifically, Ronen (2008) states that “accounting should…provide information about the realization of expectations…[which] is important for assessing the reliability of expectations and the quality of managerial performance, as well as in facilitating the improvement of the forecasting process.” 6 Consistent with evidence of functional fixation on earnings and other forms of mispricing in the literature (e.g., Lakonishok, Shliefer, and Vishny 1994; Sloan 1996; Teoh, Welch, and Wong 1998; and Penman and Zhang 2002), we also considered the possibility that investors undervalue periodic unrealized gains and losses (UGL). We regressed returns for the 12 months beginning five months after fiscal year end on UGL, Fama and French’s (1992,
-5-
To our best of our knowledge, our paper is the first to document the value relevance of RECLASS or the reclassification of any other component of AOCI. As discussed in the following section, our results significantly update and expand upon prior research on the value relevance of banks’ realized and unrealized gains and losses by Barth (1994) and Ahmed and Takeda (1995). Our results have implications for the ongoing debate about the relative and incremental usefulness of fair value versus amortized cost accounting for financial instruments. Our findings collectively support the FASB’s view expressed in the ED that it is incrementally value relevant to preserve amortized cost information based on the realization principle within a fair value accounting framework. The remainder of the paper is organized as follows. Section II summarizes relevant prior research. Section III develops our hypotheses and describes our research design. Section IV describes the sample and data sources and provides descriptive statistics of the variables. Section V contains the analysis of the value relevance of RECLASS. Section VI examines liquidity, growth, and other potential determinants of the value relevance of RECLASS. Section VII shows how RECLASS predicts future comprehensive income in a similar fashion as it explains market value. Section VIII concludes.
II. PRIOR RESEARCH Our study is primarily related to two areas of prior research: (1) studies on the incremental and relative value relevance of the fair values versus amortized costs of financial 1993) three factors, and stock return momentum (Jegadeesh and Titman 1993). We found statistically weak (10% level) evidence that banks with higher unrealized gains and losses experience economically modest higher future excess returns (0.5% higher for the highest versus lowest decile of UGL). Including the year-ahead change in RECLASS in the regression model weakened the drift and rendered it statistically insignificant, consistent with investors mispricing unrealized gains and losses in part because they do not fully incorporate their association with future RECLASS. Given the statistically weak and economically modest return drift, we concluded that investor mispricing explains at most a small portion of the high value-relevance of RECLASS.
-6-
instruments, particularly investment securities, and (2) studies on the incremental and relative value relevance of comprehensive income versus net income. We discuss these literatures in turn.
Gains and Losses on Investment Securities Studies The most related prior studies to ours, Barth (1994) and Ahmed and Takeda (1995), examine the value relevance of disclosures of unrealized and realized gains and losses on banks’ investment securities.7 Barth estimates levels and first differences models in annual crosssectional regressions and pooled regressions with fixed effects, while Ahmed and Takeda estimate first differences models in pooled regressions. In her levels model, Barth regresses the market value of equity on the book value of equity and the fair value and amortized cost of marketable securities. Her levels model estimation yields a significantly positive coefficient on the fair value of marketable securities and an insignificant or negative significant coefficient on the amortized cost of marketable securities. Barth concludes that the fair values of marketable securities provide significant explanatory power beyond amortized costs, but not vice versa. In her first differences model, Barth regresses returns on net income before securities gains and losses (alternatively, the change in that net income measure), periodic realized gains and losses, and periodic total (realized plus unrealized) gains and losses. Her first differences model estimation yields a negative coefficient on realized gains and losses and a positive coefficient on total gains and losses that usually is insignificant except for large banks holding liquid securities.
7 Barth (1994) hand collected the fair values of marketable securities for a sample of banks from 1970-1990, which appear to have disclosed the fair values of marketable securities in financial reports under industry GAAP or practice. Ahmed and Takeda (1995) obtained similar data from commercial bank holding companies’ regulatory Y9C filings from the second quarter of 1986 to the fourth quarter of 1991.
-7-
Barth interprets the weaker result for the income statement variables in the first differences model as attributable to greater noise in these variables. Ahmed and Takeda (1995) argue that the weakness of the results for Barth’s (1994) first differences models are attributable in part to omitted changes in the value of other net assets attributable to interest rate movements during the year. After adding an explanatory variable that captures the joint effect of the bank’s exposure to interest rates and the change in interest rates during the year to their first differences valuation model, Ahmed and Takeda find significant increases in the value relevance of both unrealized and realized gains and losses.8 While Barth (1994) and Ahmed and Takeda (1995) are influential and well-conducted studies, we argue it is time to revisit and update their findings for both financial reporting and economic reasons. Regarding financial reporting, these studies examine sample periods prior to the issuance of SFAS No. 130—or for that matter any FASB fair value accounting or disclosure standard (i.e., pre-SFAS No. 107)—so that fair values and related realized gains and losses were at a minimum less prominently presented in financial reports. In contrast, since the effective date of SFAS No. 115 in 1993, AFS securities have been recognized at fair value on the balance sheet with standardized tabular format disclosures of AUGL and disclosures of realized gains and losses by type of investment security required in footnotes. Since the effective date of SFAS No. 130 in 1998, RECLASS has been prominently recognized in financial statements. Hence, for well over a decade investors have been provided with highly visible information about fair values and related unrealized and realized gains and losses for AFS securities in financial reports; this presumably constitutes sufficient time for investors to have become accustomed to and learned how to use this information. In addition to the omitted variables issue raised by Ahmed
8
Ahmed and Takeda (1995) also examine the effects of income, capital, and tax management on the value relevance of unrealized and realized gains and losses.
-8-
and Takeda (1995), a possible explanation for Barth’s (1994) weak results for periodic gains and losses in her first differences model are that investors put less weight on disclosures rather than recognized amounts or that they had not yet become comfortable using fair value information. In fact, we believe this possibility is a strong likelihood based on prior research on differential investor appreciation of comprehensive income pre-and post-SFAS No. 130 described below. Regarding economics, banks have changed considerably since the early 1990s, a period notable for both the culmination of the thrift crisis and a sharp rise in interest rates in 1994. Prior to these events, many banks assumed considerable interest rate risk by holding much longer duration assets than liabilities. This fact motivates Ahmed and Takeda’s (1995) inclusion of an explanatory variable that captures the joint effect of the bank’s exposure to interest rates and the change in interest rates during the year, and it explains the significant effect this variable has on the coefficients on unrealized and realized gains and losses in their study. The vast majority of banks have avoided such significant duration mismatches since no later than 1994. This does not mean omitted variables are no longer an issue, however. In fact, we argue the opposite is true; individual banks’ various exposures typically have become more interconnected over time due to the rise of derivatives and other risk management products as well as the related development and use of sophisticated risk management models, such as Value at Risk. We attempt to capture these interconnections better than prior research by estimating more extensive empirical models based on Ohlson’s (1995) combined balance sheet and income statement valuation model, disaggregating both book value and comprehensive income into components with differential persistence or other accounting or economic meaning.
