Subprime

by hilzoy

Atrios picked up on a story that I've been watching for a couple of days now: the near-implosion of two Bear Stearns hedge funds. Since this post is long-ish, and has no obvious dividing lines, I'm putting it below the fold.

A WSJ article (subscription) from yesterday's paper gives some background:

"Bear Stearns and some of its executives had invested only $40 million in the two hedge funds, raising more than $500 million of additional investments from outsiders, including wealthy individuals and a breed of money managers known as "funds of funds," which spread their money around a collection of hedge funds. In addition to the cash raised from investors, Bear Stearns's Enhanced Leverage and Credit Strategies funds have invested billions of dollars borrowed from lenders.

As of March 31, the Enhanced Leverage fund had $638 million in investor capital and at least $6 billion in borrowings. It used the money to make $11.5 billion in bullish bets and $4.5 billion in bearish bets, according to documents reviewed by The Wall Street Journal. Its sister fund had $925 million in investor money, and made $9.7 billion in bullish bets and $4 billion in bearish bets.

A high level of "leverage," or use of borrowed money, magnifies returns for a fund if everything works as planned. If the security drops in value, that same leverage can amplify losses.

Consider the impact of an investment in a $1,000 security with $100 of a fund's own money and $900 of borrowed money. Annual interest of 6%, or $60, on that security could be amplified to returns of 15% for the fund. However, if the value of the security dropped by just 3%, to $970, the loss would translate into a 30% drop in the fund's own money. That, in turn, could trigger a margin call by the fund's lenders.

Prominent in the Bear Stearns funds' portfolio were investments in collateralized debt obligations, or CDOs, which are pools of securities that can include mortgage-backed bonds, corporate bonds or leveraged loans, among other things. Many of the CDOs the Bear Stearns funds invested in were pools of mortgage-backed bonds.

Wall Street sells pieces of CDOs to investors all over the world, with some taking risky pieces that are more prone to losses but that promise high returns, while other investors take safer, lower-yielding pieces. They are complicated investments to value because they are effectively giant bundles of securities that have been put together and taken apart again."

So: two highly leveraged funds investing in extremely complicated investment instruments, mostly based on subprime mortgages, which are, as the WSJ article tells us, "the relatively risky home loans made to borrowers with troubled credit histories." A couple of months ago, the funds posted significant losses, and investors started pulling their money out. This proceeded at an escalating pace, with larger players like Bank of America and Goldman pulling out, until a few days ago when the funds were near collapse. A rescue deal involving infusions of cash was put together and fell apart; today the whole thing seemed to be collapsing. As of tonight, JP Morgan has reached a deal with Bear Sterns to recover its money without having to seize any of the fund's assets, which means that one huge pile of securities won't be dumped on the market; however, Merrill Lynch and Deutsche Bank are trying to auction off their shares of the funds' assets, which means that another large chunk will. The WSJ and NYT are emphasize the fact that a meltdown has been avoided. The Financial Times, on the other hand, says this (sub. req.):

"The giant market for securities backed by US subprime mortgages was thrown into turmoil on Wednesday as lenders struggled to sell more than $1bn of assets seized from two Bear Stearns hedge funds that suffered heavy losses on subprime bets.

The complex securities being auctioned are rarely traded and early attempts to sell the collateral met with mixed results. The prospect of the “fire sale” knocked down prices for similar mortgage-backed assets and sent a key derivative index for the market to record lows.

The rout highlights the risks investors take when they buy illiquid and hard-to-value securities. Fire sales in times of stress can trigger dramatic changes in pricing in such markets, perhaps leading other holders of assets to mark their values down and triggering demands for additional collateral from lenders."

***

As I understand it, the reason this story matters outside the world of Bear Stearns and its lenders is as follows. First, two large funds seem to be unwinding, and that means that people are going to lose a lot of money directly. Second, this will probably make people generally skittish about subprime mortgages, and funds that invest in them. But there's another way in which this could spill over, which has to do with the fact that even leaving aside any drops in value caused by the present troubles, it's not clear to me that anyone has a clear idea of how, exactly, to value these sorts of securities. Here's the WSJ:

"Unlike stocks and Treasury bonds, whose prices are continually quoted and easily obtained, many of these derivative instruments trade infrequently and don't have clear market prices. To come up with market values for these investments -- a process known as "marking" their positions to market -- investment funds often rely on their own valuation models.

