Thanks for the educational post. We are all (okay, not the Palin crowd) learning a lot more than we ever wanted to about the financial markets - and how we are screwed. A friend sent this e-mail to me and it might help throw some more light on our situation.

"Analogies are never perfect, but here's one using horse racing. Don't expect a perfect correspondence to the banking situation, but I think it is close enough for government work.
Joe goes to the track and bets $2 on a horse.

Two guys standing nearby get into a discussion and Fred says to Sam, "I'll bet you $5 that Joe wins his bet."
Next to them are Bill and Bob. Bill says: "I'll bet you $10 that Fred welshes on his bet if he loses."

Next to them is Sally. Sally says: "For $3 I'll guarantee to Bill that if Bob fails to pay off, I'll make good on the bet."

Sally then goes to Mary and borrows the $7 needed in case she has to ever pay off and promises to pay back $8. She doesn't expect to every have to pay since she believes Bob will always make good. So she expects to net $2 no matter what happens to Joe.

A quick calculation indicates that there is now 2+5+10+3+7 = $27 riding on the outcome of the horse race.

Question how much has been "invested" in the horse race?

Answer:

$50,000 by the owner of the horse who is expecting to recoup his investment from the winnings of the horse and other future deals. Everyone else is gambling, not investing.

The issue with the home market is that the only "investor" was the person who bought the home. All those engaged in the meaningless derivatives spun off from this are gambling. You can see how quickly the face value of all these side bets can exceed the underlying investment. Who is holding these side bets - not the homeowner? It is the people at the failing investment banks, hedge funds and similar enterprises. Notice that the bailout is being directed at them not the homeowners.

The real world is, of course, even more complicated. Over the last 30 years people have been allowed to place bets on everything starting with the value of stock averages. They might as well bet on the temperature in Newark at 8:00 AM.

So when you hear everybody saying this is a crisis caused by the housing collapse, be skeptical. We are in the midst of a classic pyramid or Ponzi scheme and there is no way out except for people to lose a lot of money. All that is different this time is that it is the taxpayers who are being asked for the cash."

This a pretty good overview, but most of the discussiiions miss a big point. The accounting standard takes into account the issue of whether there is an "active market" and allows reporting entities to choose alternative methods. But those methods must be explained and justified.

Aye, there's the rub. The whole justification thingy.

See Finacial Accounting Statement 157, "Fair Value Measurements" for the rule.

I thought that was a very good explanation Hilzoy. JohnBubba's example is also intersting. The problem with mark-to-model, is that the model maker is not a disinterested party, and thus may let either conscious or unconscious biases influence the result. Even in the case where you have two traders trading derivatives with each other, even with no intent to deceive, if their models differ, a seemingly profitable trade could be constructed, that both parties would consider to be gainful to themselves. Looked at from a broader perspective however, we see that nothing is actually created or destroyed (except perhaps confidence, or distrust) by the transaction. I think a big part of the problem, is that very few people are able to recognize a perpetual motion machine when they see it. They are most likely to get sidetracked in the details, and actually come to believe that everyone can benefit from something for nothing. Of course the laws of thermodynamics don't precisely carry over to the world of money, it is quite possible, for money to be created or destroyed by some of these schemes. But since money, is a human invention meant to be a standin for something of actual physical value, we should be able to ascertain that nothing of value is being created by these games. We need to train people to recognize perpetual motion machines, and the financial equivalent of perpetual motion machines. Ponzi scheme, at their very heart are financial perpetual motion machines, which most particpants believe will spew out wealth from nothing.

Yes, thanks for the clear and informative post. I like SJRSM's idea: modeling is fine as long as there are reasonable constraints on the models employed.

bigTom,
Have you noticed that macroeconomics looks a lot like classical thermodynamics? If only they could pin down the analogue of entropy, they'd have a better chance at spotting those perpetual motion machines. (I haven't seen enough micro theory to notice if these parallels carry through there, but presumably it would look something like the passage from statistical mechanics to thermodynamics.)

Last week on the Michael Medved show, a guest claimed that "mark to market" was a Marxist principle. (no, I'm not making that up)

modeling is fine as long as there are reasonable constraints on the models employed.

Posted by: matt

The models SUCK and should be thrown out. Anyone who relies on models like this:

“We were seeing things that were 25-standard deviation moves, several days in a row,” said David Viniar, Goldman’s chief financial officer.

SOURCE:

The computer models used to assess risk and benefit in the derivative markets were terribly wrong about the sub-prime mess. In the August 13 edition of the Financial Times, we learn that Goldman Sachs' programs were ridiculously off:

“We were seeing things that were 25-standard deviation moves, several days in a row,” said David Viniar, Goldman’s chief financial officer.

