Eating your own Coco February 11, 2014 at 11:12 pm

Dealbreaker catches a deeply amusing story from the WSJ:

banks including the U.K.’s Barclays PLC, Germany’s Deutsche Bank AG and Switzerland’s UBS AG could shore up their U.S. subsidiaries… the U.S. units would issue to their parents a type of bond that converts into equity if the U.S. business’s capital falls below a certain level… The European parent companies would finance the purchases of their subsidiaries’ debt by issuing bonds to investors, these people say.

Coco double leverage: very cool. But not very chocolatey, sadly. You can call it an internal convertible if you must. I’m gonna call it something else…

Tying Loan Rates to Borrowers’ CDS Spreads February 10, 2014 at 8:23 pm

João Santos blogs about an interesting paper on Liberty Street Economics.

The development of the CDS market has provided banks both with a new way to manage their credit risk and with observable information on borrowers’ default risk. While banks do not seem to use the CDS market extensively to lay off credit risk, it appears that they are relying on information from the CDS market in their lending business. We find that since early 2008 banks have increasingly extended loans to corporations with interest rate spreads tied to the borrower’s CDS spread (or to a CDX index) over the life of the loan, a practice referred to as market-based pricing… We find that the difference between the actual and the hypothetical fixed-rate spread over the life of loan is always negative for loans priced off the CDS market, confirming that market-based pricing has lowered the cost of bank credit.

We know that CDS spreads dramatically over-state the probability of default, but clearly old fashioned bank lending spreads over-stated it even more. I do worry that this approach risks pushing the borrower into default as their spreads blows out, too – whether or not that blow out was caused by a change in fundamentals.

Nothing to see here, move right along February 6, 2014 at 9:40 pm

Dear me, can’t a billionaire write a bunch of long-dated exotic derivatives without those pesky kids sticking their noses in? Warren is getting a rough ride from Dan McCrum repeatedly, and today from Matt Levine, and he doesn’t quite deserve it. Here’s why.

  • First, level three assets are not necessarily toxic. They can just be hard to value precisely. Even if Warren had just sold 20 year plain vanilla equity index puts, they would be level three, as there is no ready market in twenty year implied vol (although you can make a pretty decent guess from where the ten year is).
  • Warren is naked short. Black Scholes and related approaches are valuation techniques which work if you are hedging (and you can indeed replicate the derivative by following your model’s hedge ratios). There is actually quite a good theoretical argument (if not an accounting standards one) for him not to mark to market – an argument that would be more convincing if he has written an insurance policy that is then transformed into a derivative via an SPV. We don’t know that he hasn’t done this. But we do know for sure that Warren’s strategy is to write insurance and invest the premiums. As long as he collects enough premium and his risks are diversified, he’s happy: for him, at least at 50,000 feet, the business model is all about collecting premiums and investing them. Writing long-dated puts is a good way to raise cash – as long as you don’t have to post collateral (which Warren didn’t).
  • Even if he has written a worst-of put, this was not a particularly exotic derivative in 2006-7. Back then people were playing with a whole range of basket options (see for instance the mountain range trades originated by Soc Gen’s traders and rapidly taken up by the rest of the street). While these trade types might not be in options 101, they haven’t been cutting edge for twenty years.
  • One of the hard things about running an equity derivatives book is getting enough long-dated vega. No one wants to sell it; everyone wants to buy it.
  • So the reason there was a trade is the age-old two people wanting different things. Warren wanted cash, and saw the premium as good compensation for the insurance he was writing. His counterparties saw cheap vega that was hard to buy any other way, and a good credit. Two well informed parties with different takes on the world trading with each other is not, I am afraid, a scandal.

Silver linings February 2, 2014 at 9:12 am

2014 Cocoa

I mean not just to clouds, but to Kit Kats too. In January the cocoa price went up precipitously, driven partly by increased demand from emerging markets, as the Nymex cocoa future price plot shows. This is profound implications for the cost of what’s inside those foil wrappers.

A global emerging market crisis could of course plunge the financial system back into crisis and cause untold harm – as it did with Russia/LTCM in recent memory – but it will at least reduce the pressure on cocoa prices. Instead of messing around with those troublesome credit indices, perhaps JPMorgan should have hedged their loan book with rolling cocoa puts…

What’s an “equity like” product in the credit market? February 1, 2014 at 3:15 pm

FT alphaville quote this chart from BAML in an article about (amongst other things) Cocos:

Frozen rain

It strikes me that leveraged loans are not much like Cocos. Think of it this way. Start with a senior bond from a good quality issuer. You can make this product more ‘equity like’ in a number of ways:

  • Make the issuer dodgier, so the credit spread goes out to reflect the increased PD;
  • Make the bond more subordinated, so the loss given default goes up;
  • Make the coupons deferrable/PIK’able, creating a risk of loss without a credit event;
  • Make the bond mandatory convertible or bail-in’able so that you can get something other than par back, again without it being a credit event.

