Felix Salmon

McNamara and model risk

Felix Salmon
Jul 7, 2009 19:47 UTC

Philip Delves Broughton notes that Robert McNamara, one of Harvard Business School’s most notorious graduates, basically did in the field of war what Wall Street quants did in the field of finance:

The journalist David Halberstam wrote that McNamara mistrusted people who did not speak his language of statistics and hard data. If it ever came down to one person saying something “just didn’t feel right” or that it “smelled wrong”, he would always go with his facts over their feeling. Fatally, in the case of Vietnam, the data he received was not accurate.

When Wall Street quants fail to account for model risk, they can end up losing hundreds of billions of dollars. But that’s an improvement over what happened when McNamara failed to account for model risk: those losses were much worse.


Based on my reading of Halberstam’s account of the War,
McNamara had reservations about America’s progress in Vietnam operations even in the Kennedy years (when the war was still at an incipient stage).

His shortcoming was not a tendency to implicitly believe modeled forecasts. Rather, it was the desire to please the boss (Johnson) with pleasant tidings.

When journalism misses the big picture

Felix Salmon
Jul 7, 2009 19:21 UTC

Robert Teitelman thinks that since he’s in charge of a publication aimed at financial-market professionals, there’s no need to spend much effort on making it easy to read:

There’s an entire world of B2Bs like The Deal and Dealscape that, in fact, are targeted at practitioners. The difficulty of the B2B game is not necessarily to write more accessibly as it is to report and write with greater sophistication and depth.

There are two problems here, as I see it. Firstly, there’s no reason that accessible journalism can’t be sophisticated and deep. It’s not necessarily easy to write accessibly about complex and sophisticated ideas, but yes, it can be done. The main problem is that it takes much more time and effort: the amount of work I put into my Wired story on the Gaussian copula function, for instance, was a good order of magnitude greater than the work that I would put into writing at that length on the blog. Maybe straitened journalistic enterprises don’t have the resources to make their stuff accessible.

But secondly I think that financial journalists are deluded if they think financial-market professionals are willing and able to wade through pages and pages of dry, jargon-heavy prose. The financial professionals I know tend to have short attention spans and have no particular eagerness to read the trades — especially any story in which they’re not quoted. Just because you’re writing for a business audience doesn’t mean your writing shouldn’t be lively and accessible. And, ideally, short.

Teitelman adds, apropos my call for more accessible financial blogging,

Salmon and his commenter skip past the hard question here, however: Can the complexity of finance (and economics) be effectively captured by the kind of simple explanations required by an audience that barely knows the basics? Let’s put it another way: In telling that “simple” story, is the journalist distorting the situation, highlighting certain aspects, accentuating certain tendencies and ignoring others?

It’s true that the mass audience does tend to be attracted by simple explanations; I got a worryingly large number of emails after my Wired piece came out essentially saying “thanks, you’ve now explained everything”. Which of course one article about one formula could never do. But I never asked for journalists to oversimplify, and there’s no reason that accessible journalism can’t show many sides to any given story. What’s more, trade journalism is also guilty of many of the sins which Teitelman enumerates.

It’s not just journalists, of course, who will highlight certain things and ignore others. Often, the journalists do that just because they rely, of necessity, on their industry sources — and their sources are doing it too. One of the problems with trade journalism is that a lot of day-to-day reporting is done via the banks’ PR departments, and the PR departments tend only to serve up managing directors and above for interviews. And when you talk to high-level people, you’re often talking to people who are genuinely ignorant. Think of AIG Financial Products, as described by Michael Lewis:

It’s hard to know what Joe Cassano thought and when he thought it, but the traders inside A.I.G. F.P. are certain that neither Cassano nor the four or five people overseen directly by him, who worked in the unit that made the trades, realized how completely these piles of consumer loans had become, almost exclusively, composed of subprime mortgages.

Or think about Bob Rubin, who famously told Carol Loomis that he’d never heard of the notorious liquidity puts which ended up all but destroying Citigroup until after it was far too late — despite the fact that Rubin, more than any other individual, was meant to be the person taking the big-picture view of the bank’s overall risk profile.