What's more interesting about this statment is that it's almost exactly what the Masters of the Universe said over 10 years ago. From Frank Partnoy's INFECTIOUS GREED, pp. 263:

Risk management was more art than science, and risk could not be boiled down to a single VAR [Value-At-Risk] number. LTCM's [Long Term Capital Management] VAR models has predicted that the fund's maximum daily loss would be in the tens of millions of dollars, and that it would not have collapsed in the lifetime of several billion universes, Askin Capital Management's VAR had been only about $15 million just before it collapsed. Barings' VAR models said its risk was zero.

Given the issues associated with mark-to-market, why not use mark-to-model, but have both counterparties agree on the model to consumate the trade? That way, both companies would be using the same model. Thus if one company was showing a positive value of x on the trade, the other would have to show -x on the trade.

Checks and balances can work great outside the political arena as well. If this were practical, it would alleviate the issues associated with mark-to-market that Hilzoy spells out, but also force the traders to remain honest with their estimated valuations.

Does anyone know whether this would be feasible or not?

JohnBubba - That's a major problem with letting the rich have all the money. They aren't investing it in anything. They're gambling it all away at the Wall Street casino.

Think of all the lost opportunities over the last 30 years. Think of what all that money could have created. Instead, they played with derivatives.

Time to bring back higher marginal tax rates on the rich, especially the super rich, and invest that money in real projects. Energy and infrastructure especially. But of course, we'll never do that. We're locked in to Reaganomics.

Mark to market. The previous basis of presentation of marketable securities was "lower of cost or market".

The theory was that it was better for a financial entity to understate its valuation (if a security or other asset that had been bought or constructed appreciated in value over time - sometimes a very long time).

My understanding was that that basis was changed because of the bias downward for firms (say like Berkshire Hathaway) that bought shares and held them over an extended period.

So, on BH's financial statements, their acquisition of an insurance company at cost may have been 1,000,000 in 1966, but is now worth to a theoretical third party 1,000,000,000. The stated net equity of the company would be grossly misrepresented, making the financial statements themselves irrelevant as information.

With more complex instruments, even cost may be indeterminate, say if it resulted from the swap of an asset for a pool of conditional assets, that resulted from the packaging of a swap of other conditional assets.

The maize becomes unravelable.

Its enough to prohibit the marketing of conditional or swapped derivatives, as the purpose of publication of financial statements is to create an informed marketplace (a willing informed seller exchanging with a willing and informed buyer), rather than a sale of fake Rolex's (on margin, with a short option as hedge).

Although US commercial law defines the dollar as legal tender, meaning that every asset is marked in dollars and must be liquidatable in dollars; the most modern investors do not have a final reference commodity.

A bushel of wheat, or a barrel of crude oil, or a kilowatt hour with no stated delivery time/location (all with real world uses as well as currency), may end up a more useful basis of financial statement presentation than dollars, and that only temporarily.

A trader who blogs at http://market-ticker.denninger.net/, commented that a big part of the rescue plan is that the government will buy derivatives that are undervalued, and hold them until they can be sold at a profit. He said, if that's really a situation that exists in the market today, why aren't savvy traders buying those derivatives themselves? If they aren't doing it, it's probably not a good idea for Paulson to spend our tax dollars doing it. But that's where we are.

It's complicated but let's make it simple:

Wall Stree has been running a big Ponzi scheme for years. (If you don't know what a Ponzi scheme is -- Google it.)

Ponzi's scheme's work as long as there are more investors coming in and not too many previous investors who wish to cash out. We are seeing a meltdown as everyone wants to cash out for good reason.

The bailout plan is an attempt to keep the scheme going. Suspending mark to market will definitely help banks and the government keep the con game going for a little longer.

Unfortunately -- it's very very hard to reflate a Ponzi scheme once everyone knows it's a con.

Aw shucks, there you go bein' all logical 'n' stuff, when real small-town Main Street people like Todd 'n' me know that it's all just a bunch o' hooey, an' what's real important is that there be strong protection for home-owners and the little guy who's just workin' and savin', and the government just gets outta the way an' lets him save his hard-earned money for his family's future!

But seriously, as one who has earned both a PhD and a living studying and creating models such as these, I can tellya fer sherr (oops, sorry) that mark-to-model is a very slippery slope indeed, Typically the most accurate models are very complex, which means there's a lot of levers that can be pulled to influence them, and the best models are transparent, which means it's very clear which levers to pull if one wants to influence them. Transparency is its own antidote, of course, in that attempts at manipulation are also transparent if and when they're revealed, but rather often output is promulgated and acted upon long before the basis for that output is fully understood.

