Streetwise Professor

January 18, 2010

Clearing and Systemic Risk: A Partial Inventory of Effects

Filed under: Derivatives,Economics,Exchanges,Financial crisis,Politics — The Professor @ 6:02 pm

In the aftermath of the crisis, clearing has been advanced first and foremost as a means of reducing systemic risk.  In my opinion, for the most part, these assertions reflect a superficial understanding of clearing, when they don’t reflect abject ignorance.

Things are actually quite complicated.  I thought it would be worthwhile to present a list (which is long, but probably not exhaustive) of some of the ways in which the adoption of a central counterparty can actually create or worsen systemic risks.  These are presented in no particular order.

  1. The adoption of a CCP affects the priority among creditors.  In essence, netting positions improves the priority of the claims of derivatives counterparties (more exactly, of clearinghouse members via the clearinghouse) of a defaulting member.  It also worsens the priority of the claims of the bankrupt/defaulter’s other creditors.  It is not clear a priori that this reallocation of priority improves welfare, or reduces the likelihood of a systemic shock resulting from the default of a large derivatives trading firm.  For instance, worsening the position of those supplying short term credit will exacerbate their losses in bankruptcy/default.  These creditors (e.g., repo counterparties, corporate paper buyers) may be systemically important.  (EG, worsening the losses of CP buyers may result in runs on money market funds that hold CP.)  Moreover, worsening the position of some creditors will tend to make them more likely to run.  This can tip a shaky but surviving firm into failure.  Changing priority will affect capital structure and transactions costs in complex and unknown ways that could have systemic implications, as market participants adjust contracting behavior in response to the change in priorities.  Of course, changing priority in favor of derivatives counterparties could reduce systemic risks by reducing their incentive to run, and reducing the losses they suffer.  But it is by no means clear a priori how this cuts.  This is related to the general point that the first order effects of clearing are to change the allocation of risk, not its total amount.  It is not trivial, to say the least, whether these reallocations reduce or increase systemic risks.
  2. A system with customer clearing (as in futures markets, and as is being proposed/advocated/adopted in some OTC clearing organizations) tends to increase the default losses that CCP members suffer, even holding total trading positions constant.  (Customer positions are likely to increase, moreover, as I discuss below.)  Since the CCP members are likely to be systemically important financial institutions, this would tend to exacerbate systemic risks. In a bilateral market, exposures to the default of firm A is limited to A’s counterparties, in relation to their trades with A.  Some end users who trade with A, who would not be members of a CCP, suffer from A’s default, as do other dealers likely to be CCP members.  In contrast, in a CCP with customer clearing, the obligations of A, as a CCP member, to non-members are transferred to other member firms.  That is, adoption of clearing tends to redistribute the burden of default losses from end users who are not CCP members to financial intermediaries who are.  (Indeed, this is one of the advertised virtues of CCPs.  For years, the Board of Trade Clearing Corporation touted the fact that no CUSTOMER had lost money as a result of default: that’s exactly because clearing insured customers by transferring the risk to clearing members, i.e., large financial intermediaries.)  Thus, clearing tends to increase the default losses borne by systemically important financial institutions.  Another example of the effects of clearing on risk allocation that plausibly increase systemic risk.
  3. CCPs create a single point of failure, and a concentration of risk.  This can have a variety of effects beyond the obvious one of the systemic shock that would result if a CCP in fact fails.  It means that there is less diversification of the risk of errors, e.g., model risk, or other risks in the evaluation of default risk.  If a CCP’s model/methodology/information for evaluating performance risk is flawed, these flaws are almost by definition systemic because the pricing of risk in all deals is wrong.  In contrast, when multiple parties are evaluating each others’ performance risks using different models or different information, errors will be less correlated.
  4. CCPs generally do not price balance sheet risk (due to informational and governance considerations), which can encourage more trading by weaker firms, and can also encourage firms to take on more balance sheet risk (e.g., by taking more risk on non-cleared positions/investments).  This tends to increase risk in the system, and not just in the derivatives market: it can increase risks in other markets too.  It also provides another mechanism by which risks in other parts of the financial system affect the derivatives markets.
