Mandatory accounting change and debt covenant violation: Additional evidence from SFAS 150

Abstract

This study examines whether managers, in response to the liability classification of mandatorily redeemable preferred stock (MRPS) under SFAS 150: Accounting for financial instruments with characteristics of both liabilities and equity, restructure MRPS outstanding to keep debt off the balance sheet and, consequently, avoid increasingly likely debt covenant violation. Findings of this study show an economically and statistically significant association between MRPS restructuring and the likelihood of firms violating leverage covenant constraints, given firms reduce, on average, reported MRPS by USD259,688 for every 1 percentage point increase in leverage covenant tightness. No evidence exists of an association between MRPS restructuring and the presence of a debt covenant constraint that will be tightened by the debt classification of MRPS, or firms’ pre-adoption proximity to covenant violation. Overall, these findings help resolve conflicting evidence on the effectiveness of proxies used in prior studies to capture managers’ concerns about covenant violation due to a mandatory accounting change.

JEL classifications: M41

Keywords

Debt covenant violation hybrid securities interest coverage covenants leverage covenants mandatory accounting change

1. Introduction

A motivation behind firms issuing financial instruments with characteristics of both liabilities and equity is the supply of quasi-debt while keeping the instrument out of the liabilities section of the balance sheet (Stokes and Whincop, 1993). One such financial instrument is mandatorily redeemable preferred stock (MRPS). MRPS is non-convertible, bears a fixed dividend rate and stipulates mandatory redemption of the nominal amount on a predetermined date/s. Traditionally, US firms have classified MRPS as neither debt nor stockholders’ equity, but mezzanine equity located between the liabilities and equity sections of the balance sheet (American Accounting Association Financial Accounting Standards Committee, 2001; Kimmel and Warfield, 1995; Nair et al., 1990; Neuhausen, 2003; Wertheim and Schneider, 1992).

This article examines firms’ response to SFAS 150: Accounting for certain financial instruments with characteristics of both liabilities (SFAS 150, 2003), which requires MRPS to be classified as a liability in the issuer’s balance sheet. Consistent with this classification, dividends to MRPS holders must be classified as interest expense in the income statement. The introduction of SFAS 150, on 15 May 2003, is an exogenous shift in accounting for MRPS and provides a natural experiment to examine the association between debt covenant violation and firms’ transaction structuring to achieve a desired financial reporting outcome.

The resultant increase in debt and interest expense when firms adopted SFAS 150 increased their likelihood of debt covenant violation (McCarthy et al., 2004; Mulford and Maloney, 2003; Reinstein, 2004; Schroeder et al., 2006). The Wall Street Journal, for example, reports that the failure to take mitigating action against the debt contracting consequences of SFAS 150 ‘could result in defaults under one or more of these agreements’ (Rapoport and Weil, 2003: C1). The purpose of my study is to examine whether firms, in response to SFAS 150, restructure their MRPS to avoid debt covenant violation. Firms can renegotiate potentially binding debt covenants, or restructure transactions, to mitigate the debt contracting consequences of the accounting change and avert violation (El-Gazzar, 1993; Godfrey and Warren, 1995; Imhoff and Thomas, 1988; Smith, 1993). MRPS restructuring occurs when firms (1) retire MRPS before the fixed redemption date, thereby removing the stock from the balance sheet; (2) remove the fixed redemption date or (3) introduce a conversion feature. Either of the two latter approaches enables mezzanine classification of the stock to be maintained (Kirchner, 2003; Sinnett, 2003). As neither restructuring (Engel et al., 1999) nor debt covenant violation (Beneish and Press, 1993, 1995; Sweeney, 1994) is costless, cost-conscious managers undertake MRPS restructuring when it is cost effective in averting covenant violation. Inter alia, variation in the cost effectiveness of MRPS restructuring is attributable to whether violation will occur upon adopting SFAS 150 and, consequently, whether violation costs will be incurred.

My study addresses three research questions:

Research Question 1: Do firms engage in MRPS restructuring following the introduction of SFAS 150?

Research Question 2: Presuming firms engage in MRPS restructuring to avoid debt covenant violation, which proxies are effective in capturing management’s concerns over the debt classification of MRPS on their likelihood of violating debt covenant constraints?

Research Question 3: Do firms engage in MRPS restructuring as a function of the costliness of debt contract renegotiation?

In relation to the first research question, I identify firms with mezzanine-classified MRPS in their most recent annual financial statements prior to SFAS 150 and trace their reported MRPS in the first financial year after SFAS 150 became effective. From a sample of 108 firms with 146 MRPS issues outstanding at the time of SFAS 150, the findings indicate both a statistically and economically significant reduction in reported MRPS. In particular, there is a USD2.767 billion reduction in reported MRPS following the introduction of SFAS 150, after controlling for naturally retiring MRPS with a fixed redemption date during the sample period.

In answering the second research question, I employ cross-sectional regression analysis to examine the association between the change in reported MRPS and managers’ concerns over covenant violation, where the likelihood of covenant violation due to SFAS 150 is measured using three proxies: (1) the existence of a debt covenant constraint that will be tightened by the debt classification of MRPS, (2) the tightness of the constraint prior to SFAS 150 and (3) the increase in tightness of the constraint due to the debt classification of MRPS. Regression results indicate that the magnitude of MRPS restructuring is negatively associated with the increase in tightness in leverage covenant constraints that would result from liability reclassification. In particular, firms restructure USD259,688 of reported MRPS for each 1 percentage point increase in leverage covenant tightness arising from SFAS 150 compliance. The remaining proxies bear no significant association with MRPS restructuring. Such regression analysis also answers the third research question, with MRPS restructuring not being a function of the costliness of debt contract renegotiation proxied by private debt syndication.

My study is motivated by literature affirming that the classification of hybrid securities by accounting standards influences issuer behaviour, with the classification of instruments as debt or equity important in an issuer’s choice of financing (Fargher et al., 2018). In light of issuers choosing financial instruments with classifications that suit their financial reporting objectives (Fargher et al., 2018), my study makes several contributions to extant literature. First, it contributes to the literature focusing on whether managers restructure transactions to mitigate the financial reporting consequences of a mandatory accounting change (Elliott et al., 1984; Horwitz and Kolodny, 1980; Imhoff and Thomas, 1988; Li, 2014; Marquardt and Wiedman, 2007; Mittelstaedt et al., 1995; Selto and Clouse, 1985; Wasley and Linsmeier, 1992), including the debt contract effects of the accounting change. In addressing whether transaction restructuring is motivated by debt contracting considerations, the literature uses firm leverage to proxy for the likelihood of incurring debt covenant violation costs. Using leverage to proxy violation likelihood introduces measurement error as debt contract details are ignored (Mohrman, 1993), and fails to differentiate those firms whose proximity to violation is protected against accounting change and those that are not.¹ My study uses the accounting-based features within sample firms’ private debt contracts to measure their proximity to violation prior to, and the magnitude by which their proximity is tightened by, SFAS 150. In doing so, I contribute to the literature by providing direct evidence on whether, in response to a mandatory accounting change, firms willingly incur transaction restructuring costs to avoid incurring increasingly likely debt covenant violation costs.

Second, prior studies examining firms’ reactions to mandated changes in generally accepted accounting principles (GAAP) traditionally emphasise those accounting changes that require the recognition of a new item within the financial statements, whether that be the income statement (Elliott et al., 1984; Horwitz and Kolodny, 1980; Marquardt and Wiedman, 2007; Selto and Clouse, 1985; Wasley and Linsmeier, 1992) or the balance sheet (Bens and Monahan, 2008; Imhoff and Thomas, 1988; Mittelstaedt et al., 1995). My study differentiates from prior research by examining responses to a mandatory accounting change that requires the reclassification of an accounting item already recognised in the relevant financial statement. The reclassification of mezzanine-classified MRPS as debt does not alter the total claims against firms’ assets. Similarly, the expense classification of dividends to MRPS holders does not reduce the profit available for distribution to common stockholders.² By analysing a mandated change in GAAP that requires within-financial statement reclassification, any evidence of MRPS restructuring reflects the economic importance that the managers of MRPS firms place on financial statement classifications.

Third, my study helps resolve how to measure managers’ concerns over the likelihood of violating a financial covenant due to a mandatory accounting change. Conflicting evidence exists as to whether managers’ decision to avoid covenant violation due to a mandatory accounting change is influenced by the existence of a financial covenant that will be tightened by the accounting change (Moser et al., 2011), the tightness of the covenant constraint prior to the accounting change (Collins et al., 1981; Espahbodi et al.,1995, 1991; Leftwich, 1981; Lys, 1984; Salatka, 1989) or the increase in tightness of the covenant constraint due to the accounting change (Begley, 1990; El-Gazzar, 1993; Watts and Zimmerman, 1986). The findings of my study help resolve such mixed evidence by testing all three measures of violation likelihood. Upon doing so, the results indicate that managers, when considering the impact of a mandatory accounting change on covenant compliance, focus on the increase in covenant tightness caused by the change rather than the existence of a covenant affected by the change or firms’ proximity to the covenant constraint prior to the change. This provides useful insights into which proxies used in prior literature are effective in capturing managers’ concerns over the impact of a mandatory accounting change on their likelihood of violating a debt covenant constraint.

Moser et al. (2011) find that firms, in response to SFAS 150, are more likely to redeem trust preferred stock (TPS) rather than reclassify as a liability when subject to financial covenants whose constraints will be tightened by recognising TPS as a liability. My study is related to, but differs from, the work of Moser et al. (2011) on several grounds. First, TPS is equivalent to straight debt for tax purposes, with dividend payments to TPS holders being deductible. MRPS, however, is regarded as equity for tax purposes. To redeem TPS and forego tax savings is a tax consideration that MRPS issuers do not face (Engel et al., 1999). Given the tax considerations TPS issuers face in deciding whether to restructure to avoid SFAS 150, which MRPS issuers do not face, it is reasonable to assume their responses to SFAS 150 may differ. Second, Moser et al. (2011) examine whether firms’ decision to redeem is associated with the presence of a debt covenant affected by SFAS 150, while I quantify the magnitude of the decision to redeem and examine whether the magnitude of redemption is also associated with either the proximity to covenant thresholds prior to, or the increase in proximity to covenant thresholds due to, SFAS 150. Third, my study separates the balance sheet and income statement effects of SFAS 150 by examining the impact of SFAS 150 on proximity to leverage and interest coverage covenant violation individually.

The remainder of the article is organised as follows. In the next section, as part of the institutional background, I discuss the requirements of SFAS 150 concerning the classification of financial instruments with characteristics of liabilities and equity. Section 3 reviews relevant literature and develops the hypotheses. Section 4 describes the research design, sample selection and data collection techniques. Section 5 discusses the main results, while the results of additional analyses are discussed in Section 6. Conclusions are drawn in Section 7.

2. Institutional background

Effective for financial periods ending on or after 15 September 1979, ASR 268: Presentation in financial statements of ‘redeemable preferred stocks’ (ASR 268, 1983) prohibited Securities and Exchange Commission (SEC) registrants from classifying preferred stock with mandatory redemption on a fixed date/s as part of equity.³ In particular, ASR 268 required the amounts applicable to MRPS be presented in the balance sheet as a separate line item and not combined with stockholders’ equity. While ASR 268 specified MRPS was not to be classified as equity, it did not require classification as a liability, stating that ‘[i]t is not the Commission’s present intention to deal with the conceptual issue of whether redeemable preferred stock is a liability’ (ASR 268, 1983: 3766).

The practical consequence of ASR 268 was that most SEC registrants classified MRPS as neither debt nor equity, but rather as a separate line item in their balance sheet in the mezzanine section located between the liabilities and equity sections of the balance sheet (American Accounting Association Financial Accounting Standards Committee, 2001; Kimmel and Warfield, 1995; Neuhausen, 2003; Wertheim and Schneider, 1992). Furthermore, dividends were accounted for as a distribution of profit and, thus, did not affect net income (Nair et al., 1990; Schneider, 1993).

