THE ECONOMICS OF THE PRIVATE MARKET
Appendix E: A Review of the Empirical Evidence on Covenants and Renegotiation
The Nature of Covenants in Private Placements
Laber (1992) has examined the frequency of different types of covenants in private placements. From a sample of twenty-five private placements issued between 1989 and mid-1991, he found that all had covenants that
Eighty-eight percent had covenants that set a maximum leverage ratio; 72 percent had covenants that restricted investments; and 48 percent had covenants that required a minimum interest coverage ratio. Laber did not report on the frequency of covenants related to working capital, although such covenants do appear in private placements. 192
According to market participants, most financial covenants in private placements are incurrence covenants; occasionally one or two maintenance covenants may be included, especially when these are designed to match maintenance covenants in other debt of the issuer, such as bank loans.
The Relationship between Covenant Tightness and Issuer Quality
Our discussion with market participants indicated that the number and tightness of financial covenants in private placements, just as in other debt markets, are a function of the quality of the issuer. Securities purchase agreements for lower-quality issuers often include many of the financial covenants seen by Laber, and the covenants are tight in that stipulated minimum values for ratios are close to current values. Contracts for moderately risky issuers often include only one or two financial covenants with minimum values set further from current values. Highly rated issues usually have no financial covenants, although A-rated issues may have a debt-incurrence covenant if their maturity is beyond seven years.
In contrast, Hawkins (1982) found in a sample of fifty securities issued in the mid-1970s no relationship between the restrictiveness of covenants in private placements and the quality of the issuer, for three types of financial covenants: working capital restrictions, cash payout restrictions, and debt restrictions. One reason for Hawkins's findings is that the market may have changed in the 1980s. Indeed, many observers have argued that covenants in private placements have become less restrictive over the past decade (Asquith and Wizman, 1990; Brealey and Myers, 1991; Brook, 1990; and McDaniel, 1988). If restrictiveness has decreased more for higherquality issues than for lower-quality issues, then the difference would be accounted for. Along this line, Asquith and Wizman found for public bonds that covenant restrictions on debt financing and dividends fell more for A-rated bonds than for lower-rated bonds. Fitch (1991), however, states that the decline in the strength of public bond indentures is characteristic of the public market and not the private placement market. The decline is also inconsistent with the interview results of Brook (1990) and the market descriptions of Chemical Bank (1992) and Travelers Insurance Company (1992). 193
Cross-Market Differences in Covenants
Market participants indicated that bank loans contain roughly the same types of covenants as those in the private placement market, with two differences. First, financial covenants in bank loans are typically maintenance covenants, whereas most covenants in private placements are incurrence covenants. Second, bank loan covenants are usually tighter. In some cases, we were informed, private placement covenants are set by loosening the covenants in an issuer's existing bank loan.
Although the types of financial covenants are roughly the same in private placements and bank loans, evidence points to subtle differences in the way they are implemented. Travelers (1992) observes that bank loan covenants tend to reflect a lending philosophy different from that motivating private placement covenants. It argues that banks, as short-term lenders, emphasize liquidity or working capital. This approach is reflected in the inclusion of working capital covenants. Banks also appear to be more sensitive to the relation between total liabilities and net worth; therefore, bank loans tend to restrict the ratio of total liabilities to net worth. In contrast, private placement investors tend to emphasize the importance of a firm's long-term assets and its long-term debt because they are long-term investors. As a result, according to Travelers (1992), private placements seldom include working capital covenants, and they tend to restrict long-term liabilities rather than total liabilities. 194
Empirical Evidence on Cross-Market Differences
Several studies that focused on the differences between covenants in privately placed debt and those in public debt indicate that private placement covenants are tighter than public bond covenants. Smith and Warner (1979) observed this fact from a 1971 American Bar Foundation study of bond indentures. Laber's (1992) conclusion that covenants in the private placement market are more restrictive than those in the public market was based on a comparison of his private placement data with the findings of other studies of covenants in the public market. DeAngelo, DeAngelo, and Skinner (1990) also presented evidence that the covenants of private debt contracts are tighter than those of public debt contracts. El-Gazzar and Pastena (1990) had similar results and also reported that private placement covenants tend to be tighter than those in large syndicated bank loans. This finding, however, probably reflects the predominance of large syndicated bank facilities in their sample as opposed to loans to the smaller bank-dependent borrowers that are the focus of our comparison. 195 Because they did not appropriately control for issuer size and quality in their analysis, their results are not inconsistent with the general proposition that bank loan covenants associated with bank-dependent borrowers are tighter than private placement covenants, which are in turn tighter than public bond covenants, controlling for the size and quality of the borrower.
