What do we know about the role of financial reporting in debt contracting and debt covenants?

Abstract

The paper examines the role of financial reporting in debt contracting and in particular focuses on the definition, measurement, and monitoring of accounting-based covenants used to manage agency relationships arising from borrowing by firms. The paper also reviews research in areas of financial reporting where the presence of accounting-based covenants provides incentives to managers, notably choice of accounting method, lobbying on standard setters' proposals, and accounting earnings management. Although US dominated and latterly increasingly focused on large datasets and quantitative and analytical methods, relevant research is available from a range of methodologies and countries and the paper reflects this variety and identifies both inter-jurisdictional differences and inter-temporal changes in debt contracting practices. Despite the extensive research which is reviewed important areas for new research remain.

Keywords:

• debt contracting
• debt covenants
• financial reporting

Notes

^†This paper was invited by the editor to complement the paper by Shivakumar (2013) in this Special Issue, in order to reflect observations of participants in the 2012 ICAEW conference on Information for Better Markets that it would be useful to have in parallel a specific review of complementary prior research on the relationship between financial reporting and debt contracting. The observations of a reviewer are acknowledged with appreciation.

Cash flow default arises when a borrower fails to meet repayments of interest or principal while technical default refers to breaches of accounting-based and other covenants including cross-default.

This key representative reference should be set in a broader context. More generally, agency theory/property rights theory and costly contracting theory are characteristic of new institutional economics and law and are based on the work of many researchers from Coase (1937) onwards (see Mercuro and Medema 1997, especially Chapter 5). Important underpinning is found in theoretical research in economics (e.g. Hart and Moore 1995, Hart 2001, the latter for a survey).

Some equivalent UK practitioner material on boilerplate debt contracts was available in this period (Lingard 1985), Encyclopaedia of Forms and Precedents (2010), or emerged later (ACT 1991).

For historical background on debt covenants see Day and Taylor (forthcoming).

Covenant-lite contracting involves relaxed (or no) accounting and finance maintenance covenants but incurrence covenants more traditionally used in high yield bonds. Covenant-lite contracting grew in popularity in the period preceding the global financial problems beginning in 2007.

This paper deals only briefly with contemporary large dataset research where its findings relate to other parts of the paper and as it affects future research possibilities.

Other accounting-based covenants appearing infrequently were: interest cover (six cases), minimum fixed assets to secured debt (two), and maximum assets acquired to capital and reserves (two). Five other accounting-based covenants appeared once each. See Citron (1995, Table 3, p. 145).

See also Begley (1994).

Four other categories related to restrictions on management behaviour and several covenants allocated to these categories had accounting elements and had been treated as de facto accounting-based covenants by other researchers. Examples are: dividend restrictions (limited to a certain percentage of net income or some other ratio); debt issuance and seniority restrictions; and asset disposal and M&A restrictions. In all, 27 types of non-accounting management restrictions were identified.

Reporting covenant incidence is not the purpose of Nikolaev's paper. However, he reports it in summary form (see Table A). Briefly, almost all contracts contained at least one covenant; the mean number of covenants was 7.42 per contract; 41% of contracts contained at least one accounting-based covenant with the modal number being two (in 33% of all contracts); dividend restrictions were present in 40% of contracts, mode two covenants (in 31% of all contracts); restrictions on M&A, investment, and asset dispositions were almost universal (in 93% of contracts) with two or four covenants being most frequent (47% and 38% of all contracts, respectively); and covenants limiting financing activities were in 72% of contracts, mode two covenants (in 39% of all contracts).

Day and Taylor (2001) report the regular occurrence of covenants restricting working capital, frequently accompanied by the use of inventory as collateral, in UK uncommitted (i.e. overdraft) loan facilities. This finding is interesting as research on private debt emphases committed (i.e. medium and longer term) debt as the assumed source of covenant effects for firms. The characteristics of restrictions in overdraft/short-term debt are under-researched.

See Citron et al. (1997) on debt contracting for MBOs.

Two further companies, whilst not having actual cash flow covenants, had covenants in contracts applicable to US loans which were similar to cash flow covenants. Additionally, two companies had contracts requiring provision of cash flow forecasts and two others contracts identifying cash flow shortfalls as events of default.

See Beatty et al. (2002) on flexibility in covenant calculations associated with accounting changes.