SFAS 150 was released in May 2003 requiring a mandatorily redeemable financial instrument to be classified as a liability in the balance sheet. Furthermore, any dividends paid or payable to holders of such stock constitute interest expense in the income statement (SFAS 150, 2003). In accordance with SFAS 150, a financial instrument is mandatorily redeemable when the issuer has no discretion to avoid a transfer of assets on the fixed redemption date/s (SFAS 150, 2003). This arises when preferred stock has a fixed redemption date/s and is non-convertible. The existence of a conversion feature that permits the holder to convert the preferred stock prior to the fixed redemption date/s introduces uncertainty as to whether redemption on the fixed date/s will occur. As such, an unconditional redemption obligation exists only for preferred stock that (1) is non-convertible with (2) a fixed redemption date/s.

The presentation requirements outlined in ASR 268 cease to apply to MRPS after the effective date of SFAS 150.⁴ This means issuers of MRPS are required to reclassify such stock from the mezzanine to the liabilities section of the balance sheet. Whereas ASR 268 prescribed non-equity classification, and thereby permitted mezzanine classification, SFAS 150 requires liability classification of MRPS.

3. Literature review and hypotheses development

3.1. Economic incentives to avoid financial covenant violation

Violation of a financial covenant imposes costs on the borrower, including a decline in firm value (Beneish and Press, 1995; Mather, 1999b). It also increases the incremental contracting costs imposed by lenders post-violation (Beneish and Press, 1993; Sweeney, 1994), including accelerating debt repayment, reducing borrowing limits, seizing borrowers’ assets, increasing interest rates, imposing more and tighter financial covenants and imposing fees to refinance and renegotiate the debt (Brown et al., 1992; DeFond and Jiambalvo, 1994; Sweeney, 1994).⁵ Lenders’ reactions to financial covenant violation also extend to forcing CEO turnovers (Nini et al., 2012) and restricting borrower access to credit markets (Roberts and Sufi, 2009a; Sufi, 2009). Borrowers, therefore, have an economic incentive to avoid financial covenant violation and the resultant costs.

3.2. Firms’ responses to mandatory accounting changes

A mandatory accounting change during the debt contract’s life may affect the accounting numbers used to determine a borrower’s proximity to covenant violation. Such change may tighten or loosen the covenant constraints imposed on the borrower. The financial covenant may become less stringent on the borrower, as slack is introduced. Conversely, the covenant may become more restrictive to the extent that the mandatory accounting change may cause an inadvertent breach by the borrower (Smith and Warner, 1979; Watts and Zimmerman, 1986).

When the introduction of an accounting standard affects a borrower’s existing debt contract a potential response by the borrower is to restructure transactions to offset the tightening financial covenant constraint caused by the mandatory accounting change (Godfrey and Warren, 1995; Imhoff and Thomas, 1988).⁶ Prior literature establishes that firms structure transactions to achieve a desired financial reporting outcome (see, for example, Ayers et al., 2002; Beatty et al., 1995; Bowman, 1980; Comiskey and Mulford, 1986; Ely, 1995; Marquardt and Wiedman, 2005; Matsunaga, 1995; Shevlin, 1987). A subsequent change in GAAP that denies firms’ desired financial reporting outcome is an exogenous event that potentially requires firms to restructure transactions.

Prior literature provides evidence of such reactions occurring. In particular, evidence exists that firms curtailed research and development outlays following changes in GAAP requiring firms to expense research and development costs that were previously capitalised (see, for example, Elliott et al., 1984; Horwitz and Kolodny, 1980; Selto and Clouse, 1985; Wasley and Linsmeier, 1992); lessees substituted operating leases for capital leases to circumvent a change in GAAP requiring the recognition of capital leases on the balance sheet that were previously disclosed in footnotes to the financial statements (Imhoff and Thomas, 1988); retiree health plans reduced health care benefits following the requirement to recognise such benefits as a liability and a corresponding expense (Mittelstaedt et al., 1995); firms with outstanding issues of contingently convertible debt (COCOs) undertook costly restructuring of their COCOs in response to a mandatory accounting change requiring the inclusion of COCOs in the denominator for diluted earnings per share (EPS) calculations (Marquardt and Wiedman, 2007); and, issuers of cash-settled convertible debt repurchased and redeemed the convertibles following an interest expense–increasing mandatory accounting change requiring the bifurcation of the proceeds received on issuance into liability and equity components (Li, 2014).⁷

3.3. The impact of mandatory accounting changes on financial covenant violation

3.3.1. Choice of GAAP

GAAP used in determining compliance may be (1) rolling over the contract’s duration or (2) fixed at a specified date. Under rolling GAAP the accounting numbers used in financial covenants are obtained from audited financial statements prepared in accordance with GAAP current at the time of calculating covenant compliance (Leftwich, 1983; Watts and Zimmerman, 1986; Whittred and Chan, 1992; Whittred and Zimmer, 1986). A change in GAAP causes a change in the accounting numbers used to calculate financial covenant compliance, which tightens or relaxes the financial covenant.

Under fixed GAAP financial covenant compliance is determined using GAAP current at the time of the contract (Fogelson, 1978; Leftwich, 1983; Smith and Warner, 1979). Any subsequent change in GAAP does not affect the accounting numbers used to determine covenant compliance as GAAP is frozen at a historical point in time, thereby protecting contracting parties against future accounting changes altering covenant stringency.

3.3.2. Accounting definitions

Accounting definitions may be (1) GAAP-consistent or (2) tailored GAAP. When GAAP-consistent definitions exist contracting parties accept the defined terms within GAAP. A change in an accounting definition due to a change in GAAP will shift the borrower’s proximity to violation (Dichev and Skinner, 2002; Frost and Bernard, 1989).

If tailored accounting definitions exist, the contracting parties tend to use GAAP as a benchmark and then undo specific features of GAAP (Citron, 1992; Leftwich, 1983; Stokes and Tay, 1988; Whittred and Zimmer, 1986). A departure from GAAP may involve having an accounting definition more expansive than GAAP (Ramsay and Sidhu, 1998). An example is a leverage covenant that already includes MRPS in the definition of debt prior to SFAS 150. The introduction of SFAS 150 does not change this definition, or tighten the borrower’s proximity to violating the leverage covenant, as MRPS is already treated as debt.

3.4. Hypotheses development

3.4.1. Hypothesis 1 (H1)

Firms may restructure MRPS by retiring or redeeming MRPS early to remove it from the balance sheet, or by altering the redemption and/or conversion clauses to remove the unconditional redemption obligation so as to fall outside the scope of SFAS 150. A motivation behind MRPS restructuring is also to offset SFAS 150’s tightening of financial covenant constraints. The resultant increase in an MRPS firm’s reported debt and interest expense due to the adoption of SFAS 150 increases the likelihood of violation (McCarthy et al., 2004; Mulford and Maloney, 2003; Reinstein, 2004), with violation likely in some cases (Rapoport and Weil, 2003; Schroeder et al., 2006). To proxy for managers’ concerns over the impact of a mandatory accounting change on firms’ violation likelihood, prior literature has used several alternatives. One proxy is the existence of a debt covenant constraint that will be tightened due to the mandatory accounting change (Moser et al., 2011). While this proxy considers the accounting-based features of debt covenants by focusing on only those covenants whose compliance is affected by changes in accounting, it presumes an equal impact of a mandatory accounting change on violation likelihood irrespective of a firm’s actual proximity to violating the covenant constraint.

Prior literature has also used the proximity of the borrower to covenant thresholds prior to the mandatory accounting change. Specifically, the more proximate a borrower is to covenant constraint thresholds prior to a mandatory accounting change, the more probable covenant violation is and, consequently, the more likely violation costs will be incurred upon adoption of the constraint-tightening accounting change (Collins et al., 1981; Espahbodi et al., 1991; Leftwich, 1981; Lys, 1984; Salatka, 1989). A borrower is more likely to restructure transactions in response to a mandatory accounting change the more likely they perceive that violation costs will be avoided by restructuring. If proximity to violation is positively associated with the likely costs of covenant violation, transaction restructuring is more likely to occur the more proximate the borrower is to covenant violation prior to the accounting change.⁸

Begley (1990), however, argues that proximity to covenant constraints does not indicate likelihood of violation, particularly when such proximity has been stable over time. Begley (1990) asserts that a better indicator of the likelihood of default is the variation in the accounting numbers used in determining covenant compliance. Prior literature that uses a borrower’s pre-adoption proximity to a covenant threshold to measure their likelihood of violating a covenant constraint fails to consider the magnitude of the effect of the mandatory accounting change on the tightness of the constraint. Accordingly, the more a mandatory accounting change tightens covenant constraints, the higher the likelihood of covenant violation and, consequently, the greater likelihood of violation costs being incurred (El-Gazzar, 1993; Watts and Zimmerman, 1986).

One strand of prior research uses the existence of an affected debt covenant constraint to proxy the likelihood of incurring violation costs due to a mandatory accounting change (Moser et al., 2011). Another strand suggests that the likely incurrence of violation costs due to a mandatory accounting change is positively associated with the borrower’s proximity to covenant thresholds prior to the accounting change (Collins et al., 1981; Espahbodi et al.,1995, 1991; Leftwich, 1981; Lys, 1984; Salatka, 1989), while a third strand argues it is the increase in proximity to covenant thresholds due to the accounting change that violation costs are positively associated with (Begley, 1990; El-Gazzar, 1993; Watts and Zimmerman, 1986). While the latter captures the likely impact of a mandatory accounting change on firms’ covenant compliance, the former have been employed by researchers (perhaps due to data and time constraints) as proxies for likely impact. As there are assertions in prior studies, supported by empirical evidence, that each proxy is able to capture management’s concerns over covenant violation due to a mandatory accounting change, the following alternative hypotheses will be expressed in the null form:

Hypothesis 1a (H1a): There is no association between the change in reported MRPS following the introduction of SFAS 150 and the existence of a covenant constraint that will be tightened due to the introduction of SFAS 150.

Hypothesis 1b (H1b): There is no association between the change in reported MRPS following the introduction of SFAS 150 and the tightness of covenant constraints prior to the introduction of SFAS 150.

Hypothesis 1c (H1c): There is no association between the change in reported MRPS following the introduction of SFAS 150 and the increase in tightness of covenant constraints due to the introduction of SFAS 150.

3.4.2. Hypothesis 2 (H2)

As an alternative to MRPS restructuring, a borrower may enter into contract renegotiations with the lender to avoid incurring increasingly likely violation costs due to SFAS 150. With the objective of preserving the position of contracting parties, the accounting-based features of the debt contract are renegotiated (Imhoff and Thomas, 1988). Specifically, the definitions of accounting terms applied in determining covenant compliance may be renegotiated to depart from GAAP, thereby insulating future compliance calculations against the impending change in GAAP (Godfrey and Warren, 1995), or the covenant constraint itself may be loosened to maintain a borrower’s pre-adoption proximity to violation (Frost and Bernard, 1989).

As contract renegotiation is not costless (Beneish and Press, 1993; DeFond and Jiambalvo, 1994; Sweeney, 1994), a firm is more likely to engage in MRPS restructuring to avoid violation costs the more costly debt contract renegotiation will be. As such, a firm’s willingness to incur MRPS restructuring costs is a function of the renegotiation costs that would be saved by restructuring. Consistent with the assertion that the costs of renegotiating public debt are higher than the costs of renegotiating private debt due to a greater number of parties to the contract (Collins et al., 1981; Holthausen, 1981), prior literature argues renegotiation costs of syndicated private loans are higher than the renegotiation costs of single-lender private loans (Duke and Hunt, 1990; El-Gazzar and Pastena, 1990). Loan syndication occurs when there is more than one lender to the private debt contract, which increases renegotiation costs because of the additional time and effort in negotiating and reaching agreement with multiple lenders compared with a single lender. MRPS restructuring is, therefore, more likely to be a cost-effective way of avoiding increasingly likely violation costs due to SFAS 150 when the firm is subject to a syndicated private loan. Based on the above, my second hypothesis is as follows:

H2: There is a negative association between the change in reported MRPS following the introduction of SFAS 150 and whether the private debt is syndicated.

