Stress Testing and CCPs – Are We Missing the Point?

Catalyst’s Stephen Loosely questions if clearers and clearing houses are going far enough with stress testing and suggests a wider testing of assets and liabilities is crucial to effect risk management.

Clearing houses stress test the cleared liabilities of their members to protect themselves against extreme losses in a default. Regulations direct what stress testing should provide for. On the surface, it looks like we’re all safe. But we’re missing something big.

Clearing houses make their money from fees and treasury income, rather than active risk taking like a bank. As part of the ‘financial plumbing’, their priority is the integrity of the clearing house and the wider financial system. 

So stress testing is an important safeguard – and clearing houses need to go stronger, deeper and wider than the banks.

Clearing houses do a lot of things well: they collect collateral against the market risk inherent in their members’ cleared positions, margin model backtesting is robust, and oversight and governance strong. They stress test a wide range of historic scenarios against combinations of members simultaneously defaulting, as per regulatory instructions. 

Now, historic scenarios are all well and good, but lightning doesn’t strike twice in the same place: the next big market shock is unlikely to be a repeat of something that’s already happened in the past. 

While legislators have been great at locking the stable door, what’s the next horse to bolt going to be? It must be said that regulators haven’t exactly been helpful here…“extreme but plausible” stress testing is ambiguous advice and the reverse stress test/stress to destruction  ploy is by definition not particularly helpful. 

So how do you decide what to stress test for?

Perhaps a sovereign bond crisis is the next big shock, perhaps the collapse of the Euro or an economic feud with China as the value of its current account surplus disappears as US treasuries fall.

If you were a clearing house these would be sensible scenarios to consider. The problem is clearers only seem to stress test the value of their member’s cleared positions, and not much else. What do I mean by that?

Think about the clearing house as a business…are the values of cleared positions the only item at risk? Are they the only thing that should be stress tested? Do economic shocks generally happen in isolation? Are Clearing Houses (systemically important institutions that they are) at risk if they’re only stress testing their cleared business? Absolutely yes. 

So where are the other weaknesses? The single largest one is collateral, i.e. the assets that members post to the clearing house to collateralise the risk the clearer has from clearing those positions.

Clearing houses generally do not stress test their assets alongside their liabilities. Sure, they haircut them, but a haircut is what it says it is: a bit off the back and sides, not a close shave.

Table 1: Stress losses on cleared positions parameterise the default fund. What, then, covers stress losses on collateral?

Let’s take an example….sovereign bond market contagion spreads to the UK, which is downgraded. This triggers the default of a UK bank that has posted £200mm of bonds as collateral at a clearing house against its £200mm short swap position. 

The bank has defaulted, leaving the clearer holding the swap book. Thus far, no worries, insofar as this is what central clearing is for. The clearer will have to act fast: the UK Government has been downgraded so bond prices are falling, interest rates are rising and that short swap book is haemorrhaging losses (remember long bonds=short swaps). 

But what resources does the clearer have with which to manage that defaulted swap book? It measured the risk as £200mm, and took £200mm collateral in UK Government bonds…but the UK’s been downgraded, the bonds are now only worth £180mm and falling and the swap book is losing money fast…the Clearer has a shortfall of £20mm and it hasn’t even started winding up the default yet…

Clearers must stress test assets and liabilities together, otherwise they are merely assuming that their initial margin collateral will still be worth sticker price in a default.

This is why the Committee on Payments and Settlement Systems and the International Organisation of Securities Commission (CPSS IOSCO) recommends that clearing houses take stress conditions into account when haircutting collateral – but it’s an inconsistent solution and makes pledging good quality collateral to the clearing house extremely expensive. If stress losses on cleared positions can be used to calibrate a default fund, why not stress losses on collateral?

And indeed the CPSS IOSCO recommendations don’t go far enough; this scenario could get much worse.

The number of large, well rated banks that clearers feel comfortable using as investment counterparties (for investing cash initial margin with) or as payments banks (for collecting margin monies from other banks) is small and getting smaller. 

This means it is really quite feasible that the bank that goes bust as a clearing member is also an investment counterparty and also a payments bank…

So as well as a default to manage with insufficient collateral and losses on the cleared book, the clearer now has to contend with an investment counterparty and a payments provider who have both gone owing money. Sure, there may well be a legal claim on this or some repo collateral but in the heat of a default it’s cash on tap the clearer needs, not more collateral that’s losing value or a favourable finding from litigation months down the line.

Needless to say, a bank default + sovereign debt downgrade is not an environment in which financial institutions want to lend to each other, so the clearing house may well have a tricky time obtaining liquidity to finance its default management also.

Could it get any worse? Actually, yes. Most clearing houses feature a slice of their own capital base as one of the resources to deal with a default in stressed scenarios. Obviously this is predicated on the clearer being well capitalised. Just as any other business, this capital relies on the clearing house not sustaining any losses from its business model. Given the competition in central clearing, fee cuts, interoperability and the “bankruptcy” remote nature of capital requirements impacting on investment income, this starts to look like a harder assumption to make.

So, what’s the answer? Clearing houses must stress test their assets alongside their liabilities to ensure that they have the resources they think they do in the event of a default. These stress tests must consider the implications of a bank default on more than just cleared business and must encompass the default of payments banks and investment counterparties as well as the potential liquidity drain.

Clearers need to holistically stress test all risks in their business model together – rather than, as regulators imply, separately. Stress testing the cleared positions only is clearly less than half the battle.