Sunday, January 25, 2009

Was Merrill Casualty #3 of The Basis Trade After DB Prop and Citadel

In a bet gone very bad, that if true would make Jerome Kerviel's $5 billion loss at Soc Gen seem like amateur hour, the WSJ reports ($$$ link with hat tip to that the main reason for Merrill's massive $15 billion Q4 loss was due to some very large basis trades gone horribly wrong. We wrote briefly about the basis trade here but now with attention turning more firmly to this topic, it is worth revisiting.

Before we get back to what potentially could be the culprit for over $30 billion in prop trading and hedge fund losses last quarter in all of Wall Street, let's reexamine the basics. As we mentioned previously, at its core, a basis trade is a hedged position where an account buys a bond (let's say with a 5 year maturity) and hedges it with a matched-maturity (or comparable duration) Credit Default Swap. In this way, the account is completely insured from default risk on the bond purchased since if the underlying company that issued the bond were to file for bankruptcy, the account would lose the principal value on the bond but would pick up the recovery from the CDS, ending up with a 0 net gain/loss at the time of default event (CDS settlements can be physical or cash as defined by the CDS OTC-clearing authority ISDA, we will write more about this at a later date).

When an account buys the bond, one also receives the cash flows associated with the coupon on the bond; when hedged in a basis trade, as CDS has a "negative coupon", or the quarterly payment associated with paying for the "insurance", the net recurring cash out/inflow to the account is known as the "basis." In the good old days, the basis would be usually positive, meaning that to hedge a position perfectly, there would be some, usually very minor quarterly cash outflow, usually to the tune of 5-10 bps on the entire notional exposure. Again, in the good old days, a "negative basis" was rare to find - these are positions that for whatever technical or fundamental reasons, would be net cash positive. In other words, an account would have no default risk thru the bond's maturity, and would be compensated to have it put on the books. It would be rare to find negative bases of -10 bps, so hedge funds and prop desks would immediately snap these up as they became available in the market and usually lever them up dramatically, sometimes to the tune of 100-to-1, using a gullible Prime Broker or other synthetic instruments, and end up with anywhere from a 5% to 10% risk free annuity for years.

Another basis trade 101: When looking at the basis of a bond and match-maturity CDS, the most relevant bond metric is its Z-spread, or the spread to Libor, not the more traditional spread to comparable treasury. When comping bonds and CDS, one cares mostly about the bond's Z-spread as that gives the most appropriate reference of whether the bond trades rich or cheap vis-a-vis a CDS. This is because a CDS spread is also relative to the appropriate metric on the Libor curve, not relative to Treasuries.

So back to the basis: the problem with the whole "risk free" concept is that it made one major assumption - that liquidity would be essentially infinite. As the Bear Stearns implosion and the Lehman bankruptcy showed, this is one assumption that would be promptly crushed, and would lead to dramatic aberrations in the basis trade. One major issue was the availability of CDS, or rather lack thereof, to hedge cash bond positions. As prime brokers rushed to conserve liquidity, they made it virtually impossible for accounts to take advantage of dropping bond prices, and increasing Z spreads. They did this by exponentially increasing funding costs on CDS: traditionally the margin requirements on CDS would be in the sub 1% range; after Lehman some counterparties raised the margin requirements as high as 20%, and others even asked for the whole margin to be paid up front. On a $10 million CDS position, which traditionally would only have cash outflows every quarter to fund the quarterly insurance payment, all of a sudden accounts would have to pony up to $2 million in margin just to put a trade on (or keep it on, leading to many basis forced unwinds). Of course, this made it prohibitive for many but the largest hedge funds to participate in the heretofore extremely liquid CDS market, thereby creating phenomenal dislocations and the great arbitrage of negative basis trades that would create 5%-10% and on rare occasions even 15% unlevered returns!

As this can be a handful to swallow at first, we have presented this graphically. We chose to demonstrate the negative basis trade currently available in CIT Group's 5.0% Notes due 2/2014, which we match with CIT 5 year CDS due March 20, 2014. On the graphic below, it is evident that the basis had been gravitating around zero for a while, initially starting off as positive around the time JPM was taking over Bear, then was roughly 0 for several months, but then became dramatically negative the day after Lehman filed. In fact the spread went from 0 to almost 1,500 bps (or 15%) almost overnight! And in the subsequent liquidity constrained market, the basis spread has fluctuated all over, and is currently roughly -500 bps.

This is just one example: most high yield and cross over names currently represent negative basis opportunities: some interesting outliers include Marriott Hotels, Home Depot, Temple-Inland and Omnicom, all of which are BBB- (or higher rated) credits yet present negative basis opportunities of 300 bps and wider. We would be very cautious with blindly purchasing these bases, as the liquidity premium in the market will likely be a key concern for along time, and as long as that is the case, there is no reason why the basis trade should collapse. In fact, if there are any more risk flaring episodes and liquidity becomes even more valuable, these spreads are likely to blow out to even wider levels.

