Accounting-based debt covenants

2005, Peasnell et al. 2005). Research findings from these studies will be referred to when relevant, but the findings have to be interpreted with some caution. In recent years, earnings management has been examined by alternative methods. To increase the power of the research design, specific reporting decisions rather than aggregate accruals are investigated. Some studies are also conducted on firms that are known to have managed earnings ex post.

Recent evidence has demonstrated that earnings-based compensation explains earnings management at a discount relative to equity-based incentives (Schipper and Vincent 2003, Graham et al. 2005, Yaari and Ronen 2008:80). Moreover, conditional stocks and stock options have become a major component of top management-compensation packages (Hall and Liebman 1998, Murphy 1999, Hall and Murphy 2002, Denis, Hanouna and Sarin 2006). This may suggest that earnings-based compensation is of less importance. The asset-impairment literature has to a limited extent investigated earnings-based compensation incentives. Some exceptions are found in Beatty and Weber (2006), Lapointe-Antunes, Cormier and Magnan (2008) and Ramanna and Watts (2009). Ramanna and Watts (2009) include an indicator variable for CEO cash-bonus payments.

They find an insignificant association between this indicator variable and goodwill-impairment losses. Similar evidence is found by Lapointe-Antunes et al.

(2008). Beatty and Weber (2006), however, report a significantly positive association between the indicator variable for bonus payment and goodwill-impairment losses.

managers and shareholders and managers and debtholders (Black and Scholes 1973, Merton 1974, Jensen and Meckling 1976, Smith and Warner 1979, Leftwich 1983). As shareholders are concerned that managers are too risk-averse, the debtholders on the other hand are concerned that the shareholders are too much of a risk-taker. Debtholders prefer low-risk projects that increase the probability of debt and interests being paid. The shareholders’ claim is analogous to a call option on the firm’s assets with an exercise price equal to the face value of debt. The debtholders’ claim, however, is analogous to a put option in that their upside is equal to the face value of debt. If the firm value falls below the face value of debt, the debtholders lose the difference between the face value of debt and the firm value (Black and Scholes 1973, Merton 1974). Shareholders can potentially transfer wealth from debtholders to themselves by investing in riskier assets than expected when the debt was issued (asset substitution). The potential of wealth transfer increases as firm value falls and the shareholders call option moves from being well in the money to being at or close to the money, and it becomes particularly actute as this option falls out of the money (Merton 1974, Kothari et al. 2010).

Debt covenants are intended and designed to restrict managers from engaging in investment and financing decisions that reduce the value of the debtholders’ claim (Smith and Warner 1979, Leftwich 1983, Guay 2008). Because debt covenants frequently are written on accounting numbers and violation of these covenants are believed to be costly for the firm, managers of firms that are close to violating debt covenants are supposed to make reporting decisions that reduce the likelihood of default (Watts and Zimmerman 1986:186-91). This leads to the debt-equity hypothesis formulated by Watts and Zimmerman (1986:216): “Ceteris paribus, the larger a firm’s debt/equity ratio, the more likely the firm’s manager is to select

accounting procedures that shift reported earnings from future periods to the current period.” In more general terms, this hypothesis can be rephrased as a debt-covenant hypothesis, where managers have incentives to make reporting decisions that reduce the likelihood of debt-covenant violation (Field et al. 2001). The strength of these incentives depends on the expected costs of violation.

Debtholders will not engage in contracting unless they expect to be better off writing these contracts. If the agency costs due to opportunism are expected ex ante, Jensen and Meckling (1976) and Watts and Zimmerman (1986:189) demonstrate that these agency costs to a large extent will be borne by managers and shareholders. The debtholders will on average be price-protected, which makes contracting unnecessary. This conclusion, however, is based on the assumption that debt is traded in an efficient debt-capital market, which is generally not the case. Without efficient markets for debt, the debtholders will not be price-protected. Under such conditions, contracts serve an important role to align the interests of managers and shareholders with those of the debtholders. An efficient contract will, therefore, minimise the agency costs. If, however, the contracting and information costs are too high, the contract may turn out to be inefficient and itself provide incentives for earnings management.