Looking back at the history of journalism over the course of the financial crisis, the problem was never too much oversimplification as it was too many journalists taking a narrow view of the market: they didn’t think nearly enough about — or push bankers to answer tough questions about — big-picture systemic risks. It’s a hugely important role of journalism to put events in large-scale perspective. Those stories should be written more often, and they should be written as accessibly as possible, by journalists and bloggers both.


“But secondly I think that financial journalists are deluded if they think financial-market professionals are willing and able to wade through pages and pages of dry, jargon-heavy prose. The financial professionals I know tend to have short attention spans and have no particular eagerness to read the trades — especially any story in which they’re not quoted. Just because you’re writing for a business audience doesn’t mean your writing shouldn’t be lively and accessible. And, ideally, short.”

This is definitely true, but at the same time, financial professionals don’t want to wade through explanations of what Libor or a credit default swap is every time they read an article. You’ve got to strike a balance.

It’s also worth noting that there are different kinds of professional/trade reader. An article aimed at the MD types you describe (or senior management in general) should take a different approach to one aimed at people working at the coalface, so to speak.

Posted by Ginger Yellow | Report as abusive

How did the automakers emerge from bankruptcy so quickly?

Felix Salmon
Jul 7, 2009 17:59 UTC

Micheline Maynard has a good 1,000-word article today on the surprising fact that both GM and Chrysler managed to exit bankruptcy in record time. But who or what should get the credit? Steve Rattner? The two judges involved? Section 363 of the federal bankruptcy code? And is this a heartening precedent for the wave of future bankruptcies which seems inevitable when all those leveraged loans mature over the next four or five years? Or is it a one-off, linked to extreme levels of government involvement, which is unlikely to be repeated?


i would give the credit to obama, without his help i think chrysler would be history. good job obama.


Felix Salmon
Jul 7, 2009 16:24 UTC

Paul Collier is worried about the skewed incentives built in to any bonus system: the upside of taking risk — a big bonus — is much bigger than the downside if the risk blows up:

The inherent problem facing shareholders is that incentive payments cannot go negative. However much damage a manager inflicts, wiping out both shareholders and depositors, the consequences cannot be remotely commensurate.

Collier has a solution: a new crime, called bankslaughter.

With bankslaughter, when the bank blows up – even if it is a decade later – a criminal investigation traces back to determine whether crucial decisions were reckless. If a reasonable banker faced with the information available at the time would not have taken those risks, the person responsible is dragged off the golf course and jailed.

Once bankslaughter was on the books, bonuses would be less dangerous. Managers would have to weigh the balance between risk and return and take defensible decisions. I doubt hyper-caution would be a problem: the overly cautious would not get bonuses. Surely we can rely on our bankers to exhibit the necessary degree of greed.

Is it reasonable to hold professionals criminally liable if they take reckless risks with other people’s money? I don’t see why not. Especially if they work at a leveraged and systemically-important institution. After all, people can be jailed for insider trading, which is far more of a victimless crime than bankslaughter.

Update: Jeff had more on this — including crediting the name to Timothy Garton Ash — last week.


Better to first unwind the fiduciary duties. Boards of Directors comp committees approve ginormous bonuses based on putting client money way far at risk, in order to benefit bank shareholders.

Choose up guys — who is it, the shareholder or the customer?

You can’t shoot the customer so the shareholders can collect the insurance. Bankslaughter.

Sheesh — this analogy is too good. Look at Turquoise dark pool — all the IBs get together and commit customer genocide. Worse, they arrange so one customer causes the “murder” of another. This needs to go to the financial equivalent of the World Court for criminal prosecution of genocidal maniacs.

Posted by Annie | Report as abusive

How to reform overdraft fees

Felix Salmon
Jul 7, 2009 15:55 UTC

A couple of very interesting comments have appeared on my blog entry on regulating bank fees in general, and overdraft fees in particular. J Mann asks what exactly I’m proposing, and what I think the consequences might be:

Are you thinking about (1) requiring banks to allow customers to opt-out of overdraft protection (or maybe requiring opt-in);

(2) setting maximum overdraft fees, but permitting banks to decline to provide ovedraft financing altogether; or

(3) requiring banks to provide overdraft financing to all checking customers and setting maximum rates for that financing?