See Shivakumar (2013) for a discussion of this US research. See also Ball et al. (2008) for additional material on performance pricing.

See Shivakumar (2013) for a discussion of this research from different perspectives.

Material loan contracts are required to be reported by the SEC and are an exhibit to 10-K or 10-Q filings or an attachment to 8-K filings (Freundenberg et al. 2012).

The classifications are as follows: debt to EBITDA, interest cover, fixed charge cover, net worth, debt to capitalisation, tangible net worth, senior debt to EBITDA, EBITDA, current ratio, debt service cover, debt to net worth, quick ratio, asset cover ratio, cash and cash equivalents, senior debt to capitalisation, working capital, and senior debt to net worth (Freundenberg et al. 2012, Table II).

In his classification income statement-based covenants are: interest cover, and fixed charge cover (earnings to inter alia interest expense, principal payments, capital expenditure, tax, and dividends); balance sheet-based covenants are: leverage (total debt to total assets or total net assets), net worth (including all assets or only tangible assets), and current ratio. The debt-to-earnings covenant is classified as income statement-based since debt was not subject to fair value rules during the sample period.

Over the study period interest cover use was stable, fixed charge cover was broadly stable in use, as was use of debt-to-earnings; in contrast, use of all balance sheet covenants declined with leverage use falling by more than half, while net worth and current ratio covenants declined more markedly.

Their classification is: capital-covenants: quick ratio, current ratio, debt-to-equity ratio, loan-to-value ratio, debt to tangible net worth, leverage ratio, senior leverage ratio, and net worth requirement; performance-covenants: cash interest coverage, debt service coverage ratio, level of EBITDA, fixed charge coverage ratio, interest cover, debt to EBITDA, and senior debt to EBITDA (Christensen and Nikolaev 2012, Appendix A, pp. 106–107).

As reported in Accountancy Age (2010), a US corporation included the following in a debt contract: ‘Audited consolidated balance sheets of the group members … [must be] reported on by and accompanied by an unqualified report from a Big Four accounting firm.’

CEO inside leverage is the ratio of debt to equity compensation received. Executive debt compensation is pensions and other deferred compensation which may be unfunded and unsecured and hence have debt characteristics. A high ratio of debt to equity compensation may align executive incentives more closely to debtholders than shareholders, especially when measured in relation to firm leverage and may signal reduced agency costs (Sundaram and Yermack 2007).

Citron concludes that the gearing finding offers support to the use of the debt/equity hypothesis in the UK not because it proxies for closeness to covenant violation but because high gearing is associated with greater likelihood of presence of accounting-based covenants.

The 18 banks included the 5 clearing banks with national coverage, 3 with smaller national networks, 6 investment banks, and UK branches of overseas banks.

These are behavioural not administrative classifications. Cash flow lending is supported by future cash flows while asset-based is based on collateral. Interviewees reported that the former was controlled by performance covenants and the latter by covenants on asset values, borrowing, and seniority; the former were likely ceteris paribus to be more stringent than the latter.

Zhang (2009) using a sample of US firms from 1988 to 2007 observed that covenants were stricter when set during downturns in the business cycle; as noted above, Demerjian (2011) did not conclude that competition contributed to change in use of balance sheet covenants.

See also Drucker and Puri (2009) for evidence that banks use tighter covenants to substitute for reputation in the secondary loan market.

Measurement rules in debt contracts may also affect management preferences for alternative accounting policies in financial statements. See Gore et al. (2000) who found that use of rolling GAAP in debt covenants was an important influence on corporate management's accounting preferences on the issue of goodwill.

The common form of information covenant stated that submission of annual accounts should be ‘ … as soon as they are available and not later than 180 days from the end of each accounting reference period’ (Day and Taylor 1995, p. 19).

Highly standardised legal forms for definitions in part resulted from the influence of a small number of large law firms advising all sample banks. For an analysis of other factors creating contract standardisation see Kahan and Klausner (1997).

Some banks sought to resolve ambiguity and disputes by recourse to auditors by inserting terms such as: ‘If there is any dispute as to any interpretation or computation for [the] paragraph above, the interpretation or computation of the auditors for the time being of the company will prevail’ Day and Taylor (1995, p. 21).