They might also ask the dealers who sell them the bonds to update them on changes in the bonds' underlying value. When there are no sales to base prices on, dealers come up with prices based on their own statistical models and an array of assumptions about what's happening in the market or the assets that back the securities.

"There's some real concern about how realistic dealer quotes are," said Andrew Lo, a finance professor at the Massachusetts Institute of Technology who is also a principal in AlphaSimplex Group LLC, an asset-management company that also runs a hedge fund. "You're talking about quotes during normal times that are very different from quotes during stress times."

There is no indication that Bear Stearns's fund managers sought to mislead lenders or investors about the value of the funds. Indeed, the firm's approach to valuing its securities seems to be in line with guidelines set up by Moody's Investors Service, which evaluates hedge-fund practices. But the crisis does point to the kinds of valuation problems hedge funds and their investors or lenders can run into, even when they follow sound practices.

A forced sale of the Bear Stearns funds' assets now could trigger a broader repricing of mortgage-backed bonds and lead to losses and margin calls -- demands for additional cash or collateral -- at other funds."

And the problem isn't just that these securities are not traded often enough to get a good sense of their value from the market; it's that they are excruciatingly complicated. Here's a portfolio manager speaking to the NYT about the assets put up for auction today:

"He said it would take time — perhaps several days — for potential buyers to drill down into some of the more complex securities in order to value them before any bids could be prepared."

That might explain why some investors didn't bother to try to understand the values:

"Until the end of last year, many large institutional investors were willing to bet on subprime bonds and C.D.O.’s without even trying to value the underlying assets, said Louis Pizante, chief executive of Mavent, a firm that helps investors analyze thousands of mortgages at a time. Many simply placed investments in assets based on their credit ratings.

Mr. Pizante recalled what one customer said last year as he asked for only a basic assessment of mortgage loans: “The broader your review is, the more reasons you are going to tell me why I can’t buy these loans.”"

Oops.

So, basically, you have a whole lot of money invested in securities that are very hard to value correctly under any circumstances. Since these securities are not traded frequently, people try to value them by other means: computer modelling, comparisons to similar instruments, etc. The "comparison to similar instruments" part will drive down the valuations of these securities directly: a whole lot of "similar instruments" just got dumped on the market and are being sold at fire sale prices. But the computer modelling part probably has issues of its own. I would imagine that even assuming that everyone involved in making these arcane valuations is acting in complete good faith, the combination of inevitable error and the temptation to come up with a model that tells you what you want to hear would mean that a decent number of these securities were somewhat overvalued to start with. And the very difficulty of valuing these securites accurately has to be a temptation to people not to act in good faith. From the WSJ:

"The issue of pricing has plagued the hedge-fund industry for several years. In 2004, the Securities and Exchange Commission accused four men involved in running Beacon Hill Asset Management LLC of intentionally shading the value of their own investment in mortgage backed bonds to inflate the value of their portfolio when reporting returns to investors. The men settled with the SEC for $4.4 million and were banned from the industry.

For hedge funds, the incentives to abuse the murkiness of the market can be huge, because they are compensated by fees based on the values of their portfolios. "Hedge funds have the incentive to mark their positions more favorably because their compensation is tied to those marks," says one senior Federal Reserve official.

"No one in the subprime business wants to ask the question of whether they need to re-mark all the assets. That would open the floodgates," said Janet Tavakoli, president of Tavakoli Structured Finance, a consulting firm in Chicago. "Everyone is trying to stop the problem, but they should face up to it. The assets may all be mispriced.""

Or, as Calculated Risk puts it: "What a mess. No one knows what many of these securities are worth; some of them may be worthless."

In any case, the value of those securities is bound to get a lot more scrutiny now, since investors will not want the fund they've invested in to be the next one to fall, and the various ratings agencies will also be examining these valuations more carefully. And if other funds that have invested in these securities have to mark down their value, whether because they had overvalued them to start with or because the Bear Stearns funds' problems drive down their value, that obviously increases the likelihood that those funds will also find themselves with insufficient assets, in which case there could be further liquidations, in a sort of horrible cascading effect.

