With commodity and energy prices regaining volatility, most players are struggling to find answers to one of their most critical risk factors – Counterparty Credit Risk. While many, especially in the oil & energy sector, have VaR based tools to measure and manage their Market Risks, few have ventured formally into expanding the scope of Risk Management to Credit Risk.
Counterparty Credit Risk can be understood better by looking at its 3 basic categories - Default Risk (Counterparty fails to honor the deal), Replacement Risk (Risk that you may not be able to replace that defaulted deal under the same conditions with any other counterparty) and Settlement Risk (Risk of a 3rd party, like the bank playing an intermediary, failing – resulting in default). Most Energy and Commodity Trading Companies are exposed to these risks, but even some of the world’s leading companies today rely on “thumb-rules” or score-card based approach for Countering Credit Risk. This kind of non-scientific approach leaves companies completely vulnerable to oncoming credit defaults which they have no way to pre-empt and prepare for.
However, this realization is pushing some of the leading companies to re-look at their credit risk management strategies and build a scientific model around it so measuring it is easier. One of the models that comes up often as a favorite choice, is PFE – Potential Future Exposure. In the banking industry, regulators have mandated that banks measure and report their PFE along with their VaRs as an integral part of their overall enterprise risk number.
What is PFE?
Potential future exposure (PFE) is the likely outstanding amount at some future time at which a default event might occur. It can be expressed as a dollar amount or as percentage of book value, current value, or notional value.
Many philosophical and practical debates have ensued regarding whether to treat PFE as a statistical measure of economic value, or a fixed number adjusted using “in-house market intelligence“. The case against “theorists” is made on the basis that no market-based models can really and reliably predict counterparty behavior the way in-house people can, since they have been dealing with those counterparties for years. However, a credit risk measure like PFE at least allows us to visualize the quantum of loss that could happen in case of a default in probabilistic terms, much like VaR. PFE is also a stochastic measure and is thus affected by statistically random factors like currency movements, interest rates, and such – and therefore carries a lot of information with it than is humanely possible (at least for a complex portfolio). Moreover, it is also important to treat all swaps and forwards consistently by using the same model to value their Present Value from their Future Value. PFE turns up pretty good results on these areas.
How to Measure PFE?
Statistical models built for measuring PFE borrow concepts heavily from Market Risk models. Measuring PFE though, is not as straight-forward. All the deals need to be revalued as their future value will be different from their current value. And as there are several risk factors that could affect the price movements, we also need to use statistical models of probability, which is where similarity between credit and market risk models appear the most.
The computational complexity involved in this can be estimated from the fact that even if each deal in a portfolio needs to be valued with 10,000 risk factors over an estimated 100 simulation time points, that’s 1 million numbers per deal ! A small(ish) portfolio of just 100 different deals can therefore generate 100 mn numbers, which is nearly impossible for any spreadsheet based solution to derive, but is not a very difficult feat to achieve for a software solution built specifically for such analytics. At RiskEdge, we use Monte-Carlo Simulations approach to determine PFE, which provides by far the most accurate picture by simulating a large number of scenarios with relative ease.
Another complexity in measuring PFE is differing perceptions on netting similar trades. Netting agreements are legal contracts between two counterparties that allows them to, in an event where one counterparty defaults, aggregate deals before settling the payments. Sometimes select trades are left out of the ambit of netting agreements as well. This complex arrangement means that netting and non-netting rules have to be carefully woven in such a way that the deals are filtered in the correct way. In its recently released PFE module, RiskEdge provides clients with PFE measures for both netting and non-netting scenarios, thus enabling them to get different perspectives on the same data.
In the end…
To summarize, here is a quote from the character “John Tuld” in the movie Margin Call (2011) that hits home - “… it’s certainly no different today than its ever been. 1637, 1797, 1819, 37, 57, 84, 1901, 07, 29, 1937, 1974, 1987-Jesus, didn’t that f**k up me up good-92, 97, 2000 and whatever we want to call this (2008). It’s all just the same thing over and over; we can’t help ourselves. And you and I can’t control it, or stop it, or even slow it. Or even ever-so-slightly alter it. We just react. And we make a lot money if we get it right. And we get left by the side of the road if we get it wrong.”