-9-
Other Fair Value of Financial Instruments Studies Advocates of fair value accounting generally claim that the fair values of financial instruments have higher or incremental value relevance compared to the amortized costs of the instruments. Numerous empirical studies have tested and generally found some support for this claim. However, the results of these studies vary depending the type and liquidity of the financial instruments considered, the type of firms involved, as well as aspects of the research design such as the use of levels versus first differences valuation models. These studies often but not always find that market values are significantly incrementally associated with fair values controlling for amortized costs, but not vice-versa. We describe only the three earliest and most influential studies below. Using disclosed fair values under SFAS No. 107, Disclosures about Fair Values of Financial Instruments, Barth, Beaver, and Landsman (1996), Nelson (1996), and Eccher, Ramesh, and Thiagarajan (1996) examine the incremental value relevance of disclosed fair values of essentially all banks’ financial instruments—not just investment securities but also loans, deposits, and debt—beyond the amortized costs of these instruments. The results of the three studies differ somewhat due to their differing model specifications. This is particularly true for loans, for which only Barth, Beaver, and Landsman find fair values to be incrementally value relevant. By and large, however, these studies find that the fair value of most financial instruments are incrementally value relevant beyond the amortized costs of those instruments.
Comprehensive Income Studies Studies investigating the value relevance of comprehensive income generally find that it is less value relevant prior to the effective date of SFAS No. 130 than afterwards, consistent with
- 10 -
investors responding more to amounts that are recognized in financial statements rather than disclosed. For a sample prior to the effective date of SFAS No. 130, Dhaliwal, Subramanyam, and Trezevant (1999) find that comprehensive income does not have a stronger association with returns than net income, except for financial firms. They also find that the AFS securities adjustment is the only component of other comprehensive income that improves the association between income and returns, again primarily for financial firms. O’Hanlon and Pope (1999) report similarly negative results for other comprehensive income items for a sample of U.K. firms. In contrast, for samples after the effective date of SFAS No. 130, Biddle and Choi (2006) find that comprehensive income dominates other income measures in explaining equity returns. Chambers et al. (2007) find that the association between market value and the non-AFS security components (e.g., cash flow hedges and pensions) of other comprehensive income is approximately dollar-for-dollar, but that the association between market value and the AFS security component of other comprehensive income is far above one (3.45, with a t statistic of 4.12 on the difference from 1). We find a similarly high association for AUGL. Chambers et al. unsuccessfully attempt to explain this higher-than-expected association by controlling for interest rate changes (which drive changes in the value of many of banks’ other financial instruments) and net debt (a proxy for the extent of banks’ asset-liability management). As described later in the paper, we find that this higher-than-expected association is fully eliminated when RECLASS is included in the valuation model. This implies that the association between the AFS security component of other comprehensive income and the realization of gains and losses drives Chambers et al.’s finding.
- 11 -
The finding that the value relevance of comprehensive income increased post-SFAS No. 130 suggests that the value relevance of unrealized and realized gains and losses examined by Barth (1994) and Ahmed and Takeda (1995) may also have changed over time. This is one reason why we believe it is time to revisit and update the findings of these studies. Somewhat analogous to the valuation relevance studies just described, Bamber et al., (2010) examine managerial choices to disclose other comprehensive income in a performance statement that includes net income rather than in another allowed format. They find that managers with stronger equity incentives and lower job security are significantly less likely to report other comprehensive income in a performance statement, consistent with these managers believing the visibility of other comprehensive income is higher disclosed in that format. To avoid striating our limited number of observations on too many dimensions, we do not distinguish the particular format that banks use to make the required disclosures.
III. HYPOTHESIS DEVELOPMENT AND RESEARCH DESIGN RECLASS and the Mechanics of Recyling Gains and Losses Under SFAS No. 115, firms initially record aftertax unrealized gains and losses on AFS securities in accumulated other comprehensive income (AOCI), a component of owners’ equity. Accumulated unrealized gains and losses (AUGL) remain in AOCI until one of the following events occurs involving the securities: a. Sale, b. Transfer to trading, or c. Other-than-temporary (OTT) impairment write-downs.
- 12 -
When one of these events occurs, the related AUGL is reclassified to net income (with pretax realized gains and losses and tax effects recorded on separate line items on the income statement) and thence to retained earnings. We refer to the aftertax reclassification as RECLASS. RECLASS changes the components but not the total amounts of owners’ equity and comprehensive income. As a result, RECLASS often is referred to as “recycling” amounts previously recognized on the balance sheet onto the income statement. The following equations illustrate the mechanics of this recycling. All amounts are aftertax in these equations. Net income (NI) during a period equals net income before realized gains and losses (NIBRGL) plus RECLASS.
NIt=NIBRGLt+RECLASSt.
(1)
The change in retained earnings (RE) during a period equals NI minus dividends (DIV) during the period, which equals NIBRGL plus RECLASS minus DIV during the period:
∆REt = NIt – DIVt = NIBRGLt + RECLASSt – DIVt.
(2)
Hence, RE increases with RECLASS. The change in AUGL during a period equals the unrealized gain or loss (UGL) during the period. UGL equals the total (i.e., unrealized plus realized) gain or loss (TGL) minus RECLASS during the period.
∆AUGLt = UGLt = TGLt - RECLASSt. Hence, AUGL decreases with RECLASS.
- 13 -
(3)
The change in owners’ equity (OE) equals ∆RE plus ∆AOCI during the period. ∆RE is given in equation (2) and ∆AOCI equals ∆AUGL plus other comprehensive income from sources other than AFS securities (OCIother) during the period, yielding
∆OEt = ∆REt + ∆AOCIt = NIBRGLt + RECLASSt - DIVt + TGLt - RECLASSt + OCIothert
(4)
= NIBRGLt + TGLt - DIVt+ OCIothert Hence, ∆OE is unaffected by RECLASS due to its perfectly offsetting effects on RE and AOCI.
Valuation Model and Related Hypotheses Because Barth (1994) finds considerable noise in first differences valuations models, our primary analyses employ levels valuation models.9 We estimate several variants of the following model that includes both balance sheet and comprehensive income statement information:
MVt = α + β1BVt + β2CIt,
(5)
where MVt denotes the market value of owners’ equity, BVt denotes the book value of owners’ equity, and CIt denotes comprehensive income. In the empirical analysis, the model variables are per share. Ohlson (1995) derives a model similar to equation (5) from three assumptions: (1) dividend discounting with a constant cost of equity r, (2) clean surplus accounting, and (3) a first
9
We estimated a returns model derived directly from equation (6) below as the change in price plus dividends per share, with all model variables deflated by beginning price, taking into account share distributions. Consistent with Ahmed and Takeda’s (1995) findings, this returns model yields similar but weaker results than our levels valuation model. For example, we obtain a less positive and significant coefficient on ∆RECLASS in this model—3.76 significant at the 5% level—than for the corresponding coefficient on RECLASS in our primary levels valuation model reported in column III of Table 3—8.10 significant at the 1% level.