The debt contracts may include different covenants. For instance, there might be covenants that constrain managers’ decisions regarding dividend payouts, future debt issuances, participation in mergers and disposition of assets (Leftwich 1983, Dichev and Skinner 2002). Debt covenants can be accounting-based or non-accounting-based and will appear more often and be tighter in private rather than public debt contracts. A variety of accounting-based ratios are used to set debt covenants. Ratios such as debt-to-cash flows, debt-to-equity and interest coverage

are intended to measure the firm’s ability to make debt-related payments (Dichev and Skinner 2002). These covenants are calculated on current GAAP or modified GAAP (Leftwich 1983). When current GAAP is the relevant basis for calculation, accounting can either be frozen at the time of the debt issuance or be allowed to follow changes in GAAP over time.

Departures from current GAAP are quite common. Generally, these departures lead to more conservative accounting as certain increases in net earnings and net-asset values are excluded when calculating the covenants. For instance, in some contracts accounting for intangibles such as goodwill are excluded (Leftwich 1983, Holthausen and Watts 2001, Watts 2003, Beatty, Weber and Yu 2008).

Debtholders are often believed to use a liquidation approach. Some assets such as goodwill are expected to have liquidation values equal to zero. This justifies exclusion of book goodwill (Holthausen and Watts 2001, Beatty et al. 2008, Kothari et al. 2010). The exclusion might also be justified on the basis of unverifiability and asset-value uncertainty (Leftwich 1983, Kothari et al. 2010).

Guay (2008) find evidence consistent with book goodwill being excluded from the calculation of net-worth covenants. He argues, however, that the exclusion of intangible assets in net worth is likely to vary across firms, depending on the intensity of recognised intangible assets and the importance of intangible assets in the business model. For firms with few intangible assets, debtholders may exclude intangible assets when calculating net-worth covenants. For firms with lots of intangible assets, however, debtholders will likely include book values of these assets, if available, when calculating the covenants. There are at least two reasons why intangible assets are included. First, tangible net worth may not be a relevant metric of financial health in these firms. When a large fraction of the assets are intangible, debtholders will probably want to obtain decision rights when

intangible assets are substantially impaired. Second, debtholders may have an interest in monitoring the covenant, intangible assets to tangible assets. As intangible assets are amortised or impaired over time, the firm must recoup those earnings effects in cash flows or tangible assets to avoid losing covenant slack (Guay 2008, Zang 2008). Lambert (2010) argues that all assets, also intangible assets such as goodwill, are relevant for debtholders. The reason is that all assets generate cash flows which can be used to pay off debt. He also argues that the liquidiation approach is only relevant for debtholders when firms are in financial distress. In all other situations, profitability and cash-generating capacity of the firms are of most interest.

Unfortunately, details of debt covenants are generally unavailable to researchers.

Empirical tests of the debt-covenant hypothesis frequently rely on variables that are supposed to be positively correlated with debt covenants. The most frequently used indicator is the debt-to-equity ratio (Duke and Hunt 1990, Smith 1993, Dechow et al. 1996). Later research generally relies on actual covenant data (Healy and Palepu 1990, Beneish and Press 1993, Smith 1993, Sweeney 1994, Dichev and Skinner 2002). Watts and Zimmerman (1986:216) argue that the debt-to-equity ratio is a reasonable approximation of most debt covenants, as the likelihood of these other covenants being violated will increase in debt-to-equity ratio. Nonetheless, they encourage researchers to increasingly rely on details of debt covenants. Duke and Hunt (1990) examine empirically the accounting-based debt-covenant details. They find that the debt-to-equity ratio captures the most common accounting-based restrictions used in actual debt covenants. They conclude that researchers are comparatively safe to use the debt-to-equity ratio as a proxy for actual covenants. Others, however, oppose this conclusion. Dichev and Skinner (2002), for instance, argue that debt-to-equity only to a limited extent