And it's not as though the housing market itself really needed something to drag it down:

"The worst is yet to come for the U.S. housing market.

The jump in 30-year mortgage rates by more than a half a percentage point to 6.74 percent in the past five weeks is putting a crimp on borrowers with the best credit just as a crackdown in subprime lending standards limits the pool of qualified buyers. The national median home price is poised for its first annual decline since the Great Depression, and the supply of unsold homes is at a record 4.2 million, the National Association of Realtors reported.

"It's a blood bath," said Mark Kiesel, executive vice president of Newport Beach, California-based Pacific Investment Management Co., the manager of $668 billion in bond funds. "We're talking about a two- to three-year downturn that will take a whole host of characters with it, from job creation to consumer confidence. Eventually it will take the stock market and corporate profit.""

Oh goody.

***

I should say that one of the many reasons why I am not an active investor in anything is because I systematically get things wrong. I am pretty decent at spotting reasons why the prices of stocks, houses, and all sorts of other things are unsustainable, and why their unravelling coud lead to calamity. And I'm often right, in the sense that eventually things do go south, for more or less the reasons I expect. But I'm generally pretty dreadful at understanding why things might be expected to improve; and so I tend to predict bad things years before they occur, since virtually all the countervailing forces are invisible to me. So take everything I say about this stuff with that in mind.

As always, I welcome corrections by people who understand this stuff better than I do.

Comments

[HankP]

You are right to be pessimistic, only more so. One of the results of this will be that the low interest, easy money policy that was started by Alan Greenspan, where all economic issues were dealt with by making debt more and more inexpensive, will start to unravel. There is so much liquidity sloshing around the system at this point that we could easily see a return to stagflation, or even worse intentionally high inflation to reduce the repayment costs of our insane deficits and debt. The only positive that I can see coming out of this is that it will make an electoral sweep by the Democrats in 2008 much more likely, and help discredit Republican supply side economics for a generation.

[Curt Adams]

You quote the WSJ as saying Bear Stearns hadn't attempted to mislead investors about the value of the funds. But they have. Bear Sterns has manipulated the subprime market to prop up prices of subprime mortgages:

http://www.cnbc.com/id/19085195

I smell a future criminal investigation.

I recall reading articles about the manipulation months ago. The WSJ has no excuse for missing this market manipulation.

[Ginger Yellow]

"Second, this will probably make people generally skittish about subprime mortgages, and funds that invest in them. "

Well, people are pretty skittish about them already. Spreads on BB rated subrime mortgage bonds had already widened by over six percentage points before the Bear fund collapse. As a journalist covering the European CDO market, among other things, I'm more interested in the knock on effects within structured credit. Back when Enron and WorldCom and all the other companies were going bust there was a big problem because many CDOs weren't just invested in those companies, but also in other CDOs exposed to those companies. There was a severe lack of transparency and a kind of domino effect as one CDO's downfall led to several more, which then affected others and so on. Supposedly arrangers and investors are more alert to that sort of risk these days, but I'm skeptical.

[OCSteve]

Hilzoy: I’m not much of an investor either. The one thing even I managed to learn though: if you hear/read the word derivative run away as fast as you can.

I agree with you that this sounds bad. I’ve been expecting it, tracking the number of foreclosures and tax sales in my area. I dislike the thought of profiting off of other people’s misery, but I’m seriously thinking I should be prepared to buy a certain brand new condo at a “fire sale” price when the developer goes bankrupt, which I fully expect to happen.

[Ugh]

OCSteve - you're not profiting off their misery, if no one stepped up to buy their assets at "fire sale" prices then they would suffer all the more.

[OCSteve]

Ugh: True, that. So I guess I won't feel too guilty. ;)

[Catsy]

It's things like this that have kept me from trying to get a home loan right now. My credit isn't horrible, but it isn't that great either, so my only real options would be a high-interest fixed rate, or a subprime. I wouldn't be able to afford the payments for a fixed rate that I'd qualify for, and getting a subprime loan would be a recipe for disaster. In addition to the high likelihood of the rate getting raised above what I can afford, I also stand a good chance of seeing my equity crater when this house of cards starts coming down.

