One of the things that has intrigued me about the financial crisis is that it is pretty clear that looting took place at the high end of the financial services industry, yet few have called it by that name, in part because it has been difficult to identify the mechanisms by which it occurred.
Looting, as described by George Akerlof and Paul Romer, occurs when business owners go broke at society’s expense (loot) rather than go broke as the unfortunate result of having gambled on success and lost. And looking at Wall Street, the fact set strongly suggests that looting took place, In the old days of private partnerships, firms might blow themselves up on an individual basis, and you’d occasionally see serious industry downdrafts thinning the herd (the back office crisis of the late 1960s, the one-two punch of the end of fixed equity commissions plus the down leg of the 1970s bear market). But having the industry run of the cliff en masse was previously limited to commercial banking.
Now when William Black, a senior bank regulator during the S&L crisis, has talked about this phenomenon, he frames it as control fraud, meaning orchestrated in a deliberate fashion at the top level of a firm. While that probably contributed, particularly if an organization was desperate, but I suspect more complicated mechanisms played a big role.
To give a very simple example: back in the ages of stone knives and bearskins, when I was young, Citibank engaged in double and sometimes triple counting to encourage different groups internally to play nicely together. Thus if a deal yielded $1 million in profit center earnings, each operation that worked on it booked $1 million in profits.
The problem that then resulted was that individual units met their targets, but at higher levels, where the double/triple counting was netted out, the band didn’t.
My suspicion is we have had a massive version of an analogous version of the same problem, in which the internal (and probably even the reported) accounting did not accurately reflect the true economics of the business. And the way this turned out to be looting (as opposed to mere stupidity of the sort practiced traditionally by commercial banks) is that on Wall Street, a very big percentage of revenues is paid out as comp. So exaggerated earnings translate very directly into bigger pay packages.
Bruce Krasting, via Institutional Risk Analytics, sets forth one example:
Simplistic example. You are a bank. You have two customers. Southwest Airlines (NYSE:LUV) and Chevron (NYSE:CVX). SW calls and says, “Lets do a deal. Every 6 months for five years we will swap money based on 1 mm barrels of oil at $60. If the then cash price is less than $60 I pay you. If it is more than $60 you pay me.”
You say “fine” and then call Chevron. You enter into the exact opposite of the first swap trade. You make money doing this. You have a matched book. As a result of your purchases and sales you will have $50,000 in profit every six months for the next five years or a total locked in gain of $500,000.
At the end of the day your boss asks, “How did you do today?” You answer, “we made $500,000”. He says, “Great. Here is your bonus”. If you went to three market makers who did exactly the same as you did on that day the profit and loss results might be significantly different. For example:
– One could report a gain of $500,000. That would be justified. But would overstate the economic results
– Another might just record a gain of $100K (the current year portion that is “locked in”). The balance of the income is realized annually for five years. This would understate income. No one like that.
– Yet another might book just $400K of gains currently. This would represent the NPV of the $500k over the five-year period based on an internal capital rate set by the banks management. This is close to economic reality and is the form that most banks use to value this type of business. While it is the most appropriate and reasonable approach, it too is badly flawed.
The next day your bank announces quarterly earnings and says, “We made $400K. We are paying a dividend based on that and we are paying bonuses based on that. Net-net, retained earnings will go up by $250,000. It was a good quarter”
In this very simple example there is no cash from this profit. The cash will come to you over five years. So you have to borrow to pay the dividend and the bonus and all the other current expenses relating to this profit.
What you have done is ‘borrow’ retained earnings from future cash flow. You look like you have retained earnings to support your tier one capital ratios. But when there is a real ‘cash call’ on your equity (2008) you do not have the cash equity to survive. So you have a TARP party.
The OTC reform proposal plan does not give the banks the ability to do this. And that is why they do not like it. This same income recognition issue exists in the OTC swap market for currencies with maturities greater than one year.
Since the big financial intermediaries claim to run netted CDS books (and we found at least one big stuffee, AIG, who was eating a ton of risk), CDS alone could have played a major role in future profits being accelerated and treated to a large degree as cash flow earnings.