Given that mark-to-model is (or at least can be) considerably less time-sensitive than mark-to-market, since market data are real-time but models can be run any time, perhaps one useful constraint on mark-to-model might be some form of required aging, e.g., you can't use results that are, say, less than a week old, by which time presumably a detailed study could be made of input and methods and suspect output discredited.

Typically the most accurate models are very complex..

Typically, they are absurd. See my post above.

it would seem like a no brainer that mark to market is a bad idea.

even to a complete neophyte like myself.

it takes real expertise to turn it into an option.

Even with Mark to Market you still need some kind of time averaging, short as it may be, true? Otherwise, company assets would wiggle around a lot even during one day. Like climate, I assume accountants can average, but what's the moving window? Could it be made longer, to protect from market fluctuations, or even medium-term trends? Maybe the accounting philosophy that there must be one unique "valuation" needs a look at too, so one could pick from a few different versions of "value"? I'm not a pro, so forgive anything I got wrong or didn't appreciate.

If a bank values a block of CDO's at $1M using mark to model, while market cost would be a half million, what benefit does he get other than deceiving investors? Is the idea that he would be able to borrow more based on overvalued collateral? And doesn't that just spread the risk even wider?

Assclowns of the Week #71: Grand Theft Autocrats Edition is up, if anyone's interested. McCain, as expected, Bogarted two spots, including #1.

I think mark to model would be fine if you could do something similar to the old cake-division rule: A cuts, B chooses the slice. You get to mark to model if you're willing to have someone who hates you choose the model.

Insurance regulators require insurance companies not to count certain assets that would be accepted under GAAP or international accounting rules. Non-admitted assets are a wonderful solution to investments whose value cannot be determined by the market. Mark it to model if you must, but then exclude it from your equity, with a note telling folks how much was excluded.

Mark to market has been, more or less, the rule for ages. The older form of it was called "lower of cost or market" which didn't allow companies to book their investment gains till sold, but required investment losses to be acknowledged immediately. Most equities and bonds are valued as of the closing bell at the end of the quarter. The problem comes with all sorts of derivatives, which Steve J has pointed out, are far more erratic than equities or bonds.

There is no reason for Treasury to buy a single derivative. The first thing they could do, either by SEC fiat or by asking Congress to come back to pass a law, is make naked derivative trading illegal immediately, forcing everyone to unwind their naked positions in things like credit default swaps and telling everyone that no one will collect on any naked positions. In JohnBubba's example, Fred, Sam, Bob and Bill are all engaged in naked derivatives because they are not party to the fundamental transaction, the bet on the race. Sally and Mary are engaging in transactions that could never happen if there were no naked derivatives. None do a bit of good for the economy.

Well, hell, I thought my brilliant idea of a rolling 30 day average "market" valuation tto which to mark was a brilliant new thought. Heh heh. 'S what I get for thinking without adult supervision...

The rules on Mark to Market are being relaxed because once TARP starts buying assets, a market for these assets will exist.

So why would this matter? Because even at the highly inflated prices the Treasury intends to pay for these assets, if banks and other financial institutions holding similar assets were forced to value them at the prices the Treasury pays, some of our largest financial corporations would be recognized to be insolvent.

$700 billion, as large as it sounds, is insufficient to bail them all out. That is why this entire amount will eventually be lost trying to prop up the banks. This whole approach is designed simply to buy time in hopes that these institutions can transfer enough of their toxic assets to the government and other parties in order to claw their way back to solvency.

In the not so distant future, look for bailout 2.0 to continue this process.

This is seriously misleading. Fair value accounting (mark to market is not really right) serves several functions. But keep in mind that it came to the fore with the support of financial institutions because it enabled more lending -- it increased capital value when markets rose (for what it was the companies held). The only real alternative is historical cost -- assets valued at purchase price. That severely restricts the flexibility of financial (and other) institutions.
Abandoning fair value accounting is an invitation to securities fraud. Exactly what is it investors -- all the people buying stocks or mutual funds -- are supposed to look to in order to determine the value of stocks? Made up values? I think you need to re-think this post.

It is the nature of a bank to borrow short and lend long. A bank needs to be able to sell assets obtain enough cash to pay back depositors' requests for funds. A mark to market requirement helps insure that the banks can, in fact, raise sufficient cash.

What happened to Wachovia, among others, is that depositors wanted their money and the bank could not sell its illiquid assets to get enough cash.