4. Research design

4.1. Empirical models

The following cross-sectional regression models are used to test the remaining hypotheses:

\begin{array}{l} Δ M R P S = α + β_{1} A F F E C T + β_{2} S Y N + β_{3} S E C + \\ β_{4} F I N R I S K + β_{5} I N C F I N R I S K + β_{6} F C F + β_{7} S I Z E + ε \end{array}

(1)

\begin{array}{l} Δ M R P S = α + β_{1} P R O X L E V + β_{2} P R O X C O V + β_{3} S Y N + β_{4} S E C \\ + β_{5} F I N R I S K + β_{6} I N C F I N R I S K + β_{7} F C F + β_{8} S I Z E + ε \end{array}

(2)

\begin{array}{l} Δ M R P S = α + β_{1} I N C P R O X L E V + β_{2} I N C P R O X C O V + β_{3} S Y N + β_{4} S E C \\ + β_{5} F I N R I S K + β_{6} I N C F I N R I S K + β_{7} F C F + β_{8} S I Z E + ε \end{array}

(3)

All variables are defined in Table 1 and elaborated below.

Table 1.

Definition and measurement of variables used in the empirical models.

Dependent variable
$Δ M R P S$	It is the difference between MRPS outstanding recognised in firms’ most recent annual financial statements after, compared with prior to, SFAS 150 adoption deflated by pre-adoption total assets. It is measured as $\frac{(M R P S_{t} - M R P S_{t - 1})}{T A_{t - 1}}$ (4)
	$M R P S_{t}$ = recognised MRPS outstanding as at the most recent financial year-end following SFAS 150 adoption
	$M R P S_{t - 1}$ = recognised MRPS outstanding as at the most recent financial year-end prior to SFAS 150 adoption
	$T A_{t - 1}$ = total assets as at the most recent financial year-end prior to SFAS 150 adoption (Compustat Annual Data Item AT)
Test variables
$A F F E C T$	It is 1 if a private debt contract contains leverage or coverage financial covenants whose constraints are tightened upon SFAS 150 adoption, 0 otherwise
$P R O X L E V$	It is the tightness of a leverage covenant constraint, measured as actual leverage divided by permitted maximum leverage, as at the most recent financial year-end prior to SFAS 150 adoption
$P R O X C O V$	It is the tightness of a coverage covenant constraint, measured as required minimum coverage divided by actual coverage, as at the most recent financial year-end prior to SFAS 150 adoption
$I N C P R O X L E V$	It is the percentage increase in the tightness of a leverage covenant constraint upon adoption of SFAS 150, measured as $I N C P R O X L E V = (P R O X L E V p o s t / P R O X L E V) - 1$ (5)
	$P R O X L E V p o s t$ = leverage if classifying reported MRPS outstanding as debt divided by permitted maximum leverage, as at the most recent financial year-end prior to SFAS adoption
$I N C P R O X C O V$	It is the percentage increase in the tightness of a coverage covenant upon adoption of SFAS 150 and is measured as $I N C P R O X C O V = (P R O X C O V p o s t / P R O X C O V) - 1$ (6)
	$P R O X C O V p o s t$ = required minimum permitted coverage divided by coverage if classifying dividends to MRPS holders as interest expense, as at the most recent financial year-end prior to SFAS 150
SYN	It is 1 if a private debt contract has more than one lender, 0 otherwise
Control variables
SEC	It is 1 if a private debt contract is a secured loan, 0 otherwise
FINRISK	It is total liabilities divided by total assets, as at the most recent financial year-end prior to SFAS adoption (Compustat Annual Data Items LT/AT)
INCFINRISK	It is the percentage increase in leverage upon the debt classification of reported MRPS outstanding at firms’ financial year-end immediately preceding the introduction of SFAS 150, measured as $I N C F I N R I S K = (F I N R I S K p o s t / F I N R I S K) - 1$ (7)
	FINRISKpost = firms’ leverage ratio upon the debt classification of pre-existing MRPS outstanding
FCF	It is operating cash flows minus capital expenditure divided by total assets, as at the most recent financial year-end prior to SFAS 150 adoption ( $Compustat Annual Data Items OANCF - CAPX / AT$ )
SIZE	It is the natural log of total assets, as at the most recent financial year-end prior to SFAS 150 adoption (Compustat Annual Data Item AT)
$ε$	It represents the error term

MRPS: mandatorily redeemable preferred stock.

ΔMRPS equals zero to the extent to which any decline in MRPS outstanding is due to ‘normal’ maturation or retirement, which arises when MRPS has a fixed redemption date coinciding with the introduction of SFAS 150.

4.1.1. Dependent variable

The change in reported MRPS (∆MRPS) due to the introduction of SFAS 150 is measured as the difference between MRPS outstanding recognised in firms’ most recent annual financial statements after and before SFAS 150 adoption, scaled by pre-adoption total assets. Being a continuous variable, a negative value for ∆MRPS indicates a decline in reported MRPS following SFAS 150 adoption, while a positive value for ∆MRPS represents an increase in reported MRPS outstanding arising from the issuance of additional MRPS by the firm.

Neuhausen (2003) argues that firms will restructure their MRPS to avoid the liability classification of SFAS 150. There are three ways to do so. First, firms may retire MRPS before the fixed redemption date, thereby removing MRPS from the balance sheet. Second, firms may, subject to the agreement of stockholders, amend the redemption clause to remove the unconditional redemption obligation (Kirchner, 2003; Sinnett, 2003). By removing the fixed redemption date, the outstanding MRPS becomes conditionally redeemable and outside the scope of SFAS 150. Third, firms may introduce a conversion feature to the preferred stock, thereby also creating a conditional redemption obligation and circumventing SFAS 150. While the first approach eliminates the MRPS, under the second and third approaches the preferred stock is still outstanding but, as an unconditional redemption obligation no longer exists, it is no longer recognised as MRPS. All three approaches were adopted by firms to restructure their MRPS, with Appendix 1 providing examples of each approach undertaken by specific sample firms. In measuring $Δ M R P S$ all three approaches are captured.

The fixed redemption date of certain MRPS issues may fall within the same reporting period as firms’ adoption of SFAS 150. In these instances, a decline in reported MRPS is not in response to SFAS 150, but rather the maturity (in part or in full) of MRPS on their fixed redemption date. ASR 268 requires a firm’s disclosure of redemption terms. These are examined to identify those MRPS issues with fixed redemption dates that coincide with firms’ adoption of SFAS 150. ΔMRPS equals zero for any decline in reported MRPS outstanding due to such ‘normal’ maturation.

Given the explanatory variables in Models 1, 2 and 3 are at the firm level, for sample firms with multiple MRPS issues it is necessary to convert their ΔMRPS to the firm level. To achieve this, I combine multiple MRPS issues into one firm-specific single issue.

4.1.2. Test variable: Presence of covenant constraint tightened by SFAS 150 (H1a)

Model 1 relates to testing H1a. Consistent with the test variable of Moser et al. (2011), AFFECT captures the presence of a debt contract adversely affected by SFAS 150. In particular, AFFECT is coded 1 for sample firms subject to private debt containing leverage or covenant constraints that will be tightened upon SFAS 150 adoption, 0 otherwise. A constraint will be tightened upon SFAS 150 adoption if compliance is measured using rolling GAAP and it contains GAAP-consistent accounting definitions, as it is only when both features are present that covenants are not insulated from mandatory accounting changes.

4.1.3. Test variables: Proximity to covenant thresholds (H1b)

Model 2 relates to H1b, where proximity to covenant thresholds is separated into leverage (PROXLEV) and interest coverage (PROXCOV) covenant constraints. PROXLEV refers to the tightness of leverage covenant constraints prior to the adoption of SFAS 150, while PROXCOV measures a borrower’s pre-SFAS 150 proximity to coverage covenant constraints. Consistent with prior literature (Duke and Hunt, 1990; El-Gazzar and Pastena, 1991; Press and Weintrop, 1990), I construct a tightness ratio so that as the ratio approaches a value of 1 firms are closer to covenant violation. The tightness of a covenant constraint is measured at a borrower’s most recent financial year-end prior to SFAS 150. A value of 1 (or above) indicates violation.⁹

4.1.4. Test variables: Increase in proximity to covenant thresholds (H1c)

Model 3 relates to H1c, where increases in proximity to covenant thresholds are separated into leverage (INCPROXLEV) and interest coverage (INCPROXCOV) covenant constraints. Consistent with the approach adopted in El-Gazzar (1993), INCPROXLEV measures the percentage increase in leverage covenant constraint tightness that would result from recognising reported MRPS outstanding as debt, where reported MRPS outstanding is measured at the most recent financial year-end prior to SFAS 150 adoption. As such, INCPROXLEV captures the change in firms’ proximity to leverage covenant thresholds upon reclassifying existing levels of MRPS due to impending MRPS adoption. Similarly, INCPROXCOV measures the percentage increase in coverage covenant tightness from recognising dividends to MRPS holders as interest expense. Dividends to MRPS holders are measured using the dividend rate (as specified in the notes to the financial statements) payable at the time of SFAS 150 adoption.

4.1.5. Test variable: Costliness of debt contract renegotiation (H2)

Consistent with prior literature, SYN is a dichotomous variable coded 1 if the loan is syndicated, 0 otherwise (Asquith et al., 2005; Beatty and Weber, 2003). A loan is treated as syndicated if there is more than one lender to the contract. A negative coefficient on SYN is expected.

4.1.6. Control variables

I also control for debt contract and firm characteristics that may influence a firm’s decision to restructure their MRPS. A lender is more likely to waive a borrower’s violation if a collateral requirement exists within the debt contract (Chen and Wei, 1993; El-Gazzar, 1993; Mather, 1999a). MRPS firms, therefore, will be less likely to engage in MRPS restructuring if the affected debt contract is secured. Consistent with prior literature, SEC is a dichotomous variable equal to 1 if a debt contract is secured, 0 otherwise (Asquith et al., 2005). A negative coefficient on SEC is expected.

I control for a firm’s financial risk prior to, and increase in financial risk due to, the introduction of SFAS 150. Prior literature acknowledges that a firm’s incentive to keep debt off the balance sheet includes managing firm risk. Bowman (1979) and Hamada (1972) illustrate that the use of debt financing increases firm risk. This association captures increasing financial risk (Bowman, 1980; Chance, 1982; Hill and Stone, 1980), including greater risk of default as leverage increases. A firm’s debt rating is enhanced by off–balance sheet debt financing (Engel et al., 1999), with Frischmann et al. (1995) asserting that the liability classification of MRPS will lead to credit analysts reassessing the financial risk of MRPS firms. FINRISK and INCFINRISK control for firms’ levels of financial risk prior to, and increase in financial risk upon, SFAS 150 adoption, respectively. FINRISK and INCFINRISK are measured as a firm’s debt-to-assets ratio prior to, and the increase in debt-to-assets ratio following the introduction of, SFAS 150, respectively.¹⁰ A negative coefficient on FINRISK and INCFINRISK is expected.

I control for a firm’s free cash flows (FCF), as a firm’s cash reserves affect their ability to commit to cash settlements upon the early retirement of MRPS (Marquardt and Wiedman, 2007). A negative coefficient on FCF is expected, and is measured as a firm’s operating cash flows less capital expenditure scaled by total assets in the financial year prior to SFAS 150. Finally, I control for firm size (SIZE), measured as the natural log of total assets as at the financial year-end prior to SFAS 150. Conflicting literature exists on the association between MRPS restructuring and SIZE, with one strand asserting that large firms’ financial strength and stability lead to fewer covenant violations (Dichev and Skinner, 2002), thereby lessening the need for MRPS restructuring, while another strand asserts that as the costs of capital restructuring are relatively less burdensome for larger firms, they are more willing to incur them (Marquardt and Wiedman, 2007). I make no prediction of the sign of the SIZE coefficient.