Do readers have any sense of how prevalent this was? I would assume it was close to universal for CDS. And what other products would have been subject to it besides CDS?
If you are not set up to comment, please e-mail me at yves@nakedcapitalism.com. Thanks!
23 comments
Comments are closed.
I can't answer the specific question.
However, this is one of the truly fine posts. It present a new way for a non-financial industry observer to think about the opaque accounting and cash flow aspects of the current mess.
It's time to apply RICO. These companies were run like fiefdoms, each capo with his own turf. The apple is rotten to the core.
YS:
Bank accounting is terrible, internally and externally. You ask "how prevalent this was". I suspect this went on at all banks. Krasting's example in which a bank records $400,000 profit on the "oil price swaps" immediately brought an old friend to mind: Enron! This was Enron's accounting method for electic power swaps! Enron recoreded the PDV of the "profit" from a 20-year electric power swap up front. The Big 87654 apparently learned nothing from Arthur Andersen's demise.
Even the $400,000 "profit" is questionable as the number can be manipulated by using inappropriate "internal capital" charges.
Yves,
With all the too-big-to-fail nonsense (and no changes in that being thought about) and the guaranteed tax-payer funded bailout, bonus tied to such appropriated earnings, where is the incentive for this change?
I don't see any!!!
Thanks for laying out the different possibilities that might have been used. The accounting textbook answer would seem to be to use the most conservative method; of the ones listed recording the $100k gain would seem the best fit. Of course, the traders are not going to like this. But it would be in the best interest of the shareholders.
Using an NPV calculation doesn't make sense because the discount rate chosen is a subjective figure.
The OTC reform won't change anything about this, at least not in the way B. Krasting is presenting it. Actually, from the point of view of his analysis, immediate profit recognition would be easier. Once a product is cleared in an exchange, the profit of the market maker on the bid/ask spread is recognized immediately.
This being said, there are other issues :
The first one is the presence of margin calls. When a product is cleared on a exchange, there are regular margin calls that trickle down from the broker to the final customer. The advantage of OTC products is that margin call procedures are flexible (or, in some cases, absent). It entails of course a counterparty risk (imagine SW defaulting when oil prices are low…), and it explains why the upfront margin is not 500,000k. From the point of view of Banks, this is legit business : After all lending is what are banks made for !
It is true that valuing such counterparty risk is tricky (notional is dependent of a market variable, and the variability itself can be correlated to the occurrence of default). Therefore the cash flow receivable are level 2 (or 3) assets where there is a degree of discretion in the valuation procedure. But, guess what ? This feature is shared by ANY banking/insurance business, even your good ol' mom and pop community banking. Loan provisioning (or Reserver determination for insurers) is and has always been a "mark-to-believe" exercise. The 100,000 "capital cost" discount is no more, and no less, arbitrary and prudent than such activities.
The second issue is the one of price transparency (usually, cleared price enjoy much wider dissemination) and the ability for clients like SouthWest and Chevron to bypass completely the Banks's balance sheets (provided South West and Chevron agree to pay margin calls…). I would guess that THIS is the main problem for the Dealers. If one doesn't need a strong balance sheet to be a broker in a market, the universe of competitors expands massively, and margins are compressed. All of a sudden, it is customers that can afford yachts…
If one wanted to be fairly precise about the value of this transaction, one would incorporate:
a) Discount factors for cash flows (e.g. PV of 400K not 500K)
b) Cost of credit for both counterparties (assuming to be conservative that the transaction increases the credit usage for both counterparties)
c) Cost of capital (including borrowings if bonuses are paid out based on asset valuations, rather than cash flow)
What is permitted in accounting terms depends on the derivative. IAS 39 / FAS 157 tightened up the P&L recognition, such that what are now called "level 3" (mark-to-model) assets do not qualify for up-front P&L recognition. This reduced the attraction of putting 30 year Bermudan trades on the books for example. Note that whilst a level 3 asset does not qualify for up-front P&L recognition, it does qualify for loss-deferral- if a derivative is put on the books at a Mark-to-Market of zero, but subsequently "in reality" has a loss of $100MM, it can continue to be reported at a Mark-to-Market of zero. Hence the potential attraction of reclassifying losing trades as level 3 assets.