- 14 -
order autoregressive process for abnormal comprehensive income.10 Ohlson shows that perfectly transitory (e.g., fair value accounting based) abnormal comprehensive income yields a pure balance sheet model with β1=1 and β2=0. He shows that perfectly permanent abnormal comprehensive income yields a pure income statement model with β1=0 and, assuming a constant rate of dividend payout of comprehensive income λ, β2=(1+r)/r-λ.11 Our regression models begin with equation (5) and decompose both BV and CI into components. Specifically, we decompose BV into the aftertax book value of AFS securities (BVafs) plus the aftertax book value of other net assets (BVother). We further decompose BVafs into the amortized cost of AFS securities (COST) plus AUGL. We decompose CI into net income before extraordinary items and discontinued operations (NIBEX), extraordinary items and discontinued operations (EX), and other comprehensive income (OCI). We further decompose NIBEX into RECLASS and NIBEXother. We further decompose OCI into ∆AUGL (which equals TGL-RECLASS or equivalently UGL) and OCIother. Incorporating these variable decompositions into equation (5) yields our most extensive valuation model: MVt = α + β1COSTtafs + β2AUGLt + β3BVothert + β4RECLASSt + β5NIBEXothert + β6EXt + β7∆AUGLt + β8OCItother.
(6)
In addition to the extensive model, we estimate a nested version of equation (6) that includes only the balance sheet variables COST, AUGL, and BVother, as well as a nested version that includes only the comprehensive income statement variables RECLASS, NIBEXother, EX, 10
Equation (7) in Ohlson (1995) differs in four primary respects from our equation (5). First, his comprehensive income term is a capitalization of comprehensive income reduced by dividends. Second, the weights on the book value and income terms are inversely related. Third, he allows for non-accounting information. Fourth, his equation does not include an intercept. 11 If dividend payout is not constant, then dividends should be incorporated in Ohlson’s (1995) capitalized comprehensive income term. However, prior research by Hand and Landsman (1998) finds that including dividends in the Ohlson model empirically appears to capture dividends signaling future income rather than dividend payout.
- 15 -
∆AUGL, and OCIother. In untabulated robustness tests, we include a large number of control variables motivated by Ahmed and Takeda (1995)—in particular, their interactive interest rate exposure/interest rate change variable—and the broader banking literature. We also further decompose the book value and comprehensive income variables by line item in equation (6). These robustness tests yield substantively the same results as those reported in the paper. We develop five sets of null and alternative hypotheses to investigate whether and why RECLASS is value relevant. The first pertains to whether RECLASS is incrementally value relevant beyond AUGL and the other explanatory variables in equation (6). The null hypothesis (H1N) is that RECLASS is not value relevant because the fair value of AFS securities and the cash received from selling those securities conveys all relevant information, leaving no room for RECLASS to convey incremental information. The alternative hypothesis (H1A) is that RECLASS conveys incremental information due to the importance of realization. RECLASS appears both directly as a separate variable (its effect on NI) as indirectly as a decrease in ∆AUGLt (its effect on OCI) in equation (6). Hence, we take both the direct and indirect effects of RECLASS into account and state these hypotheses as the restrictions on the “total” coefficient β4- β7 on RECLASS:12 H1N: β4- β7=0. H1A: β4-β7>0.13 12
RECLASS also affects AUGL and, depending on the type of realization of gains and losses, one of COST or BVother in equation (6). For example, for economic realizations of gains, RECLASS also appears as an increase in BVother (its effect on cash) and a decrease in AUGL (its effect on AFS securities). However, unlike RECLASS’s direct and indirect effects discussed in the text, these two additional effects would exist in a clean surplus fair value accounting system that does not report RECLASS or otherwise preserve amortized cost information about realized gains and losses. We conduct our tests using the total coefficient β4-β7 because it better corresponds to our focus in this paper. However, we have also conducted all tests using the alternative total coefficient (β3+β4)-(β2+β7), which yields the same conclusions as for our tests using β4-β7, albeit with slightly lower significance levels (e.g., 5% instead of 1% in our primary tests reported in column III of Table 3). 13 We do not consider the effect of taxes on the total coefficient on RECLASS (an aftertax variable). Warfield and Linsmeier (1992) examine the effect of cross-sectional differences in taxes on the coefficient on realized gains and
- 16 -
Liquidity of AFS Securities Our second set of hypotheses examines one possible reason for the value relevance of RECLASS. Critics of fair value accounting often allege that fair values are unreliable. If this lack of reliability is the reason for the positive association between MV and RECLASS, then this association should be stronger for banks that hold less liquid securities for which fair values likely are measured less reliably. We again state null and corresponding alternative hypotheses as restrictions on the total coefficient on RECLASS:
H2N: β4-β7 does not vary with the liquidity of banks’ AFS securities. H2A: β4-β7 is higher for banks holding less liquid AFS securities.
Growth Our third set of hypotheses examines the possibility that RECLASS is value relevant because the realization of gains and losses interacts with or indicates bank growth. For example, this could occur if bank managers realize gains and losses to smooth income—say by offsetting recognized losses or gains, respectively, on economically matched positions14—thereby increasing investors’ ability to predict future income or cash flows (Ronen and Sadan 1981, Ronen 2008).15 We provide descriptive evidence that RECLASS smoothes income and
losses. They argue and provide evidence that realized losses for tax-paying firms and realized gains for non-taxpaying firms convey good news, at least for the first three quarters of the fiscal year. In principle, these differences could affect the coefficients on RECLASS in our empirical models. Due to the almost universal good health of the banking industry during our sample period—far better than the 1980-1985 period examined by Warfield and Linsmeier—our sample banks invariably paid taxes. 14 Banks have considerable ability to realize gains (losses) on securities to offset recognized losses (gains) on economically hedged exposures, because most of their AFS securities are debt instruments that are involved in assetliability management. 15 We do not posit or investigate whether realization of gains and losses smoothes net income in a fashion deemed credible by investors. We note, however, that costs that might yield such credibility include regulatory costs
- 17 -
unrealized gains and losses below. Any such effect would be enhanced to the extent that future income or cash flows are more important for higher growth banks. We again state null and corresponding alternative hypotheses in terms of the total coefficient on RECLASS:
H3N: β4-β7 does not vary with banks’ growth. H3A: β4-β7 is higher for higher growth banks.
Prediction of Future Comprehensive Income To investigate the aforementioned possibility that realized gains and losses enhance investors’ ability to predict future income, we regress year-ahead comprehensive income (“CI”) on the same explanatory variables in equation (6). CIt+1 = δ + γ1COSTtafs + γ 2AUGLt + γ3BVothert + γ4RECLASSt + γ5NIBEXothert + γ6EXt + γ7∆AUGLt + γ8OCItother.
(7)
We also estimate equation (7) with two components of CIt+1 as the dependent variable: (1) NIBEXt+1, because investors likely view RECLASS as more value relevant if it indicates future NIBEX rather than the other components of future CI; and (2) RECLASSt+1, because banks realize gains and losses over a fairly long period, as we show later. Regardless of the dependent variable, the null and alternative hypotheses parallel H1N and H1A.