So back to our original topic. How could Merrill lose $15 billion on basis trades? And not just Merrill: Boaz Weinstein's group at Deustche Bank lost over $1 billion on this same trade, and basis trades are the main reason why Citadel has lost over 50% in 2008. Anecdotally, basis trades on CDOs are the reason why AIG, and most of the U.S. insurance industry is in its current deplorable state.

How would one go about estimating the P&L impact to these asset managers? It is not difficult: as the basis explosion resulted in a mismatch of DV01, or dollar equivalent change in 1 bps point in both bonds and CDS, or, netted out via the basis trade itself, one can calculate what the adverse MTM impact was on any notional position. If we take the CIT example above, and we assume that Merill had a $10 billion notional basis position in the name (this is an oversimplification but it was probably true for their overall basis portfolio), and the spread blew out from 0 to 1,500 bps around the time of the Lehman events, Merrill would have experienced a roughly $6 billion hit on the position (an average DV01 of $4MM), which implies that a $15 billion loss could have been created as simply as experiencing a blow up on $25 billion on basis trades. And this assumes no leverage which is naive for the prop desk model: if ML had leveraged its prexisting basis trades even 10x, the total basis trade notional needed to create this loss would have been only $2.5 billion. Is it inconceivable that ML had $25 billion in basis trades? Not at all - after all they were a preeminent CDS trading powerhouse and had one of the most active basis trade prop desks.

This is merely another amusing anecdote of what happens when you have a very popular trade in which everyone had piled in, from hedge funds to prop desks to insurance companies, and one of the assumptions that had been taken for granted disappears i.e., liquidity. The outcomes are only now starting to emerge: so far they have cost the jobs of John Thain (while we are amused by his office decoration choices, if he had not lost $15 billion in Q4, we are confident he would still have his job), "prodigy trader" Boaz Weinstein, and soon possibly Ken Griffin. It seems nobody ever learns from the Black Swan parable even though Nasim Taleb has been pounding the table on this for over 2 years now. Just as the Volkswagen short squeeze caused Adolf Merkle his life, and many hedge fund managers their jobs, every time you encounter this type of overhyped, "hedge fund hotel"-type trade, you will inevitably see casualties.

We at Zero Hedge would venture to surmise that the current bubbly purchasing of Treasuries and corporate loans will be the cause for the next 2 black swan events. We do not know in what form yet (by definition), but would caution all investors from getting involved at this point. Sphere: Related Content
Print this post


Anonymous said...

Looking at the spread, if they put the trade before Lehman and then took at loss on the 1500bp jump, then if they held up to the position they would have recoup some of the loss, as the basis tightened in jan 09. Did they unwind the trades at a loss then? not clear...

Anonymous said...


1. Bear Stearns not Lehman in the phrase "JPM was taking over Lehman" in the paragraph above the chart?

2. JPM not Morgan Stanley in the phrase "we assume that Morgan Stanley had a $10 billion notional" in the 3rd paragraph below the chart?

Love your site -wish I were smart enough to understand all of it.

PS Please feel free to delete this after you've read it.

Tyler Durden said...

Not only did it tighten, it overshot on the way down as well... however as in the case of volkswagen, most people were covering at €1000, not when it dropped back to 250. while there you had margin calls as an issue, being a prop desk for a bank you cud probably circumvent that concern, so it was just the psychological matter of covering before you incurred even more losses. The all time basis wides were just after lehman, then they retreated after hope that TARP v1 would be buying debt and CDS (check out's recommendation on how the government may take advantage of the basis trade), and then have blown out again as liquidity has been rearing its ugly head.

Tyler Durden said...

@2 thx for corrections, noted. our collective stream of consciousness needs editorializing sometimes

Timothy said...

In arriving at the 6Bln figure you leave out convexity which must be pretty large in a 1500Bp move (nitpicking I know) so notionals must have been larger.

On a more substantial note, if basis blew out (or is it in in this case) and massive stops were triggered after which it calmed down somewhat should we expect someone (GS maybe) to report huge windfall profits?

Tyler Durden said...

extrapolated convexity by just checking what a 1500 bps MTM change per CDSW is (should be there, check). you are right that the DV01 ranges from 6mm to 2mm the more convex you go.

and yes, someone will likely benefit from thain's downfall, although i hear GS prop was also axed wrong in basis

Anonymous said...

Can you explain why would someone looking to buy protection via a CDS would have to put collateral, I thought it would be the seller of protection. In AIG's case, it sold protection on billions, and didn't have enough cash to post on margin calls. Why would buyer need to post collateral - they lose protection in case they don't pay; whereas if a seller of protection cannot pay collateral, that insurance is worthless. Am I missing something??

Tyler Durden said...

Dont think of it as an out of the money margin call (which you are right that AIG is impacted by) but as the cost to the PB to lend free (or not so free) leverage to a risky counterparty.

Anonymous said...

Edit needed in the first para (as of 9PM EST 1/25): "the main reason for Morgan Stanley's massive $15 billion Q4 loss was due to some very large basis trades gone horribly wrong."

You mean Merrill of course.