4.2. Sample selection, data collection and sample distribution

4.2.1. Identifying firms with MRPS

Compustat Annual Data Item PSTKR (Preferred/Preference Stock – Redeemable) is used to identify firms that recognised redeemable preferred stock in their balance sheet at financial year-end immediately preceding SFAS 150 promulgation. Compustat Annual Data Item PSTKR, however, extends beyond preferred stock with a fixed redemption date to also include, as per ASR 268, preferred stock redeemable at the option of the holder and preferred stock with redemption conditions not solely within the control of the issuer. Having identified firms with redeemable preferred stock, the EDGAR database on the SEC website (www.sec.gov) is used to obtain these firms’ audited annual financial statements (lodged as a 10 K filing). The conversion and redemption features of firms’ redeemable preferred stock, as disclosed in the notes to the financial statements, are viewed to identify the initial sample of firms with preferred stock that is non-convertible with a fixed redemption date/s (i.e. MRPS). As detailed in Table 2, an initial sample of 144 firms with 191 outstanding MRPS issues are identified. MRPS issues are excluded from the sample when there is no financial reporting reason for the issuing firm to react to SFAS 150 or the issuing firm’s reaction to SFAS 150 cannot be measured. Specifically, seven MRPS issues with a debt classification are excluded as it is only those firms with mezzanine-classified MRPS whose balance sheets are affected by the liability classification required under SFAS 150, and not those that already recognise MRPS as a liability. Lack of financial statement data post-SFAS 150 results in the exclusion of six issues because the issuer’s reaction to SFAS 150 cannot be quantified. Eighteen MRPS issues are eliminated as the issuing firm is a non-public entity. A non-public entity has no financial reporting incentive to react to SFAS 150 during the time period examined, as the introduction of SFAS 150 was delayed for these firms until no earlier than 15 December 2004. Finally, 14 issues are excluded for being cancelled by the time of the issuing firm’s post-SFAS 150 filing for reasons not related to MRPS restructuring.¹¹ This process generates a final sample of 146 MRPS issues (by 108 firms).

Table 2.

Identifying outstanding MRPS issues and their issuers at the time SFAS 150 was introduced.

Sample selection
		Issuers (number)	Issues (number)
Less	Non-convertible fixed redemption date redeemable preferred stock issues (i.e. MRPS)	144	191
	Debt-classified MRPS		7
	MRPS with insufficient financial statement data post-SFAS 150		6
	MRPS issued by non-public entities		18
	Cancelled MRPS		14
Equals	Final sample of MRPS firms and issues, respectively	108	146

MRPS: mandatorily redeemable preferred stock.

4.2.2. Identifying private debt contracts

The debt contracting consequences of SFAS 150, which form the basis of H1 and H2, are examined in the context of private debt contracts with a duration that spans the financial years preceding and following SFAS 150 adoption. My study is limited to private debt contracts for several reasons. As private debt is the predominant source of corporate debt finance (Cotter, 1998, 1999; Dichev and Skinner, 2002; Nini et al., 2009; Roberts and Sufi, 2009b), it is more likely MRPS firms are subject to private than public debt covenant constraints. While the use of financial covenants in public debt has declined in recent times (Begley and Freedman, 2004), financial covenants within private debt remain common (Bradley and Roberts, 2015; Dichev and Skinner, 2002). Moreover, private debt covenants are more restrictive than public debt covenants (Mather and Peirson, 2006; Smith, 1993; Smith and Warner, 1979; Watts and Zimmerman, 1986), with Dichev and Skinner (2002) noting that private lenders set debt covenants tightly and use them as ‘trip wires’ to monitor borrower compliance. Given their frequency, and tightness, private debt covenants are more likely to be violated than public debt covenants upon SFAS 150 adoption. As such, using private debt contracts facilitates testing whether firms willingly incur transaction restructuring costs (by engaging in MRPS restructuring) to avoid the costs of debt covenant violation.

Consistent with prior literature (Beatty and Weber, 2003; Mohrman, 1996; Press and Weintrop, 1990), the Exhibit Index of MRPS firms’ 10 K filings are viewed to identify ‘material’ private debt contracts active at the introduction of SFAS 150.¹² For MRPS firms subject to several current private debt contracts at any point in time, the largest private debt contract (expressed in terms of the principal amount of debt issued) is the designated contract for analysis. This prevents firms with multiple private debt contracts dominating the analysis (El-Gazzar and Pastena, 1990), and will best capture the debt contract effects of SFAS 150 given the number of financial covenants and the tightness of their constraints is an increasing function of the size of the loan (El-Gazzar and Pastena, 1991).¹³ The accounting-based features of each debt contract are viewed, including the choice of GAAP and accounting definitions used within to determine covenant constraints affected by SFAS 150. A firm’s proximity to covenant thresholds prior to, and the increase in proximity due to, SFAS 150 adoption is then calculated. A firm’s permitted leverage or coverage is obtained from the debt contract, while actual leverage or coverage is ascertained using Compustat, ensuring that the contract-based accounting definition is applied. Dealscan is used to identify whether the debt was syndicated (SYN).

4.2.3. Control variables

Dealscan is used to identify whether the debt was secured (SEC), while Compustat is used to gather data for the remaining control variables that comprise Models 1, 2 and 3.

5. Results and discussion

5.1. Descriptive statistics and univariate analysis

Panel A of Table 3 presents descriptive statistics of the 146 MRPS issues outstanding when SFAS 150 was introduced. The aggregate dollar value of the 146 MRPS issues outstanding prior to the introduction of SFAS 150 is USD10.586 billion, with the mean (median) issue dollar value being USD72.510 (USD18.175) million.¹⁴ After excluding any decline in MRPS outstanding attributable to ‘normal’ maturation or retirement (i.e. MRPS with a fixed redemption date coinciding with the introduction of SFAS 150), the aggregate dollar value of MRPS outstanding post-SFAS 150 declined to USD7.818 billion, with the mean (median) dollar value of each issue being USD53.553 (USD8.272) million. These values are less than their pre-SFAS 150 counterparts, indicating an overall decline in the reported MRPS outstanding subsequent to SFAS 150. ΔMRPS($M) quantifies the reduction in outstanding MRPS upon the introduction of SFAS 150. The economic significance of SFAS 150 is reflected by the USD2.767 billion reduction in reported MRPS during the financial reporting period of SFAS 150 promulgation. When expressed as a percentage of firms’ pre-SFAS 150 total assets (ΔMRPS(%)), the mean change in reported MRPS is −0.0264%.

Table 3.

Descriptive statistics and comparative univariate analysis for 146 MRPS issues outstanding at the time SFAS 150 was introduced.

Panel A: Descriptive statistics.
Variable	Obs.	Sum	Mean	Std. dev.	Min.	25%	Median	75%	Max.
Issue specific
Pre-SFAS 150 MRPS ($M)	146	$10,586.474	$72.510	$137.170	$0.026	$4.000	$18.175	$73.234	$1110.110
Post-SFAS 150 MRPS ($M)	146	$7818.769	$53.553	$124.157	$0.000	$1.000	$8.272	$48.356	$1110.110
$Δ M R P S ($ M)$	146	−$2767.700	−$18.956	$79.096	−$471.000	−$1.925	$0.000	$0.126	$56.241
$Δ M R P S (%)$	146		−0.0264	0.1325	−0.7679	−0.0003	0.0000	0.0006	0.1786
Panel B: Univariate analysis.
	Obs.	Paired t test (one tailed)
		t statistic				p
Pre- and post-SFAS 150 MRPS ($M)	146	2.89				.002***
Pre- and post-SFAS 150 MRPS (%)	146	2.41				.008***

MRPS: mandatorily redeemable preferred stock.

Pre-SFAS 150 MRPS ($M) is the dollar value of MRPS outstanding preceding the introduction of SFAS 150 in millions; Post-SFAS 150 MRPS ($M) is the dollar value of MRPS outstanding post-SFAS 150; $Δ M R P S ($ M)$ is the difference between post-SFAS 150 MRPS ($M) and pre-SFAS 150 MRPS ($M) excluding any change due to ‘normal’ maturation; $Δ M R P S (%)$ is the difference between post-SFAS 150 MRPS ($M) and pre-SFAS 150 MRPS ($M) scaled by total assets pre-SFAS 150 adoption (Compustat Annual Data Item AT) multiplied by 100.

***

indicates significance at the 1% level, one tailed.

Table 3, Panel B, summarises results of tests of a significant difference between sample firms’ reported MRPS pre- and post-SFAS 150. Given the expected decline in reported MRPS post-SFAS 150, one-tailed testing is undertaken. The paired t test indicates a significant difference in reported MRPS pre- and post-SFAS 150.¹⁵ In particular, the difference in means in pre- and post-SFAS 150 reported MRPS is significant, whether expressed as $M (t = 2.89, p = .002) or as a percentage of total assets (t = 2.41, p = .008).^16,17 This decline in reported MRPS is consistent with Levi and Segal (2015), whose time-series analysis finds that the dollar value of new MRPS issuances as a percentage of firms’ total financing increases from 1998 until the Financial Accounting Standards Board (FASB) announces its intention to enact SFAS 150, after which it falls to its lowest levels in 21 years and continues to fall until 2006. The results in Panel B of Table 3 provide preliminary evidence that firms with existing MRPS responded to the introduction of SFAS 150 by reducing their reliance on MRPS as a source of finance.

Recall from Table 2, there are 38 more MRPS issues (n = 146) than there are MRPS firms (n = 108), indicating some sample firms (n = 25) had more than one MRPS issue. In untabulated findings, prior to SFAS 150 these 25 firms with multiple MRPS issues reported total MRPS outstanding of USD$5.373 billion. In the year following SFAS 150, these firms reported total MRPS outstanding of USD$4.070 billion, indicating a decline in reported MRPS (excluding ‘normal’ maturations) of USD$1.303 billion (or 0.1726% of their pre-SFAS 150 total assets). While representing only 23.15% of sample firms (25 out of 108), multiple issuers, reported total MRPS outstanding prior to SFAS 150 (i.e. USD$5.373 billion) represents 50.76% of the USD$10.586 billion reported by all sample firms. Moreover, 47.10% of the decline in MRPS reported by all sample firms (i.e. USD$2.767 billion) is accounted for by the USD$1.303 billion decline in MRPS reported by multiple MRPS issuers. Univariate analysis indicates that the decline in reported MRPS by multiple MRPS issuers is significant whether using a parametric paired t test (t = 1.78, p = .080) or a non-parametric Wilcoxon signed-rank test (p = .029). This analysis highlights the importance of firms with multiple MRPS issues in the findings of whether, and the extent to which, firms engaged in MRPS restructuring in response to SFAS 150.

5.2. Descriptive statistics and multivariate analysis (H1 and H2)

Panel A of Table 4 presents contract- and firm-specific descriptive statistics relating to H1 and H2. Descriptive statistics on sample firms’ debt contracts show 26 contain leverage or coverage constraints that use rolling GAAP-consistent accounting definitions and, thus, will be tightened upon SFAS 150 adoption (measured as AFFECT). Mean (median) tightness of leverage covenant constraints prior to SFAS 150 (measured as PROXLEV) is 0.7991 (0.7827), indicating that sample firms are operating at 79.91% (78.27%) of the maximum leverage permitted within their debt contract prior to the debt classification of MRPS. Proximity to leverage covenant thresholds ranges from 35.55% to 168.41%, indicating instances of leverage covenant violation prior to SFAS 150.¹⁸ The median tightness of coverage covenant constraints (PROXCOV) prior to SFAS 150 indicates sample firms are operating at 89.76% of their required minimum coverage.¹⁹ Proximity to coverage covenant thresholds ranges from 49.08% to 1562.50%, showing occasions of coverage covenant violation prior to SFAS 150.