H4N: γ4-γ7=0. H4A: γ4-γ7>0.
associated with realization of losses, tax costs associated with realization of gains, and reputational consequences for managers who manage income along an ex post unsmooth path.
- 18 -
We also test sets of hypotheses analogous to H2 (distinguishing banks on liquidity of AFS securities) and H3 (distinguishing banks on growth) using equation (7) instead of equation (6).
IV. SAMPLE, DATA, and DESCRIPTIVE STATISTICS We obtain most accounting and market value data from Compustat and stock return data from CRSP. We hand collected AUGL and RECLASS from disclosures of the components of (accumulated) other comprehensive income in banks’ annual reports. In the Appendix we provide several examples of these disclosures. SFAS No. 130 required firms to prominently report other comprehensive income and its components in financial statements for fiscal years beginning after December 15, 1997. For this reason, our sample period begins in 1998, and it covers the nine fiscal years through 2006. Because of the time required to hand collect AUGL and RECLASS, we restrict our sample to the 200 largest U.S. commercial banks based on total assets in 1998. We also require the sample banks to be traded on NYSE, AMEX, or NASDAQ and to have all necessary data on the variables in equation (6). The market capitalization of the 200 banks ranges from $50 million to $230 billion, indicating that our sample selection based on size is not particularly restrictive. Our final sample contains 1,033 bank-year observations with complete data on the equation (6) variables. Recall that RECLASS results from three different events: sales of AFS securities, transfer of the securities to trading, and OTT impairment write-downs of the securities. The first event involves economic realization and the second and third events involve realization for accounting purposes only. Analysis of our hand-collected disclosures indicates that only 11 of the sample had reclassifications that were due in part or whole to AFS securities being transferred to trading
- 19 -
or other-than-temporary impairment write-downs, i.e., realization for accounting purposes only.16 The low frequency of transfers to trading likely results from the proviso in paragraph 15 of SFAS No. 115 that such transfers should be “rare.” The low frequency of OTT impairment writedowns likely results from our sample period preceding the financial crisis and from the fact firms generally do not record impairments for declines in value driven by movements in interest rates unless they have decided to sell the affected securities. Hence, almost all of the variation in RECLASS reflects sales of securities, i.e., economic realization. Table 1, Panel A reports descriptive statistics for the variables AUGL, TGL, and RECLASS for our sample banks. The mean AUGL is 10 cents per share, with a sizeable standard deviation of 70 cents per share. The likely reason why AUGL is positive on average is that interest rates decreased significantly and fairly steadily for almost two decades prior to our sample period. Equity prices also rose considerably, albeit less steadily, over this period. The mean of RECLASS is 3 cents per share, with a standard deviation of 16 cents per share. The mean ratio of the absolute value of RECLASS to the absolute value of net income equals 6%, indicating that the realization of gains and losses on AFS securities typically has a sizeable effect on net income. The mean of TGL is -2 cents per share, with a standard deviation of 57 cents per share. The mean ratio of the absolute value of TGL to the absolute value of net income equals 26%, over four times larger than the corresponding ratio involving RECLASS. Hence, total gains and losses are far more variable than realized gains and losses. Table 1, Panel B reconciles the means of the beginning and ending balances of AUGL, TGL and RECLASS for the sample observations for which all four variables are available, as
16 Possibly this low frequency reflects nondisclosure by banks is due to the immateriality of some realizations for accounting purposes only, because SFAS No. 115’s provisions need not be applied to immaterial items. Partly for this reason, we do not remove these 11 observations from the sample. Our results are not noticeably affected by the removal of these observations from the sample.
- 20 -
expressed in equation (3) above. To maintain consistency, all variables are deflated by the end of year number of shares outstanding for the purposes of constructing this panel. Mean AUGL decreases by 6 cents per share per year, from 13 cents per share to 7 cents per share. This decrease is attributable to mean TGL of negative 3 cents per share and mean RECLASS of 3 cents per share. The opposite signs of mean TGL and mean RECLASS indicate the disconnect between total and realized gains and losses. Table 2 reports two descriptive analyses regressing RECLASS on current NIBEXother and either current AUGL (Panel A) or TGL for the current and three prior years (Panel B). The purpose of these analyses is to provide insight into two issues: (1) whether banks use RECLASS to manage income—income smoothing would yield a negative association of RECLASS with NIBEXother, while big baths would yield the opposite—and (2) the strength and lag structure of the relationship between realized and unrealized gains and losses. The two panels of Table 2 provide consistent results. There is evidence of a low level of income smoothing through realization of gains and losses, with a small negative coefficient of -0.01 on NIBEXother that is significant at the 1% level in Panel A and 10% level in Panel B. There is also evidence of gradual realization of gains and losses, with statistically significant but relatively low positive associations between RECLASS and AUGL (coefficient=0.04, significant at the 1% level) and between RECLASS and current and lagged TGL (coefficients smoothly decline from 0.12 for current TGL to 0.05 for three-year-lagged TGL, all significant at the 5% level or better).
- 21 -
V. VALUE RELEVANCE OF RECLASS Table 3 presents the results of estimating equation (6). To mitigate heteroscedasticity, we deflate all variables by number of shares outstanding. In untabulated robustness tests, we alternatively deflate by total assets and obtain substantively the same results. To incorporate clustering of observations by time period and firm, we include year fixed effects and report t statistics that incorporate firm-level clustering of observations (Petersen 2009). As benchmarks for the estimation of the full equation (6) in column III of Table 3, we first discuss the estimation of two nested versions of the equation: a pure balance sheet model (column I) and a pure comprehensive income statement model (column II). We maintain the same decompositions of BV or CI, respectively, in the two nested models as in the full model. The results of estimating the pure balance sheet model reported in column I of Table 3 indicate that MV is significantly positively associated with AUGL at the 5% level. Moreover, similar to Chambers et al. (2007), the coefficient of 3.25 on AUGL is over twice as high as the coefficients on COST and BVother; in fact, this coefficient is too high for transitory gains and losses. As discussed below, this coefficient decreases to a normal level once RECLASS and the remaining comprehensive income variables are added to the model. These results are consistent with investors assigning at least normal value relevance to unrealized gains and losses. The results of estimating the pure comprehensive income statement model reported in column II of Table 3 indicate that MV is significantly positively associated with RECLASS. The total coefficient on RECLASS is β4-β7=9.88-0.42=9.46, significant at the 1% level. The total coefficient is closer to the coefficient of 13.34 on the relatively permanent NIBEXother than it is to the coefficient on the relatively transitory comprehensive income components: 2.05 on EX, 0.42 on ∆AUGL, and 1.49 on OCIother. These results are consistent with investors putting
- 22 -
considerably greater valuation weight on realized than unrealized gains and losses. This likely is attributable in part to RECLASS having greater persistence than ∆AUGL, as suggested by the results reported in Table 2. The results of estimating the full equation (6) reported in Column III of Table 3 indicate that the value relevance of RECLASS diminishes only slightly when AUGL and other balance sheet variables are added to the model. In contrast, the coefficient on AUGL becomes insignificant once RECLASS and other comprehensive income statement variables are added to the model. Specifically, the total coefficient on RECLASS is β4-β7=7.84-0.56=7.28, significant at the 1% level. This total coefficient is somewhat closer to the coefficient of 10.84 on the relatively permanent NIBEXother than to the coefficients on the relatively transitory comprehensive income components: 1.82 on EX, 0.56 on ∆AUGL, and 1.29 on OCIother. These results are consistent with investors deeming RECLASS to have considerable incremental value relevance beyond AUGL and the other explanatory variables in equation (6).
