بند های اقدام جمعی: چگونه آنها بر بازدهی اوراق قرضه مقتدر تاثیر می گذارند؟

We study the effects of the adoption of collective action clauses (CACs) on government bond yields by exploiting secondary market data on sovereign bonds quoted in international markets from March 2007 to April 2011. CACs are assessed security by security. Using a panel data approach, we find a U-shaped effect of CACs on yields according to the credit rating of the issuer. While the impact is negligible for the highest ratings, a significant yield discount emerges for mid-range ratings, which is smaller for bad ratings and possibly insignificant for the worst ratings. This relationship appears fairly robust across a number of checks. This evidence may reflect the fact that CACs are valuable because they help with orderly restructuring unless the perceived probability of default is too small. Nevertheless, at low ratings, this effect can be weakened by an increasing risk of moral hazard.

Collective action clauses (CACs) are contractual provisions that, inter alia, allow a qualified majority of bondholders to change the terms of a bond in a way that is legally binding for all bondholders.2 CACs are by no means a new instrument for the regulation of debtor–creditor relationships (Weidemaier and Gulati, 2012): some similar contractual provisions were already present in English corporate bonds as early as in the 19th century, and were later also applied to sovereign bonds issued under the jurisdictions of the UK, Luxembourg and Japan. By contrast, CACs have long remained uncommon in the US. The resurgence of interest in CACs began in the 1990s, when they were suggested as a device to curb the need for unpopular public bailouts (Group of Ten, 1996), as another option to the statutory approach, which would implement a treaty-based mechanism for coping with unsustainable debts (Krueger, 2001). However, CACs only became common in the US after an important issuance by Mexico in 2003, possibly because they were preferred to the statutory approach as more market-friendly and easier to implement (Roubini and Setser, 2004 and Eichengreen, 2003). Whether and how CACs impact sovereign bond yields is essentially an empirical issue, as contrasting effects can be surmised from a theoretical point of view. On the one hand, CACs help creditors' coordination, limit disordered default, curb holdout risks, hinder prisoner dilemma outcomes and reduce delays detrimental both to the debtor and to the majority of creditors. Thus, as CACs should be valuable ex-post if a default occurs, the market may acknowledge their value through a yield discount.3 On the other hand, CACs represent a limitation on individual bondholder's rights with respect to unanimous clauses, and they can reduce the incentive for the issuer to fully repay the debt by enabling opportunistic declaration of default. According to this argument, by making a default easier, CACs might also make it more likely. This ex-ante moral hazard channel should tend to increase the yields demanded by the market.4Ghosal and Thampanishvong (2013) formalise the trade-off that may arise between ex-ante incentives for the debtor and the ex-post benefits of creditor coordination: while the former tend to increase yield, the latter tend to reduce yield when compared to unanimity clauses. Since the end of the 1990s, a relatively large number of papers have empirically addressed the relationship between the adoption of CACs for sovereign bond issues and their yields. As yet, no consensus has emerged on the sign of this linkage or even about the conditions under which such a relationship could exist. Indeed, disagreement among authors arises even as to the methodology that should be followed when conducting the empirical analysis and regarding the nature and structure of the dataset. This is a rather problematic situation, as one of the largest experiments in the field – the inclusion of standardised mandatory CACs in all new euro area sovereign bonds beginning in January 20135 – is being implemented. In this context, the goal of this paper is to take advantage of various lessons learned, related to both methodology and datasets, and to offer a broader approach to testing the relationship between the adoption of CACs and bond yields. We exploit a dataset spanning from March 2007 to April 2011 with yields on 292 securities listed on major international markets. Thanks to a new feature added by Bloomberg, we are able to determine for each bond whether or not a CAC is in place, overcoming one of the main pitfalls of many earlier studies, which relied on the bond's governing law of New York or London as a proxy to gauge CACs adoption. The sample is large enough to allow us to focus on sovereign bonds, enhancing comparability. We choose not to enlarge the dataset with corporates, which could give rise to spurious correlations. Our study encompasses a relatively large number of countries at various stages of development, thus focusing not only on emerging market issuers as done in most previous studies. Some previous studies (e.g., Becker et al., 2003) have stressed the need to use secondary rather than primary market data, as a way to mitigate instances of endogeneity, structural breaks and omitted variable bias. We follow suit compiling our dataset with secondary market yields taken monthly, for a total of 50 time periods. The exceptional market patterns that occurred throughout the sample should make it easier to identify any effect of CACs on yields. The bulk of our empirical analysis consists of the estimation of a panel model. This is a relatively novel approach in this area of the literature, as most previous papers focused on a snapshot at a certain time. The extension of the period under scrutiny offers two clear advantages: (i) it renders the analysis less dependent on the idiosyncrasies in the data at any specific point in time and (ii) it allows us to check whether and how the link under examination has evolved with market developments (e.g., the impact of downgrading the country issuing the bond). The results of this study show that credit ratings do matter for the impact of CACs on yields. The inclusion of CACs lowers most yields for bonds whose issuers fall in the middle of the rating scale. For very good ratings, no statistically significant difference in yields is observed to result from the use of CACs, while for bad ratings the yield discount is smaller than that for mid-range ratings, to the point of becoming insignificant for the lowest ratings. This relationship appears to hold across several robustness checks. These results suggest several points. First, the ex-post beneficial effect of CACs for orderly restructuring is, indeed, valued by the market, but it requires the probability of default to be non-negligible. The lack of this characteristic helps to explain why CACs do not seem to have any effect for the best-rated countries. Second, the effectiveness of the ex-ante moral hazard channel is likely to be affected by the issuer's rating. One of the greatest costs of default for debtors willing to maintain access to markets is in terms of reputation and the risk of being denied access (see, among others, Eaton and Gersovitz, 1981 and Sturzenegger and Zettelmeyer, 2006). These constraints are much weaker for poorly-rated debtors, who have a low reputation anyway and are typically less reliant on international bond markets for funding. In this respect, Reinhart and Rogoff (2009) remark that poorly-rated countries usually have less access to international bond markets, because their funding sources mainly consist of subsidies and loans from the official sector, and they also have a higher propensity to debt repudiation. Consistent with these characteristics, the market might exhibit greater moral hazard fears about CACs for poorly-rated countries. Third, issuers in the middle of the rating scale are afforded the largest discount by the market because the probability of default is concrete, but the incentive for the debtor country to meet its obligations and maintain access to international markets is sufficiently high. Fourth, there is no evidence, irrespective of ratings, that the use of CACs increases borrowing costs: even for the worst rated issuers, we find that yield-increasing components never significantly overwhelm the yield-decreasing components. The rest of the paper is organised as follows. In Section 2 we review the empirical literature on the effect of collective action clauses on bond yields; in Section 3 we present the dataset and highlight some descriptive statistics. Section 4 reports the econometric analysis of the panel data and comments on the main findings, while Section 5 addresses several issues related to sensitivity analysis. Finally, Section 6 summarises the conclusions of this work.