But back to the example, which focusses on paying cash bonuses based on Mark-to-Market profits (instead of being paid out of positive cash flow). There is nothing wrong with doing so, although if done on a large scale the effect is to inflate the balance sheet & to make the bank vulnerable to MTM fluctuations on the derivatives. This was previously known as Metalgesellschaft risk, although now presumably people will look at how AIG was brought down by counterparty marks.
However, to be paying cash bonuses (whether appropriately or otherwise), there still needs to be an accounting profit (we assume here the managers will not pay bonuses if the accounting shows a loss). So where is this accounting profit coming from?
a) For exotic trades, the dealers sell a derivative at a wide spread, and hedge with derivatives trading at tight spreads. The wide spread is reported as an economic profit. The IAS 39 / FAS 157 rules restricted the ability to book this as profit, although historically taking such profit was quite common
b) There are certain trades that consume large capital charges, which charges don't get picked up by the accounting. Prudent banks would not pay large bonuses on such trades, given the potential for large capital calls. Imprudent banks would be less worried. Selling deep out-of-the-money options is the classic example here (e.g. S&P puts or super-senior CDS risk)
c) Some derivative trades such as oil swaps will be credit-intensive. Therefore if the bank is undercharging the trading desks for such credit usage, the trading desks will be showing a reported profit, when it is not economically profitable. The clearest recent example of this issue was hedging the super-senior risk of mortgage securities with the (then-AAA) monolines- in retrospect this was not hedging at all.
I'm not sure how the OTC reform proposal prevents banks from making cash payouts from MTM profits. My guess, however, is that the exchange would have a more conservative credit policy than the banks (at least those which are squawking).
Charles suggestion also makes sense- that it may cost less for customers to use the balance sheet of an exchange rather than the balance sheet of a bank, and that banks don't like being disintermediated.
Ordinary old swaps can be abused in the same way. It is the usual thing – understate the risk and get the bonus out ASAP. Can't remember the details but could pick brains if you like. Some deals had 20 year tenors, by which time the trader has had time not just to trouser his bonus, but retire.
Oh, looking at Recovering Banker's comment properly he is nailing it. What he said.
Although the exercise of describing compensation practices for securities in existence will be illuminating, it is important to keep in mind that the most egregious excesses come from new product development. Accounting departments are woefully slow in adapting to new market instruments, and I would hazard a guess that a large amount of the incentive to develop new products is to game the accounting system in a drive for bigger bonuses.
What an excellent piece. When do the trials and executions begin?
Lest we crick our necks while holding our noses so high at such odious practices, kindly remember that this is the same way the bulk of people use their credit cards:
"Sure, I can buy that $1200 t.v. set. I don't have the cash now, but I am certain I will have income from my job over the next 3 years, and since wages always go up (cough) I might as well realize the gains of my clever plan now!"
Same story, bigger budget.
Meli:
You are correct. Having lived in the accounting world, I can tell you many "New Financial Products" are accounting creations that make no economic sense, but yield an accounting result at variance with the underlying economics. And one more thing: a big fee for the investment banker who dreams them up.
Yves, Great piece. The looting trickles up throughout all m-t-m financial businesses. Your readers have already touched on most of the large-scale processes, so here are just a few additions:
1) Even in the 25 year-old interest rate swap business, traders and trading managers still manipulate p&l reports by taking advantage of poorly constructed yield curves. Lack of adequate granularity in a simple rate-swap curve, for example, will misrepresent the value of FRA's on adjoining business days. Think of 3-month libor set on December 31st versus January 2nd. If year-end borrowing pressure is high, the Dec 31st FRA will have in reality a markedly higher rate than the Jan 2nd one, but a poorly constructed m-t-m curve may price those FRAs as if they were identical. Hence a trading team that wants to boost apparent profit can sell Dec 31st, buy Jan 2nd, and book the difference in yields as if it were a real profit. Yep, still goes on.
2) The more complex the derivative the more likely that any two companies use different inputs and curve-construction methodologies to value the transactions on their books. So a Bear Stearns and JPM, for example, could execute a multi-legged deal and simultaneously BOTH trading departments could book a profit.