VI. DETERMINANTS OF THE VALUE RELEVANCE OF RECLASS In this section, we examine two possible reasons for the value relevance of RECLASS. First, we consider the possibility that unrealized gains and losses are unreliable, as critics of fair value accounting often allege. This possibility is inconsistent with our findings discussed above that AUGL is value relevant in a pure balance sheet model and also with the high liquidity of most AFS securities. Second, we consider the possibility that RECLASS interacts with or is an indicator of future bank growth.
- 23 -
Liquidity To test the liquidity explanation for the value relevance of RECLASS, each year we estimate the full equation (6), as in column III of Table 3, for two equal-sized subsamples formed based on a proxy for the liquidity of AFS securities. Following Barth (1994), we use the percentage of AFS securities that are U.S. Treasuries as our proxy. In untabulated robustness tests, we include broader measures of liquid securities, with substantively the same results. The results are reported in Table 4. The coefficient on RECLASS is higher and more significant for the high liquidity subsample. Specifically, the total coefficient on RECLASS is β4-β7=11.41-3.77=7.64 in the high liquidity subsample compared to 3.18+1.19=4.37 in the low liquidity subsample, with the difference significant at 5% level. These results are consistent with hypothesis H2N that the value relevance of RECLASS does not result from the unreliability of unrealized gains and losses recorded in AUGL.
Growth To test the growth explanation for the value relevance of RECLASS, each year we estimate the full equation (6), as in column III of Table 3, for two equal-sized subsamples formed based on a proxy for bank growth. We use growth in net interest income as our growth proxy. In untabulated robustness tests, we include various alternative proxies for bank growth, with substantively the same results. The results are presented in Table 5. The coefficient on RECLASS is higher and more significant for the high growth subsample. Specifically, the total coefficient on RECLASS is β4β7=11.63-0.31=11.32 in the high growth subsample compared to 2.75+0.39=3.14 in the low
- 24 -
growth subsample, with the difference significant at the 1% level. This result is consistent with hypothesis H3A that RECLASS interacts with or is an indicator of future bank growth. Our measures of liquidity and growth are essentially uncorrelated, so that the two analyses capture different phenomena. In untabulated analyses, we estimated equation (6) for the four subsamples formed by the intersections of the two liquidity and two growth subsamples. We find that the coefficient on RECLASS is highest (essentially the same as the coefficient on NIBEX) and significant at the 1% level in the high liquidity and high growth subsample, about half the size but still significant at the 10% level for both the high liquidity and low growth subsample and the low liquidity and high growth subsample, and approximately zero and insignificant in the low liquidity and low growth subsample.
Other Determinants Considered In untabulated analyses, we considered two other possible determinants of the value relevance of RECLASS. First, as discussed above, in our sample RECLASS almost always reflects economic realization of gains and losses. Such realization may occur because the bank needs cash. Second, banks’ Tier 1 regulatory capital is increased by their realization of gains (Moyer, 1990). During our sample period, however, almost all U.S. banks had ample access to liquidity and were well capitalized, and so these possible determinants are a priori unlikely to explain the value relevance of RECLASS. Consistent with this, we found no evidence that RECLASS is associated with cash shortage or regulatory capital status, or that partitioning equation (6) on these variables affects the value relevance of RECLASS. 17
17 It is possible that banks choose to realize gains and losses in conjunction with their other accounting decisions, in particular, in setting allowances and provisions for loan losses, the major accrual estimates for most banks. Because Beatty, Chamberlain and Magliolo (1995) provide evidence that realized gains and losses are not correlated with these accrual estimates, we did not pursue this possibility.
- 25 -
VII.
PREDICTING FUTURE COMPREHENSIVE INCOME USING RECLASS
To the extent the observed value relevance of RECLASS is attributable to rational market pricing, RECLASS should be associated with future valuation attributes. We investigate whether that is the case by regressing year-ahead CI on RECLASS and the other explanatory variables in equation (6). Because RECLASS may be associated with many future years’ CI, to maximize the power of the test the dependent variable ideally would be future CI summed over more than one year in these analyses. We do not do this to avoid losing observations due to the limited number of years of data available. We conduct similar analyses replacing year-ahead CI as the dependent variable with year ahead NIBEX and year-ahead RECLASS. We do this in part because RECLASS does not affect current CI; for example, positive RECLASS increases NIBEX and reduces OCI, with no effect on CI. We also do this in part because the descriptive analyses reported in Table 2 suggest that banks smooth net income using RECLASS, and that RECLASS is persistent because it reflects gradual realization of the current and prior years’ TGLs. We conduct the analyses described above for the overall sample as well as for the high and low liquidity subsamples examined in Table 4 and the high and low growth subsamples examined in Table 5. As in the value-relevance analyses, we examine the total coefficient γ4-γ7. Because RECLASS is more value relevant for the high liquidity and growth subsamples, it should also have higher total coefficients for those subsamples. We report the results of the analyses described above in Table 6. Panels A, B, and C of the table reports the results for the overall, liquidity, and growth samples, respectively. In the overall sample analyses reported in Table 6, Panel A, the total coefficient on RECLASS is
- 26 -
significantly positive at the 5% level in the CI and NI regressions and at the 1% level in the RECLASS regression, consistent with hypothesis H4A. The results are somewhat weaker than the levels valuation model results reported in Table 3 due to the lower power of the tests. In the high liquidity subsample results reported in Panel B1, the total coefficient on RECLASS is more positive and similarly significant for this subsample as for the overall sample in Panel A, despite the halving of the number of observations. In contrast, in the low liquidity subsample results reported in Panel B2, the total coefficient on RECLASS is less positive and is insignificant in the CI and NI regressions. These results are consistent with hypotheses H2A and H4A. In the high growth subsample results reported in Panel C1, the total coefficient on RECLASS is more positive and at least as significant for this subsample as for the overall sample in Panel A, despite the halving of the number of observations. In contrast, in the low growth subsample results reported in Panel C2, the total coefficient on RECLASS is negative and insignificant in the CI and NI regressions and less positive in the RECLASS regression than for the overall sample. These results are consistent with hypotheses H3A and H4A. In summary, RECLASS predicts banks’ year-ahead CI, more so for banks with more liquid AFS securities and higher growth. These results are consistent with the results indicating the value relevance of RECLASS reported in Tables 3-5.