Table 4.

Contract- and firm-specific descriptive statistics for 108 sample firms.

Variable	Obs.	Mean	Std. dev.	Min.	25%	Median	75%	Max.
Contract specific
AFFECT (1/0)	108	Of the 108 sample firms, 26 (24.1%) have private debt containing covenants tightened by SFAS 150
PROXLEV (%)	108	0.7991	0.3125	0.3555	0.5400	0.7827	0.9806	1.6841
PROXCOV (%)	108	1.6091	3.2206	0.4908	0.7626	0.8976	1.1660	15.625
INCPROXLEV (%)	108	0.5767	1.3874	0.0017	0.1492	0.0737	0.3958	6.5999
INCPROXCOV (%)	108	0.0721	0.1381	0.0001	0.0032	0.0176	0.0785	0.5498
SYN (1/0)	108	Of the 108 sample firms, 22 (20.4%) are subject to syndicated private debt
Lenders (number)	108	9.461	7.658	1.000	3.000	7.500	14.500	27.000
SEC (1/0)	108	Of the 108 sample firms, 19 (17.6%) have secured private debt
Firm specific
FINRISK (%)	108	0.8985	0.6465	0.0143	0.6464	0.7930	0.9405	5.6818
INCFINRISK (%)	108	0.7141	4.9886	0.0003	0.0078	0.0459	0.2760	51.7820
FCF (%)	108	−0.0066	0.1381	−1.0854	−0.0175	0.0181	0.0525	0.1404
SIZE (log)	108	6.5128	2.4417	1.1268	4.8344	6.5911	8.1840	11.0180
Total assets ($M)	108	$5902.530	$11,849.927	$3.086	$125.765	$729.389	$3583.615	$60,966.000
Net income ($M)	108	$25.653	$397.206	−$2324.000	−$4.143	$4.043	$66.881	$1386.000
Book value of equity ($M)	108	$1158.301	$3059.534	−$1776.510	−$59.651	$22.796	$882.850	$19,697.000

MRPS: mandatorily redeemable preferred stock.

Lenders (number) is the number of lenders to the debt contract; Total assets ($M) is total assets in millions pre-SFAS 150 adoption; Net income ($M) is net income (loss) in millions pre-SFAS 150 adoption (Compustat Annual Data item NI); Book value of equity ($M) is stockholders’ equity in millions pre-SFAS 150 adoption (Compustat Annual Data Item CEQ). All other variables are as defined in Table 1.

The mean (median) percentage increase in proximity to leverage covenant thresholds due to the debt classification of MRPS (measured as INCPROXLEV) is 57.67% (7.37%). The mean (median) percentage increase in proximity to coverage covenant thresholds due to the classification of dividends to MRPS holders as interest expense (measured as INCPROXCOV) is 7.21% (1.76%). The descriptive statistics for INCPROXLEV and INCPROXCOV are consistent with prior literature that the magnitude of the balance sheet effect of SFAS 150 exceeds the income statement effect (Loud, 2004; McCarthy et al., 2004; Schroeder et al., 2006). There are 22 instances when firms have syndicated private loans, thereby having lower debt renegotiation costs than other firms, while there are, on average, 9.461 lenders per private loan.

Firm-related descriptive statistics indicate that, on average, sample firms are levered to 89.85% of total assets prior to SFAS 150 (measured as FINRISK).²⁰ The median percentage increase in sample firms’ leverage from the introduction of SFAS 150 (measured as INCFINRISK) is 4.59%. Mean (median) total assets of MRPS firms are USD5.902 billion (USD729.389 million), while MRPS firms are generally profitable given their mean (median) net income is USD25.653 (USD4.043) million.

Table 5 reports the estimated coefficients and significance levels from the multivariate continuous regression analysis.^21,22 The base model results, where all test variables are excluded, are presented in Column 1, while the results for Models 1, 2 and 3 are reported in Columns 2, 3 and 4, respectively. With respect to Model 1, there is no association between the change in reported MRPS (ΔMRPS) and the presence of a debt covenant constraint that will be tightened due to SFAS 150 (AFFECT). Similarly, the results for Model 2 indicate a firm’s proximity to either leverage or coverage covenant thresholds prior to SFAS 150 (PROXLEV and PROXCOV) is not associated with MRPS restructuring in response to SFAS 150. The results for Model 3 show MRPS restructuring is not associated with the percentage increase in proximity to coverage covenant constraints that would result from SFAS 150 adoption (INCPROXCOV). However, ΔMRPS is negatively associated with the percentage increase in proximity to leverage covenant constraints (INCPROXLEV) that would otherwise result if restructuring had not occurred (t = −3.45, p < .001). This result is both statistically and economically significant, as a 1 percentage point increase in leverage covenant stringency from compliance with SFAS 150 is associated with firms restructuring USD259,688 of reported MRPS.²³ These findings indicate that managers’ concerns about incurring violation costs, and resultant MRPS restructuring to avoid these, are most effectively captured not by their proximity to covenant thresholds prior to SFAS 150, or the presence of an affected financial covenant, but rather the increase in proximity to covenant thresholds due to SFAS 150. More specifically, managers’ decisions to avert liability reclassification by restructuring MRPS appear motivated by the desire to avoid increasingly likely leverage covenant violation, and the resultant costs, that would otherwise ensue.

Table 5.

Regression estimates of the cross-sectional association between sample firms’ change in reported MRPS and the presence of a leverage or covenant constraint that will be tightened by the debt classification of MRPS (Column 2), the tightness of the leverage and covenant constraint prior to SFAS 150 (Column 3) and the increase in tightness of the leverage and covenant constraint due to the debt classification of MRPS (Column 4), and debt contract and firm characteristics.

	Pred. sign	(1)	(2)	(3)	(4)
INTERCEPT	?	0.061(1.38)	0.063(1.41)	0.059(1.31)	0.068(1.55)
AFFECT	?		−0.028(−0.43)
PROXLEV	?			0.040(1.40)
PROXCOV	?			0.002(0.68)
INCPROXLEV	?				− 0.044(–3.45)***
INCPROXCOV	?				0.100(1.02)
SYN	-		0.040(0.62)	0.004(0.14)	0.000(0.03)
SEC	-		0.005(0.17)	−0.019(−0.54)	0.031(1.43)
FINRISK	-	−0.091(−2.28)**	−0.092(−2.30)**	−0.092(−2.31)**	−0.093(−2.35)**
INCFINRISK	-	−0.020(−7.95)***	−0.020(−7.65)***	−0.020(−7.73)***	−0.020(−7.76)***
FCF	-	−0.231(−2.12)**	−0.226(−1.96)**	−0.248(−2.13)**	−0.240(−2.07)**
SIZE	?	−0.000(−0.08)	−0.000(−0.23)	−0.000(−0.13)	−0.001(−0.35)
R ²		.3425	.3453	.3495	.3768
N		108	108	108	108

MRPS: mandatorily redeemable preferred stock.

All variables are as defined in Table 1. Coefficients are reported with t statistics in parenthesis. Coefficients are White-adjusted for heteroscedasticity.

***

and ^** indicate significance at the 1% and 5% levels, respectively, one tailed.

The results for Models 1, 2 and 3 indicate there is no association between ΔMRPS and SYN, meaning that H2 is not supported. The lack of findings in support of H2 may indicate that the costs of engaging in MRPS restructuring are less onerous than the costs of debt contract renegotiation, irrespective of whether or not the private debt is syndicated. As such, ΔMRPS would not vary according to the presence of syndicated private debt. Of the control variables, in Models 1, 2 and 3 ΔMRPS is negatively associated with FINRISK (t = −2.30, −2.31 and −2.35, respectively; p < .05), indicating that a firm is more likely to restructure its MRPS to avoid its debt classification the more highly levered the firm is. Reported results across all models also show that the greater the increase in firm financial risk due to the debt classification of reported MRPS (INCFINRISK), the greater the decline in reported MRPS following the introduction of SFAS 150 (t = −7.65, −7.73 and −7.76, respectively; p < .001). Finally, a negative association exists between ΔMRPS and FCF (t = −1.96, −2.13 and −2.07, respectively; p < .05), consistent with the prediction that the magnitude of MRPS restructuring is influenced by the cash reserves available to meet redemption requirements.²⁴ There is no association between ΔMRPS and other control variables.

6. Additional analysis

6.1. Pre-SFAS 150 covenant violation

Descriptive statistics reported in Table 4 indicate that some MRPS firms are in violation of leverage and interest coverage constraints prior to the introduction of SFAS 150.²⁵ To the extent violators seek to overcome their covenant violation, they will engage in MRPS restructuring to avoid a further shift in covenant compliance calculations beyond the covenant threshold. Alternatively, a violator may accept the debt contracting consequences of SFAS 150 as they are already in violation (akin to ‘taking a bath’ on the violation). Each response depends on whether violation costs vary with the degree of violation. If violation costs are an increasing function of the degree to which covenant thresholds are exceeded, borrowers have an incentive to engage in MRPS restructuring to avoid moving further beyond the covenant threshold. Conversely, if violation costs are imposed on a violation occurring, irrespective of the degree of violation, borrowers have less incentive to restructure their MRPS to avoid violation.

Univariate statistical analysis is undertaken to determine whether the response of violators to SFAS 150 is significantly different from non-violators. On average, both categories report a decline in reported MRPS. Mean ΔMRPS for violators is −0.0369, providing preliminary evidence that violators engage in MRPS restructuring to avoid further covenant violation, while mean ΔMRPS for non-violators is −0.0356. An independent samples t test indicates no significant difference in these averages (t = 0.024, p = .981). Moreover, a chi-square test of independence finds violators do not engage in MRPS restructuring more than other sample firms (χ² = 1.977, p = .160).

I also perform additional multivariate regression analysis by including a dichotomous variable in Models 1, 2 and 3 that equals 1 if the sample firm is a violator, 0 otherwise (measured as VIOL). Given the univariate evidence suggests violators respond to SFAS 150 in a manner similar to non-violators, I do not predict an association between ΔMRPS and VIOL. With respect to Model 1, I find (untabulated) that the coefficient on AFFECT remains insignificant (coefficient = −0.047, t = −0.58), as is the coefficient on VIOL (coefficient = −0.024, t = −0.50). With respect to Model 2, I find the coefficients of both PROXLEV (coefficient = 0.049, t = 1.03) and PROXCOV (coefficient = 0.000, t = 0.25) remain insignificant, as does the coefficient on VIOL (coefficient = −0.059, t = –1.11). Finally, in Model 3 the coefficient on INCPROXLEV remains negative and significant (coefficient = −0.040, t = –3.07, p = .002), the coefficient on INCPROXCOV remains insignificant (coefficient = 0.032, t = 0.23), while the coefficient on VIOL is also insignificant (coefficient = −0.015, t = −0.53). As such, the main findings across all three models are robust to the inclusion of a variable controlling for pre-existing covenant violations. Moreover, the additional analysis provides evidence that violators are not significantly different from non-violators in their MRPS restructuring following the introduction of SFAS 150.

I also re-run Models 1, 2 and 3 after excluding violators from the sample. On doing so, the (untabulated) coefficient on AFFECT is negative and significant (coefficient = −0.220, t = –6.21, p = .000) in Model 1, the coefficients on PROXLEV (coefficient = −0.058, t = −0.83) and PROXCOV (coefficient = −0.098, t = −0.99) remain insignificant in Model 2, while the INCPROXLEV coefficient remains negative and significant (coefficient = −0.031, t = –5.59, p = .000) and the INCPROXCOV coefficient remains insignificant (coefficient = −0.050, t = −0.47) in Model 3. As such, the results for Model 1 indicate that non-violators engage in MRPS restructuring as a function of the presence of a covenant constraint that will be tightened due to the introduction of SFAS 150. When read in conjunction with the findings of Models 2 and 3, it is the presence of a leverage covenant that will experience more pronounced constraint tightening due to SFAS 150 that is driving non-violators to engage in MRPS restructuring.