Doubtless there are others. The point though is that the first direction – drifting down the credit spectrum towards leveraged loans – slowly moves the loss distribution; whereas the last two add an unlikely but catastrophic event. Cocos or bail-in bonds from good quality banks are unlikely to suffer losses, but when they do, they are bad. This is a different kind of ‘more like equity’ than declining credit quality.

Sausage makers like sausages January 31, 2014 at 9:53 am

An important catch from Tracy Alloway at FT alphaville, this: Ing-Haw Cheng and colleagues looked at how the personal portfolios of mid-level staff involved in the securitisation industry in 2006 did, and found that they were even worse than the ordinary’s Joes’. Why? A job environment that fosters “groupthink, cognitive dissonance, or other sources of over-optimism” perhaps?

From the abstract:

We find that the average person in our sample [of mid-level insiders] neither timed the market nor were cautious in their home transactions, and did not exhibit awareness of problems in overall housing markets. Certain groups of securitization agents were particularly aggressive in increasing their exposure to housing during this period, suggesting the need to expand the incentives-based view of the crisis to incorporate a role for beliefs.

Martin Taylor is seriously out of date January 30, 2014 at 7:38 pm

He says:

In the 1990s IT did not enjoy especially high status within banks. Real men, then as now, wanted to meet property company bosses at the Savoy Grill and lend them huge amounts of money that were unlikely to be repaid.

This is of course wrong. Real bankers now want to meet property company bosses at Petrus and lend them huge amounts of money that is unlikely to be repaid.

Risk management quote of the week January 27, 2014 at 5:30 am

Well, of six weeks ago actually, but it is still good:

Safety is a product, not a process.

It is being said in the industrial accident context. I’ll let the Ranter explain:

In general, effective safety measures are usually something you do, and scattering costly “devices” around an unchanged process is a classic failure mode. Not least because they might instil a false sense of safety and lead people to take risks…

Accidents cost money, in the same way that quality failures cost money. At the very least, in the most cynical 19th century Yorkshire mill-owner’s view, they cause downtime, quality problems, and damage to expensive equipment. In a less cynical and more general sense, accidents are just one of the sources of excessive variability in the production process, like late change requests, tools whose tolerances are too large, or a virus outbreak among the Windows boxen. If accidents are happening, this is a symptom of problems with the process.

Reworking production processes to eliminate the sources of variability is precisely what industrial managers are meant to do all day.

In other words, risk management is not an overlay, it is what you should be doing all the time.

When the rain freezes on the trees… January 26, 2014 at 7:17 pm

… it’s kinda magical.

Frozen rain

(Click to enlarge.)

One chart to explain it all January 25, 2014 at 7:18 am

Well, perhaps not quite that much, but still, this, from Business Insider is insightful:

Global Income Growth

Pricing far out of the money derivatives in the P measure January 24, 2014 at 12:03 pm

OK, not the most attractive title in the world I know but bear with me.

Most derivatives are priced in the Q, or risk neutral measure. This is the right thing to do when you are willing and able to hedge. Essentially in the Q measure the cost of a derivative is identified with the price, according to your theory, of hedging it.

You can’t do that sometimes, often because the necessary hedge instruments are not available. This is one practical distinction between derivatives and insurance: derivatives are hedgeable, insurance isn’t. Therefore you shouldn’t price insurance in the Q measure: instead you care about the real world (as opposed to the risk neutral) distribution of outcomes.

This is important when we think about things like Warren Buffett’s basketball trade. Buffett has written insurance (not a derivative*) to protect Quicken Loans against the risk that someone will correctly predict the winner of every game in the National Collegiate Athletic Association’s men’s basketball tournament – something Quicken has offered a billion dollars for.

This is actually an interesting thing to price. If you view the tournament as IID coin tosses, then of course the insurance is worth nothing. But of course it isn’t, partly because some teams are better than others, and partly because there will be entrants to the competition with private information. It isn’t much of an edge to know that a star player is off form, but excluding a favourite team skews the distribution for very unlikely outcomes, especially if you let insiders collude by assuming, say, that someone knows the outcome of every game involving, say, ten specific teams. Moreover, you don’t just need enough premium for this insurance to pay for the expected loss; you also need enough to pay for your cost of capital supporting unexpected loss, which will be a lot as the distribution is very fat tailed. Add in the cost of actually doing all this analysis, and pretty quickly you get to a multi-million dollar premium. Indeed, something I call (after a long-retired trader) Stavros’ law probably applies: never sell any put for less than a cent of premium, no matter what the model says it is worth.