VII.
CONCLUSION
In this study, we find that realized gains and losses on banks’ AFS securities are value relevant beyond accumulated unrealized gains and losses and detailed decompositions of book value and comprehensive income. Moreover, we find that the coefficient on realized gains and
- 27 -
losses is closer to the coefficient on the relatively permanent net income before extraordinary items rather than to the coefficient on the relatively transitory accumulated unrealized gains and losses and other components of comprehensive income. We also provide evidence that the value relevance of realized gains and losses is attributable to realized gains and losses interacting with or indicating bank growth rather than to the limitations of fair value accounting for AFS securities. Importantly, we examine a sample of large commercial banks for the period 1998-2006, which is after the effective date of SFAS No. 130. This standard required firms to prominently present realized gains and losses and other components of comprehensive income in their financial statements. Our results regarding the value relevance realized gains and losses are far stronger than the corresponding prior results in Barth (1994) and, to a lesser extent, Ahmed and Takeda (1995). These studies examine sample periods ending prior to the effective date of SFAS No. 130 in 1998 or any other fair value accounting or disclosure standard issued by the FASB. The difference between our findings and those of Barth (1994) and Ahmed and Takeda (1995) mirror the difference between the findings of studies examining the value-relevance of other comprehensive pre-SFAS No. 130 (Dhaliwal, Subramanyam, and Trezevant 1999 and O’Hanlon and Pope 1999) versus post-SFAS No. 130 (Biddle and Choi 2006 and Chambers, Linsmeier, Shakespeare, and Sougiannis 2007). Collectively, these studies strongly suggest that prominent presentation in financial statements increases the value relevance investors assign to comprehensive income. Our results may also reflect investors’ increasing familiarity with fair value accounting numbers over time. Our findings suggest that the FASB should continue to require information about realized gains and losses, an amortized cost accounting construct, to be prominently presented in
- 28 -
financial statements within the fair value accounting framework for AFS securities. In a May 2010 Exposure Draft, the FASB proposes to go further and require dual presentation on the balance sheet of the amortized costs and fair values of many financial instruments. While this proposal likely has political motivations, we find it a sensible way to preserve amortized cost accounting while still broadening the fair value accounting framework for financial instruments.
- 29 -
APPENDIX SFAS 130 requires U.S. firms to disclose comprehensive income and its components, including the reclassification of accumulated other comprehensive income associated with realization of gains and losses for AFS securities, in their financial reports. This appendix provides representative examples of banks’ disclosures.
Other comprehensive income disclosures Banks typically disclose the components of comprehensive income either in the statement of owners’ equity or in a separate table. We provide an example of each of these formats below.
Statement of Owners’ Equity Example:
(Source: Amcore Financial Inc., 2005 Annual Report)
- 30 -
Separate Table Example:
(Source: National City Corp., 2007 Annual Report, p. 107)
Reclassification Disclosures In addition to the disclosures above, banks typically disclose reclassifications in a separate table, as shown in the following example.
- 31 -
(Source: JPMorgan Chase & Co., 2007 Annual Report, p.163)
Some banks also provide this reclassification information along with a rollforward of the balances of accumulated other comprehensive income in a table.
(Source: Fifth Third Bancorp, 2007 Annual Report, p. 75)
As evidenced in both of these sample disclosures, most banks report both the pretax and aftertax reclassifications. For those that report only the pretax reclassifications, we use the standard federal tax rate of 35% to calculate the aftertax reclassifications.
- 32 -
Table 1 Descriptive Statistics
The sample includes the 200 largest (based on total assets in 1998) U.S. commercial banks traded on NYSE, AMEX and NASDAQ for the years 1998-2006. Stock return data are obtained from CRSP and most financial data are obtained from COMPUSTAT. Accumulated unrealized gains and losses on AFS securities (AUGL) and reclassification of AUGL upon realization of gains and losses (RECLASS) are hand collected from banks’ annual Form 10-K filings. Sample observations must have non-missing AUGL, RECLASS, and total gains and losses (TGL) for AFS securities. SIZE denotes the natural logarithm of the market value of equity at the end of the fourth month after fiscal year end. NI denotes net income. AUGL is deflated by end of year shares outstanding, while RECLASS, TGL, and NI are deflated by the shares outstanding used in calculating earnings per share. Panel B reconciles the change in mean AUGL during the year with the means of TGL and RECLASS for the year; to maintain consistency in this panel all variables are deflated by end of year shares outstanding.
Panel A: Statistics (1,033 observations) Variable SIZE AUGL per share RECLASS per share TGL per share |RECLASS|/|NI| |TGL|/|NI|
Mean 6.95 0.08 0.03 -0.02 0.06 0.26
STDEV 1.49 0.68 0.16 0.57 0.33 0.95
Q1 5.84 -0.16 0.00 -0.25 0.01 0.05
Median 6.71 0.04 0.01 0.01 0.01 0.13
Q3 7.70 0.30 0.05 0.23 0.05 0.26
Panel B: Reconciliation of Mean AUGL, TGL, and RECLASS (1,004 observations) beginning mean AUGL per share 0.13
plus mean TGL per share -0.03
minus mean RECLASS per share 0.03
- 33 -
ending mean AUGL per share 0.07
Table 2 Reclassifications, Income Smoothing, and Gradual Realization of Total Gains and Losses
This table reports the results of regressing the reclassification of AUGL upon realization of gains and losses (RECLASS) on other net income before extraordinary items (NIBEXother) and accumulated unrealized gains and losses (AUGL) for the current year (Panel A), and on current NIBEXother and total gains and losses (TGL) for the current and prior three years. AUGL is deflated by the number of shares outstanding at fiscal year end. NIBEXother and UGL are deflated by the number of shares used in calculating earnings per share. Year dummies are included in all regression, with t-statistics adjusted for clustering among observations for the same firm (Petersen 2009). *, **, and *** indicate statistical significance at 10, 5, and 1 percent levels in two-tailed tests.
Panel A: Using Current Level of Accumulated Unrealized Gains and Losses Dependent Variable: RECLASS Intercept 0.03* -0.01*** NIBEXother AUGL 0.04*** N 1,033 2 0.07 R
Panel B: Distinguishing Timing of Total Gains and Losses Dependent Variable: RECLASS 0.07** -0.01* 0.12*** 0.09*** 0.07** 0.05*** 614 0.18
Intercept NIBEXother TGL, year TGL, year-1 TGL, year-2 TGL, year-3 N R2
- 34 -
Table 3 Value Relevance of Reclassifications
This table reports the results of pooled estimations of various nested versions of equation (6), in which the market value of owners’ equity (MV) is regressed on the aftertax components of book value of owners’ equity (BV) and comprehensive income (CI). The components of BV are: the amortized cost of AFS securities (COST); the accumulated unrealized gains and losses on AFS securities at the end of each fiscal year (AUGL); the book value of available-for-sale securities (BVafs=COST + AUGL); and the net book value of assets and liabilities other than AFS securities (BVother=BV-BVafs). The components of CI are net income before extraordinary items and discontinued operations (NIBEX); reclassification of previously unrealized gains and losses on AFS securities upon realization (RECLASS); other net income before extraordinary items (NIBEXother); extraordinary items (EX); other comprehensive income (OCI); the change in AUGL; and other comprehensive income other than ∆AUGL (OCIother). All stock variables (e.g., book value) are deflated by the number of shares outstanding at fiscal year end. All flow variables (e.g., RECLASS) are deflated by the number of shares used in calculating earnings per share. β4-β7 denotes the difference of the coefficients on RECLASS and ∆AUGL. Year fixed effects are included in all regressions and t-statistics are adjusted for clustering of observations by firm (Petersen 2009). *, **, and *** indicate two-tailed statistical significance at 10, 5, and 1 percent levels, respectively.