From interviews with staff of investment banks, rating agencies, corporate general counsel departments, and law firms, Brook (1990) found that private placements typically had tighter covenants than public bonds had but that the pattern differed by type of covenant. For three classes of covenants-disposition of assets or maintenance of capital, limitations on debt, and restrictions on dividends-private placements were the most restrictive, non-investment-grade public bonds were somewhat less restrictive, and investmentgrade public bonds were the least restrictive.
Brook also found that restrictions on investments were not present in public bonds but were common in private placements. However, negative pledge clauses and restrictions on sale-leaseback transactions were common to all three. 196 Antimerger provisions were present in all three types of securities. Some interviewees in the Brook study, however, indicated that antimerger covenants tended to be somewhat stronger in the non-investment-grade market and stronger yet in the private market. He found that in the public bond market the typical anti-merger provision did not prevent a merger but only required that the acquirer assume the acquiree's debt.
The Value of Covenant Protection
Another empirical issue is the value of covenant protection. Unfortunately, data limitations and the lack of market prices make examination of this issue problematic for the bank loan and the private placement markets. However, several studies have been conducted on the degree of protection provided by covenants in public bonds for leveraged buyouts. These studies focus on the relation between covenants and bondholder returns in leveraged buyouts (LBOs). (A loss, or negative return, to existing bondholders due to an LBO indicates a lack of covenant protection.) Marais, Schipper, and Smith (1989) found that existing bondholders did not suffer losses in LBOs. However, their results contrast with anecdotal evidence, such as the RJR-Nabisco leveraged buyout, as well as other academic studies, such as Warga and Welch (1990). Asquith and Wizman's (1990) results are particularly relevant: They found that, although existing bondholders on average incurred significant losses in LBOs, these losses were related to the strength of the covenants. They found that ''bonds with strong covenant protection gain value, whereas those with weak or no protection lose value.'' Similar results were obtained by Cook, Easterwood, and Martin (1992). Crabbe (1991b) analyzed the value of covenants in the public market ex ante by examining the pricing of super poison puts. 197 He found that public bonds with super poison put covenants paid a lower yield than those without. It is difficult, however, to extrapolate from these studies of the value of event risk protection in the public market to the value of credit quality protection provided by covenants in the private market. Of course, the ubiquity of covenants in private placements and commercial bank loans itself suggests that they are valued in these markets.
Empirical Evidence on Renegotiation
While virtually all sources, including market participants interviewed for this study, agree on the ranking of markets with respect to covenant tightness and renegotiation, some evidence suggests that public bond covenants can provide a measure of protection and that the cost of renegotiation in the public market is not necessarily prohibitive. Kahan and Tuckman (1992) studied covenant renegotiation in public bonds. They found a sample of sixty-nine firms that sought to renegotiate bond covenants during 1988 and 1989. The authors argue that their finding so many firms seeking renegotiation is inconsistent with the assumption of most researchers (for example, Berlin and Loeys, 1988; Bulow and Shoven, 1978; and Lummer and McConnell, 1989) that renegotiation is limited to the information-intensive markets (for example, commercial bank loans and private placements) and is prohibitively expensive in the public market. Covenant renegotiation in the public bond market involves the issuer's sending a ''consent solicitation'' to each bondholder requesting an alteration in one or more of the covenants in the bond indenture agreement. Except for an alteration of interest and principal provisions, most indenture agreements require a two-thirds majority of the outstanding face value of the bond issue. (The Trust Indenture Act of 1939 requires the consent of all bondholders when principal and interest are to be modified.) Kahan and Tuckman (1992) found that most solicitations are ultimately successful but noted that many of the solicitations have a coercive element in that the consenting bondholders receive a fee (typically $10 per $1,000 of face value) if the solicitation is successful, whereas those not consenting receive no fee. They also found, however, that bondholders enjoyed, on average, significant positive abnormal returns around the announcement of ''potentially'' coercive solicitations. This finding suggests that issuing firms ''cannot, or do not, exploit the coercive nature of their solicitations.''
The evidence presented by Asquith and Wizman (1990) and Kahan and Tuckman (1992) may appear inconsistent with the view that covenants are not binding in the public market, but it is not necessarily so. The results in both of these studies were driven by extraordinary events. In the Asquith and Wizman study, the events were LBOs. Similarly, the largest category by far in the Kahan and Tuckman sample also involved firms that were targets of an LBO (and most others involved relatively unusual events). The results of these studies suggest that covenants in public bonds can impose meaningful limitations on event risk. When more routine types of actions by firms are likely to trigger bank or private placement covenants, however, public bond covenants are normally not binding. 198 In a sample of 128 firms that violated accounting-based covenants, Chen and Wei (1991) found that only 4 involved public debt. The remainder were privately placed securities. 199