Discover the A-Z list of margin terminology
We outline key areas of margin management, define margin terminology and provide links to our expert insights.
The most important types of margin are Initial Margin and Variation Margin. However, additional margin is also charged to cover specific risks. Examples include:
- Liquidity Margin
- Default Margin – to cover any insufficiency in the size of the default fund.
- Credit Margin – extra margin where a member has fallen below the required credit rating.
- Model Risk – to cover potential issues with the formulas used within the algorithm, for example the option pricing model.
- Wrong Way Risk Margin – see Wrong Way Risk
Discretionary Margin – CCP’s always include in their rules the right to ask for additional margin for any reason.
This is the main component of any Initial Margin algorithm. It is a measure of the market risk of a portfolio under normal market conditions. Put another way, it is an estimate of the amount that the portfolio might be expected to lose over the holding period.
Or Central Counterparty. For cleared products, the institution that guarantees the performance of contracts and steps in if there is a default. To cover their risks they collect margin, in particular Initial Margin and Variation Margin, from their clearing members.
The purpose of delivery margin is to cover the risks between the last trading day and final settlement. There is no single way in which the delivery margin is calculated as it is dependent on the underlying delivery mechanism. Examples include taking full value, continuing to calculate Initial Margin for short positions only, or applying a specific charge based on penalties for non-delivery.
Many Initial Margin algorithms are VaR based. Normal VaR estimates how much a portfolio might lose (with a given Quantile or Confidence Level) by taking a specific scenario (depending on the number of scenarios). Expected shortfall takes into account a “fat tail” (high potential losses in the most extreme scenarios) by averaging all losses from the VaR scenario upwards. It is used in the Initial Margin algorithms of a number of CCP‘s to ensure the margin is more robust.
Or Filtered Historical VaR. This is the base for most CCP VaR Initial Margin algorithms (for example Eurex Prisma or LCH PAIRS). It is based on the following work: Giovanni Barone-Adesi, Kostas Giannopoulos, Les Vosper VaR without correlations for portfolio of derivative securities Quaderno. This first suggested the use of FHS VaR as a way to calculate margin. The difference between this and standard historical VaR is that the returns or scenarios are scaled to current volatility.
Global Netting Agreement
A Global Netting Agreement (GNA) is an agreement a firm has with their broker to net the risk for certain asset classes between their centrally cleared (see CCP) and bilateral book when calculating margin, resulting in a reduced Initial Margin being paid to the broker. The asset classes covered are determined by the various portfolio margin models that are in effect, for example it might cover netting of fixed income products.
Otherwise known as Margin Period of Risk (MPOR). The number of days that the CCP believes it will take to close out the portfolio in the event of a default. This is usually 1 or 2 days for exchange traded derivatives and 5 days for OTC products. Used in the calculation of Initial Margin, where it is used to determine the scenarios for the Base Margin calculation.
Variation Margin ensures that profit and losses are up to date. Initial Margin is used to estimate potential losses following a default, in the period between the last margin collection and the close out of the position. This will be calculated based on:
- An assumed Holding Period
- Likely price movements under normal market conditions
- To an assumed Confidence Level
Initial Margin algorithms also cover some additional risks, such as Liqudity or Model risk. Initial Margin is calculated based on algorithms designed by the CCPs, for example SPAN.
Jump to Default
A big part of the risk of trading CDS products. This explains why CDS Initial Margin algorithms tend to differ from those used for fixed income OTC products. They usually include components that look specifically at Wrong Way Risk credit spreads and interest rate risk as well as Jump to Default. In place of VaR they may use statistical modeling of credit spread and recovery rate fluctuations via a Monte Carlo Framework.
Knowledge is key when it comes to margin. Understanding the different algorithms, as well as the risks that the various components are covering is really important if you are trading both bilateral and cleared markets.
If you want to know more about how to manage you margin efficiently, including comparisons between cleared and bilateral margin, then look here: Margin Best Practices: A How-To Guide
Also Concentration Margin. In a default the CCP will be responsible for hedging and then closing (either by auction or trading) the clearing member positions. This is likely to cost more if the member has a large position. To allow for this, a liquidity component is included in the Initial Margin.