It seems to me that the likely consequences would be some combination of

(a) banks declining to offer overdraft financing, which would leave people paying bounced check fees to their payees;

(b) banks reinstuting minimum balances for checking accounts; and/or

(c) banks removing interest and other benefits from checking accounts.

Those consequences might be worth it, but I’m curious which reform you think would get us the maximum benefit/cost ratio.

My proposal would be that banks be given a choice: they can offer automatic overdraft protection, but only if it’s free. (They can charge an annualized interest rate on the overdraft, but no set fees.) If they want to charge fees as well as an interest rate for overdraft protection, then that protection would have to be opt-in rather than opt-out, and the fees should be prominently disclosed at the opt-in stage. And yes, fees would be capped: I would say a $20 cap was reasonable, with a limit of one such fee per day.

What would the consequences be? Yes, for sure there would be more bounced-check fees. (Which should also be capped.) But checks can and should be increasingly rare things. We’re moving into a world where debit cards are replacing checks, and transactions simply don’t go through if funds aren’t available: there’s no such thing as a bounced-debit-transaction fee. Similarly ATM withdrawals would simply be declined, rather than triggering overdraft fees.

Would banks reinstitute minimum balances or otherwise stop banking the kind of poorer customers who currently generate lots of fees but who otherwise aren’t very profitable for the banks? That’s a danger, yes. We don’t want these regulations to result in a large increase in the unbanked. Maybe banks should be required by law to offer simple no-frills checking accounts for customers who can’t meet minimum-balance requirements and don’t want to pay monthly checking-account fees.

As for interest-bearing checking accounts, those beasts are rare enough to begin with that I doubt many people would notice their passing altogether. So yes, there will be costs, but I’m pretty sure the benefits would be much greater.

J Mann’s comment was followed up by one from Argel, who got very excited about the fact that a previous commenter had revealed his account number at Citibank. Is this a particularly dangerous thing to do? In Europe, people give out their account number all the time — if I want to pay you some money, I just wire the money into your account, which is free, but does require my having your account number. In the US, by contrast, people are very protective indeed of their bank account numbers. Is that for good reason? Or is it something which will just change slowly if and when we move from checks to electronic transfers?


Banks have used technology to make it very easy for consumers to track their accounts and make smart decisions about their own spending. They have online statements 24 hours a day with mobile options on your phone for many. You can also call anytime and go to the ATM to get your current balance. They also offer a line of credit to just about anyone that would help you avoid all of this but no one ever signs up for it leaving you vulnerable to overdraft fees.

The problem is we have people with $150 to their name and they expect the bank to take all the risk of covering their short term financial needs without any risk to them. Do you expect the grocery store or dry cleaners to just let you do business with them if you can’t afford it or you forgot to make a deposit…NO..they would all go out of business and banks are no different

These people are lucky their not eating lunch with their family and then getting the charge on their debit denied instead of the bank paying for it and saving them the embarrassment. I don’t think they let people wash dishes anymore.

Here’s my solution. If you can’t keep your balance above $200 you have no need for a bank account. No bank account means NO FEES!! Do yourself a favor and don’t spend what you don’t have!!

Posted by DB | Report as abusive

Why Mark Thoma doesn’t accept advertising

Felix Salmon
Jul 7, 2009 14:42 UTC

David Warsh has a good profile of Mark Thoma, who doesn’t get paid for blogging:

“I lose money on the blog,’ says Thoma. “The state pays me to do this, to be an economist. It would be wrong to take money for it. And if you take advertising, it just feels as though you’re captured.”

I don’t think it’s wrong to take money from outside sources for being an economist, just because the state pays you to be an economist. Is Mark implying that Tyler Cowen is wrong to run ads on his blog? That Mankiw is wrong to make millions of dollars writing textbooks?

