GlobalCapital: How has the OCC operated under remote working conditions?
John Davidson, OCC: We have been able to operate quit seamlessly. We have well over 98% of our staff working from home and have been able to handle record volumes and quite a bit of volatility without any issues. Many of our clearing members and exchanges are operating in the same mode.
How extreme have recent months of volatility been for the clearing and options industry?
The last six days of February saw record trading volume for the options industry and the OCC — 43% of the total trading volume from February happened over its last six trading days. March was considerably volatile and had high volume too. Even through April the volume continued to be well above our averages for last year.
The other important factor from the perspective of the central counterparty is the volatility in the markets. I have been in the industry for quite some time and ran CME Clearing during the 1987 crash. This was a much longer, sustained period of volatility. And we are quite pleased with the way our margin models operated during the period. They performed exactly as we expected.
Changes to the margin models we made and approved during 2018 and 2019 were quite important, as was the change to our clearing fund stress testing, which also performed quite well. I would also say our clearing members performed extremely well. That was a function of our financial risk management group being in constant communication with clearing members, operations officers and settlement banks.
In the US, we interact with the banking system through commercial banks. We don’t have access to central bank money for settlements. But all our obligations were met on a timely basis. I would say that the market infrastructure performed very, very well and continues to perform well, notwithstanding the fact that we are working from home.
How did you set up your infrastructure to handle the volatility?
We made a number of investments in OCC legacy systems in 2019, notwithstanding a three year project to completely rebuild our system’s infrastructure.
We had realised that we needed to invest in our legacy Encore system to make sure it was sufficiently resilient. So we purchased two new mainframes during 2019, one in our main data centre and one in our back-up data centre. We made some very significant software enhancements to the legacy, so that it processed information in a more efficient way.
We also conducted business continuity testing through a tabletop exercise. It wasn’t through any real insight, but fortuitously our tabletop exercise in November was for a pandemic. So we had a good sense of how to operate our environment on a remote basis. We have also had weather events where most of our Dallas and Chicago staff have had to work from home, so testing of those parts has been pretty seamless.
We have also been seeking regulatory approval for changes to our margin model, capital management policy and clearing fund processes. Changes to our clearing fund were approved very early this year and we have just received approval to formally seal the next step for changes on how stress test liquidity, to make sure the size of our clearing fund is ready to cope with credit risks and liquidity demands that would arise from various extreme but plausible events.
Could you describe in more detail how you have set up your infrastructure for widescale working from home?
We have a dedicated VPN network and people sign on to the same virtual desktop as machines in the office, so the infrastructure is identical. We didn’t have to transport equipment to people’s homes. Everyone has a laptop or desktop at home. We also have a stipend in place for some of the team to purchase compatible technology.
Our volume is completely transparent. It comes in on direct feeds from all 16 of the exchanges directly into our primary data centre, which is connected to our back-up data centre, and gets logged into the system. The human interaction with that is exception reporting and monitoring of issues, as well as changes in the market output of risk models.
The size of what you are looking at, with potential collateral flows, changes dramatically as the market volatility increases. But we still have the same 100 clearing members so the actual amount of data that we are interacting with doesn’t really change. A few people will monitor this from home but it’s not like a trading operation, where you need a stack of monitors taller than your head.
We would only need to go into our dark data centres if we had to fix the wire configuration in a particular process. That is extremely rare and because we are critical infrastructure in the view of the Department of Homeland Security and the US Treasury, our staff would be able to travel to the data centre to do what needs doing and go home.
Could you describe your margin methodology during this time?
Our overall margin methodology is called the STANS methodology. It is a collection of about 24 models that react to different product types and market conditions. The basic model is a 10,000 scenario Monte Carlo simulation of expected future price changes using, rather than any value at risk methodology, an expected shortfall methodology. It is very sensitive to tail events that happen in the distribution of returns.
Essentially, we haven’t really fundamentally added very many new models, but we have significantly revised the operation of a certain [number] of the models. The exception is the clearing fund, where we have introduced stress testing as the major determination of sizing the clearing fund, whereas previously we had on a fairly ad hoc basis trebled the margin variance methodology, which was not very sensitive to actual distribution of changes in a market environment.