6.2. Firms’ decision to engage in MRPS restructuring

I re-run Models 1, 2 and 3 as logit models by replacing ΔMRPS with a dichotomous variable that equals 1 if the sample firm engaged in MRPS restructuring, 0 otherwise. A firm is deemed to have engaged in MRPS restructuring if ΔMRPS is negative (excluding ‘normal’ maturations and retirements). The results (untabulated) are as follows. First, the coefficient on AFFECT in Model 1 is insignificant (coefficient = 0.186, t = 0.53), indicating that the decision to engage in MRPS restructuring is not associated with the presence of a covenant constraint that will be tightened by the introduction of SFAS 150. Second, in Model 2 the coefficient on PROXLEV is insignificant (coefficient = 0.157, t = 0.87), reflecting firms’ decision to engage in MRPS restructuring is not associated with their proximity to leverage covenant thresholds pre-SFAS 150 adoption. Surprisingly, the coefficient on PROXCOV is negative and significant (coefficient = −0.033, t = –2.46, p = .015), suggesting that firms are less likely to restructure their MRPS the closer their pre-SFAS 150 proximity to interest coverage thresholds. These results, however, may be due to violators choosing not to restructure their MRPS as they are already in violation of their covenant constraints. That is, MRPS restructuring is futile as violation costs have already been incurred. As further analysis, on re-running Model 2 (as a logit model) excluding violators from the sample, the coefficients on PROXLEV (coefficient = 0.269, t = 0.73) and PROXCOV (coefficient = −0.145, t = −0.58) are insignificant, which indicates that non-violators’ decision to engage in MRPS restructuring is not associated with their proximity to covenant thresholds before the introduction of SFAS 150. Finally, in Model 3 the coefficient on INCPROXLEV is positive and significant (coefficient = 0.114, t = 3.40, p = .001), while the coefficient on INCPROXCOV is insignificant (coefficient = 0.216, t = 0.28), which shows that firms are more likely to engage in MRPS restructuring the greater the increase in leverage covenant constraint tightening that would arise from SFAS 150 adoption. Overall, the findings of this study indicate that the decision to engage in MRPS restructuring, and the amount of MRPS restructuring, is not associated with the existence of a debt covenant constraint tightened by SFAS 150, or proximity to covenant thresholds prior to SFAS 150, but is associated with the increase in tightness of leverage covenant constraints that would arise on the introduction of SFAS 150.

6.3. Loan syndication agents

It was argued that the renegotiation costs of syndicated private loans are higher than the renegotiation costs of single-lender private loans because of the additional time and effort in renegotiating with multiple lenders compared with a single lender. No evidence exists from the main findings to support this assertion. Prior literature acknowledges that syndicated private loans may designate one lender to act as an agent for the lender group, with responsibilities including contract renegotiations upon violation (Dennis and Mullineaux, 2000; Francois and Missonier-Piera, 2007). While unanimous lender consent is required for decisions such as any changes in principal, interest or fees, the lender group is not involved in contract negotiations with the borrower. A borrower entering renegotiations with a sole syndication agent is analogous to renegotiating with a single lender, as only one lender is involved in the renegotiation process. As such, the cost of renegotiating syndicated loans with a designated agent may not differ from single-lender loans.

In additional analysis, I perform multivariate regression analysis that includes a dichotomous variable in Models 1, 2 and 3 that equals 1 if the private loan is syndicated with either no syndication agent, or more than one syndication agent, 0 otherwise (measured as SYNAGENT). To the extent the syndicated private loan does not nominate a syndication agent, the borrower enters into renegotiations with the multiple lenders that comprise the syndication group. Similarly, where more than one syndication agent has been nominated, the borrower renegotiates with multiple lenders. The dichotomous variable equals 0 if the private loan has a single lender, or is syndicated with one syndication agent, as the borrower renegotiates with one lender in either instance. Replacing SYN in the regression model, a negative association is predicted between ΔMRPS and SYNAGENT. In untabulated findings, ΔMRPS is not associated with SYNAGENT in any of the models which, when read in conjunction with the main findings for H2, indicates that MRPS firms’ willingness to engage in MRPS restructuring is not influenced by the number of lenders or syndication agents to the debt contract.

6.4. Negative book value firms

As Table 4 indicates, some sample firms have negative book value of equity. In particular, 36 of the 108 sample firms (33.33% of sample firms) report negative book value of equity (measured as Compustat Annual Data Item CEQ) in their most recent annual report prior to SFAS 150.²⁶ Traditionally, prior literature has excluded firms with negative book value of equity from empirical analysis. Justification for exclusion is on the basis that firms with negative book value of equity are in financial distress with high default risk (Fama and French, 1992; Griffin and Lemmon, 2002; Vassalou and Xing, 2004). Moreover, as firms with negative book value of equity have been few, it is argued their omission facilitates empirical findings that are generalisable to the wider population (Brown et al., 2008). Recent literature, however, documents an increase in the frequency of firms reporting negative book value of equity. Jan and Ou (2012) find that the proportion of Compustat firms reporting negative book value of equity has increased from 5.53% during 1976–1985 to 14.86% during 1996–2005. To the extent that negative book value of equity is becoming more pervasive, firms’ exclusion from test samples because of this feature limits the sample’s representativeness. Moreover, the increasing prevalence of negative book value firms is for reasons other than financial distress, including research and development intensity and the mandatory expensing of research and development expenditure (Darrough and Ye, 2007; Jan and Ou, 2012).

As part of additional analysis, I undertake univariate testing to determine whether a firm’s response to SFAS 150 depends on whether positive or negative book value of equity is reported. Splitting the sample into firms with positive and negative book value of equity, both categories report a decline in reported MRPS, providing evidence of MRPS restructuring. Mean ΔMRPS for firms with negative book value is −0.0687, and −0.0192 for firms with positive book value. An independent samples t test indicates no significant difference in these averages (t = 1.234, p = .224). Multivariate analysis is also undertaken to control for a firm’s book value of equity on MRPS restructuring by including a dichotomous variable in Models 1, 2 and 3 that equals 1 if the firm has negative book value of equity, 0 otherwise (measured as BVE). In untabulated findings, the main regression results remain qualitatively similar on controlling for the sign of firms’ book value of equity. In particular, ΔMRPS remains significantly negatively associated with INCPROXLEV (t = –3.46, p < .001) in Model 3, while the coefficients on all other test variables remain insignificant across the various models. Moreover, BVE is not associated with ΔMRPS in any of the models. These results indicate that results are insensitive to the inclusion of negative book value firms in the test sample.

6.5. Industry concentration effects

Descriptive statistics (untabulated) on industry distribution indicate that sample firms are most concentrated within three industries: electric, gas and sanitary services; communication; and business services. The electric, gas and sanitary services industry (Standard Industrial Classification (SIC) = 49) contains the most firms with MRPS, with 30 of 108 sample firms. This is consistent with prior literature documenting utilities’ strong reliance on preferred stock as a source of finance (see, for example, Kimmel and Warfield, 1993, 1995). The communication (SIC = 48) and business services (SIC = 73) industries contain the second and third most number of firms with outstanding MRPS, with 15 and 9 of 108 sample firms, respectively. As both industries invest most in research and development (Chan et al., 2001; Darrough and Ye, 2007), this is consistent with growth firms issuing hybrid securities, including preferred stock, to fund their investment opportunities (Barclay and Smith, 1995). I examine the effects of industry concentration on the main findings by re-running Model 3 on three separate occasions, where on each occasion I exclude firms within one of the industries. For the sake of brevity, my focus is on Model 3 given the main findings indicate it is the increase in proximity to leverage covenant thresholds that is most effective in capturing managers’ concerns about covenant violation due to SFAS 150 adoption. Untabulated results for each of the three separate regressions are consistent with the main findings, indicating that the main multivariate results for Model 3 are not driven by communication, utility or business services firms.

6.6. Credit rating effects

Prior literature argues that a firm’s perceived credit worthiness is enhanced by off–balance sheet debt financing (Engel et al., 1999), and that the reclassification of MRPS as a liability will lead to credit analysts reassessing the financial risk of MRPS firms (Frischmann et al., 1995) and potentially altering their credit rating. As additional analysis, I test this assertion by examining whether firms that engage in MRPS restructuring experience an improvement in credit rating following their capital restructure relative to firms that do not undertake MRPS restructuring. In particular, univariate analysis is undertaken by splitting sample firms with credit ratings (n = 60) into those that report a decline in MRPS outstanding post-SFAS 150 and those that do not. I then examine whether those that report a decline in MRPS outstanding post-SFAS 150 are significantly different from those that do not in two ratings-related areas: (1) credit rating in the financial year preceding SFAS 150 adoption (measured as RATINGpre) and (2) 12-month change in credit rating surrounding a firm’s response to SFAS 150 (measured as RATINGchange, and calculated as RATINGpre – RATINGpost/RATINGpre). A positive (negative) value for RATINGchange indicates an improvement (deterioration) in credit rating. The Standard & Poor’s division of the McGraw-Hill Companies Inc. (S&P) is one of the largest credit rating agencies in the United States. Similar to Ashbaugh-Skaife et al. (2006) and Cheng and Subramanyam (2008), firm-level credit ratings are measured using data on S&P Domestic Long-Term Issuer Credit Rating from Compustat Monthly Data Item SPLTICRM. As ratings are issued on a monthly basis, the rating issued in the last month of the financial year is used to represent a firm’s annual S&P credit rating. The S&P ratings are coded from 1 to 20 with a value of 1 representing a rating of AAA (which indicates an extremely strong capacity of the firm meeting its financial commitments) and a value of 20 representing a rating of D (which indicates the obligor has failed to meet one or more of its obligations when it became due).²⁷ Thus, a higher value of RATINGpre or RATINGpost represents a more unfavourable rating.

The results of the univariate analysis are reported in Panel A of Table 6. Mean RATINGpre for restructuring (non-restructuring) firms is 11.4000 (11.7500), with an independent samples t test finding no significant difference in these averages (t = 0.33, p = .745). Following MRPS restructuring, however, restructuring firms experience, on average, an improvement in credit ratings (RATINGchange = 0.0038) while non-restructuring firms experience, on average, a deterioration in credit ratings (RATINGchange = −0.0389). RATINGchange is significantly different across both sub-samples (t = 1.74, p = .087), and provides initial evidence that rating agencies reassessed MRPS firms’ credit risk in light of their response to the introduction of SFAS 150.

Table 6.

Univariate differences in mean credit ratings pre-SFAS 150 and the mean change in credit ratings post-SFAS 150 between firms engaging in MRPS restructuring and those that do not (Panel A) and regression estimates of the cross-sectional association between sample firms’ change in reported MRPS and their debt contract, credit rating and firm characteristics (Panel B).

Panel A: Univariate statistics.
	Restructure	No restructure	Difference
RATINGpre	11.4000	11.7500	−0.3500
RATINGchange	0.0038	−0.0389	0.0427*
Panel B: Regression estimates.
	Pred. sign	Coefficient	t
INTERCEPT	?	−0.173	−1.63
INCPROXLEV	?	−0.097	−2.56**
INCPROXCOV	?	0.249	0.176
SYN	-	−0.016	−0.73
SEC	-	0.003	0.07
FINRISK	-	0.076	1.21
INCFINRISK	-	−0.710	−2.36**
FCF	-	−0.346	−1.19
SIZE	?	0.008	1.10
RATINGpre	?	0.008	1.33
RATINGimprove	-	−0.160	−2.19**
R ²		.5343
N		60

MRPS: mandatorily redeemable preferred stock.

RATINGpre is a firm’s credit rating in the last month of the financial year preceding SFAS 150 adoption; RATINGimprove is a dichotomous variable that equals 1 if a firm experiences a credit rating improvement post-SFAS 150, 0 otherwise. All other variables are as defined in Table 1. Coefficients are White adjusted for heteroscedasticity.

and ^*indicate significance at the 5% and 10% levels, respectively, one tailed.

