feed

Article: CCPs and Systemic Risk | Are We Really Safer?

06 December 2014 | Jon Gregory

The clearing mandate, requiring central clearing of standardised OTC derivatives, is well and truly upon us. Despite this, there is clear concern as to whether OTC clearing will really make financial markets safer. Regulators appear to view central counterparties (CCPs) as a panacea, and their advantages are commonly stated, yet little consideration is given to the potential disadvantages and inherent weaknesses in channelling a large volume of OTC products through CCPs. A typical statement regarding the benefits of central clearing is [Note 1]  “The primary advantage of a CCP is its ability to reduce systemic risk through multilateral netting of exposures, the enforcement of robust risk management standards, and mutualization of losses resulting from clearing member failures”. Whilst comments such as this may indeed prove to be generally correct, there are a number of important points to discuss.

First, due to the reduction of bilateral netting between clearable (CCP) and non-clearable (bilateral) trades spread over more than one CCP, exposure reduction through multilateral netting is not an inevitable consequence of the clearing mandate (for example, Duffie and Zhu 2011 [Note 2]). Mandatory clearing can actually increase exposure due to this bifurcation which would presumably then imply a commensurate increase in systemic risk. Of course, multilateral netting efficiency can be improved by having only a small number of CCPs and, potentially, also with interoperability arrangements. However, the resulting too-big-to-fail implications are clearly less than ideal. Furthermore, multilateral netting is not a magical technique for reducing exposure but instead merely redistributes exposure from one place to another [Note 3] . In a default scenario, multilateral netting improves the claim of OTC derivatives creditors at the expense of other creditors. Hence, reduction, if any, of systemic risk is confined to the OTC market and not necessarily the financial markets in general.

Regarding “robust risk management standards” (the sort of expression often applied to large banks in the run up to the last financial crisis), it is still to be defined what these really are. Taking the initial margins of CCPs as one example: a typical calculation might require that initial margin is sufficient to cover the average of the worst 6 losses in the last two and a half thousand days (10 years). In technical terms this can be defined as a 99.8% expected shortfall (VAR’s inconsiderably more intelligent cousin). Being 99.8% confident of having enough initial margin sounds like robust risk management. However, the problem with the thousands of days of market data changes that are analysed in order to define the initial margin is that on virtually none of them have any CCP members (i.e. banks) actually defaulted. Predicting the market volatility in the aftermath of a default event using data when defaults don’t happen is dangerous. The worst six days in the above example are actually pretty much the only days of interest, given that at least some of these represent the last significant OTC default scenario (Lehman Brothers). However, taking the average is less than “robust” and would imply significant probability of losses exceeding initial margin and spilling over into default funds. At such a point, whilst the CCP may not fail, clearing members will be subject to losses and therefore this does constitute some form of default scenario. 

Second, loss mutualisation in the default fund can indeed be an efficient risk sharing arrangement but also has some problematic side effects due to moral hazard issues. For example, when executing an OTC transaction that will be central cleared, the identity and credit quality of the original bilateral counterparty is almost [Note 4] completely insignificant since the effective counterparty will be the CCP. Bilateral trades will therefore be executed without fear of counterparty risk, and weaker counterparties will not be clearly penalised for their relatively low credit quality. The auctions held by CCPs in the aftermath of a default are crucial for OTC portfolios which are relatively illiquid and subject to complex multidimensional risk factors. The mutualisation of losses from the default management process also relies on some combination of loss allocation methods that are fair and create the right incentives and discourage a prisoner’s dilemma. This leads to some quite extreme methods, such as variation margin gains haircutting (VMGH), forced allocation and tear-up which, whilst preventing a potential CCP failure, can hardly be viewed as creating any stability in the aftermath of a significant default(s). Luckily, we have recently been reminded that CCPs “probably” have enough of their own capital at risk [Note 5] (by the way, it probably won’t rain tomorrow).

Of course, CCPs have advantages such as reducing the interconnectedness in OTC derivative markets. None of the above points prove that the clearing mandate will be inefficient or not lead to a reduction in systemic risk. However, views are often heavily divided between whether an OTC CCP will be something closer to a panacea or nostrum. Surely there needs to be more attention paid to the correct assessment of the relative advantages and disadvantages of central clearing in different regions, markets and product classes. Only then can we be sure that one of the most significant aspects of regulatory reform since the last crisis will indeed make financial markets safer. 

  1. International Monetary Fund (IMF), 2010, “Making over-the-counter derivatives safer: the role of central counterparties”, Chapter 3, Global Financial Stability Report (GFSR) April, http://www.imf.org/External/Pubs/FT/GFSR/2010/01/pdf/chap3.pdf
  2. Duffie, D. and H. Zhu, 2011, “Does a Central Clearing Counterparty Reduce Counterparty Risk?”, Review of Asset Pricing Studies, 1(1), pp 74-95
  3. For example, see Pirrong, C., 2013, “A Bill of Goods: CCPs and Systemic Risk”, working paper, Bauer College of Business, University of Houston, http://www.bauer.uh.edu/spirrong/pirrong_bdf_boe_ecb_clearing_130506_pdf.pdf
  4. There is a small possibility that through the loss allocation rules of the CCP that this may be relevant.  
  5. LCH and CME have enough capital, says Isda's O'Connor”, Risk Magazine, 24th September 2014. 