Intercept BV BVafs COST AUGL BVother CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 β4-β7
I 9.09***
Dependent Variable: MV II 4.39*
1.44*** 3.25** 1.37***
III 2.35 0. 40*** 0.72 0.37***
1,033 0.55 --
- 35 -
9.88*** 13.34*** 2.05
7.84*** 10.84*** 1.82
0.42 1.49*** 1,033 0.72 9.46***
0.56 1.29*** 1,033 0.75 7.28***
Table 4 Effect of AFS Security Liquidity (Reliability of Fair Values) on the Value Relevance of Reclassifications
This table reports the results of estimating equation (6) (the model in column III of Table 3) for banks with above and below median liquid AFS securities (more and less reliable fair values, respectively). See Table 3 for description of the model and variables. Liquidity is measured as the percentage of available-for-sale securities invested in U.S. Treasury securities. β4-β7 denotes the difference of the coefficients on RECLASS and ∆AUGL. Year fixed effects are included in all regressions and t-statistics are adjusted for clustering of observations by firm (Petersen 2009). *, **, and *** indicate two-tailed statistical significance at 10, 5, and 1 percent levels.
Intercept BV BVafs COST AUGL BVother CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 β4-β7 (β4-β7)High - (β4-β7)Low
High Liquidity 3.89***
Low Liquidity 4.10*
0.64*** 1.90 0.62***
0.66*** 2.21* 0.65***
11.41*** 7.64*** -2.84
3.18* 8.17*** 3.48
3.77** 0.85 462 0.84 7.64***
-1.19 1.64** 509 0.75 4.37 3.27**
- 36 -
Table 5 Effect of Bank Growth on the Value Relevance of Reclassifications
This table reports the results of estimating equation (6) (the model in column III of Table 3) for banks with above and below median growth in net interest income. See Table 3 for description of the model and variables. β4-β7 denotes the difference of the coefficients on RECLASS and ∆AUGL. Year fixed effects are included in all regressions and t-statistics are adjusted for clustering of observations by firm (Petersen 2009). *, **, and *** indicate two-tailed statistical significance at 10, 5, and 1 percent levels.
Intercept BV BVafs COST AUGL other BV CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 β4-β7 (β4-β7)High - (β4-β7)Low
High Growth 4.44
Low Growth 1.12
0.39*** 0.94 0.32*
0.43*** 0.63 0.43***
11.63*** 11.10*** -3.91
2.75* 10.35*** 8.14*
0.31 1.69** 516 0.72 11.32***
-0.39 0.37 516 0.80 3.14 7.18***
- 37 -
Table 6 Association between Year-Ahead Comprehensive Income and Reclassifications
This table reports the results of regressions of year-ahead comprehensive income (CI) and two components of year-ahead CI—net income before extraordinary items and discontinued operations (NIBEX) and reclassification of gains and losses on AFS securities upon realization (RECLASS)—on the same explanatory variables as in equation (6) (the model in column III of Table 3). See Table for description 3 of the model and explanatory variables. γ4-γ7 denotes the difference of the coefficients on RECLASS and ∆AUGL. Panel A reports results for the overall sample, Panel B1 (B2) reports results for the high (low) AFS security liquidity subsamples, and Panel C1 (C2) reports result for the high (low) growth in net interest income subsamples. Liquidity is measured as the percentage of AFS securities invested in U.S. Treasuries. Year fixed effects are included in all regressions and t-statistics are adjusted for clustering of observations for the same firm. (Petersen 2009) *, **, and *** indicate two-tailed statistical significance at 10, 5, and 1 percent levels, respectively. Panel A: Overall Sample
Intercept BV BVafs Costafs AUGL BVother CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 γ4-γ7
Dependent variable NIBEXt+1 -0.35*
CIt+1 -0.15
RECLASSt+1 0.00
0.01 -0.23* 0.01
0.03*** 0.03 0.03***
-0.00 0.06*** -0.00
0.61* 0.90*** -0.08
0.29* 0.78*** -0.35
0.39*** 0.01 0.11**
-0.23 0.24** 973 0.54 0.84**
-0.12 0.03 973 0.69 0.41*
0.05** -0.01* 959 0.19 0.34***
- 38 -
Table 6 (Continued)
Panel B1: High AFS Security Liquidity Sub-sample
Intercept BV BVafs Costafs AUGL BVother CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 γ4-γ7
Dependent variable NIBEXt+1 -0.44
CIt+1 -0.25
RECLASSt+1 0.01
0.01 -0.26* 0.00
0.02*** 0.05 0.02***
-0.00 0.04** -0.00
0.58* 1.11*** -0.66
0.72* 0.92*** -1.20
0.66*** -0.00 0.15*
0.11 0.33** 430 0.63 0.47*
0.05 0.01 430 0.74 0.67*
0.08* -0.00 424 0.37 0.58***
Panel B2: Low AFS Security Liquidity Sub-sample
Intercept BV BVafs Costafs AUGL BVother CI NIBEX RECLASS NIBEXother EX OCI ∆AUGL OCIother N R2 γ4-γ7
Dependent variable NIBEXt+1 -0.19
CIt+1 0.11
RECLASSt+1 0.02
0.01 -0.11 0.01
0.03*** 0.13 0.03***
-0.01** 0.07*** -0.00
0.25 0.75*** -0.53
0.22 0.66*** -0.53
0.24*** 0.02 0.10*
-0.47* 0.22** 484 0.54 0.72*
-0.25* 0.07** 484 0.69 0.47
-0.00 -0.01 477 0.13 0.24***
- 39 -
Table 6 (Continued)
Panel C1: High Interest Income Growth Sub-sample Dependent variable CIt+1 NIBEXt+1 Intercept -0.22 -0.17 BV BVafs Costafs 0.02 0.02** AUGL -0.36* 0.00 other BV 0.01 0.02** CI NIBEX RECLASS 1.13* 0.79** other *** NIBEX 0.99 0.81*** EX -0.59 -0.97* OCI ∆AUGL 0.11 -0.15 OCIother 0.29** 0.10** N 487 487 2 R 0.59 0.70 γ4-γ7 1.02* 0.94***
Panel C2: Low Interest Income Growth Sub-sample Dependent variable CIt+1 NIBEXt+1 Intercept -0.05 -0.42*** BV BVafs Costafs 0.02 0.03*** AUGL -0.12 0.04 BVother 0.02 0.03*** CI NIBEX RECLASS -0.11 -0.11 other *** NIBEX 0.75 0.75*** EX 0.42 -0.39 OCI ∆AUGL -0.47* -0.14 OCIother 0.14 -0.08** N 485 485 2 R 0.54 0.71 γ4-γ7 0.36 0.03
- 40 -
RECLASSt+1 0.04
-0.00 0.07** -0.01* 0.57*** 0.01 0.07* 0.07** -0.01 482 0.27 0.64***
RECLASSt+1 -0.00
-0.00 0.05** 0.00 0.22*** 0.01 0.26* 0.01 -0.01 476 0.17 0.21***
REFERENCES