Presently I'm working on slowly rebuilding my credit, and when the market really craters I'll reevaluate my prospects. Unfortunately real estate is really expensive in my area of WA.

[hairshirthedonist]

I am pretty decent at spotting reasons why the prices of stocks, houses, and all sorts of other things are unsustainable, and why their unravelling coud lead to calamity. And I'm often right, in the sense that eventually things do go south, for more or less the reasons I expect.

This should make you an excellent short-seller. Knowing when things will go down is just as good as knowing when they'll go up. You just have to reverse the order of you buying and selling. (I know this only from having watched Trading Places some number of times. Man, that Eddie Murphy's funny.)

[hilzoy]

Hairshirt: that's the thing: I don't know when they'll go down; just why. I started expecting the colapse of the tech bubble early in 1997; it was a complete mystery to me that it didn't burst until several years later.

[hairshirthedonist]

Yeah. Timing would be a problem. But you just need to get your hands on the orange-crop report before it goes public and give the Dukes a phony, erroneous report. It's that simple.

[J Thomas]

There was a severe lack of transparency and a kind of domino effect as one CDO's downfall led to several more, which then affected others and so on. Supposedly arrangers and investors are more alert to that sort of risk these days, but I'm skeptical.

Lots of businesses have this trouble, though it seems like investors shouldn't.

If your business is building subdivisions, and there's reason to think there won't be a market for awhile, should you stop? Should you lay off the workers and put your equipment in mothballs and hope you can start over again when conditions improve?

Not usually. Usually you keep doing what you know how to do, and you compete with the other survivors to see who can lose his shirt slowest, and the ones who lose that competition aren't around for the next boom any more than the ones who quit early.

It's one of the ways our economic system is inefficient, and I don't see a general solution.

[Sebastian Holsclaw]

"It's things like this that have kept me from trying to get a home loan right now. My credit isn't horrible, but it isn't that great either, so my only real options would be a high-interest fixed rate, or a subprime. I wouldn't be able to afford the payments for a fixed rate that I'd qualify for, and getting a subprime loan would be a recipe for disaster."

If I were you, I would contact a good loan broker and look into the situation anyway. I got a surprisingly low interest rate fixed loan on my recent first purchase, and my credit is decidedly non-stellar. Some might call it earthy. ;)

[Slartibartfast]

Pork belly futures. Pork bellies, which is used to make bacon, which you might find in a bacon, lettuce and tomato sandwich. It's that simple!

[Ara]

Hilzoy: On the tech bubble, isn't everything explained merely by the fact that people are skeptical but *not* as skeptical as you are? After all, most tech was just as unprofitable in 1997 as in 2001, but there was still Tulipomania going around and plenty of investor exuberance.

From what I know nobody can really account for why NASDAQ plunged when it did. I've heard that one of the big theories is that a coincidence of a few big sell orders executed simultaneously, and that led to a feedback cycle of panic and liquidation.

[hilzoy]

Ara: my theory about why it didn't collapse in 1999 is that the entire tech sector was buoyed by a wave of Y2K fixing, and fell when that wore off. But I have no clue whether this is true.

[Ginger Yellow]

"Lots of businesses have this trouble, though it seems like investors shouldn't."

Well in a way it wasn't their fault. Unless you were an equity investor you wouldn't know which precise bonds were in the pool, and in 2000-2002 most of the time the equity holders were the arrangers and/or managers of the pools. In a number of cases (some of which went to court), these banks were able to swap deteriorating assets from their balance sheets into the deals.

These days the equity is almost always sold to third parties and the documentation has (supposedly) been tightened to prevent this sort of abuse and increase transparency, but I'm hearing a lot of stories that the sheer volume of deals being done means that the lawyers don't have time to go through the documents carefully.

[Model 62]

A little OT, but an intriguing diversion, nonetheless: housing costs 1890 to present, adjusted for inflation, depicted as a roller coaster [google video].

[J Thomas]

"Lots of businesses have this trouble, though it seems like investors shouldn't."

Well in a way it wasn't their fault. Unless you were an equity investor you wouldn't know which precise bonds were in the pool, and in 2000-2002 most of the time the equity holders were the arrangers and/or managers of the pools.

