Counterparty Credit Risk

1. Introduction

ACES+ is a credit risk system which measures, analyses and manages counterparty risk of financial institutions and corporations by using simulation methodology. ACES+ along with DATA+, VAR+ and SWAP+ is developed by CD Financial Technology.

ACES+ builds on the development work of the SWAP+ system, on the basis of which several important extensions have been added. ACES+ covers extensively both conventional credit risk and derivatives credit risk positions. It forecasts credit exposures and rating changes, and allows inclusion of credit mitigation techniques. Technically, ACES+ system is ready for parallell computing; hence it can deal with the largest and most complex of credit risk portfolios.

Credit exposure in a cash instrument transaction arises when the institution has a net receivable mark-to-market position vis a vis a counterparty. On the other hand, credit exposure of a derivatives transaction typically arises whenever the institution is in a profitable mark-to-market position and the counterparty faces a mark-to-market loss. If the counterparty defaults, the institution loses some or all of the value of the transaction. The potential loss for the institution in default is credit exposure. Credit risk depends on credit exposure and the creditworthiness (i.e. rating) of the counterparty.

ACES+ can separately distinguish between contracts (transactions), counterparties and guarantors and associated risks caused by each such entity. Also, various netting structures and other credit mitigation techniques such as collateral and recovery possibilities can be taken into account in detail.

The software uses the path-dependent Monte Carlo simulation technique in order to simulate the credit exposure of any contract involving counterparty risk. In each simulation step, the full distributions of both pure market risk and credit risk for each entity (with appropriate netting rules) is determined. Finally, the development of different risk profiles of each entity over time can be observed.

ACES+ simulates both changes in market variables and changes in creditworthiness of counterparties (rating migration).

2. Analysis of contracts

ACES+ supports a wide range of contracts and products including cash flow based contracts, fixed cash flows, index linked cash flows, floating rate notes, swap contracts and many types of option linked repayments.

In loans and other cash flow based contracts it is typical to enter into agreements in which part or all of the credit exposure and risk is covered by credit mitigation techniques, e.g. collateral or guarantee arrangements.

In long term derivatives markets, it is common to enter into contracts with initial zero or near zero mark-to-market values. Be they conventional loan agreements or derivatives transactions, subsequent cash flows and market variable movements will inevitably make the market value of any long term contract (including its collateral and/or guatantee) positive or negative. This means the contract introduces risk which is either market or credit related. A further complication is that a bundle of contracts always contains 'structural' credit mitigation to a certain degree, as all the market factors affecting the mark to market values of all the positions seldom move simultaneously against the institution. In this way, portfolio effects play an important role in credit risk as well, independently of netting or grossing.

The complexity of credit analysis increases when we have many contracts with the same counterparty or if more than one contracts are guaranteed by the same counterparty. ACES+ combines credit risks so that we ‘net’ a contract with negative market value with a contract with positive value. If netting may not be applied, then ‘grossing’ is done, i.e., summing up each contract’s credit risk into aggregate risks without the netting effect.

The credit exposure involved in any contract depends on the type of cash flows, volatility of the underlying market variables and maturity of the contract. The potential credit risk is determined by the characteristics of the contract while the potential credit loss involves also creditworthiness of the counterparty. ACES+ determines both market value and creditworthiness with path-dependent simulation.

When counterparties enter a contract, collateral may be posted or guarantees may be given by guarantors of counterparties, depending on the type of transaction.
ACES+ allows the user to build the netting rules quite freely. The following diagram illustrates how the netting and grossing is done for a sample case.

ABC Bank is the guarantor of ABC Derivatives. Contracts A and B, done with ABC Derivatives, can be netted. Also, C and D can be netted. E and F may not be netted due to certain legal restrictions.

When we browse the results of ABC Derivatives, its risk is reported on mark-to-market basis where contracts A and B have been combined in ‘one contract’. The same combination is applied on C and D. Contracts E and F will carry their individual risks and values when aggregated to the total risk level as it is not legal for them to net each other.