What Warsh and Thoma don’t mention here is the legal reason that Thoma doesn’t run ads. Thoma tends to quote other people’s writings at substantial length, often with little or none of his own commentary attached. The vast majority of people picked up by Thoma are very happy about it, but inevitably there are going to be a few who get antsy about copyright. And if Thoma doesn’t make any money from his blog, it becomes virtually impossible for anybody to claim damages.

Some bloggers are much more cavalier when it comes to fair use than others; Thoma and Yves Smith spring to mind as bloggers who tend to quote at great length. But Smith’s blog has lots of ads on it, which means that she’s much more likely to find herself the recipient of takedown notices, C&Ds, and other nastygrams. Thoma, I think, has a much easier life, with much less tail risk, by making the decision to accept no advertising at all. And as an added benefit he has one less thing to worry about on the blog.


She has apparead on TV a few times so you can google for her. “Yves” is pronounced like Eve (i.e. “eev”).

Posted by Argel | Report as abusive

Regulatory arbitrage attempt of the day

Felix Salmon
Jul 7, 2009 13:51 UTC

Patrick Jenkins gives a good example of why insurance is not a sensible way to think about or regulate financial products:

Investment banks, including Goldman Sachs and Barclays Capital, are inventing schemes to reduce the capital cost of risky assets on banks’ balance sheets…

The schemes, which Goldman insiders refer to as “insurance” and BarCap calls “smart securitisation”, use different mechanisms to achieve the same goal: cutting capital costs by up to half in some cases…

Under Goldman’s idea, it would sell an insurance product to a bank with a toxic portfolio, effectively shifting the risk of the underlying assets off the balance sheet. The insurance would require far less capital to be carried against it than the original assets.

The insight here (I believe Goldman considers it “financial innovation”) is that insurance is fundamentally more leveraged than finance. Rolfe Winkler is wrong when he says that regulating financial products as insurance would force banks to reduce leverage — quite the opposite.

Think about a pair of banks, A and B. Each has $1 billion in loans on its books, and needs $80 million in capital to be held against those loans. But then B insures A against any losses on its loan book, while A insures B against any losses on its loan book. Presto, each bank is now fully insured against loss, and needs much less capital. Each bank also, of course, has a large contingent liability should the other bank’s loans go bad. But the amount of capital that an insurer needs to hold against such contingent losses is much smaller than the amount of capital that a bank needs to hold against its own loans.

The fact that it’s Goldman coming up with this bright idea is particularly ironic since it was Goldman which revealed the way in which banks could use securitization to reduce their capital requirements. The bank even proposed a very sensible new principle:

Securitized loans should, in aggregate, face the same capital requirements as the underlying loans would if they were held on bank balance sheets.

I wonder whether Goldman’s financial innovators got the memo.


Horse donkey and Anglo-Saxon hybrid anvil? Family now occupied 贠 60% of the earth? Rich, and has been a world with guns to bark? Others to plunder the wealth. Then use to dig? Take the money to support those who come for the next rob his wealth provides more 䠼 attaching machine? The clown, this is why the dalai lama and hot 栯? Maria is the source of the market.
This past or being plundered all the people should be awake, such as Germany, boycott, these hybrids

Posted by youpijiufang | Report as abusive

Commercial real estate datapoint of the day

Felix Salmon
Jul 7, 2009 13:28 UTC

Worldwide Plaza is being sold after all (a previous deal fell through), and at what looks like a seriously knock-down price:

Deutsche Bank AG has agreed to sell Worldwide Plaza, a 1.8 million square-foot skyscraper in New York City, for $600 million to developer George Comfort & Sons and partner RCG Longview…

The sale price works out to roughly $330 a square foot.

Worldwide Plaza is a very high-class office building, home to, among other tenants, the swanky offices of Cravath, Swaine & Moore. It also has what until recently would have been something extremely attractive: a huge amount of unleased space (709,000 square feet, to be exact), vacated by the departing Ogilvy & Mather.

A year or two ago, long-term leases were poison for commercial real-estate valuations, since they reduced landlords’ ability to hike rents. Vacant space, by contrast, was like gold dust: prime midtown office space was leasing at well over $100 a square foot.