Suggestions to reduce but not eliminate looting at the trading desk level:
1) require exchange-trading or clearinghouse netting,
2)or hold back m-t-m profit claims and hence bonuses until the cashflows underlying reported profits are actually realized,
3)either way, apply a small tax to each transaction to discourage point-shaving (like in the FRA example) and to fund better-informed regulatory oversight.
Great post.
How to deal with accounting or regulatory motivated financial "innovations"?
Require every transaction to be assigned a pre-existing accounting method and regulatory regime.
If banks want to offer a new product, they can wait for regulators to rule on the appropriate regulatory and accounting treatment.
MAybe it's time to give customers shares in the banks they deal with, so the continued theft of the customer's money can't continue.
At least at my credit union I feel I have some sense of ownership. I would never, ever, put money into a bank again the way they are currently structured.
What, no mention of the flagarant century-old looting from the "last lender of resort"?
ftm says:
"Require every transaction to be assigned a pre-existing accounting method and regulatory regime. "
An excellent suggestion.
meli
What has happened in the last 3-5 years can best be described as a massive fraud. Fraud by borrowers, fraud by lenders, fraud by wall street, fraud by rating agencies and fraud by the CEO's, fraud by government regulators (yes, they had the tools to catch and stop this action but our executive at the time essentially told them to do nothing).
Unfortunately people are going to point at financial products and try and blame them for the problems. Mortgage backed securities have been around since 1979, CDO's since 1991 and CDS since the late 90's.
The products are sound until they are implemented with fraudulent actions. Like with Enron, off balance sheet transactions serve a stated purpose, but Enron employed them fraudulently.
What is unfortunate is this financial mess could have been handled quite deftly, the banks should have been taken over by the FDIC, broken up and spun out. That is what they have been doing for years and they are very good at it. Expand the FDIC power to take over insurance companies and since all brokers are now banks essentially they already have that power.
Excellent post Yves.
Fraud and Consequences — the lack of consequences for the perpetrators. Until this is rectified nothing really else matters. The existing players (both government and private sector) are trying to pretend that this crisis does not have perpetrators and there was no willful fraud. Naturally they are doing this to protect themselves and attempt to dupe the public into believing that this was all some unforeseeable accident and that we will all be back to "normal" soon if you would all just accumulate more debt, consume beyond your means, and believe. Nothing here to see — move along.
As others have colorfully noted — until there is a "day of reckoning" for those behind this episode society will retain the cancer and ultimately it will metastasize. Civilization's advancement depends upon Truth and Justice….we've had very little of either so far from Washington or Wall Street.
William displays his prowess with hammer and nail.
Yves,
Thanks for the explanation for the financially illiterate like me :)
What I don't understand is how this can even be legal. When I worked for a big company (non-financial), we basically had two basic rules about booking revenue and profit:
1. You can't book it unless it's LITERALLY in the bank.
2. You can't book it if there's a chance that we'll reimburse it back to the customer for one reason or another (warranty etc).
Why can't the banks follow the same rules? If the customer hasn't even paid them yet, how can they book the profits? In your example, what if both Chevron and Southwest go bankrupt the next day (or the day after the bonuses are paid)?
A number of responders describe the problem of accounting when there is a gap as a result of long dated transaction. Because of the mismatch it is an imperfect hedge.
A different, clearer example: You are now an FX dealer at the bank. IBM calls and says, " We want to hedge our European cash flows. We want to sell 100mm Euros two Years forward.
You say, "1.4600 value June 7,2011.
They say, "Done".
You call up your pals at Deutche Bank London for a price.
They say, "1.4700 for that date and amount.
You say, "Done"
You just made $1mm. It is locked in. There is nothing that can change the fact that on 6/7/11 you will collect your profit of $1mm.
Now, What is the acounting for that? How much of that $1mm do you book today?
Keep in mind that the phones ring all day long for deals like this at the large firms.Big numbers.
Bruce Krasting
Bkrasting@gmail.com
Bruce,
This is still fallacious. First, Bankers Trust shows that counterparties sometimes do get out of trades, or succeed in getting a haircut.
Second, the organization still has expenses applicable over the life of the transaction.