The Forever Protocol

Fixing the settlement backlog the irritating way

Dick Frase, Dechert LLP
Originally published in the October/November 2005 issue

In the third week of October 2005, much to their surprise, an “ISDA novation protocol” was unceremoniously dumped on the UK’s hedge fund managers, or at any rate on those of them who dealt in credit default swaps (CDSs). The protocol was part of a series of US initiatives to reduce the CDS settlement backlog, and it rewrote the whole way the market closed out its open positions. No one was quite sure how it would work in practice, because the things it was planning to do had never been done before. And it was “forever” – once a firm had signed it, it was not allowed to withdraw, no matter how disastrous the effects of the protocol might be.

Some hedge fund managers had heard about the protocol. ISDA, the specialist derivatives documentation organisation, had been working on it for months. And they knew it was meant to cover CDS and interest rate trading. But it seemed to be very much an inter-bank matter, and the 50 firms that had signed up to the protocol so far were all banks. If there was a knock on effect for the buy-side, no doubt the banks would let them know what it was in due course.

Then, at the last moment, everything changed. A week before the final adherence deadline of 24 October, the banks suddenly started ringing round and twisting the managers’ arms to sign up. But sign up to what? No one really knew. The implications for the funds’ business were still not clear, and everyone had a different story. As the week of 17 October ticked away the stories changed daily, with new announcements from ISDA and the banks, new discussions, new uncertainties. (To rub salt in the wound, the protocol included a representation that anyone who signed up understood what they were doing.) In the US, the Managed Funds Association had come out strongly against the protocol at the beginning of the month, complaining that ISDA had ignored its concerns and that the protocol’s proposals were unworkable. In the UK there was no equivalent trade body to represent the hedge fund sector, and the investment managers were on their own.

By the end of the week, Friday 21 October, the big managers had come to the conclusion that they would probably have to sign, but were reluctant to do so until the position became clearer. Although the protocol was still open for adherence till the end of November, the banks had agreed to abide by it from 24 October. The discussions the week before had all been on the basis that, if a manager chose not to adhere, it could still trade under the rather cumbersome fall-back process specified in the ISDA Master Agreement. So on Monday, 24 October, funds which had not adhered should still have been able to trade under this fall-back process. However, notwithstanding this, it quickly became apparent on Monday that some banks were refusing point blank to trade with any fund which had not signed the protocol. This was the clincher. By the end of the day the main London managers were (not very happily) signing their funds up, and the adherents to the protocol had reached 700.

What was it all about? And why was the process so unsatisfactory?

Hedge funds and the CDS market

The traded CDS market is a bank-centred market for writing and buying credit protection – effectively a form of insurance. In recent years, however, hedge fund managers who specialise in debt have started to use credit default swaps as a form of investment. By writing protection against the default of underlying names they take a long position, earn a regular “premium” payment, and assume the credit risk of the name defaulting. To close out the position they buy matching protection under another CDS. The two contracts can then be held to expiry, but this has various disadvantages. The fund continues to have credit exposure to two counterparties, it has to make payments and settlements on an ongoing basis under both contracts, and it may have to post collateral on both deals. Given that the life of the contract may be several years, this can be quite a commitment.

So the fund will normally want to liquidate the matched positions rather than hold to expiry. If both trades are with the same bank (bank A) this will be relatively easy. But ideally the fund should be able to seek a better quote from another bank (bank B), and use that bank B quote to close out with bank A. On a futures exchange this would be achieved by the automatic novation of bank A and B’s matched positions to the clearing house; and their replacement with a corresponding, netted-out position in the customer’s account with its clearing broker. In the OTC market, the novation process is more cumbersome, and has to be achieved by a series of bilateral negotiations between the three parties involved.

The Confirmation backlogs

Under the standard terms of clause 7 of the ISDA Master Agreement, the prior written consent of Bank A is required before the fund can novate its position to bank B. This is perfectly logical from a legal point of view, but in practice it was apparently too cumbersome for use when liquidating positions. Instead, more informal practices grew up, under which novation would be agreed orally, and the written confirmation would, or should, come along later. But the process was not operated properly, and the confirmations were woefully behind. Close-outs went undocumented or unmatched; sometimes the remaining party (bank A) was not even aware that its counterparty had been changed. In a speech on 18 October Callum McCarthy, the Chairman of the Financial Services Authority, expressing general dissatisfaction at the settlement backlog, adding pointedly that: “this [backlog] has been exacerbated by the growth in the assignment of trades by hedge funds without obtaining the consent of the original counterparty – such that firms have to ‘fly blind’ in not knowing who their counterparty exposures are to for significant periods. This creates legal uncertainty with regard to default events,”

A month earlier in the US, on 15 September, the Federal Reserve Bank of New York hauled 14 of the major US bank dealers over the coals for their poor market practices in this area and asked what they were going to do about it. In response, on 4 October, the banks made a number of commitments designed to reduce the confirmations backlog. One of these was that, with effect from 24 October, they would only accept novations if they either (a) came from protocol adherents or (b) followed the letter of the novation requirements in the main ISDA Master. The time frame for this, from 4 to 24 October, was thus just 13 working days. And so it was that the novation protocol went from being a slow, not very satisfactory, ISDA project to becoming a hustled up, not very satisfactory, central plank of the Fed initiative to beat the backlogs.

