May 15, 2013

Orange County: Story of a Risk Management Disaster

Robert Citron

I read this very interesting article about the role played by models in finance. By now, everyone should be aware that science and mathematics-based investment calculations applied to Finance may lead to enormous errors. However, models are not futile, their value is highly dependent on the art of their use. Therefore, I totally agree with the premise in the article that people should stop lying about models. 

But this is not the topic of today's post. I wanted to start a series of posts on risk management disasters. History has provided us with many examples to learn from. One of them that struck me the most is the Orange County bankruptcy. The biggest municipal bankruptcy ever. OC lost approximately $1.6 billion on a $7.6 billion portfolio. In the time period 1982-1994, Robert Citron, the longtime Treasurer-Tax Collector of Orange County, earned double the return on California's portfolio: 7.8% VS 4.8%. The first lesson I learned in graduate school is that you can NOT get more return without getting more risk. So this should already be a warning signal. But I will get to the composition of the portfolio in a bit. As in (almost) every RM catastrophes, lack of management supervision is one of causes of the disaster (see JP Morgan, London Whale case). In the OC case, the board wasn't willing to challenge Mr. Citron record of success. In fact, Thomas Riley, who was the OC Supervisor, on April 2nd tried to defend Citron by saying"This is a person who has gotten us millions of dollars. I don't know how in the hell he does it, but he makes us all look good." OC portfolio was a mix of customized structured products (that make Quants rich :D), interest rate swaps and fixed income securities. I want to emphasize that you can do only three things when you use complicated financial instruments such as derivates, swaps and options: HEDGING, SEMI-HEDGING (or hedging only partially) and SPECULATING. And you should always try to discern what a financial institution is saying it is doing and what it is actually doing. Most of the times, when a financial institution says it is hedging, it is indeed speculating. OC was clearly speculating! As an evidence of that, I would like to bring as an example a particular interest rate swap OC was using before declaring bankruptcy. An interest rate SWAP (vanilla) is an OTC (over-the-counter) contract where one counterparty pays a fixed rate of interest while the other pays a floating rate of interest, both based on a notional principal amount. What is the purpose of such a contract? The counterparty that pays a fixed rate of interest has usually an exposure to a floating interest rate, therefore by committing to pay a fixed rate of interest, it offsets the risk of an high volatile floating interest rate (usually loans to companies have a floating interest rate, for this reason they entry in contracts like this one). The second counterparty (which commits itself to pay a floating interest rate) is usually a big investment bank. I will not go into the specific details of the contract (cash exchanges, expiration, etc) as not relevant for this post. Now, let's consider a $100M contract, where A pays a fixed rate of 4% to B in return for receiving the 6-month LIBOR + 100bp (as compensation for the risk). A is long the swap. 

If A wants to hedge its exposure it would need a $100M swap contract. 

The net income for A and B, using some simple arithmetic would be: (LIBOR + 100bp) + (4% - LIBOR) = 5%

This is a simple hedge. A is hedging 100% of its exposure. '(LIBOR + 100bp)'represents the amount of interest 'B' pays to 'A', while (4% - LIBOR) is the payout for 'A', it can be positive or negative, it depends on the fluctuation of the LIBOR. If 4% < LIBOR, A is happy as the fixed rate it pays is less than the LIBOR, plus A also gets more interest payments from B (don't forget that B pays (LIBOR + 100bp) to A). If 4% > LIBOR 'A' is unhappy as the fixed rate 'A' pays is more than the LIBOR. Also A gets less interest payment in the form of (LIBOR + 100bp). The total net exposure for the two part is 5%. 

Now, let's consider the case where A has a $100 million exposure and it would need a $100M swap. Unfortunately, the only security available with a floating rate is $200 million. Thus A pays a fixed rate on a $200M security while B pays a floating interest rate on a $100M security.

The net income would be: (LIBOR + 100bp) + [2*(4%) - 2*(LIBOR)] = 9% - LIBOR

The only change in the equation above is the multiplier 2 (or gearing factor), as the security is worth twice the amount A is trying to hedge. This is an inverse floater and A here is speculating! This trade can be good or bad (live every one else) for A. It depends on the fluctuation of the interest rate: 9% - LIBOR. If 9% < LIBOR, A is happy as it pays a fixed rate to B which is less than the LIBOR plus A gets more interest payments from B in the form of  (LIBOR + 100bp). If 9% > LIBOR, A is losing money. 

We can go on and on with other examples. Let's assume now that A has a $100 million exposure but the only security available with a floating rate is $500 million. Yet, B pays the floating interest rate on a $100M security while A pays the fixed interest rate on a $500M security. 

The net income would be: (LIBOR + 100bp) + [5*(4%) - 5*(LIBOR)] = 21% - 4*LIBOR

The multiplier (or gearing factor) is 5 this time. This is called a leveraged inverse floater. Needless to say, the company here is speculating! Again, this trade can be good or bad for A. If 21% < 4*LIBOR, A is happy as it pays a fixed rate to B which is less than the LIBOR plus A gets more interest payments from B in the form of (LIBOR + 100bp). If 21% > 4*LIBOR, A is losing money.

OC had $5.3 billion in super charged inverse floater swaps, betting on falling US interest rate. OC was A's counterparty in the example above. OC portfolio also had $11.8 billion in fixed income securities (that lost value as interest rates raised) and some leverage (minimum common denominator for each RM catastrophes). As rates began to rise in February 1994 the losses begun to accumulate. In 10 months, the one year yield rose from 3.54% to 7.14%!!! The funny part is that Mr. Citron was relying on a mail-order astrologer and a psychic for interest rate forecasts!!! What really stroke me about the OC case was that the math above is just 7th grade arithmetic! We are not deriving Black and Scholes or doing complicated simulations. Mr. Citron could have seen the effects of the positions he entered in, both for a favorable or an unfavorable market. Also, market factors changed a lot and pretty fast and the intrinsic market risk exacerbated liquidity risk (like in a bank run). Finally, a lack of management and portfolio supervision were important elements for the disaster. For the record, OC had a AA rating by S&P and Moody's in 1994 before the disaster occured...

*Correction: OC was the biggest municipal bankruptcy until 2011. Alabama followed.


No comments:

Post a Comment