correlate with firms’ actual closeness to covenant restrictions. They conclude that the debt-to-equity-ratio is a fairly noisy proxy for managers’ reporting incentives triggered by debt covenants. Despite the mixed evidence on the validity of the debt-to-equity ratio, this proxy is widely employed in the asset-impairment literature (Lo and Tan 2002, Segal 2003, Sellhorn 2004, Kvaal 2005, Lapointe-Antunes et al. 2008, Zang 2008).

The incentives to avoid covenant violations will be a function of the probability of violation and the expected default costs imposed on the firm given violation. The probability of violation will be determined by the debt-covenant slack and the choices made regarding the calculation of the debt covenant (Dichev and Skinner 2002). The expected default cost is in focus here. If a firm is technically default, this may result in significant default costs (Sweeney 1994). Beneish and Press (1993) demonstrate that the cost associated with technical defaults is quite significant. They estimate that refinancing costs resulting from interest-cost increases vary from 0.84% to 1.63% of the market value of the borrower’s equity.

Gopalakrishnan and Parkash (1995) identify six potential debtholder responses to covenant violations: termination of the debt contract, demand for immediate repayment, increased collateral, increased interest rate, imposition of additional covenant constraints and a waiver. Immediate repayment is rare. The common response is a waiver. Dichev and Skinner (2002) for instance, demonstrate that violations occur rather frequently. They find that approximately one-third of the loans violated covenants. In addition, most loans with debt covenants had multiple violations. The same results are also demonstrated by Gopalakrishnan and Parkash (1995). Both lenders and borrowers indicated a waiver as the most likely response to the violation of an accounting-based debt covenant. This suggests that other information sources than financial statements are used to decide whether to waive

or not (Lambert 2010). Still, a violation will likely impose some default costs on the firm.

This provides the managers and shareholders with incentives to try to avoid violations by making earnings-increasing reporting decisions. Sweeney (1994) examines the time series of reporting decisions prior to firms violating accounting-based debt covenants. She investigates whether managers change accounting methods, which type of accounting methods managers change, when they make these changes and to what extent these changes affect the restrictiveness of accounting-based covenants. Her findings demonstrate that firms approaching violations of accounting-based covenants are more likely to make earnings-increasing accounting changes and early adopt earnings-earnings-increasing mandatory accounting changes relative to a sample of control firms matched on industry and size. Beneish and Press (1993) find evidence in line with Sweeney (1994). They find that debt-covenant violators make earnings-increasing reporting decisions in the year of violation and up to five years prior to the violation. Using accrual-estimation models for investigating earnings management, DeFond and Jiambalvo (1994) find that managers use discretionary accruals to avoid debt-covenant violations. They examine firms that report debt-covenant violations in their financial statement, and their findings suggest that there are positive discretionary accruals in the year prior to the violation and the year of the violation. Some evidence, however, contradicts these findings. DeAnglo, DeAnglo and Skinner (1994) argue and find evidence inconsistent with the debt-covenant hypothesis.

They state that “(…) managers of troubled firms have incentives to take discretionary write-offs that signal to the lenders their willingness to acknowledge and deal with the firm’s problems” (DeAnglo et al. 1994:134). They find that in the year of the dividend reduction 40 out of 76 firms report impairment losses or

restructuring charges. The incentives to do so are found in the desire to affect the renegotiation outcomes. More than 87% of the firms were renegotiating contracts with lenders or labour unions, had changes in top management and/or lobbied for governmental support, all of which plausibly motivated managers to reduce reported earnings. In contrast, Dichev and Skinner (2002) find a significantly higher proportion of firms reporting accounting numbers slightly above current ratio and tangible-net-worth constraints. Taken together, it is reasonable to predict that managers make reporting decisions to avoid covenant violations.