The Forever Protocol

“Protocol” is ISDA jargon for a form of multi-party contract where anyone who wants to be a party to the contract (an “adherent”) serves a one-way notice on ISDA to this effect. Past protocols have usually related to specific issues or events, rather than general market practice, and adherents have always had the right to withdraw from the protocol if they wished to. “Adherence” is achieved by emailing ISDA a letter in a prescribed form promising to observe the terms of the agreement. The names of the parties are then posted on the website. Thereafter, when any one of those parties deals with another, the terms of the protocol will automatically apply.

The novation process laid out in the protocol is relatively straightforward. When the fund agrees its novation with bank B, that oral agreement is binding on the parties subject only to the condition that bank A gives its written consent (by Bloomberg, email etc) to the substitution of bank B. ISDA did not want this consent process to drag on forever because of the uncertainty this would cause. So it imposed an “aggressive” (i.e. not very practical) cut-off time of 6 pm in the location of Bank B. If consent was not received by then the novation failed and the fund would be left with two matching and opposite trades, one with bank A and one with bank B. There would be no opportunity to correct this later, even if all three parties agreed. Once a failure, forever a failure.

Clearly, a “drop dead” cut-off of this sort is going to generate problems. What if someone traded with bank B at 5.59? Their novation would be doomed before it had begun. International time differences alone could render a trade void. A fund in New York that opened a trade with bank A in New York and then agreed to close out a trade with bank B in London at 1pm New York time would be unable to obtain bank A’s consent afterwards, because it was then 6pm in the location of bank B and the cut-off time was already past.

How far in advance would the fund need to deal to beat the cut-off deadline? Nobody could tell. Historically the players in the bank A role had given consent orally, but had probably taken days to produce written evidence of that consent. If that continued under the new system there would be continuous novation failures, with a corresponding proliferation of matching trades. The settlement process would become worse, not better. And the buy side – the hedge funds – would be taking the pain.

In the US the importance of this issue to the hedge fund community had been spotted at a relatively early stage by the Managed Fund Association. In the UK, it hadn’t. The legalistic novation language used by ISDA did not suggest any obvious connection with the trade close-out issues described above. Nor was it immediately obvious, at the ops level, what the down-side was to staying out of the protocol and following the standard ISDA master process. This had to be identified in discussion with the actual traders, by comparing notes with other managers, and raising queries with ISDA and the banks. And since none of the managers had expected to have to sign the protocol until the banks started ringing them up, this whole process was crammed into the week of 24 October.

More generally, and stating the obvious, the hedge fund managers are small operations, and do not have the staff back up of large banks, which would enable them to devote resources to monitoring something like the novation protocol which did not appear to be especially relevant in the first place.

A number of other features of the protocol process did not inspire confidence.

The protocol is designed to apply to ongoing market practice generally, yet is only open for adherence until 30 October (extended recently to 30 November). What if someone changed their mind later and wanted to join? Supposedly this would not be allowed. But actually ISDA are going to have to open up the protocol again to allow newly-formed funds and banks to join, and at that point it is hard to see why other adherents should not be allowed in as well. So the real reason for the tight window for adherence, was to stampede everyone into signing up. An effective short term tactic, but not one designed to elicit long term confidence.

Would a fund have to disclose that it was novating when it asked bank B for a quote? The protocol was not specific, but there were various suggestions in ISDA guidance that this might be the case. If so would this enable bank B to price against the fund, rather than giving an arms’ length two-way price? The issue is still open, but there have been indications that this is not necessarily intended, and that ISDA may clarify the position in due course.

Was the 6 pm cut-off really forever? Wouldn’t it be possible for the fund to start over again the next day, but this time novating two trades instead of one? Logically this would seem possible, but at the time of writing the position is still uncertain.

The rushed ISDA timetable made no allowance for fund governance or tax issues. If anything went wrong it would of course be the managers’ problem, not the banks’. ISDA and the banks’ general assumption appeared to be that managers must surely have an unfettered right to sign on behalf of their funds, and any suggestion that they might want fund board approval was just the managers being difficult. But entering the forever protocol is not a routine matter, and boards should at least have had a chance to understand the issues if this was what they wanted. Also, under the UK tax regime, a manager signing on-shore on behalf of its off-shore fund raises tax issues.

What next?

So what conclusions do we draw from this? First, there is no doubt that the confirmations backlog is unacceptable and it is in everyone’s interests to find a solution. It is also true that the solution is likely to come a lot quicker if it is given a strong push by the US regulators.

What is less happy is the failure by ISDA and the banks to engage the hedge fund community in the protocol process; the failure to translate what they were doing into terms the buy side could readily understand, the last minute bullying exercise, and the general impression that the protocol was designed to dump any residual problems on the hedge funds.

This is a taste of things to come. Other US initiatives in the pipeline include requiring trades to be matched through the US Depository Trust Clearing Corporation and requiring UK hedge fund managers to register with the SEC. If there is going to be more hedge fund regulation in the UK one would have expected it to come from the FSA rather than the Fed and the SEC; but as it is there is now something of a free for all. What is apparent, is that there will be little sympathy for the hedge fund community just because it has been pushed around a bit. If London hedge fund managers want to avoid a repeat of this protocol debacle, they will need to have a more organised approach, and a better representation of their interests in the international regulatory arena.