Mar 12Liked by Russell Clark

I think you're barking up the wrong tree slightly here. The issue with LCH isn't compression trades, which are just a tidying up operation and have a small (and reducing) effect over time. The risk models have the usual issues with backward looking windows etc., but every major bank (and most HFs) do the same thing. In my experience, the risk models in the clearing houses are actually more sophisticated than those in the banks.

They're still not 'correct', if there is such a thing, but no worse than for bilateral trades, and for the regulators value at risk is the only game in town anyway. They do the typical stress tests the banks do, and they're pretty heavily monitored, again more than any bank I've worked at.

The real issues are the transformation of market risk into liquidity risk, which will be the cause of any eventual failure. A market shock leads to cash margin calls which need to be paid almost immediately and potentially causing a downward spiral as clearing members need cash fast. The LCH is not obliged to pay out margin on 'winning' portfolios in this way.

During Brexit Friday they drained massive liquidity from the market without returning any on the other side. In fact they made multiple intra day calls if I remember correctly and got a slap on the wrist from the bank.


The second issue is one you mentioned in the comments, the management of a default. I have spoken to several bank 'risk managers' with some responsibility in this area, and none of them had the foggiest idea of how that will operate, or what the clearing house rule book says they are entitled to do (anything they like). They are in for a massive shock if it ever happens, that kind of event for a large bank will be carnage.

If the dangers of clearing is your thing, Craig Pirrong is the go-to guy. He's been writing about the possible liquidity spiral since it was mandated. He's pretty spot-on I think.


Expand full comment

Hi Russell, interesting article, and your point on the LCH "repricing" the risk in gilts -- that is, changing margin requirements based on historical vol lookbacks -- intuitively makes sense for why it would cause a dearth of buyers and a market dislocation.

However, based on my experience trading IR derivates at a major NY hedge fund and interacting with LCH, I don't think your point on compression trades is accurate. Imagine there are 3 banks that all trade GBP IRS with each other. Bank A pays fixed with Bank B. Bank B pays fixed with Bank C. And Bank C pays fixed with Bank A. All for the same notional, say $1tn. Then each bank has, on it's books, a payer and receiver of $1tn and, therefore, net 0 interest rate exposure. However, they have immense ($1tn) notional exposure and counterparty exposure and thus would have to post margin with each other based on the large ($1tn) notional. Now comes along LCH, which steps in as counterparty to all trades. So each bank has a payer and receiver versus not different counterparties, but a single counterparty being LCH for the exact same amount, and both LCH and each of the banks agrees to just tear up the contracts and return the margin. Therefore the total outstanding notional in the market goes from $3tn down to 0, and all the associated margin is returned to the individual banks. Therefore, compression trades do, in fact, reduce outstanding notional exposure, which is why margin is returned.

Please tell me if I am missing something in your argument on compression trades.

Essentially this process used to take place prior to LCH as well when hedge funds wanted to "clean up" their books. We would send out large packages of net zero risk trades (but against different counterparties) for pricing so that a single bank would step in as the single counterparty and effectively tear up the excess notional that came from facing different counterparties on offsetting risk.

Expand full comment

Great piece. This centralization creates enormous systemic risk fat out in the tail. It looks ‘nicer’ and more ‘efficient’ but it’s risky nonetheless

Expand full comment

With all due respect as I am new to your newsletter, I don’t believe this is accurate. The reduction of gross notional in clearinghouses was a response GFC changes to how banks are capitalized, notables SLR as a backstop to risk based capital and potential GSIFI capital add ons based on a battery of metrics based on complexity, reach and size. Keeping gross notional sitting around for the sake of attracting capital add ons doesn’t make sense and starves the productive deployment of that capacity to other parts of the global economy that needs it. The clearinghouse facilitated this reduction, but didn’t force it. To put in simple terms if I owe you $20 and you owe me $20 should we just call it even? Unwinding Lehman Brothers derivative book was very complicated, even know their net risk to the street was negligible (I did it for a bank). If this had been cleared it would have been much simpler even without compression. Now, clearinghouses can cause forced selling if a member doesn’t have the liquidity to maintain its exposure, as was the case with MF Global. If you can’t make a margin call your collateral will be sold to protect the clearinghouse and its members. If the assets posted are the same ones causing your financial stress or stress in the market, they will be sold into a weak market, thus potentially causing more stress.

Sorry for the long post.

Expand full comment

ex post facto information. where else are their systemic risks to monitor for now and in the future ?

seems this weak spot is similar to 2008 - no way to unwind 30:1 .

Expand full comment
Oct 12, 2022·edited Oct 12, 2022

The problem with all of this is 4-5% yield on a 10 year bond is not that high. 1-2% was anomalously low. What private sector people did when yields were 1-2% to gear up returns is hardly the BoE's concern as long as the market is clearing and there is no systemic collapse risk.

Expand full comment

What do you think is the solution Russell? Are we going back to OTC trading and counter-parties?

Expand full comment