Today, everything has been turned on its head: those 709,000 square feet aren’t generating any income, and therefore have very little value. As a result, the 1.8 million square feet of Worldwide Plaza are worth just $600 million: by contrast, the $1.5 million square feet of 666 Fifth Avenue sold for $1.8 billion — or $1,200 per square foot — in 2006. On a price-per-square-foot basis, that’s a decline of more than 70% from the peak of the market.


That\’s right, I forgot the retail …actually Carlyle bought it last year for an even crazier amount:

http://www.reuters.com/article/pressRele ase/idUS125811+02-Jul-2008+BW20080702

Posted by peter | Report as abusive

Why insurance commissioners should not regulate CDS

Felix Salmon
Jul 4, 2009 23:38 UTC

There’s a meme doing the rounds — I fear it may have been caught by my colleague Rolfe Winkler — that credit default swaps are insurance products, and that therefore they should be regulated by insurance regulators. So before this nonsense spreads any further, it’s worth explaining just why that’s a very bad idea.

First, credit default swaps are not insurance, they’re swaps. A lot of journalists talk about them being “like an insurance contract” when they try to explain what they are, and that’s true, as far as it goes — they do share certain characteristics with insurance. But that doesn’t mean they are insurance. It doesn’t mean that some foolish law should be passed forcing buyers of protection to have an “insurable interest” in some underlying debt instrument, and it certainly doesn’t mean that all CDS should be regulated by some insurance commissioner somewhere.

Many swaps can be thought of as being like an insurance contract, should one be so inclined. For instance, when Larry Summers entered into a massive interest-rate swap while president of Harvard, he essentially locked in a fixed rate on the university’s debt: he was insuring Harvard against the risk that rates would rise. But insurance contracts can’t blow up in the way that the Summers swap did: when rates fell, the university ended up losing a cool $1 billion.

Now let’s say that Harvard had bought credit protection on the state of Massachusetts instead — perhaps Summers, worried about the state running out of money, feared that in an attempt to make up a budget gap, it would start (sensibly) taxing the university’s endowment. Just like with the interest-rate swap, if the spreads on Massachusetts’ CDS had tightened in, then the CDS contract would have cost Harvard a lot of money — protection buyers are just as subject to margin calls as protection sellers are (but as buyers of insurance products are not). It’s the same reason why airlines take large one-off gains and losses from their jet-fuel hedges, even if the purpose of putting on those hedges is to smooth out their fuel expenses.

It’s also worth noting that Harvard might well not be considered to have an “insurable interest”, if it didn’t hold any Massachusetts bonds — even though buying credit protection on Massachusetts might make sense. Defaults have repercussions far beyond the narrow circle of bondholders, and there are many people who might want to hedge against a certain entity defaulting, even if they don’t directly hold that entity’s debt.

A CDS is simply a deal whereby two counterparties promise to pay each other a certain income stream. The buyer of protection commits to paying a fixed amount of money every six months; the seller of protection commits to paying an uncertain amount of money (to be determined via an auction mechanism) should certain events happen in the future. Yes, insurance policies work in a similar manner. But so do, say, office lottery pools — a bunch of co-workers all commit to paying a small amount every week on the understanding that if something improbable happens in the future, they will share a large but unknown amount of money.

Even if some bright spark determines that CDS are insurance contracts, however, that doesn’t mean that they should properly be regulated by insurance commissioners. After all, even if they are insurance contracts they’re also financial derivatives, and it’s pretty clear that financial-derivative regulators are much more likely to be able to effectively regulate financial derivatives than insurance regulators are. What’s more, the SEC looks as though it’s going to be given explicit responsibility for regulating CDS; as we’ve seen in the banking sector, all hell tends to break loose when multiple regulators share responsibility for regulating the same companies or financial instruments.

To make matters worse, there is no national insurance regulator in the US: insurance is regulated on a state-by-state level. Why should a credit default swap entered into between two Delaware counterparties be regulated by the New York State insurance commissioner? It makes no sense at all

And let’s not forget that the New York State insurance commissioner — the only insurance regulator even remotely capable of regulating credit default swaps — was the regulator responsible for regulating MBIA, Ambac, and all the other monoline insurers who blew up as a result of writing far too many underpriced credit-default swaps. The SEC may or may not be an effective CDS regulator, but New York State has proved itself an ineffective CDS regulator.