ACES+ analysis captures the following objectives:

  1. Foreseeing the different characteristics of credit exposure that may occur
  2. Estimating the expected level of credit losses of a counterparty's or guarantor's default
  3. Pricing incremental credit risk generated by an incremental transaction in an existing credit risk portfolio, compared with a benchmark transaction

3. Methodology

The basis of ACES+ analysis is the distribution of the mark-to-market risk value of an individual contract. This is combined with the distribution of credit rating of the counterparty of the contract. The result is a realistic distribution of the value of the contract, and its forecasted evolution over time.

For example, a bond that matures in five years has after one year an uncertain market value because of the uncertainty in interest rates. Due to the stochastic nature of interest rates the probability distribution of the market value resembles that of a normally distributed random variable. However, rating migration introduces the possibility of a credit rating change, which has a much more dramatic effect on the discount factor used to price the bond. The true bond value distribution has then greater volatility than volatility induced by interest rates alone. Also, the distribution could be skewed or multimodal.

We use two types of measures for the credit risk: the Value-at-Risk (VAR) measure and the expected value of the credit risk exposure.

The diagram below shows the probability distribution of the market value of a contract at future point of time.

3.1. Expected credit exposure

Assuming a zero initial value contract, the expected credit exposure is represented by the right hand side of the above diagram, as we consider positive values of the distribution. Further, if denotes the market value of a contract from the analyst’s viewpoint at time t and denotes the corresponding counterparty credit risk, the expected credit risk, , is defined as .

If the contract were, instead, say a conventional business loan with no collateral or guarantee, then the zero point of the distribution would be at the initial market value of the contract, around which the above fluctuation would take place. The expected credit exposure would be a sum of the initial market value and , as defined above.

3.2. VAR measure

VAR measure is a probabilistic tool where risk measurement is based on the distribution of mark-to-market values. VAR gives the possibility to compare and combine different types of portfolios and contracts correctly. For example, a portfolio of conventional loans can be compared with a portfolio of long maturity derivatives and they can be combined to a portfolio whose true total market and credit risk can be measured.

VAR levels can be formally defined with a probability measure P, such that fulfils the condition , where p denotes the probability level. VAR measure has an intuitively clear content. For example, we can calculate the possible mark-to-market values of a contract for tomorrow and we mark the point where 95% of all possible values have lower credit risk (lower market value) than the remaining 5%. The amount of credit exposure defined with this procedure is the 95% VAR level for credit risk.

The VAR measure and the expected credit loss are determined for a specified time t because, in order to report the respective figures over the total life cycle of the contract, it is necessary to apply the measures over time.

The potential credit exposure and its maximum are illustrated in the figure below. The mark-to-market (MTM) VAR level measures the downside potential in a contract value, whereas the credit risk (CR)VAR level measures the upside potential.

In standard reporting, ACES+ provides four different (customizable) VAR levels, two on the market risk and two on the credit risk. As the entire distributions of portfolios are being computed, any distribution fractiles can be computed, if needed. The MTM VAR is measured from the left side tail of the portfolio value distribution, which is obtained by applying netting when contracts are combined. CR VAR is measured from the right side tail of the portfolio value distribution, where the value of the portfolio is obtained by applying grossing (when applicable).

A typical choice would be to request MTM VaR levels of 1% and 5% and CR VaR levels of 95% and 99%. (At this point, it should be noted that the MTM VAR for market risk reporting is always based on the netting of individual contracts while MTM for credit risk purposes involves potentially both netting and grossing.)

Rating migration in ACES+ is modeled applying the techniques of Markovian processes. Estimates for rating transition probabilities are typically obtained from rating agencies or other sources, including internal scoring procedures. This information are usually estimates of annual transition probabilities. The information is mathematically transformed so as to be compatible with flexible simulation time steps.

Collateral can be dealt with in ACES+ so that the actual contract (or portfolio of contracts) and its collateral constitute a portfolio, which is processed in a regular manner. The only special feature and difference of assets provided as collateral is that they can only be taken into account in credit risk analysis up to the level in any given market situation, where they cover 100 % of the underlying risk. The creditor will not benefit from any excesses beyond that. With this very general collateral treatment ACES+ can deal with virtually any type of collateral situation.