  5. CCPs mutualize default risk.  Mutual arrangements tend to encourage moral hazard and risk taking.  These can be controlled, but only at a cost.  Moreover, monitoring and control of risk are delegated to an agent typically subject to low powered incentives, and monitoring of the agent by the principals (CCP members) is subject to free riding because the benefits of monitoring accrue to all members but the costs fall completely on the member that incurs them.  In a CCP, deficient monitoring tends to have systemic consequences for reasons discussed above.
  6. Clearing can result in the capture of scale economies, but sacrifice scope economies, in the allocation of default risk.  It is not evident that this trade-off, and its effects on the allocation on default risk, is beneficial.
  7. The adoption of clearing is likely to have a tendency to increase total positions and trading volume.  This tends to increase overall market risks, and default risks along with them.  Position sizes are likely to rise for several reasons.  First, netting tends to reduce the capital required to support a given size of net position.  Expanding netting through clearing therefore tends to reduce the costs of maintaining positions, leading market participants to hold bigger ones–thereby taking on more risks.  (Another equilibrium effect.)  This cost reducing-position increasing effect is mentioned by advocates of clearing (I can provide cites) as one of its advantages.  More efficient use of capital is generally desirable, but it is important to note that bigger markets pose greater risks–including greater risks of default.  (The effects of netting on position sizes was actually advanced as a reason for not adopting clearing at the CBT in the 1910s-1920s: lack of clearing was said to “encourage conservatism” by limiting the size of positions–and hence risks–that firms could hold with a given level of capital.)  Second, by redistributing default risks from CCP non-members to CCP members, the adoption of a CCP tends to increase the derived demand to trade by end users.  This also encourages growth in market size, and via the channel discussed above, tends to increase the risks faced by systemically important institutions.  Put differently: do advocates of clearing really want derivatives markets to be bigger?  That’s a very likely consequence of clearing.
  8. In times of market stress, and large price moves, the rigidity of CCP collateral mechanisms can exacerbate liquidity problems.  Daily and intra-day variation margin payments are financed to a considerable degree through the extension of credit.  Large variation margin obligations resulting from large market price moves therefore translate into large increases in credit demand in stressed market conditions.  As was demonstrated during the Crash of ’87, under these conditions banks may decline to extend credit due to heightened uncertainty and lack of information about who is solvent and who is not.  Failure to obtain credit to make a margin call may force closure of a clearinghouse (again, as almost happened to the CBT and the CME in October, 1987).  Given the systemic, bottleneck nature of a large CCP, this would almost certainly spark a systemic crisis; it would certainly put pressure on central banks to provide liquidity, just as would occur from the failure of a large dealer in the OTC market.  Moreover, the traders who need to make variation margin payments are likely to sell assets.  This can create a positive feedback mechanism with very negative effects, by accelerating and exaggerating asset price declines, thereby necessitating additional margin payments, etc., etc.

Like I said at the outset, this list isn’t exhaustive.  And I certainly don’t know how, on net, all the various effects of clearing would play out from a systemic risk perspective.

But the point is that the advocates of clearing don’t either.  I haven’t seen evidence that they’ve even inventoried the possible effects, let alone made reasoned judgments about the net effect, or if that is even possible given the complexity of the system, the likely complicated equilibrium effects (i.e., how people adjust their behavior in response to the new regime), and the inevitable lack of information necessary to estimate the effects.

In brief, the mandated adoption of clearing will have big effects on the allocation of risk, the total amount of risk in the system, and the incentives of firms and individuals to take on, manage, and monitor risk.  Many of these effects plausibly increase systemic risk.  There are plausibly offsetting effects, but suffice it to say that anyone (and that means you, Timmy!, Gary, et al) who sells mandated clearing as a surefire means to reduce systemic risk are guilty of false advertising, not to say outright fraud.

Print Friendly, PDF & Email

1 Comment »

  1. I thought I was the only one that was nuts. This certainly, to me, concentrates the risk into one or two locations (for commodities) the CME and ICE. Talk about too big to fail. They haven’t seen anything yet.

    Comment by Michael Berry — January 2, 2011 @ 7:55 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress