Every quarter the Office of the Comptroller of the Currency (the OCC) releases a report on U.S. bank derivatives activities. Derivatives trading (i.e. regarding financial instruments not held for hedging purpose) is an especially risky part of universal banks' activities. Yet the accounting of derivative contracts is relatively opaque in the bank quarterly financial reports, as the information is scattered over several sections and these items are partly taken off-balance sheet and therefore remain hidden from the casual observer or the investor uninterested in seeping through piles of abstruse documents. This lack of transparency remains despite the fact that credit derivatives like collateralized debt obligations (CDO), credit default swaps (CDS), and asset-backed securities, played a big role in the financial crisis of 2007-2008 and resulted in huge losses for many banks. Derivatives trading is risky, and it is still hard to estimate the amount of risk taken by individual banks. Consequently, the OCC quarterly reports are a gold mine for U.S. investors who want to know more about this. In this article I will compare the credit risk some banks incur through derivatives and how this risk has varied with time, based on OCC reports going as far back as 1996. I focus on JPMorgan Chase (NYSE:JPM), Bank of America (NYSE:BAC), Citigroup (NYSE:C), and Wells Fargo (NYSE:WFC), four of the largest U.S. banks. All these groups have very different risk profiles.
Derivatives and Risk
The OCC identifies two main risks when trading derivatives: credit risk and market risk.
Market risk is usually measured by a value-at-risk --- VaR ---, which is typically disclosed by large banks in their quarterly financial reports. It is, very roughly, a statistical estimate of how much a bank can lose during a typical bad trading day (where typical and bad are defined in probabilistic terms). At the end of Q3 2015, the total VaR reported by the 5 largest U.S. banks was $357 million, which is pretty low compared to credit risk.
Credit risk is an issue for banks trading in over-the-counter (OTC) derivatives. It is the risk that the expected cash flow or payment from a derivative contract will not happen, because the counterparty defaults on its obligations. Exchange-traded derivatives are not --- or less --- subject to this counterparty risk. A default by a counterparty may result in the bank having to write-off its derivatives at a significant loss. For a credit-default swap for instance, the risk to the buyer is that the issuer will default when the credit event is triggered and it will not make the expected payment: the buyer thought he/she was being insured, but when the house is on fire he/she finds out that the insurance cannot pay the bill. The risk to the issuer of the CDS is that the buyer won't pay its regular premiums. Here I focus mostly on OTC derivatives because they are inherently more risky: first, it is more difficult to estimate the fair value of an OTC derivative than an exchange-traded one because they are more illiquid; second, there is this counterparty risk. Exchange-traded derivatives are mainly "vanilla" options and futures contracts, while OTC derivatives include forwards, CDS, interest-rate swaps, CDO, swaptions, etc. According to the OCC report for Q3 2015, interest-rate derivatives represented 76.9% of all derivatives traded in the U.S. (in term of notional amount). Therefore, the OTC derivatives market is significantly larger than the exchange-traded one.
In recent years, the proportion of OTC derivatives that is centrally cleared has steadily increased: a single counterparty, the CCP --- central clearing party, or central counterparty clearing house --- clears the transactions, which reduces counterparty risk (while adding some concentration risk). The Dodd-Frank Act of 2010 restructured the derivatives market by requiring that certain standardized swaps be cleared through a CCP, and by imposing strict margin standards. Still, not all OTC derivatives can be centrally cleared. While most interest-rate swaps go through a CCP, as of Q3 2015, the notional value of centrally-cleared derivatives was only 61% of the notional value of all OTC derivatives.
Are the risks linked to derivatives visible on a bank balance sheet? No. The U.S. GAAP (Generally Accepted Accounting Principles) allows banks to offset (net) their derivative contracts when they report them at fair value on their financial statement. The idea of offsetting/netting is that rather than reporting several payment streams between issuer and buyer (if they entered into several contracts), a single consolidated net value is reported. This is different from the International Financial Reporting Standards (IFRS) accounting that most of the world follows, and this explains why the assets and liabilities reported by U.S. banks are deceptively small compared to, say, European banks. IFRS does allow some bilateral netting of derivatives, but to a lesser extent than the U.S. GAAP. In recent years, there have been attempts at closing the gap between GAAP and IFRS. The rationale behind netting is that it reduces credit exposure. However, it somewhat hides the full extend to which a bank is involved in derivatives trading. Reporting net derivatives at fair value does not give the reader a useful indication regarding credit risk. Even when notional derivatives amounts are reported in the quarterly reports, they are not that useful as often these amounts are not exchanged, and therefore bear little relevance to the actual credit risk.
So, how to measure OTC derivatives credit risk? There is a standardized and widely used approach to compute credit risk, called the Current Exposure Method (or CEM). Under the CEM, the credit exposure of a bank is the sum of its current credit exposure (the CCE) and of its potential future exposure (the PFE). The OCC also computes the net current credit exposure (NCCE; the gross negative fair value of all derivative contracts is subtracted from the gross positive fair value). For a given contract, if the fair value is positive, then this is the current credit exposure of the bank. If it is negative, then the current credit exposure is 0: the bank poses a risk to its counterparty, but its counterparty pose --- in theory --- no credit risk to the bank. The total CCE is the sum of each contract's CCE. Then, the PFE is computed by multiplying the notional amount of derivative contracts by a conversion factor that depends on the perceived riskiness of each contract and their years to maturity. For instance, the conversion factor for an interest-rate swap with a 1 year or less maturity is 0, implying no future risk, while for some CDS it can reach 0.1. This basic PFE can then be "corrected" by a net-to-gross ratio, to take into account the reduction in credit risk that bilateral netting agreements offer. Prior to Q2 2007, the OCC was not correcting the PFE by this net-to-gross ratio, but it decided to change its methodology (which significantly lowered the reported credit risk exposures).