Collective action clauses are contractual provisions included in the issues of sovereign bonds to ensure orderly debt restructuring. The impact of CACs on borrowing costs is an empirical issue over which research has not yet reached a consensus, with respect to methodology, dataset construction or results. The first wave of studies in the early 2000s identified CACs with the bonds' governing law and disputed the use of primary versus secondary market data on grounds of sample homogeneity, endogeneity issues, and data accuracy. Contrasting results were found, indicating that CACs could depress or enhance yields, or have no effect at all. The pace of research then slowed, in part because the governing law was revealed to be a poor proxy for the inclusion of CACs. This topic has since gained renewed momentum in recent years, following the sovereign debt crisis, in which CACs have been advocated as a tool to facilitate debt restructuring (see Committeri and Spadafora, 2013). A systematic and quantitative study of their effect on yields still appears to be needed. In this paper we update previous research encompassing the financial and the sovereign debt crisis; we contribute to the literature with a more homogeneous sample and more accurate bond-specific tracking for the inclusion of CACs. Our focus on monthly data from the secondary markets allows us to take advantage of panel data techniques. Our empirical analysis suggests that the effect of CACs on yields may vary in a non-linear way according to the issuer's rating. In particular, a U-shaped impact of collective action clauses on yields seems to emerge, with a discount provided to issuers in the middle of the ratings scale and no effect for issuers at the extremes. Our interpretation of this finding is that the advantages of CACs are greater for the creditors of mid-rated issuers, because the probability of default is not negligible (making CACs actually valuable) while there is less suspicion of opportunistic behaviour on the part of the debtor. In contrast, the chance of default is low for very well-rated issuers, thus reducing the value of ordered restructuring, while poorly rated issuers face lower reputational costs and are suspected of moral hazard to a greater degree if they choose to include CACs that favour debt restructuring. Notably, at any rating none of the specifications and robustness checks finds an association between the inclusion of collective action clauses and an increase in sovereign bond yields with all else remaining equal. Though our analysis is strictly descriptive and based on sovereign bonds issued in international markets, our research is partly motivated by the recent introduction of CACs in euro area domestic bonds. The European Council decided on 24–25 March 2011 that standardised and identical CACs would be included in all new euro-area government bonds from 2013 onwards.41 The conclusions of the March meeting state that “the inclusion of CACs in a bond will not imply a higher probability of default or of debt restructuring relating to that bond”. It is important to recall that the introduction of CACs in all euro-area bonds is part of a comprehensive set of measures aimed at strengthening the economic governance of the European Union and to safeguard the stability of the euro area. In particular, financial assistance will be granted by the European Stability Mechanism given three main elements: a macroeconomic adjustment programme; stringent policy conditionality; an analysis of public debt sustainability (see ECB, 2011). The inclusion of CACs contributes to uphold the no-bail-out principle among countries in the European monetary union. The very existence of a framework for involving the private sector in restoring debt sustainability will lead investors to carry out more careful risk assessment; a proper role for market discipline on government spending would be restored. The introduction of CACs hence contributed to make it politically viable to set up a framework to provide financial support to liquidity-constrained countries, dampening the pressure to redistribute the burden of a default among countries.