RDP 2003-02: Do Collective Action Clauses Influence Bond Yields? New Evidence from Emerging Markets 2. Background on Collective Action Clauses

Collective action clauses have been commonly used in international sovereign bond documentation since becoming market practice under English law for corporate bonds in the nineteenth century.[8] The CACs included in modern-day sovereign bonds generally contain two main types of provisions:

  • majority restructuring clauses: enabling changes to be made to the financial terms of a bond instrument by a qualified majority (say 75 per cent) of bondholders, thus binding all holders of that bond issuance to the new terms; and
  • majority enforcement clauses: allowing a majority of bondholders to limit the ability of an individual creditor to enforce its rights against the sovereign (e.g., to declare the bond payable, or to initiate litigation) in the event of default.

Although the design of these provisions may vary amongst bonds, they typically include – in the case of majority restructuring clauses – rules related to the conduct of bondholder meetings, quorum requirements and voting rules, while majority enforcement clauses often specify that the proceeds of any litigation against the sovereign be shared amongst all creditors.

While CACs are traditionally included in bonds issued under English governing law, and also in bonds issued under Luxembourg and Japanese governing laws, they are not traditional in sovereign bonds issued under New York or German governing laws. In the case of the US, this has been a matter of market practice rather than legal requirement. In particular, although the Trust Indenture Act of 1939 precludes the modification of payment terms without unanimous approval of all holders of corporate bonds, the law does not cover sovereign bonds, which could contain CACs. Instead, sovereign bonds issued into the US market have – at least until the path-breaking Mexican issue of 26 February 2003 – used similar documentation to the corporate indentures that US investors are familiar with, thereby not including CACs.[9]

Because CACs are traditional in some markets but not in others, it turns out that sovereign borrowers frequently switch between the use and non-use of CACs in their bond issuance. Table 1 provides data from Dealogic's Bondware database for the 20 largest emerging market sovereign issuers over the period January 1991–January 2003. We separate issuance in international markets into three groups, based on the governing law of the bond contracts. The first includes bonds issued under English, Luxembourg and Japanese governing laws, for which the use of CACs is customary. The second includes the two major governing laws, New York and German, for which the use of CACs is not customary. The third group includes other governing laws. Table 1 indicates that 17 of the 20 most active sovereign borrowers have switched between issuing bonds with and without CACs over this period.[10] Furthermore, 96 per cent of the total issuance of these 20 borrowers is from issuers who have used both types of contractual form.

Table 1: Use of Collective Action Clauses in International Bond Issuance
January 1991–January 2003 – US$ billion
Country CAC governing law Non-CAC governing law Other governing law Total issuance
Argentina 18.1 38.6 0.6 57.3
Mexico 4.9 33.0 0.1 38.0
Turkey 12.9 24.0 36.8
Brazil 12.3 20.3 32.6
Hungary 10.1 6.1 1.1 17.3
Russia 14.0 1.8 15.8
Colombia 0.8 12.0 0.1 12.9
Philippines 5.1 7.3 12.4
Lebanon 3.1 6.5 9.6
Venezuela 0.3 8.3 0.1 8.7
South Africa 2.9 4.7 7.6
China 1.9 5.7 7.5
Poland 4.6 2.0 6.6
Panama 4.2 4.2
Malaysia 0.8 3.3 4.0
Korea 4.0 4.0
Uruguay 0.4 3.3 3.7
Croatia 3.3 0.2 0.1 3.6
Ukraine 3.5 3.5
Tunisia 1.9 1.4 3.3
Total 100.9 186.4 2.1 289.4

Notes: The data are calculations by the authors based on data from Bondware and include all bond issuance into international markets by those emerging market sovereigns that are among the 20 largest issuers over January 1991–January 2003. The first column of data includes issuance under English, Japanese and Luxembourg governing laws, while the second column includes issuance under New York and German governing laws (subject to the adjustments noted in Footnote 9). Columns and rows may not add due to rounding.

At one level, the data in Table 1 might be viewed as comprehensive evidence that the existence or absence of CACs cannot be a major influence on borrowing costs. In particular, if the presence or absence of CACs were a major influence on borrowing costs, we would expect borrowers to decide which type of contractual form provided the cheapest (or otherwise most appropriate) form of financing for them, and then to always use those contractual terms. By contrast, borrowers switch frequently between contractual forms, suggesting that the presence or absence of CACs is not a major influence on borrowing costs.

However, this view might be somewhat simplistic, given that issuance in certain currencies or certain markets may in practice constrain the use or non-use of CACs based on what is traditional for that particular market. For example, issuance in Japanese yen is invariably under Japanese governing law and therefore inevitably includes CACs. Accordingly, in the analysis that follows, we concentrate on a subset of issuance where one can plausibly make the case that there is some choice in the use or non-use of governing law. In particular, we focus on issuance in US dollars and euro (and some of its predecessor currencies). In each case, an analysis of governing law usage indicates that borrowers have the choice between issuance in the Euromarket with CACs, or the Euromarket or US (or global) market without CACs.


See IMF (2002) for further details of the nature and use of CACs. [8]

However, there are exceptions to the correspondence between governing law and existence of CACs in emerging market bonds. One exception is the case of Brady bonds, which are typically under English governing law, but exclude CACs for historical reasons. Another oft-cited exception is a (non-sovereign) US$300 million Euromarket issue by a Thai public utility in October 1998 with New York governing law, but which nonetheless included CACs. We have investigated further as to whether there may be additional Euromarket issues with New York governing law that include CACs, and searched for the offering documents for all Euromarket issues under New York governing law with an issue size of at least US$400 million. Of the 15 bonds where we were able to check offering documents, we found 6 cases where a bond issued under New York governing law appears to include CACs. These included 2 issues from each of Egypt, Lebanon and Qatar. Accordingly, the analysis that follows adjusts for this additional information on the use of CACs.
The results of this exercise are something of a caveat (albeit probably not a major one) to the series of papers that have studied the pricing of CACs in sovereign issuance, assuming a perfect correspondence between governing law and the presence of CACs. However, we suspect that the overall conclusion from most earlier work – that CACs do not affect bond pricing – remains quite robust. Indeed, we suspect that this evidence that CACs are used more frequently in the Euromarket than had earlier been assumed strengthens the case for the proposition that CACs have not previously been noticed by investors and a fortiori have not increased borrowing costs. [9]

The number of sovereigns to have used both types of contractual form would also include Korea if we also included quasi-sovereign issuance in the table. In particular, the Republic of Korea has made only two bond issues in its own name (in April 1998) but the state-owned Korea Development Bank was a frequent borrower prior to this (sometimes using CACs), and was viewed as representing a sovereign credit. [10]