Supporting SIMM

Q: How can firms ensure that their collateral management operations and technology are able to support SIMM? 

Rahba: Firms need to embark on quite a long and complex journey to get their organisation and collateral management procedures up to speed with the phase-in of initial margin regulations. Such a project covers the following dimensions: legal negotiation, model validation and implementation, back-testing and collateral operations:

• Legal negotiation: firms need to sign new legal documents, IM CSAs, that will govern the mandatory exchange of initial margin. Such a process involves several steps: identify those counterparties which are subject to IM rules, compare trading relationships with active CSAs, negotiate contract clauses, including applicable jurisdictions, choice of calculation method (ISDA SIMM vs Schedule-based), choice of custodians, etc.

• Model validation and implementation: firms can either use a schedule-based method, or a model with margins that meets a 99% confidence level of cover over a 10-day standard margin period of risk. The industry has widely adopted the ISDA Standard Initial Margin Model, which is a parametric Value-at-Risk model that involves the aggregation of sensitivities across various risk factors (similar to the FRTB standardized approach). Firms need not only to implement these models, but also to go through a regulatory approval process across national regulators.

• Back-testing / Benchmarking: as part of this validation process, firms need to provide back-testing results for portfolios in scope, i.e. simulate SIMM on past data to gauge its accuracy and effectiveness. Similarly, they need to provide benchmarking results, comparing SIMM to another model (e.g. VaR).

• Collateral operations: once firms are equipped with capabilities to compute regulatory-compliant initial margins, they need to update their collateral processes to ensure these margins get settled on time.

We can see that multiple departments are thus impacted including front office, collateral management operations, risk and legal departments. 

In short, to ensure that collateral operations and technology both support these regulatory margin requirements using the SIMM, and to be ready by the deadline, financial institutions need to look at the collateral management process front to back and holistically, across these several dimensions, and make sure an appropriate level of governance is put in place. 

Of course, there are variations with readiness and the degree of work required to get up to speed with margin rules and SIMM 2.0 depending on the size and shape of the firm itself. Larger firms, for whom initial margin requirements have been phased in since September 2016, tend to have a more decentralised risk infrastructure. These firms might therefore need to have a decentralised way of managing SIMM, generating risk sensitivities across these systems. Smaller firms that will phase in between now and 2020, generally have a more centralised way of managing the process from sensitivity generation across asset classes down to margin calculation and so on. 

Q: How different is version 2.0 of SIMM compared to the previous version?

Rahba: The new version of the SIMM, 2.0 introduces risk factors for three product types – volatility indices, quanto CDS and municipal swaps – and includes a full recalibration of parameters (correlations and risk weights). 

This new version has been in effect since the end of 2017. Looking ahead, the model will need to be calibrated on a regular basis as part of the governance of SIMM. We are expecting updates of this model on a yearly basis, meaning new parameters as part of this calibration exercise, and feedback from market participants, as well as a methodology review. Firms will need to be able to comply with these basic SIMM calculation models, and methodology reviews, and maintain these calculation calibrations on a yearly basis. 

Q: Based on the experience of those who've already implemented SIMM, what are some of the tricky implementation points to be handled with care?

Rahba: Let’s look at the end-to-end process of getting IM calls settled on a daily basis. It is a three-step process:

• Sensitivities generation, across various asset classes and products. 

• IM computation, by aggregating these sensitivities, using risk weights and correlations.

• IM call processing.

We see most challenges arising from the first and last piece of the process. Sensitivity generation requires the ability to project and compute sensitivities across various predefined risk factors (rates, credit, fx, equity and commodities). This is a challenging task because of the variety of products that needs to be covered and validated, and it requires trading and risk platforms to:

• Represent and model key collateral data from the CSA agreements, such as scope of product and applicable jurisdictions.

• Provide capabilities to output various types of sensitivities – delta, vega, curvature.

• Be flexible enough to cater for SIMM specifics on top of an existing usage (e.g. cross-currency swaps).

• Classify these sensitivities, according to the SIMM rules, possibly relying on an external utility providing such classification services.

• Format results as per the standard ISDA CRIF (Common Risk Interchange Format).

The last piece of the process, margin call processing, has similarities with the variation margin process, but it comes with some specificities: in particular, once firms have agreed with their counterparties on the amounts of initial margin and additional collateral, they are then generally relying on tri-party agents to fulfil their collateral requirements. While the VM process is mostly a cash-based bilateral process, IM is security-based, and involves a third-party.

It highlights the need for collateral management systems to:

• Extend the VM process to IM, including the ability to process in real-time margin calls on an electronic messaging platform such as Acadiasoft Marghinsphere.

• Extend the usage of securities as collateral.

• Connect with key tri-party agents to get collateral allocated.

We have helped our clients overcome these challenges. Larger firms that were part of earlier phases, might be using our platform from a front-office perspective, where we are focusing on making sure they can generate sensitivities across asset classes. Smaller firms use our platform front-to-back, and need not only to generate CRIF-formatted sensitivities, but also rely on our capabilities to compute initial margin, and use our collateral operations to generate margin calls.  Across all our clients, we are seeing collateral data reinforced as being key across various departments: not only from a VM and IM calculation and processing perspective, but also for pricing, trading and risk management. 

On top of these implementation challenges which focus on the daily calculation and processing of IM calls, back-testing and benchmarking requires dedicated attention as part of the model validation exercise. From an implementation perspective, market risk capabilities (P&L, Value at Risk) need to be leveraged to pass the validation step, and monitoring procedures need to be established. So it should not be underestimated.

In a nutshell, a SIMM project involves various parties (front-office, risk, operations, legal) and different technology capabilities along those functions, so one of the challenges is to set up the right project governance model.

Q: Are firms looking at initial margin as a post-trade down-stream process, or is it influencing trading decisions?

Rahba: First and foremost, firms need to be regulatory compliant by exchanging margins on a daily basis. We have detailed how complex such a project is. As IM regulations are being phased-in until 2020, we see a lot of institutions focusing on initial margin as a post-trade down-stream process.

As the dust starts to settle for firms that have put the deadline behind them, these firms are looking more and more at the impact of such margins on the value chain of trading. For example, firms might want to put in place some risk control procedures and limits to closely monitor how far they are from the initial margin threshold. Initial margin costs can also be used pre-trade to feed a trading decision, for example to optimize counterparty choice. Some clients are also looking at extending valuation adjustment capabilities to incorporate funding effects related to IM, in the Margin Valuation Adjustment (MVA).