I hate to go off topic here, but looking at McCain's health plan, from his website it says:

While still having the option of employer-based coverage, every family will receive a direct refundable tax credit - effectively cash - of $2,500 for individuals and $5,000 for families to offset the cost of insurance. Families will be able to choose the insurance provider that suits them best and the money would be sent directly to the insurance provider. Those obtaining innovative insurance that costs less than the credit can deposit the remainder in expanded Health Savings Accounts.

Even Obama is spewing out misleading information about this "plan".

First off, it does not appear to be a tax credit, it is just cash. This is actually good for low income individuals/families who intend to buy insurance. But it is also a huge give-away to people who choose not to purchase insurance because they are young, stupid, or under the impression that they are blessed with good health.

Under this plan, you could refuse to buy any insurance and put $2500 into a Health Savings Account.

In effect this would reward the young and stupid by allowing them to bank their premiums, which are paid by the government.

How long will it take for someone to figure out that they can use this money to pay for "health club" memberships, medical vacations, etc.?

"The first thing they could do, either by SEC fiat or by asking Congress to come back to pass a law, is make naked derivative trading illegal immediately, forcing everyone to unwind their naked positions in things like credit default swaps and telling everyone that no one will collect on any naked positions."

Freelunch,

If I understand them correctly, most CDSs are a form of insurance - a guarantee to pay the purchaser of the contract in the event of default by a certain issuer.

If that is indeed the case, that would explain why the New York State Insurance Commissioner was interested in these contracts, if I recall correctly, specifically for companies like AIG that were selling a lot of them.

The commissioner's concern was with "insurable interest". This is the idea that the beneficiary must have an economic interest in the continued viability of the insured, a requirement that insurers are legally obligated to confirm. In the case of contracts like life insurance, this is usually rather pro forma since the beneficiary is typically a spouse or other financial dependent of the insured.

In the case of a "covered" CDS this is also not an issue because the beneficiary (most often also the purchaser of the insurance) is a lender to the insured and thus has an insurable interest in the borrower's continued solvency. However, in the case of a "naked" CDS, the purchaser of insurance is simply placing a side bet on the insured's solvency, not protecting himself against financial loss due to failure of the insured. Hence, there is no insurable interest.

As the precedent is there with other types of insurance, it certainly seems logical to apply the same strictures requiring insurable interest to CDSs, but since they're largely unregulated (and treated as securities rather than insurance, I think?) I don't think it's even been attempted until the NY insurance commissioner looked into it. I would guess that one of the hazards people fear about naked CDSs is a similar one to that mentioned here:

http://law.freeadvice.com/insurance_law/business_insurance/1insurable_interest.htm

Hedge funds instead use mathematical models of their own to estimate and report the value of their CDO holdings to investors -- a practice known as "marking to model."

It doesn't take a Buffet (or a Feynman with a glass of ice water and a ten cent o-ring) to see that this system's morality and rationality break down right there.

Hilzoy's conclusion is spot on:

But why anyone thinks that the best response is to let companies go back to mark to model is a mystery to me. This is especially true now...

Mark-to-market was at the heart of Enron's accounting; they used it to overvalue assets and report the gain as income in their SEC filings. But of course, we've fixed that problem, it can't ever happen again.

But the problem with mark-to-market that just won't go away is that it replaces a solid number, obtained by applying a universally understood formula to data in the books, with an estimate calculated by the company itself. The choice is between an incorrect estimate, e.g. book value, of perfect pedigree and a potentially better estimate subject to tweaks, wishful thinking, and Fastows. After the past few years, I'm leaning toward the former, myself.

Ahh, Hilz, it's part of the duopoly's sellout. You didn't really believe the Republican-Democrat duopoly was going to change that much, did you?

A computer model that cannot be tested against reality -- in the industry, we call that a video game.

Only the self-deluded, ignorant, or dishonest would rely on a computer model that cannot be tested against real world data.

Enron's accounting gaffs were related but different than the current maize of multiple flaws.

Enron functionally valued its own stock as an asset (rather than as net equity), creating a spiral effect within its own boundaries.

The housing and financial markets create the same spiraling effect (values increasing or decreasing in a self-perpetuating momentum), but in the market as a whole. In that way, the spiral up or down, is more insidious as it is ubiquitous.

Spirals exist everywhere in markets and in nature. The thing that allows them to occur without distorting the functioning of the system as a whole, is that there are countervailing fundamentals that can step in.

For example, when market psychology (feeding on itself) pushes the price of a security below what a rational investor thinks that security is worth in common financial logic (fundamentals), then there is a buyer and the spiral lessens or stops.