Recovery rates can be modeled in ACES+ on a flexible and fully customized basis assuming either deterministic or stochastic rates of recovery.

4. Risk measures in ACES+

ACES+ provides the following risk measures:

  1. Market risk (with contract netting)
  2. Expected mark-to-market value
  3. Volatility of market value
  4. MTM VAR (1% level)
  5. MTM VAR (5% level)
  6. Credit risk (with contract netting / grossing)
  7. Expected credit exposure at time t
  8. CR VAR (95% level)
  9. CR VAR (99% level)
  10. Average credit exposure over contract lifetime
  11. Maximum potential credit exposure (95% VAR)
  12. Credit loss estimates (with contract netting / grossing and rating migration)
  13. Expected credit loss at time t
  14. Credit loss forecast over contract lifetime

5. Simulation model

ACES+ uses stochastic Monte Carlo simulation. Scenarios of market prices and credit ratings are generated randomly for every time step. Contracts and counterparty exposures are priced against these scenarios. The market values of the contracts as well as counterparty and guarantor exposures are stored for reporting.

The user can analyze the market with flexible, and if needed, varying time steps, applying a desired number of simulated scenarios per time step. Typically, feasible risk profiles can be generated using 10 000 simulations for each time step.

The advantages of Monte Carlo simulation are: portfolio effects between market variables are correctly shown; individual contracts can be combined correctly to account for netting and grossing rules and non-linear instruments are treated consistently.

6. Advantages of ACES+

  1. Market variables are simulated with true correlations so risk statistics reflect portfolio effects
  2. Credit migration simulation gives a realistic distribution of contract's market value
  3. Path simulation allows long term changes (both market variable and credit rating) to take effect
  4. Multiple yield curves for one currency represent the credit spreads of different rating classes
  5. Flexible simulation steps to get more accuracy for desired maturity brackets, while still including the entire future time horizon
  6. Credit risk reporting can be combined to counterparty and guarantor levels in order to take the true netting rules into account
  7. Non-linear risks are correctly analysed with options and other derivative instruments
  8. Yield curves are simulated with a two-factor model, five parameter model or with principal components. These methods allow realistic spot and forward rate dynamics
  9. Yield curves can be built on risk-free interest rates with arbitrarily shaped constant or stochastic spread curves
  10. The variable coverage is large since ACES+ allows many types of stochastic processes to drive the market (Geometric Brownian motion, mean reversion and diffusion)
  11. Reporting includes graphical representations of credit exposure and mark-to-market distributions

ACES+ simulation engine can be used to price loan, deposit and various long term derivatives contracts against relevant benchmark contracts to ensure that the increased credit exposure in the operation being considered is fully compensated by the lower funding costs

7. Hardware and software environment

User interface of ACES+ is a Microsoft Access application. Portfolio data, market variables and calculation results are stored into a relational database server (MSDE or MS SQL Server 7.0). All the simulations and statistical work is done on a separate simulation engine which is implemented as a Microsoft COM object written in C++. The engine is capable of parallel simulation in a multiprocessor server.

These are the hardware and software requirements for ACES+ installation:

  • Workstation with Windows NT 4.0, 700 MHz CPU, 512 MB RAM and 100 MB of free hard disk space
  • Microsoft Access 2000, MSDE or MS SQL Server 7.0

Simulation engine need not be on the same workstation as the user interface is on, but both computers must be on the same network for COM architecture to work.

ACES+ can have hundreds of market variables and thousands of contracts with millions or cash flows in them. The larger and more complex the model, the more memory the calculation server should have.

8. Note

Depending on the client’s needs, VAR+ and the ACES+ applications or elements thereof can be combined and/or complemented to form a comprehensive risk management system. Alternatively, parts of them can also be used to support either in-house developed risk management systems or other systems available in the market.

For all our products, full demonstrations along with case studies are available on an individually tailored basis. For more information please contact:

CD Financial Technology
Mikonkatu 8

Tel: +358 9 612 3322

Related Products
Related Solutions