Li (2010) argues that negotiated measurement rules in debt contracts aim to make accounting variables more precise so as to increase contracting efficiency. He examines definitions of covenants involving net income and net worth in a sample of US private debt contracts and concludes that usefulness of accounting variables in debt contracting depends on trading-off ‘noise’ in accounting numbers with contracting relevance. See Shivakumar (2013) for a discussion of Li's findings in relation to fair value accounting and debt contracting.

PAT has been contested from various perspectives from the outset. See Christenson (1983), Tinker et al. (1982), Lowe et al. (1983), Whittington (1987), Sterling (1990), Chambers (1993), Sue (1997), Neu (1997), and Milne (2002) for critiques and commentary.

For discussion of financial reporting management and contracting for management compensation see Shivakumar (2013) and political costs Milne (2002).

Reviews of empirical research on inter alia financial reporting and debt contracts are available in Watts and Zimmerman (1986, 1990), Whittington (1987), Fields et al. (2001) Armstrong et al. (2010).

See Shivakumar (2013) for a review and discussion of implications of such recent research.

For example, the SEC has such disclosure requirements for US firms (Nini et al. 2012), and UK GAAP and IFRS require specific disclosures on liquidity risk and other risks having a bearing on going concern assessment, including defaults and covenant breaches (FRC 2009).

Roberts and Sufi (2009b) find covenant violations are used as contingencies to accelerate renegotiations.

Zhang's (2009) findings for a sample of US firms for 1988–2007 suggest that increase in covenant strictness is associated with increase in expected default recovery rates with other explanatory variables constant.

For example, under Chapter 11 procedures of the US Bankruptcy Code debtor-in-possession (DIP) finance has an important function in providing short-term liquidity to facilitate orderly work outs. DIP is used because firms are prevented from making use of pre-insolvency lines of credit during the bankruptcy process. The automatic stay provisions of Chapter 11 eliminate covenant protections of pre-petition debt which may be replaced by standard positive and negative covenants in DIP loans (Maxwell and Delbridge 2010). Transition across covenant schemes is an interesting area of research for various bankruptcy jurisdictions.

For methods of estimating discretionary accruals see Peasnell et al. (2000).

Examples are present of commentary and research on accounting and finance innovations as in Briys and Crouhy (1988), on foreign currency options, O'Hanlon and Peasnell (1998) on EVA, and Stumpp (2000) on EBITDA.

Prior to the 1980s high yield debt generally consisted of existing debt instruments which had been reduced from investment grade but from the mid-1980s bonds began to be issued as high yield (Maxwell 2010). In addition, credit default swaps emerged, protecting against risk of default on debt obligations issued by an entity, and which may be considered as leveraged finance as they are an alternative investment form to underlying firm assets (Mahadevan 2010).

Examples of helpful distinctions are: Haw et al. (1991) and Duke et al. (1995) who choose public debt to reflect debt contracting costs as they associate it with higher covenant costs than private debt due to higher violation and renegotiation costs; Begley (1994) who noted that, amongst non-convertible US public debt, subordinated debt contained fewer covenants and firms issuing subordinated debt were smaller and financially weaker; and Citron (1995) who distinguishes his results on covenant types in public debt by secured fixed, secured floating, subordinated and convertible, unsubordinated and convertible debt, or bonds.

The Loan Pricing Corporation's DealScan reports detail from syndicated and bilateral loans collected by staff reporters from lead arrangers and SEC filings; the Mergent Fixed Investment Securities Database provides data on public debt contracts; and EDGAR is a database of SEC filings of private debt contracts. All are databases which have been used by empirical researchers.

Drucker and Puri (2009) observe that Dealscan sometimes under-reports number of covenants in deals and that deals with no covenants reported may be data errors.

Some indirect evidence on debt contracting across a wide range of countries is available. For example, Kim et al. (2011b) study the effect of IFRS adoption on loan contracting in 40 countries. They use data derived from Dealscan involving non-US companies/non-US loans over the period 1997–2005. Their sample consists of 5178 facility years (including 205 for IFRS adopters) of which 4076 are for 37 countries other than Australia, Canada, and the UK; see Table 1 for the range of countries represented. Kim et al. (2011b) use dummy variables for the presence of financial covenants and of general covenants and a variable for number of covenants in debt contracts so their research gives only general indications of the presence of covenants in debt contracts for companies from countries other than Australia, Canada, UK, and the USA. They conclude inter alia that adoption of IFRS is associated with banks imposing less restrictive non-pricing terms on companies, including less restrictive covenants.