Multivariate regression analysis is also undertaken by introducing two new variables into Model 3: RATINGpre, as defined above, and RATINGimprove, which is a dichotomous variable that equals 1 if a firm experiences a rating improvement post-SFAS 150 (i.e. RATINGchange has a positive value), 0 otherwise. The results are reported in Panel B of Table 6 and are consistent with the univariate findings of Panel A. In particular, while RATINGpre is not a determinant of MRPS restructuring, MRPS restructuring is associated with a subsequent improvement in credit rating given the significant negative coefficient on RATINGimprove (t = –2.19, p < .05). Overall, these findings indicate that rating agencies reassessed restructuring firms’ credit risk favourably upon reducing their reliance on MRPS as a source of finance post-SFAS 150.

6.7. Characteristics of non-restructuring firms

As evident from the descriptive statistics reported in Panel A of Table 3, the median value of ΔMRPS ($M) (ΔMRPS (%)) is USD$0.000 (0.000), indicating that half of the sample firms do not report a decline in MRPS outstanding following the adoption of SFAS 150. Moreover, a quarter report an increase in MRPS outstanding post-SFAS 150. These findings warrant investigation of the contract-, firm- and industry-specific determinants of reporting an increase in MRPS post-SFAS 150 and, in particular, whether non-violation considerations are paramount. To achieve this, I create a dichotomous variable (measured as MRPSINC) coded 1 (0) if the firm reports an increase (decrease) in MRPS outstanding post-SFAS 150 and perform two additional tests. First, I examine the correlation between MRPSINC and contract, firm and industry characteristics used in earlier analyses, with the results reported in Panel A of Table 7. Panel A of Table 7 indicates that MRPSINC is not correlated with any covenant violation-related variables, suggesting that reporting an increase in MRPS post-SFAS 150 did not pose covenant violation concerns for these firms. The findings do, however, indicate that firms reporting an increase in MRPS post-SFAS 150 are smaller (SIZE: correlation coefficient = 0.267, p < .05) loss-making firms (NI: correlation coefficient = 0.217, p < .10) with negative book value of equity (BVE: correlation coefficient = 0.318, p < .01) that are less likely to be utility firms (UTIL: correlation coefficient = 0.383, p < .01) but more likely to be business services firms (BUSSERV: correlation coefficient = 0.295, p < .05). Second, univariate statistical analysis is undertaken to determine whether firms reporting an increase in MRPS post-SFAS 150 (i.e. MRPSINC = 1) are significantly different from those reporting a decrease (i.e. MRPSINC = 0). As MRPSINC = 0 firms (MRPSINC = 1 firms) constitute the control (treatment) sub-sample, the t-statistic sign is the inverse of the correlation coefficient sign. The results are reported in Panel B of Table 7, and are consistent with those reported in Panel A. Overall, the findings indicate that firms reporting an increase in MRPS post-SFAS 150 are not driven by covenant violation concerns, but rather reflect firm- and industry-specific considerations.²⁸

Table 7.

Correlation between sample firms reporting an increase in MRPS post-SFAS 150 and their debt contract, firm and industry characteristics (Panel A), and univariate differences between those sample firms reporting an increase and those reporting a decrease in MRPS post-SFAS 150 according to their debt contract, firm and industry characteristics (Panel B).

Panel A: Pearson correlation coefficients.
	AFFECT	PROXLEV	INCPROXLEV	PROXCOV	INCPROXCOV	SYN	SEC	FINRISK	INCFINRISK	FCF	SIZE	NI	BVE	COMM	UTIL	BUSSERV
MRPSINC	−0.091	−0.052	−0.138	0.122	−0.031	−0.068	−0.075	0.056	0.165	−0.000	−0.267**	−0.217*	0.318***	0.027	−0.383***	0.295**
Panel B: Univariate statistics.
	AFFECT	PROXLEV	INCPROXLEV	PROXCOV	INCPROXCOV	SYN	SEC	FINRISK	INCFINRISK	FCF	SIZE	NI	BVE	COMM	UTIL	BUSSERV
MRPSINC	0.67	0.37	1.19	−0.96	0.23	0.49	0.55	−0.51	−1.40	0.03	2.43**	1.81*	−2.83***	−0.27	3.74***	−2.43**

MRPS: mandatorily redeemable preferred stock.

MRPSINC is a dichotomous variable that equals 1 if a firm reports an increase in MRPS post-SFAS 150, 0 if a firm reports an decrease in MRPS post-SFAS 150; BVE is a dichotomous variable that equals 1 if a firm reports negative book value of equity pre-SFAS 150, 0 otherwise; Net income ($M) is net income (loss) in millions pre-SFAS 150 adoption (Compustat Annual Data item NI); COMM is a dichotomous variable that equals 1 if a firm is in the communication industry, 0 otherwise; UTIL is a dichotomous variable that equals 1 if a firm is in the electric, gas and sanitary services industry, 0 otherwise; BUSSERV is a dichotomous variable that equals 1 if a firm is in the business services industry, 0 otherwise. All other variables are as defined in Table 1.

***

^{, **}and ^*indicate significance at the 1%, 5% and 10% levels, respectively, two tailed.

7. Summary and conclusion

This study examined whether managers, in response to the liability classification of MRPS mandated under SFAS 150, restructured MRPS outstanding to keep debt out of the liabilities section of the balance sheet and, consequently, avoid increasingly likely debt covenant violation. As the financial reporting benefits of issuing mezzanine-classified MRPS were eliminated, firms reduced their reliance on MRPS financing post-SFAS 150 by USD2.767 billion, indicating managers’ initial motivations in issuing MRPS was to manage their financial statements. The findings also show an economically and statistically significant association between MRPS restructuring and the increase in proximity to leverage covenant thresholds due to SFAS 150, given firms reduced, on average, reported MRPS by USD271,492 for every 1 percentage point increase in leverage covenant tightness. Managers are, therefore, mindful of the costs of debt covenant violation relative to the costs of transaction restructuring when attempting to circumvent a mandatory accounting change. Overall, these findings indicate that it is the likely impact of a mandatory accounting change on firms’ covenant compliance, as measured by the increase in leverage covenant tightness due to the debt classification of MRPS, which is most effective in capturing managers’ concerns about covenant violation due to a mandatory accounting change. The results are robust to numerous additional tests, including controlling for pre-existing debt covenant violation, firms’ book value of equity and industry concentration effects. Additional analysis also reveals that MRPS restructuring firms experience improved credit ratings following the restructure, while non-restructuring firms’ decision not to restructure is attributable to firm- and industry-specific considerations and not covenant violation concerns.

This study is, however, subject to several limitations. First, this study is limited to those firms that held MRPS at the time of SFAS 150. As such, this study is unable to assess the attractiveness of MRPS as a financing tool post-SFAS 150, as this would require identifying firms that issued MRPS following the introduction of SFAS 150. Second, this study examines managers’ response to a specific mandatory accounting change, namely SFAS 150 and, as such, the findings may not be generalisable in predicting how managers will respond to other accounting changes requiring the recognition of new debt on the balance sheet. Finally, I acknowledge that the main findings are not applicable to all sample firms, but rather only a subset of sample firms whose private debt covenant constraints will be tightened upon SFAS 150 adoption. That is, those covenants whose compliance is measured using rolling GAAP and contain GAAP-consistent accounting definitions. As such, while the control variables are relevant to all sample firms, whether with or without private debt affected by SFAS 150, it is only those sample firms with debt covenants tightened by SFAS 150 that consider the debt contract effects of SFAS 150 when engaging in MRPS restructuring.

Footnotes

Appendix 1

Final transcript accepted 19 November 2018 by Michael Badbury (AE Accounting).

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

American Accounting Association Financial Accounting Standards Committee (2001) Commentary: Evaluation of FASB’s proposed accounting for financial instruments with characteristics of liabilities, equity, or both. Accounting Horizons 15: 387–400.

Ashbaugh-Skaife

Collins

LaFond

(2006) The effects of corporate governance on firms’ credit ratings. Journal of Accounting and Economics 42: 203–243.

Asquith

Beatty

Weber

(2005) Performance pricing in bank debt contracts. Journal of Accounting and Economics 40: 101–128.

ASR 268 (1983) Presentation in financial statements of ‘redeemable preferred stocks’.

Ayers

Lefanowicz

Robinson

(2002) Do firms purchase the pooling method? Review of Accounting Studies 7: 5–32.

Barclay

Smith

(1995) The priority structure of corporate liabilities. Journal of Finance 50: 899–916.

Beatty

(2007) How does changing measurement change management behavior? A review of the evidence. Accounting and Business Research 37: 63–71.

Beatty

Weber

(2003) The effects of debt contracting on voluntary accounting method changes. Accounting Review 78: 119–142.

Beatty

Berger

Maglioli

(1995) Motives for forming research & development financing organizations. Journal of Accounting and Economics 19: 411–442.

10.

Begley

(1990) Debt covenants and accounting choice. Journal of Accounting and Economics 12: 125–139.

11.

Begley

Freedman

(2004) The changing role of accounting numbers in public lending agreements. Accounting Horizons 18: 81–96.

12.

Beneish

Press

(1993) Costs of technical violation of accounting-based debt covenants. Accounting Review 68: 233–257.

13.

Beneish

Press

(1995) The resolution of technical default. Accounting Review 70: 337–353.

14.

Bens

Monahan

(2008) Altering investment decisions to manage financial reporting outcomes: Asset-backed commercial paper conduits and FIN 46. Journal of Accounting Research 46: 1017–1055.

15.

Bowman

(1979) The theoretical relationship between systematic risk and financial (accounting) variables. Journal of Finance 34: 617–630.

16.

Bowman

(1980) The debt equivalence of leases: An empirical investigation. Accounting Review 55: 237–253.

17.

Bradley

Roberts

(2015) The structure and pricing of corporate debt covenants. Quarterly Journal of Finance 15: 1–37.

18.

Brown

Izan

Loh

(1992) Fixed asset revaluations and managerial incentives. Abacus 28: 36–57.

19.

Brown

Lajbcygier

(2008) Going negative: What to do with negative book equity stocks. Journal of Portfolio Management 35: 95–102.

20.

Chan

Lakonishok

Sougiannis

(2001) The stock market valuation of research and development expenditures. Journal of Finance 56: 2431–2456.

21.

Chance

(1982) Evidence of a simplified model of systematic risk. Financial Management 11: 53–63.

22.

Chen

Wei

(1993) Creditors’ decisions to waive violations of accounting-based debt covenants. Accounting Review 68: 218–232.

23.

Cheng

Subramanyam

(2008) Analyst following and credit rating. Contemporary Accounting Research 25: 1007–1043.

24.

Citron

(1992) Accounting measurement rules in UK bank loan contracts. Accounting and Business Research 23: 21–30.

25.

Collins

Rozeff

Dhaliwal

(1981) The economic determinants of the market reaction to proposed mandatory accounting changes in the oil and gas industry. Journal of Accounting and Economics 3: 37–71.

26.

Comiskey

Mulford

(1986) Investment decisions and the equity accounting standard. Accounting Review 61: 519–525.

27.

Cotter

(1998) Utilisation and restrictiveness of covenants in Australian private debt contracts. Accounting and Finance 38: 181–196.

28.

Cotter

(1999) Asset revaluations and debt contracting. Abacus 35: 268–285.

29.

Darrough

(2007) Valuation of loss firms in a knowledge-based economy. Review of Accounting Studies 12: 61–93.

30.

DeFond

Jiambalvo

(1994) Debt covenant violation and manipulation of accruals. Journal of Accounting and Economics 17: 145–176.

31.

Dennis

Mullineaux

(2000) Syndicated loans. Journal of Financial Intermediation 9: 404–426.

32.

Dichev

Skinner

(2002) Large-sample evidence of the debt covenant hypothesis. Journal of Accounting Research 40: 1091–1123.