Abdel-Khalik, R. 2008. “The Case Against Fair Value Accounting.” SEC Comment Letter 4573. Ahmed, A. and C. Takeda. 1995. Stock Market Valuation of Gains and Losses on Commercial Banks’ Investment Securities: An Empirical Analysis. Journal of Accounting and Economics 20: 207-225 Bamber, L., J. Jiang, K. Petroni, and I. Wang. 2010. Comprehensive Income: Who’s Afraid of Performance Reporting. The Accounting Review 85 (October): 97-126. Barth, M. 1994. Fair Value Accounting: Evidence from Investment Securities and the Market Valuation of Banks. The Accounting Review 69 (January): 1-25. Barth, M, W. Beaver and W. Landsman. 1996. Value-relevance of Banks’ Fair Value Disclosures under SFAS No. 107. The Accounting Review 71 (October): 513-537. Beatty, A., S. Chamberlain, and J. Magliolo. 1995. Managing Financial Reports of Commercial Banks: the Influence of Taxes, Regulatory Capital and Earnings. Journal of Accounting Research 33: 231-261. Biddle, G. and J. H. Choi, 2006. Is Comprehensive Income Useful? Journal of Contemporary Accounting and Economics 2: 1-32. Chambers, D., T. Linsmeier, C. Shakespeare, and T. Sougiannis. 2007. An Evaluation of SFAS No. 130 Comprehensive Income Disclosures, Review of Accounting Studies 12: 557-593. Dhaliwal, D., K. Subramanyam, and R. Trezevant. 1999. Is Comprehensive Income Superior to Net Income as a Measure of Firm Performance? Journal of Accounting and Economics 26: 43-67. - 41 -
Eccher, A., K. Ramesh, and S. Thiagarajan. 1996. Fair Value Disclosures by Bank Holding Companies. Journal of Accounting and Economics 22: 79-117. Fama, E. and K. R. French. 1992. The Cross-Section of Expected Stock Returns. Journal of Finance 47: 427-465. Fama, E. and K. R. French. 1993. Common Risk Factors in the Returns on Stocks and Bonds. Journal of Financial Economics 33: 3-56. Financial Accounting Standards Board. 1991. Disclosures about Fair Value of Financial Instruments. Statement of Financial Accounting Standards No. 107. Stanford, CT. Financial Accounting Standards Board. 1993. Accounting for Certain Investments in Debt and Equity Securities. Statement of Financial Accounting Standards No. 115. Stanford, CT. Financial Accounting Standards Board (FASB). 1997. Reporting Comprehensive Income. Statement of Financial Accounting Standards No. 130. Stanford, CT. Financial Accounting Standards Board (FASB). 2010. Accounting for Financial Instruments and Revisions to the Accounting for Derivative Instruments and Hedging Activities. Exposure Draft. Stanford, CT. Forbes, S. 2009. End market-to-market. Forbes.com, March 29, 2009. Hand, J. and W. Landsman. 1998. Testing the Ohlson Model: v or not v, that is the Question. Working paper, University of North Carolina at Chapel Hill. Holthausen, R., and R. Watts. 2001. “The Relevance of the Value-Relevance Literature for Financial Accounting Standard Setting,” Journal of Accounting and Economics September: 3-75. International Monetary Fund (IMF). 2008. “Fair Value Accounting and Procyclicality.” Chapter 3 of Global Financial Stability Report, October.
- 42 -
Jegadeesh, N., and S. Titman. 1993. Returns to buying winners and selling losers: implications for stock market efficiency. Journal of Finance 48: 65-91. Johnson, L., and R. Swieringa. 1996. Anatomy of an Agenda Decision: Statement No. 115. Accounting Horizons 10: 149-179. Khan, U. 2010. Does Fair Value Accounting Contribute to Systemic Risk in the Banking Industry? Working paper, Columbia Business School. Lakonishok, J., A. Shleifer, and R. Vishny. 1994. Contrarian Investment, Extrapolation, and Risk. Journal of Finance 49: 1541-1578. Nelson, K. 1996. Fair Value Accounting for Commercial Banks: An Empirical Analysis of SFAS No. 107. The Accounting Review 71: 161-182. Nissim, D., and S. Penman. 2008. Principles for the Application of Fair Value Accounting. Columbia Business School Center for Excellence in Accounting & Security Analysis White Paper No. 2, July.
Ohlson, J. 1995. Earnings, Book Values, and Dividends in Equity Valuation. Contemporary Accounting Research 11: 661-687. O’Hanlon, J. and P. Pope. 1999. The Value Relevance of U.K. Dirty Surplus Accounting Flows, British Accounting Review 31: 459-482. Penman, S., and X. Zhang. 2002. Accounting conservatism, the quality of earnings, and stock returns. The Accounting Review 77(2): 237-264. Petersen, M. A. 2009. Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches. Review of Financial Studies 22:435-480. Ronen, J. 2008. To Fair Value or Not to Fair Value: A Broader Perspective. Abacus 44: 181-208. Ronen, J., & Sadan, S. (1981). Smoothing Income Numbers: Objectives, Means, and Implications. Boston, MA: Addison-Wesley Publishing Company. - 43 -
Schipper, K. 2007. Required Disclosure in Financial Reports. The Accounting Review (82): 301326. Schultz, S., and J. Hollister. 2003. Lobbying FASB on Accounting for Investments. The Journal of Applied Business Research 19: 85-99. Sloan, R. 1996. Do Stock Prices Fully Reflect Information in Accruals and Cash Flows about Future Earnings? The Accounting Review July: 289-315. Teoh, S., I. Welch, and T. Wong. 1998. Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics 50, 63-99. Wallison, P. 2008. Judgment too Important to be Left to the Accountants. Financial Times, May 1, 2008.
Warfield, T., and T. Linsmeier. 1992. Tax Planning, Earnings Management, and the Differential Information Content of Bank Earnings Components. The Accounting Review, 67 (July): 546-562. Watts, R., 1993. A Proposal for Research on Conservatism, University of Rochester Financial Research and Policy Working Paper Series FR 93-13, May.
- 44 -