 


Comments

I essentially wrote that quote from the IMF's Global Financial Stability Report (Note 1) but it's important to keep in mind that we did not advocate mandatory central clearing. Also that statement was based on the assumption (perhaps naive) that there would be much less nationalistic legislative and regulatory bumbling and fumbling.

I think the introduction of IM alone is enough to make cleared world safer than the no-IM bilateral one. It's hard to say whether it would still be safer than regulator imposed IM-bilateral world but I'm sceptical on how well IM will work in bilateral trades when it does kick in. So my vote is for cleared markets.

The article is trying to examine whether clearing as a component of the entire financial system, increases or decreases the stability of the whole market in a crisis. The case isn't proven by any empirical means, and this article plus others shows that the priority that CCPs take for assets in a crisis could serve to worsen conditions for firms overall, rather than improve them.

But with what should we compare cleared markets in terms of safety? IM- bilateral or non-IM?

Maybe we should compare the three scenarios: 1) All markets cleared, 2) Some Markets Cleared, and 3) No-Cleared.
Where the Market means all asset classes, not just OTC.

1) Might be the most stable, but would rely totally upon the risk management and default waterfall assets of CCPs
2) Is now, and is not provably more stable than 3)
3) Dates back to 100 years ago before futures.

Politicians have invented the new market structure (Item 2), so politicians will have no-one but themselves to blame when mandated clearing doesn't prevent another banking crisis - something they believe CCPs will avert.

Transparency is superior to perfection. Cleared markets pave way to increasing transparency in terms of Notional outstanding ( Open interest/Volume) which was and still hard to figure out on a daily basis completely for OTC products. GTRs ( Global trade repositories) are aggregating this information and distributing to some extent. This is a right step in right direction.
Cleared markets will also improve price transparency. For instance at what level Interest rate swaption (Deep Out of the money) has traded can be gleaned by second tier and third tier players other than inter dealer broker and Big banks. They have already had their impact in terms of reduced margins and revenues in Fixed income trading.
Central clearing with multi lateral netting will bring safety and confidence in the market. Bilateral market for vanilla and semi exotic derivative products is a dying animal.
thanks
Chandra Khandrika

I wouldn't be so quick to hard link central clearing with transparency. Maybe it makes it easier but you can mandate trade reporting without mandating central clearing. Mind you national authorities are screwing that up too. Maybe one can argue against the Duffie and Zhu (2011) single mega CCP concept on TBTF/TITF grounds, but I can't see why we can't have work towards a single mega trade repository (TR), or at least one for each product class.

Maybe efforts to create and enforce the use of legal entity identifiers (LEIs), unique trade identifiers (UTIs), and unique product identifiers (UPIs), coupled with some sort of aggregation mechanism, will overcome many obstacles, but everything would be so much easier with fewer TRs. And then there is that nationalistic flag waving and anal privacy laws making it difficult to impossible for any one body to see the whole global picture.

But my main point is that we can have transparency without central clearing... However, we need to knock some heads together to make it work effectively on both the input and output ends.

Bill, and Jon (author of this piece)

It is not so much about whether the CCPs make the markets MORE safe or LESS safe. I ignore any efforts to quantify whether the CCPs > or < risk; it is just more practical and very obvious to say that we are just shoving the same amounts of risk around all over the globe. The risk - the same risk as before 2008 - has been bulldozed to the CCPs, and for now while everything is ramping up, increased activity at CCPs seems like a panacea but just wait. I agree with this colorful Mr. K above regarding 'transparency'. Take a look at what each risk officer at the FCMs receives from the CCPs in terms of 'transparency reports' (referring to the suggested requirements discussed during the FEDs PRC gatherings on due diligence banks should do on CCPs). If THAT is all there is, I will wait for improvements. Use an example of a crisis you lived through. Amaranth for example. If all of A's OTC commodity swaps had been cleared back then, would the risk officers have better transparency today than 8 or 9 years ago about that specific market? Not from what I have seen, in terms of 'transparency reports'.
And as for global trade repositories. Nice idea. But who in the end is ultimately accountable for good data; where does the buck stop?

Rae, on the trade repository issue it's hard not to conclude that some mega institution such as the BIS be the place where the data collection and dissemination buck stops. The BIS comes to mind because they have so much experience dealing with massive amounts of financial data, although the higher frequency nature of this data may be orders of magnitude beyond what they could handle right away. But we're so far away from that on account of the cross border and privacy issues, and I fear that we'll never be able to achieve that holy grail centralized data hub. And ultimately that hub should include other transactions if the goal is to map out all interconnections. But, instead we're heading in the opposite direction, and like you say, with lots of open questions about where the buck ultimately stops...