33.

Duke

Hunt

(1990) An empirical examination of debt covenant restrictions and accounting-related debt proxies. Journal of Accounting and Economics 12: 45–63.

34.

El-Gazzar

(1993) Stock market effects of the closeness to debt covenant restrictions resulting from capitalization of leases. Accounting Review 68: 258–272.

35.

El-Gazzar

Pastena

(1990) Negotiated accounting rules in private financial contracts. Journal of Accounting and Economics 12: 381–396.

36.

El-Gazzar

Pastena

(1991) Factors affecting the scope and initial tightness of covenant restrictions in private lending agreements. Contemporary Accounting Research 8: 132–151.

37.

Elliott

Richardson

Dyckman

et al . (1984) The impact of SFAS No. 2 on firm expenditures on research and development: Replications and extensions. Journal of Accounting Research 22: 85–102.

38.

Ely

(1995) Operating lease accounting and the market’s assessment of equity risk. Journal of Accounting Research 33: 397–415.

39.

Engel

Erickson

Maydew

(1999) Debt-equity hybrid securities. Journal of Accounting Research 37: 249–274.

40.

Espahbodi

Tehranian

(1995) Equity price reaction to the pronouncements related to accounting for income taxes. Accounting Review 70: 655–668.

41.

Espahbodi

Strock

Tehranian

(1991) Impact of equity prices of pronouncements related to nonpension retirement benefits. Journal of Accounting and Economics 14: 323–346.

42.

Fama

French

(1992) The cross-section of expected stock return. Journal of Finance 47: 427–465.

43.

Fargher

Sidhu

Tarca

et al (2018) Accounting for financial instruments with characteristics of debt and equity: Finding a way forward. Accounting and Finance (in press).

44.

Fogelson

(1978) The impact of changes in accounting principles on restrictive covenants in credit agreements and indentures. Business Lawyer 33: 769–787.

45.

Francois

Missonier-Piera

(2007) The agency structure of loan syndicates. Financial Review 42: 227–245.

46.

Frischmann

Kimmel

Warfield

(1995) Innovative forms of preferred stock: Debt or equity? Commercial Lending Review 10: 14–27.

47.

Frost

Bernard

(1989) The role of debt covenants in assessing the economic consequences of limiting capitalization of exploration costs. Accounting Review 64: 788–808.

48.

Godfrey

Warren

(1995) Lessee reactions to regulation of accounting for leases. Abacus 31: 201–228.

49.

Griffin

Lemmon

(2002) Book-to-market equity, distress risk, and stock returns. Journal of Finance 57: 2317–2336.

50.

Hamada

(1972) The effect of the firm’s capital structure on the systematic risk of common stocks. Journal of Finance 27: 435–452.

51.

Hill

Stone

(1980) Accounting betas, systematic operating risk, and financial leverage: A risk-composition approach to the determinants of systematic risk. Journal of Financial and Quantitative Analysis 15: 595–637.

52.

Holthausen

(1981) Evidence on the effect of bond covenants and management compensation contracts on the choice of accounting techniques: The case of the depreciation switch-back. Journal of Accounting and Economics 3: 73–109.

53.

Holthausen

Leftwich

(1983) The economic consequences of accounting choice implications of costly contracting and monitoring. Journal of Accounting and Economics 5: 77–117.

54.

Horwitz

Kolodny

(1980) The economic effects of involuntary uniformity in the financial reporting of R&D expenditures. Journal of Accounting Research 18: 38–74.

55.

Imhoff

Thomas

(1988) Economic consequences of accounting standards: The lease disclosure rule change. Journal of Accounting and Economics 10: 277–310.

56.

Jan

(2012) Negative book value firms and their valuation. Accounting Horizons 26: 91–110.

57.

Jiang

(2008) Beating earnings benchmarks and the cost of debt. Accounting Review 83: 377–416.

58.

Kennedy

(2003) A Guide to Econometrics. Oxford: Blackwell Publishing.

59.

Kimmel

Warfield

(1993) Variation in attributes of redeemable preferred stock: Implications for accounting standards. Accounting Horizons 7: 30–40.

60.

Kimmel

Warfield

(1995) The usefulness of hybrid security classifications: Evidence from redeemable preferred stock. Accounting Review 70: 151–167.

61.

Kirchner

(2003) Will FASB 150 have a chilling effect on entity buy-sell plans? National Underwriter, 15 September, 42.

62.

Leftwich

(1981) Evidence of the impact of mandatory changes in accounting principles on corporate loan agreements. Journal of Accounting and Economics 3: 3–36.

63.

Leftwich

(1983) Accounting information in private markets: Evidence from private lending agreements. Accounting Review 58: 23–42.

64.

Levi

Segal

(2015) The impact of debt-equity reporting classifications on the firm’s decision to issue hybrid securities. European Accounting Review 24: 801–822.

65.

(2014) Impact of mandatory changes in accounting: Evidence from FASB APB 14-1. PhD Thesis, University of Toronto, Toronto, ON, Canada.

66.

Loud

(2004) Statement of financial accounting standards No. 150: Accounting for certain financial instruments with characteristics of both liabilities and equity. Insights: Willamette Management Associates Winter, 29–31. Available at: http://willametteinsights.com/04/Winter04article2.pdf (accessed 21 June 2016).

67.

Lys

(1984) Mandated accounting changes and debt covenants: The case of oil and gas accounting. Journal of Accounting and Economics 6: 39–65.

68.

McCarthy

Schneider

Seese

(2004) FAS No. 150: The effect on debt covenants. Commercial Lending Review 19: 36–40.

69.

Marquardt

Wiedman

(2005) Earnings management through transaction restructuring: Contingent convertible debt and diluted earnings per share. Journal of Accounting Research 43: 205–250.

70.

Marquardt

Wiedman

(2007) Economic consequences of financial reporting changes: Diluted EPS and contingent convertible securities. Review of Accounting Studies 12: 487–523.

71.

Mather

(1999a) Financial covenants and related contracting processes in the Australian private debt market: An experimental study. Accounting and Business Research 30: 29–42.

72.

Mather

(1999b) Financial covenants in Australian bank-loan contracts: Incidence, measurement rules and monitoring. Australian Accounting Review 9: 63–72.

73.

Mather

Peirson

(2006) Financial covenants in the markets for public and private debt. Accounting and Finance 46: 285–307.

74.

Matsunaga

(1995) The effects of financial reporting costs on the use of employee stock options. Accounting Review 70: 1–26.

75.

Mittelstaedt

Nichols

Regier

(1995) SFAS No. 106 and benefit reductions in employer-sponsored retiree health care plans. Accounting Review 70: 535–556.

76.

Mohrman

(1993) Debt contracts and FAS No. 19: A test of the debt covenant hypothesis. Accounting Review 68: 273–288.

77.

Mohrman

(1996) The use of fixed GAAP provisions in debt contracts. Accounting Horizons 10: 78–91.

78.

Moser

Newberry

Puckett

(2011) Bank debt covenants and firms’ responses to FAS 150 liability recognition: Evidence from trust preferred stock. Review of Accounting Studies 16: 355–376.

79.

Mulford

Maloney

(2003) SFAS No. 150 and Mandatorily Redeemable Preferred Stock. Atlanta, GA: DuPree Financial Analysis Lab, Georgia Institute of Technology. Available at: https://www.scheller.gatech.edu/centers-initiatives/financial-analysis-lab/files/2003/ga_tech_SFAS150_2003%20SFAS%20No.%20150%20and%20Mandatorily%20Redeemable%20Preferred%20Stock.pdf (accessed 21 June 2016).

80.

Nair

Rittenberg

Weygandt

(1990) Accounting for redeemable preferred stock: Unresolved issues. Accounting Horizons 4: 33–41.

81.

Neuhausen

(2003) FASB statement No. 150 brings big changes. Insights: The Corporate & Securities Law Advisor 17(9): 23–27.

82.

Nini

Smith

Sufi

(2009) Creditor control rights and firm investment policy. Journal of Financial Economics 92: 400–420.

83.

Nini

Smith

Sufi

(2012) Creditor control rights, corporate governance, and firm value. Review of Financial Studies 25: 1713–1761.

84.

Press

Weintrop

(1990) Accounting-based constraints in public and private debt agreements: Their association with leverage and impact on accounting change. Journal of Accounting and Economics 12: 65–95.

85.

Ramsay

Sidhu

(1998) Accounting and non-accounting based information in the market for debt: Evidence from Australian private debt contracts. Accounting and Finance 38: 197–221.

86.

Rapoport

Weil

(2003) More truth-in-labelling for accounting carries liabilities. The Wall Street Journal, 28 August, C1.

87.

Reinstein

(2004) SFAS No. 150: Accounting for financial instruments with liabilities and equity characteristics. The RMA Journal 86(5): 53–55.

88.

Roberts

Sufi

(2009a) Control rights and capital structure: An empirical investigation. Journal of Finance 64: 1657–1695.

89.

Roberts

Sufi

(2009b) Renegotiation of financial contracts: Evidence from private credit agreements. Journal of Financial Economics 93: 159–184.

90.

Salatka

(1989) The impact of SFAS No. 8 on equity prices of early and late adopting firms: An event study and cross-sectional analysis. Journal of Accounting and Economics 11: 35–69.

91.

Schneider

(1993) RIP for RPS? Will FASB bury redeemable preferred stock? Corporate Cashflow 14(5): 45–46.

92.

Schroeder

Sevin

Schauer

(2006) The economic consequences of the statement of financial accounting standards (SFAS) No. 150. International Advances in Economic Research 12: 498–504.

93.

Securities Exchange Act of 1934.

94.

Selto

Clouse

(1985) An investigation of managers’ adaptations to SFAS No. 2: Accounting for research and development costs. Journal of Accounting Research 23: 700–717.

95.

SFAS 150 (2003) Accounting for certain financial instruments with characteristics of both liabilities and equity.

96.

Sheskin

(2007) Handbook of Parametric and Nonparametric Statistical Procedures. Boca Raton, FL: Chapman & Hall/CRC.

97.

Shevlin

(1987) Taxes and off-balance sheet financing: Research and development limited partnerships. Accounting Review 62: 480–509.

98.

Sinnett

(2003) SFAS 150’s unintended alchemy: Turning positive into negative. Financial Executive 19: 40–43.

99.

Smith

(1993) A perspective on accounting-based debt covenant violations. Accounting Review 68: 289–303.

100.

Smith

Warner

(1979) On financial contracting: An analysis of bond covenants. Journal of Financial Economics 7: 117–161.

101.

Stokes

Tay

(1988) Restrictive covenants and accounting information in the market for convertible notes: Further evidence. Accounting and Finance 28: 57–73.

102.

Stokes

Whincop

(1993) Covenants and accounting information in the market for classes of preferred stock. Contemporary Accounting Research 9: 463–478.

103.

Sufi

(2009) Bank lines of credit in corporate finance: An empirical analysis. The Review of Financial Studies 22: 1057–1088.

104.

Sweeney

(1994) Debt-covenant violations and managers’ accounting responses. Journal of Accounting and Economics 17: 281–308.

105.

Vassalou

Xing

(2004) Default risk in equity returns. Journal of Finance 59: 831–868.

106.

Wasley

Linsmeier

(1992) A further examination of the economic consequences of SFAS No. 2. Journal of Accounting Research 30: 156–164.

107.

Watts

Zimmerman

(1986) Positive Accounting Theory. Englewood Cliffs: Prentice Hall.

108.

Wertheim

Schneider

(1992) Proposed reclassification of redeemable preferred stock: The impact on financial statement ratios. Journal of Corporate Accounting and Finance 4: 161–168.

109.

Whittred

Chan

(1992) Asset revaluations and the mitigation of underinvestment. Abacus 18: 58–74.

110.

Whittred

Zimmer

(1986) Accounting information in the market for debt. Accounting and Finance 26: 19–33.