The breadth of how the spiral aspects of the mortgage backed securities markets functioned, included almost everyone, represented in securities that were not able to be assessed even, so that there were no outside "fundamental" buyers, and won't be really for a while.

That is really the critical shift in any proposed or executed legislation, that there is a means to restore the fundamental value of a security, so that a rational liquid investor can act.

Something unknowable is a great exageration of risk. Whereas a rational investor might pay $95 for a risk free return of $100 in a year (time value of money), a rational investor that doesn't know whether he/she/it will get $25 or $95, can't even make a proposal.

They can't even derive relative information from what is disclosed. They can't even note, what to monitor closely (to slice up risk factors into the small portion of indeterminate factors that are addressed by attention and intervention).

Even with a cost-based, or lower of cost or market based, valuation of securities on balance sheets, there is still an annual revaluation of the securities, and phantom income.

For example, if last year I wrote down $1,000,000 of securities to $500,000, realizing a stated $500,000 loss, and this year the market price for those securities was $750,000, I would realize a change up in the allowance for securities mark-down from last year, INCOME.

But, then again, to state securities at current market value is a distortion, as there is only an inference that the market value will be the same in six months (between published financials).

There is no sure-fire method for an entity that records securities prominently on its face financial statements.

The balance sheet of a manufacturer includes inventories (also not clearly describable), equipment, etc. Real assets (of a sort).

The balance sheet of GE or GM for example, consolidated to include their financing divisions, includes more securities than hard assets.

Not to mention the effects of things like pension obligations, that become under-funded when their pension investment portfolio tanks.

Or, GM's recent enormous write-downs of future tax credit benefits, lost in a setting of no income to take the credits against.

Even with a cost-based, or lower of cost or market based, valuation of securities on balance sheets, there is still an annual revaluation of the securities, and phantom income.

For example, if last year I wrote down $1,000,000 of securities to $500,000, realizing a stated $500,000 loss, and this year the market price for those securities was $750,000, I would realize a change up in the allowance for securities mark-down from last year, INCOME.

But, then again, to state securities at current market value is a distortion, as there is only an inference that the market value will be the same in six months (between published financials).

There is no sure-fire method for an entity that records securities prominently on its face financial statements.

The balance sheet of a manufacturer includes inventories (also not clearly describable), equipment, etc. Real assets (of a sort).

The balance sheet of GE or GM for example, consolidated to include their financing divisions, includes more securities than hard assets.

Not to mention the effects of things like pension obligations, that become under-funded when their pension investment portfolio tanks.

Sorry if anyone mentioned this.

What if a company's debt/equity ratio is worse than 95%? (2 statistical Std. deviations)

Could the requirements be to report when your company is substantially sicker than the rest so extreme debt can be a live wire to company wants to touch?

This could have the benefit of weeding out the weakest companies which seems to have been a serious problem this go around. Insufficient will to let economic Darwinism do its unfortunate, necessary work.

Chris Cox of the SEC changed the debt/equity rules for investment banks in 2004. The investment banks figured we have a smooth economy and they didn't need to hold so much in reserve.

Whoops!

Here's the problem. Nobody wants to buy mortgage backed securities right now. That doesn't mean 100% of the mortgages in the mortgage backed securities are or will ever go into default. Right now only about 4% are. It also doesn't mean the actual houses backing those mortgages all burned to the ground and the land they sit on is poisoned. These are not worthless assets but without anybody buying they look like it. If we continue to force these institutions to value mortgage backed securities based on what they can get for them right now (mark to market) we are basically saying that your house and your equity in your house are worth nothing because the guy down the street lost his house to foreclosure. Chances are you if you want to sell your house right now you can't. It's not because it's worth nothing it's because prospective buyers can't get a mortgage. The same goes for those mortgage backed securities. If the government buys those securities for pennies on the dollar we taxpayers might get a real bargain. It'll also make the banks' balance sheets look better for awhile. But it won't solve the credit crisis because it won't provide enough recapitalization for those big banks. They still won't have money to lend and the credit crisis is what is killing us.

Wheat farmers can't wait til you buy that box of Wheaties to get paid. Chavez and Saudi sheiks won't wait til you fill up your tank at the pump to get paid. Your county can't wait til you pay your taxes to fill pot holes. The whole economy runs on credit and if there is none, well good luck saving up enough cash to buy a car or a house for starters.

Ira Glass's "This American Life" had another great radio show on this yesterday which should be up on the website soon. I recommend everyone listen to it and the previous one from this summer.

Tremendous post & comments here! Be better if there was any good news. What riles us long time opponents of this type of speculative financial frenzy is that the truth was known long ago(decades, at least), and yet the march into folly couldn't be stopped.