Anonymous said...

Do you think that the European insurers could be casualty no.4 of the basis trade? I think that they have universally been hedging their corporate bond exposures by buying CDS in 3Q08 - 4Q08. I assume they are all intending to hold to maturity so not likely to be fatal, but could cause a lot of red ink when they report Q4?

Anonymous said...

Love your stuff but wanted to clarify one point regrading the basis. It is true that many 'naive' basis traders use the bond's z-spread as a comp for CDS spreads to judge the relative richness/cheapness and basis opportunity.
This, however, is a flawed analysis that can (and likely did for MER) get investors in big trouble. The z-spread is only comparable to CDS spreads when the bond is trading at Par (since CDS are a par-spread instrument).
As bonds have deteriorated in price, the z-spread has risen 'faster' than CDS spreads given the convexity. So a bond trading at say $85 with a z-spread of 150bps does not have a 50bps basis to a CDS of 100bps.
The trick to trading the basis is to 'judge' the real value f the bond based on CDS spreads and then work backwards to the implied spread (or carry differential).
Fruthemore, one needs to adjust the notionals of the two legs to account for recovery differentials...if the bond trades at $85 then if default happens and recovery is $40, the bond holder loses $45, but an equal notional CDS makes $60 in that case.
As with everything, the devil is in the details BUT you are dead right that the basis (defined by me as difference in valuation of credit in cash and synthetic markets) is indeed wide and offers great opportunities for man real money accounts.

Tyler Durden said...

agree but there is no other shorthand way to compare. the real question is why is every bulge bracket FI salesguy pitching their basis packages to the moon

Anonymous said...

I know - made me laugh too - but I think answer is simple...risk managers/execs have said 'reduce risk' - meaning make losses go away! Given my thoughts on the squeeze on bond inventory (due to the basis trades) these guys can only unload basis book thru selling 'basis packages' and giving up some of their upside for guaranteed outs...i.e. they can pitch the basis trade to unload their basis books and potentially take advantage of end-users misunderstanding of recovery risk as well as relative mis-pricing.
I tend to thnk that most of these FI guys (at least the ones that sell to me) are only running their book day-today with little or no big picture view on it...volume is still king for some of these old school guys...maybe they can lock in gains on CDS legs and 'put off' (read Level 3) losses on bond legs...

Anonymous said...

Dumb question:

I understand the attractiveness of the negative basis trade - but what is attractive about a positive basis trade? These positive basis trades are presumably the majority of executed trades? My (mis) understanding is that a positive basis trade would involve insuring a bond using a CDS so that the "insurance costs" are more than the bond coupon?

If "losing money safely " is the goal then why not a low yield treasury that lags inflation?

Obviously I am not a financial whiz.

Anonymous said...

@Anonymous with the not so dumb question...

The norm is positive basis in which case you can make money by selling the Bond and CDS. This positive basis is yours to keep (minus the cost of funding the Bond sell, e.g. repo)

In the case of the negative basis you buy the Bond and CDS.

james1071 said...

What constitutes a default and triggers the CDS counterparty to buy the bonds? Specifically, does an agreed reduction in coupons or other change in terms class as a default.

If it does not, then the negative basis trade is not a guaranteed profit.This would explain the negative basis.

If it does, then bondholders with CDS insurance have an incentive to agree to a reduction in terms.(when their bonds are below par and would be bought back for a capital gain)>

does anyone know the answer?

Anonymous said...

Good day !.
You may , perhaps very interested to know how one can reach 2000 per day of income .
There is no need to invest much at first. You may commense to get income with as small sum of money as 20-100 dollars.

AimTrust is what you need
The firm incorporates an offshore structure with advanced asset management technologies in production and delivery of pipes for oil and gas.

It is based in Panama with affiliates everywhere: In USA, Canada, Cyprus.
Do you want to become an affluent person?
That`s your chance That`s what you really need!

I feel good, I began to take up income with the help of this company,
and I invite you to do the same. It`s all about how to select a proper companion who uses your funds in a right way - that`s AimTrust!.
I earn US$2,000 per day, and my first investment was 500 dollars only!
It`s easy to start , just click this link
and lucky you`re! Let`s take this option together to get rid of nastiness of the life

Anonymous said...

finally, I could find this post again. You have few [url=]useful tips[/url] for my school project. This time, I won't forget to bookmark it. :)

Anonymous said...

You may probably be very interested to know how one can make real money on investments.
There is no initial capital needed.
You may commense to get income with a money that usually is spent
for daily food, that's 20-100 dollars.
I have been participating in one company's work for several years,
and I'm ready to share my secrets at my blog.

Please visit my pages and send me private message to get the info.

P.S. I make 1000-2000 per daily now. [url=]Online Investment Blog[/url]

Anonymous said...

I share a love-hate relationship with virtual memory because of how prices are always dropping. I absolutely hate buying SD Cards for my R4 / R4i at (seemingly) a cheap price only to see it become 10% cheaper a few months later.

(Posted on Nintendo DS running [url=]R4i[/url] BB)