*Don’t get me started, though, on the hypocrisy of Buffett’s public statements about derivatives vs. what Berkshire actually does.

The FED’s collateral policies during the crisis January 21, 2014 at 7:43 pm

I can’t believe that I missed this two years ago when it came out, but I did, so herewith find Bloomberg’s carefully-researched and probably not at all conjectural expose of the FED’s collateral policies in the crisis:

Last week the Federal Reserve bravely released 894 PDF files containing 29,346 pages that detailed its heroic actions during the financial crisis.

These documents revealed how open-minded the Fed can be when it needs to be. Local governments in Belgium, Japanese fishing cooperatives, the Libyan government and many other unlikely parties received the Fed’s financial aid. Failing U.S. banks, such as Citigroup and Morgan Stanley (MS), were of course handed whatever they wanted, and permitted to post as collateral pretty much anything they could get their hands on: junk bonds, defaulted debt, volatile equities…

A team of Bloomberg investigative reporters, led by Kram Namttip, was allowed to spend a day examining what remains of the collateral collected by the Fed during the crisis. What follows is a brief summary of their findings. To wit:

— A vault in the Fed basement filled with young women, who claimed, in broken but excited English, they had been repo-ed by the Italian government.

If Italy has weathered Europe’s sovereign debt crisis so much better than its fellow deadbeats, here is why: the Fed’s nervy decision to extend credit to the Italian government against its prime minister’s social assets…

Should banks reserve through-the-cycle? January 20, 2014 at 9:56 am

The elephant in the accounts is often loan loss reserves, those oh-so-easy-to-manipulate, oh-so-big (if not actually big-eared) amounts that often drive bank earnings. If you though derivatives valuation was dodgy, welcome to the loan book. The most recent, if not the most egregious examples are surveyed in a recent Bloomberg post:

More than 31 percent of JPMorgan’s 2013 earnings, or $5.6 billion, and about 10 percent of Wells Fargo’s, $2.2 billion, weren’t really earned last year. That money came instead from the banks’ so-called loan-loss reserves… [Bank of America] has received the biggest boost from releasing reserves: The move helped it turn $11.8 billion in losses since 2010 into $11.4 billion in profit. Citigroup, which reported $40.4 billion in net income over that time, would have booked about half that amount without the accounting benefit.

This cuts the other way, too. Bank of America would have reported almost $55 billion in profit in 2009 if it weren’t for the $48.6 billion it put back into reserves that year.

This happens of course because loan loss reserves are annual estimates, and things change from year to year. There are proposals to move to provisions which would reflect losses expected over the life of the loan. This ‘through the cycle’ approach might be more stable, and would certainly result in higher levels of provisions, but they won’t hit any time soon*. Perhaps supervisors should give up on the accounting standards setters and set robust standards for regulatory through the cycle EL provisions?

*The current state of play from FASB is : “The Board discussed the next steps on the credit impairment project and decided to continue to refine the Current Expected Credit Loss (CECL) model in the proposed Accounting Standards Update, Financial Instruments—Credit Losses (Subtopic 825-15)… The Boards will continue redeliberations on the CECL model, considering feedback received through comment letters and outreach activities on Exposure Drafts issued.” Dynamic, huh?

The heart of winter January 19, 2014 at 5:27 pm

Winter 2013

Bubble me do January 17, 2014 at 11:17 am

‘Bubbles’ are much in debate – by Gavyn Davies here and here, Daoud & Diaz here, Fama here, Stein here, and so on.

A key issue, obviously, is what is a bubble. The Brunnermaier definition is

Bubbles are typically associated with dramatic asset price increases followed by a collapse. Bubbles arise if the price exceeds the asset’s fundamental value. This can occur if investors hold the asset because they believe that they can sell it at a higher price to some other investor even though the asset’s price exceeds its fundamental value.

I think this is a terrible definition. There are two big things wrong with it:

  • It implies that an asset price rise is only a bubble if it is followed by a collapse, in other words that, by definition, slow bubble deflation is impossible.
  • It assumes that there is a single ‘fundamental’ price which we can measure deviations from. In reality of course ‘fundamentals’ are just as socially constructed as market prices, and just as arbitrary. (I have used the catchphrase ‘prices are a Schelling points’ in the past.)

We can’t reasonably hope to address the financial stability implications of bubbles until we have a better definition of what a bubble is.