The Liquidity Margin is a non linear component of the margin, usually only calculated once a position goes above a particular size. It can have a significant impact on the margin requirement for certain positions. Find out more here: How liquidity add-ons can impact margin.
There is the margin number calculated by the margin algorithm, and then there is the number that is the actual requirement. This is because there are standard multipliers that are applied to the margin for particular accounts.
The most common multiplier is what used to be called the Speculator v Hedge multiplier, but now known as High Risk v Non High Risk, used by, for example, CME. A 1.1 multiplier is applied to the Initial Margin for High Risk accounts, meaning that the margin is 10% higher.
For OTC cleared products a higher margin is applied to client accounts. This is to allow for porting (moving positions to a new clearer following a default). The normal margin assumes a Holding Period of 5 days, but it might take 2 days to try and find a new clearer, and if this fails another 5 days will still be required to close out the position. Therefore the margin has to be scaled up to make it last for 5 days – generally this is achieved by multiplying the market risk parts of the client account margin by √(7/5).
Prices are assumed to follow a Normal Distribution in order to calculate VaR based margins. However, the distribution isn’t perfect which means that, unlike for scenario based algorithms like SPAN, the margin for a long position will no longer be the same as that for a short position. Some CCP‘s , such as Eurex, may also apply additional components in their Initial Margin algorithms to compensate for the fact that prices and curves are not normally distributed, adding to the complexity of the calculations.
Or Net Liquidating Value or Premium Margin.
For premium paid up front options, the Option Value is included in the overall margin calculation. Credit Option Value can be used to offset Initial Margin, whereas Debit Option Value must be covered by collateral in the same way as Initial Margin.
This doesn’t apply to all options. Some options are what is called “futures style”, with daily profit and loss being paid on them (Variation Margin), and these do not have Option Value. As an example ICE Financial, Commodity and Energy options are generally futures style.
Find out more here: How to use option value to optimise your margin in current market conditions.
Procyclicality occurs when an increase in margins creates further volatility in the market, requiring further increases in margin rates. The CCP margin requirements are meant to include buffers that limit increases in margins during periods of high volatility to avoid procyclicality. This is usually achieved by including a stress component within the Initial Margin methodology to create a lower band on the margin requirement. Alternatively, a percentage buffer is added to the margin rate that the CCP‘s believe should be set based on their statistical analysis.
Or Confidence Level. This is a measure of the probability that the initial margin will be sufficient to cover any losses following market moves. There are differences in the Confidence Level used by different CCP‘s and for different products, varying between 99% and 99.75%. This is decided by a combination of regulation and the risk appetite of the CCP.
These different Confidence Levels are then applied when calculating the VaR or Expected Shortfall as part of the margin algorithm.
The key component of SPAN parameter files. They consist of an array of potential losses per lot based on each of the 16 SPAN scenarios. The rest of the SPAN parameter file consists of all the other data needed to replicate the margin calculation, for example the Intercommodity Offsets allowed between different contracts.
The Risk Array concept is also used in the data provided by CCPs for replicating VaR based ETD margin algorithms; it’s just that they are a lot bigger – based on the number of VaR scenarios.
The most common ETD Initial Margin algorithm. Developed by CME who license it to exchanges around the world. The main component of SPAN is the Scanning Loss – a “worst case loss” from 16 scenarios, based on shifts in the underlying price (Scanning Range) and option volatility.
The SPAN algorithm includes Intracommodity Spreads, which add charges for expiry spread position which would attract little or no margin in the Scanning Loss calculation. There are also Intercommodity Offsets, which reduce the margin based on spreads between correlated products, for example Brent v WTI or between different Bond contracts.
There are other ETD margin algorithms, for example Prisma, TIMS and STANS. Some of these are based on FHS VaR rather than fixed scenarios, and many CCPs are looking to move to this. Find out more here: Impact of Moving From SPAN to VaR.