In any event, if you want strong and effective regulation, the last person you want to turn to is an insurance commissioner. Insurance companies are the most highly-leveraged financial institutions in the world, if you look at the ratio of their contingent liabilities to their book value. That’s one reason why most insurers end up blowing up: they’re generally massively exposed to tail risk. Consider life insurers, for instance: they dodged one bullet, when AIDS ended up disproportionately hitting the kind of people (gay men, intravenous drug users) who don’t tend to have children and therefore normally don’t have much in the way of life insurance. But if a pandemic does start scything down a lot of rich people with children, expect a lot of life insurers to go bust — just as property insurers would disappear en masse if a hurricane were to hit Miami or New York. In theory, insurers hedge their catastrophe risk in the reinsurance market; in practice, they don’t, or not completely. And reinsurers can go bust, too.

What’s more, insurance companies are pretty much the last outpost of extreme opacity when it comes to the effects of the financial crisis. Most of the time, the way that insurers work is that they pay out in claims slightly more than they bring in, in terms of insurance premiums; their profit comes from investing those premiums before they’re paid out. If the investment returns are negative rather than positive, as they almost certainly were last year, then the total losses can be enormous. And I’ve been hearing rumors for months that there are lots of insurance companies which are failing to come clean on their investment losses, especially with respect to their securities-lending operations. AIG was not the only firm investing repo proceeds in subprime-backed securities and losing a fortune in the process — who were the others?

One of the reasons I’m still very bearish on the markets and the broader economy is because I’m convinced that there are lots of enormous financial institutions which haven’t even started to come to terms with their losses yet. Some of those institutions are European and other non-US banks; others are large insurers and reinsurers. I have no faith in the insurers’ regulators’ ability to reassure me that such losses don’t exist or are manageable. And I certainly have no faith in their ability to regulate the CDS market. So let’s not go there.


I’ve been doing swaps since ISDA was in diapers, as well as every other kind of financial surety known to the galaxy. Your premises are all wrong here. A CDS on a portfolio of securities is economically the same thing as financial insurance, with a fee paid in exchange for a guaranty. In fact the way you draft a CDS is by taking the standard ISDA swap form, ignoring everything in it, and stapling to the back a financial guaranty agreement that has “Swap Confirmation” written across the top. I don’t know where you got the idea that it’s just the swap of two cash flows (say, like a fixed-for-floating interest rate swap), but your analysis can’t recover from a fundamental mistake like that. If it were a real swap, there would be a theoretical structure where both parties would be willing to do the swap for nothing. That doesn’t happen with either financial insurance or CDS. You can call it a guaranty, a surety, a back-up letter of credit or a CDS, but they are economically the same thing. A CDS is no more a swap than a back-up letter of credit is a letter.

Here’s another easy way to tell the difference between a real swap and a financial guaranty: If it’s drafted by a junior banker with a B.A. in psychology who happens to be sitting at the swap desk this afternoon, it’s a swap. If it takes a team of economists, senior bankers and four Wall Street law firms several months to draft, it’s not a swap.

In any case, it makes sense to regulate financial sureties, however designated, with some consistency. Unfortunately for our economy, swaps were chosen for credit protection transactions precisely because they were an unregulated form, not because of the substance.

As you correctly point out, the financial insurance companies that got into trouble did so mostly because of unregulated CDS transaction, and not because of their regulated insurance policies. It was decision of the financial institutions themselves to evade regulation that got them into trouble, and their self-inflicted wounds can hardly be blamed on the authorities who lacked the power to regulate. AIG (regulated as a thrift institution) was profitiable in its regulated business, and disastrously unprofitable in its unregulated business. That is true of many, many financial institutions that got caught up in the crisis. You could make an argument based on the recent history of regulated financial institutions and CDS that regulation needs to be consistent, or over-exploitation of a loophole (like CDS) can lead to problems. But that’s an argument for consistency of regulation, not against regulation itself.

Posted by RobNYNY1957 | Report as abusive