The key consideration is the way those models respond to changes in volatility, either in the CBOE Volatility Index (VIX) itself, or its underlying, the S&P 500. But you also have to keep in mind that there are about 3,500 options on different stocks that may behave differently from that. One of the key issues is how does a straight line respond to an increase in volatility — assuming that the same increase will happen the next day, or is it more moderated than that?
What you don’t want to do is be procyclical, so we have a number of controls to make sure we have a conservative amount of coverage, so that coverage slowly expands as volatility expands. We don’t want extreme calls for quickly increased margin requirements in a volatile situation that might have an adverse impact on the financial liquidity in the market.
We have also always had the models look at correlations, which are critical in this marketplace — particularly when you have stock indices, individual components of stock indices and relationships between particular securities in a sector. The correlations are not rigid, they change over time. So our system looks at perfect correlations between different instruments, no correlation between instruments and then a 500 day historical look-back on actual correlations between those instruments.
It chooses as a margin requirement the highest of those three possibilities. So the change in correlation is well built into the margin model infrastructure.
We have also enhanced the concentration calculations so that positions that are concentrated, and might have an adverse impact on the marketplace if they need liquidating, might have a higher margin charge than unconcentrated positions.
The relationship between the various indices and underlyings and the VIX index get more sophisticated treatment than they did in some of our past models.
CME Group auctioned off Ronin’s portfolios in March, when the firm could no longer meet its capital requirements at the CCP. How did that affect you?
Ronin is not a direct member of the OCC, but the clearing member through which it processed the options-related transaction which we clear performed very well during entire period. It kept us well informed about what was going on and met all of its obligations on a very timely basis.
Are you seeking a credit facility to guard against clearing member default, as other clearing houses have done?
The OCC has a $2bn facility in place that has a $1bn accordion feature and we are in process of a renewal that will extend into June, when we anticipate the finalisation of the renewal.
The OCC is also going through the regulatory process of approval for a non-bank credit facility. Bank credit facilities are good things and we have a very diverse group of 19 banking institutions that participate in our facility. It allows us to draw within one hour of asking for a draw or the default of a member.
But many of those banks are also parts of holding companies that have clearing members, so there is some risk of contagion. So we limit the size of the commitments that we allow from banks with holding companies that have a broker dealer as one of our clearing members.
With the non-bank credit facility we can get to public pension funds and access a committed repo facility from them, so that we don’t have that potential for contagion.
How has the clearing industry held up during this real time stress test?
I think that across the board CCPs have handled the volatility in a very robust manner.
I don’t think one would have said, before the experience of this spring, that all of the asset classes would be subject to extreme volatility at the same time: the extreme equity market volatility, the extreme situation with oil and the extreme changes in interest rate sensitivity, given the actions that various central banks took in response to the coronavirus wind-down of significant parts of the economy. The CCP infrastructure has behaved in a very appropriate and successful manner.
The performance of CCPs over the recent bout of market volatility has demonstrated their importance and the much greater strength of financial market infrastructure compared with 2008.
Finally, how are you planning to move the OCC out of lockdown and back to the office?
We are spending quite a bit of time on that question. Our view is that, unlike those parts of the economy that require face-to-face interaction, we are open for business. The exchanges that we support have been performing, so we don’t need to be first movers with respect to coming back to work. The health and safety of our employees is the most important consideration. We will not return to work in a big bang manner. Schools are not going to be open and summer camps will not open this year, so there are childcare issues to consider.
We also have to account for people with sensitive medical conditions. We have just completed a survey of employees on their concerns and on issues such as social distancing in an open plan office and how many people should be in a conference room at once. We may do rotating assignments, so one group across sectors works one week in the office and two weeks at home. We are developing increasingly sophisticated plans in that regard, but it is not yet time to exercise that.
One thing is sure: we won’t be first back to the office and it won’t be everyone showing up on Monday willy-nilly. It will be gradual and we will follow the lead of the scientists and the authorities.