So what exactly is the rate you are borrowing at to fund that derivative? January 16, 2014 at 10:14 pm

As everyone who has been paying attention knows, JPM had a $1.5B FVA hit in their most recent results. Matt Levine riffs amusingly if sometimes a little inaccurately* about a couple of aspects of this, my favourite part being:

there is… some gap between “my funding cost” and “FVA.” It’s unclear to me how much of JPMorgan’s model is based on their own funding costs and how much is based on some “market” funding cost; the earnings deck talks about “market funding rates” and “the existence of funding costs in market clearing levels,” so it seems that they’re thinking more about a market price of funding than they are about their own cost of funding.

Oh one fun fact about that. That earnings deck says that FVA “represents a spread over Libor”; based on [JPM CFO] Marianne Lake’s comments you can guess that that spread is around 50 basis points. That is, banks fund at around Libor plus 50 basis points.

Libor, you’ll recall, is supposed to be the rate at which banks can fund themselves.

I will resist the temptation to add a smilie.

*Hint: when a lawyer rights about how exactly Black Scholes works, you might want to apply a pinch of salt. Or read a careful account of the story, for instance here or here (where the key role of the replicating portfolio is explained – although I buy the Albanese ‘not fungible with debt’ argument).

Early candidate for the best politics article of 2014 January 13, 2014 at 6:38 am

It’s this piece by Ezra Klein on the role of motivated reasoning in political posturing (I was going to call it ‘debate’, but all so often, as Klein points, there is no desire to exchange views).

More fun with less colour January 12, 2014 at 11:16 am

After seeing one output from Silver Efex Pro recently, I have been playing around with it some more. This ice image is a perhaps a little too contrasty:

Winter 2013

But it is to my eyes a lot better than the B&W you get out of Photoshop’s native converter:

Winter 2013

(Click for bigger.) Clearly in niche areas like a colour to B&W converter there is room for a product that does better than Photoshop.

It’s the liabilities, stupid January 11, 2014 at 11:15 am

Jeremy Stein makes three good points:

Banks are almost always and everywhere largely deposit financed…

The asset side of banks’ balance sheets – and, in particular, their mix of loans versus securities – is considerably more heterogeneous… One interpretation of this is as follows: While lending is obviously very important for a majority of banks, it need not be the case that a bank’s scale is pinned down by the nature of its lending opportunities. Rather, at least in some cases, it seems that a bank’s size is determined by its deposit franchise, and that, taking these deposits as given, its problem then becomes one of how best to invest them.

Within the category of [banks’ investments in] securities, they appear to have well-defined preferences… it looks as if banks are purposefully taking on some mix of duration, credit, and prepayment exposure in order to earn a spread relative to Treasury bills.

With these liability-centric spectacles on, banking is a business of (1) attracting deposits and (2) figuring out something to do with them that reliably earns a spread to their cost. Credit extension is a consequence of this activity, not a core part of banks’ business model.

Cost Benefit Analysis is so bad we need more of it January 9, 2014 at 5:35 pm

John C. Coates says:

quantified CBA on those rules [in the Dodd Frank Act and elsewhere in financial regulation] amounts to no more than “guesstimation,” entailing (a) causal inferences that are unreliable under standard regulatory conditions; (b) use of problematic data, and/or (c) the same kind of contestable, assumptionsensitive macroeconomic and/or political modeling used to make monetary policy.

This is entirely fair. Coates however makes too much of this, claiming that

While CBA… is a useful conceptual framework, and quantified CBA a worthy long-term research goal, it is not capable of disciplining regulatory analysis [of financial rules] in its current state.

Let me explain why. Any cost benefit analysis of significant financial regulation is plagued with difficulty: not only is it hard to know what both sides are, there is also the significant difficulty that the industry will change in likely unpredictable ways to any regulation. With the best will in the world, it is impossible to quantify either costs or benefits accurately.

However, agencies and others should still conduct CBA. There are a number of reasons for this. First, building a CBA models focusses the mind and brings up hitherto unforeseen issues. Even if you throw away the output, it is still worth doing.

Second, a good CBA model will not produce simple cost and benefit analysis, but ranges of estimates, probability fan charts, or similar. This encourages the modeller to be more honest about model risk and often highlights the fact that, for many policy proposals, they are not certain to work. That doesn’t mean that you shouldn’t enact them, but rather that you should be alert to the possibility of failure, and ready to change policy if the evolving evidence indicates that you should.

Finally, both sides of an argument can and should do CBA. This would focus the discussion on those aspects of the problem that the two sides disagree on, and hence provide a hopefully small set of questions whose answers determine the efficacy of the proposal, or which suggest modifications to it. Wouldn’t it be great if instead of the industry going up in front of a judge and saying ‘the agency didn’t do a CBA, so you should stop them enacting the rule’, both sides instead went up with their own detailed models, and were forced to justify their differing appraisals of a policy proposal?