U.S. Banks and Derivatives Credit Exposure
Let's have a look at the credit risk four major U.S. banks are subjected to. The following figure based on 19 years of OCC reports shows the proportion of credit exposure to capital (tier 1 + tier 2) of the financial institutions studied here: BAC, JPM, C, and WFC.
The first thing to notice is that each curve is separated into 2 parts, one before Q2 2007 (dashed line), and one after (solid line). Again, the reason is that the OCC changed its methodology during Q2 2007 when computing the PFE. This figure raises several interesting points: 1) overall credit risk has significantly decreased since its peak of roughly the end of 2008, across all banks studied; 2) of the lot, JPMorgan Chase is the bank most exposed to credit risk, followed by Citigroup and Bank of America, while Wells Fargo is the safest one. Nothing very surprising here, as Wells Fargo has the reputation of being more of a brick-and-mortar retail bank than its competitors. JPMorgan Chase is much more of an investment bank. The credit exposure of JPM prior to the financial crisis of 2007-2008 seems, in retrospect, insane. Only Goldman Sachs (NYSE:GS) currently has a higher credit exposure than JPM: 530% of its capital as of Q3 2015. This is expected as GS is a pure investment bank. I wonder how many investors in JPM are aware of the bank exposure to such risk. Moreover, 94.5% of the derivative contracts that JPM enters into are OTC versus 92% for GS, and 12.65% of the gross revenues of JPM during Q3 2015 were from derivatives trading, surprisingly close to the 13.32% of GS.
This level of credit exposure is not without consequences: JPM is currently the U.S. bank with the highest capital requirements. In July 2015, the Federal Reserve assigned it a capital surcharge of 4.5% of its risk-weighted assets, compared to 3.5% for C, 3% for BAC, and 2% for WFC. Clearly, it's going to cost a lot to JPM to remain big and to continue taking as much risk as it does. JPM being a GSIB (Global Systemically Important Bank), its total required CET1 capital ratio is 7% (standard) + 4.5% (surcharge), i.e. 11.5% . As of Q3 2015, the Basel III CET1 ratio of JPM was already a healthy 11.4%. So some good progress has already been made in meeting the regulator's requirements. Still, it's not clear that even a 4.5% surcharge is an appropriate level for the kind of risks JPM seems to be taking.
Case in point: JPMorgan Chase and the London Whale
JPM is painfully aware of derivatives risks. At the end of 2011, it held a portfolio of CDS with a notional amount of $51 billion allegedly for hedging purpose, in its CIO division. This amount rose to $157 billion by the end of Q1 2012. In the meantime, the average value-at-risk reported by the division holding these CDS was $129 million during Q1 2012. $129 million seems like a lot compared to the latest reported VaR ($54 million for the bank in Q3 2015), but it's pretty tame compared to the actual loss of $2 billion reported in April 2012 and attributed to bad trades the London whale (a trader) made on these CDS. This loss was later updated to $5.8 billion. What to conclude of this?... Do not trust the reported value-at-risk. VaRs are not the maximum loss a bank can incur, just a reasonable estimate of a potential loss based on many assumptions. OTC derivatives are inherently risky, and most of the mathematical models used to price them and evaluate their risks are based on assumptions that do not always apply or are too simplistic.
Of the four banks considered here --- JPMorgan Chase, Citigroup, Bank of America, and Wells Fargo --- JPM is by far the riskiest financial institution in term of derivatives risk. Wells Fargo is the safest. This is known by most investors, and resulted (amongst other considerations) in the Federal Reserve imposing a higher capital surcharge to JPM than to any other U.S. bank. The question is: is this surcharge (4.5% of its risk-weighted assets) appropriate? I'm personally not convinced considering the high level of credit risk the bank is exposed to through its derivative contracts, and I find JPM a bit too risky for me. Only Goldman Sachs is more exposed to derivatives credit risk. In terms of trading risk, Citigroup reported a larger VaR: $116 million in Q3 2015 versus $54 million for JPM. Bank of America reported $60 million. Here too, GS is more risky than JPM with a VaR of $74 million. I do not put much faith in these VaRs, as their computation requires many assumptions that may break in unfavorable market conditions. The main risk though is the credit risk, and with a total credit exposure equal to 219% of its capital, JPM seems a much riskier proposition than WFC and its 36% exposure.
The Dodd-Frank Act, the standardization of numerous derivative contracts, and the higher capital requirements imposed by the Federal Reserve, have all made banks safer now than prior to the financial crisis. Still, risk remains and the resurgence of Collateralized Debt Obligations under the name "bespoke tranche opportunity" sold by Goldman Sachs and other banks worries me as it is a sign that the lessons of the financial crisis might start to lose some of their impact.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.