Trade v Position
Or Net v Gross. In general positions in a proprietary or house account will be held net, whereas positions in an omnibus client account will be held gross. Net v Gross becomes important when looking at the Initial Margin calculation because of the difference between US and European regulation.
US regulation allows a 1 day Holding Period to be used, as long as this is calculated based on Gross positions for client accounts. European regulation requires a 2 day Holding Period using net positions. This has led to the EMIR add-on, used to scale up Initial Margin from a 1 day to 2 day Holding Period. A factor of √2 is used to increase the Initial Margin requirement – usually by increasing the parameters by this scalar and then recalculating.
Uncleared Margin Rules
The Uncleared Margin Rules require margin to be exchanged between counterparties dependent on their Aggregate Average Notional Amount (AANA) – basically the amount of trading they conduct.
The main algorithm used to calculate the UMR margin requirement is SIMM. This takes as its inputs a set of provided parameters and the sensitivities of trades to specific risk factors. These sensitivities are provided in a standard format called CRIF.
A regulatory threshold of $50 million allows counterparties to only exchange margin when they breach this level. The main difference between UMR and the Initial Margin calculated by CCP‘s is that margin is a two way exchange for UMR, whereas cleared margin is only paid to the CCP.
You can find out more about UMR here: UMR e-book
The purpose of Variation Margin is to ensure that any profits or losses on a portfolio are “up to date” by “marking to market”.
There are two styles of Variation Margin:
- Realised – profits and losses are paid in cash, in the currency of the contract. This is applied to all futures and future style options, as well as cleared OTC products.
- Contingent – theoretical profits and losses are calculated. Examples include Option Value and Discount Contingent Variation Margin for Forwards. Losses must be covered in the same way as other margin requirements. Profits can be used to offset margin requirements in the same way as collateral.
Whether Variation Margin is Realised or Contingent is determined by a combination of the contract specification and the CCP rules.
Wrong Way Risk
For collateral, Wrong Way Risk will occur when the value of the collateral falls at the same time as either the clearing member has financial difficulties or the losses on the products that the margin is covering increases.
In practice this means that there will be specific restrictions on use of collateral, such as a clearing member not being able to use their own shares even if these are on the list of valid collateral. Coverage for Wrong Way Risk may also be included as Additional Margin as part of the Initial Margin methodology.
Or Cross Margin. Generally, this means that cleared OTC and ETD positions are margined together in a single Initial Margin calculation which takes into account the hedges between the two parts of the portfolio.
The most common Cross Margin offerings are provided by the big Interest Rate Swaps CCPs; CME, Eurex, LCH, JSCC. Fixed Income futures and options to be included in the standard swap margin algorithm are determined based on the benefit of offsets with the swaps compared with the impact of moving them from a 2 day Holding Period to 5 days to be in line with the OTC trades.
Yesterday v Today
Variation Margin is calculated by looking at the change in prices or rates between yesterday and today. CCPs will monitor these, and their impact on margin, on a continuous basis.
- If the total margin increases too much then there will be an intraday margin call – based on any change in Initial Margin and the potential end of day Variation Margin.
- Big price moves will result in a review of the margin parameters for scenario based Initial Margin algorithms such as SPAN. For FHS VaR based algorithms the price moves will have a direct impact on the scenarios used in the calculation.
You can find out more about unexpected changes in margin here:
An important number in margin calculation.
Net Margin is the sum of Initial Margin and Contingent Variation Margin (this will be Option Value) for Options or the Contingent Variation calculated for Forward contracts). This is the amount that needs to be covered by collateral (as opposed to Realised Variation Margin which is a cash flow). The Net Margin will never be greater than Zero. Any excess Contingent Variation Margin will be wasted as the profits cannot be realised.
In SPAN for certain portfolios it is possible to calculate a Zero margin. However, for short options this does not leave the CCP with any money to close out the position if there is a default. To allow for this, a Short Option Minimum Charge is calculated (a charge per short option position) meaning that the margin is no longer Zero.