I see, it's the same sort of thing. You don't shift your money out of pharmaceuticals and into something less risky unless you're expert enough in the other field to know you aren't doing even worse. Once you specialise, you can't just get out of your specialty when it looks risky, because investing outside your specialty is even riskier. And if you park your money waiting for things to improve, then you can't win.

Somebody who got out of tech stocks in 1997 would have lost out on 4 years of profits, probably more than he'd have lost in 2001. If it had lasted to 2002, even better. So instead of trying to guess which year the correction will come, maybe better to just watch the market carefully and try to get out in the first few hours after you think the correction is starting.

[Francis]

two comments.

1. I believe, with virtually no evidence, that the investors who are going to get shellacked as this market collapses are not your everyday 401(k) fund but instead high-risk, high-net worth individual investors.

2. The likely effect felt by ordinary people will be more expensive mortgages and higher barriers to getting mortgages. If mortgage originators have to (horror of horrors) buy back their non-performing loans and/or not sell certain categories of mortgages in the first place, they're going to be much more interested in ensuring that you will pay the damn thing off.

(okay, three points)

3. A slowdown in the primary mortgage market may put builders in high-growth areas in a bind because they will sell fewer houses. (See Toll Brothers stock price.) But since a lot of these builders use subcontractors who in turn use a lot of illegal labor, it may be difficult to accurately measure the impact of slower construction rates on the overall economy.

[Jim Parish]

I started expecting the colapse of the tech bubble early in 1997; it was a complete mystery to me that it didn't burst until several years later.

hilzoy, Robert Reich showed how to take advantage of this sort of situation twenty years ago:

"My reputation as a soothsayer began in October 1987, two weeks before the Dow Jones Industrial Average plunged five hundred points, when I appeared on a nationally televised talk show opposite a conservative economist who assured viewers that the bull market would continue. I demurred. My advice to viewers was to get out of the market before the coming crash. I had the temerity to predict that the crash would occur within the month. In the weeks following the crash I was inundated with letters from people wanting to subscribe to my investment letter. Politely but regretfully, I informed each of them that I did not publish an investment letter. I did not tell them, however, that I had been making precisely the same prediction for five years."

(from the preface to "The Resurgent Liberal", 1991 edition.)

[Jonas Cord Jr.]

With all these valuation problems, I'm really at a loss as to why these various instruments aren't standardized in some fashion so this sort of problem is far less likely to occur. (I don't see why any of these derivitives are a bad idea in and of themselves.)

But then again, I'm just a guy who wants markets to work well, not a Wall Street guy. I guess that might explain why.

[tzs]

Ah yes, CDOs. Almost managed to get hired at Moody's to do the statistical evaluation of such beasts, based on a background in theoretical physics.

That's the problem--the gaussian distribution which supposedly shows the spread of returns/defaults/whatever with the individual units definitely does not work when things get "sticky" and the market panics one way or the other. At that point, correlations go to 1 and whoops, breakdown of the theory.

Derivatives and Black-Scholes aren't terrible, it's just that they have to be taken (both) with a large grain of salt and a realization that in most areas there's just not enough trading to really "test" the risk analysis. Black-Scholes was derived from arguments about stochastic processes. Works great for gas molecules, where you have 10^23 of the little buggers in 22.4 cubic liters of air and they don't all panic and flee to one corner of the box. Not so great for derivatives, where you're lucky if you have a thousand points of data

And the problem with all the slicing and dicing of the different tranches into different products of risk is the large error bars in the estimation of risk levels. Plus the fact that if risk is going to commodified and sold to someone else, heck, it's not your own money on the line....

[ken]

As a general rule when everyone, including your hairdresser, says a bloodbath is coming in a market, it is already long over.

The best time to buy a house is when no one else wants to buy a house.

This does not mean to go out and just buy any old thing but it does mean that this is the perfect time to search out for the perfect house you would not otherwise get to see cause it would, in a normal market, get snapped up the day it was put on the market.

The trouble with waiting is that all the best homes get bought first, usually by someone with foresight, leaving others with the run of the mill bland cookie cutter homes that are still left unsold.

This is true also for mortgage securities. You can be sure that there is plenty of money waiting to see what price these 'unwanted' securities will come to the market. When priced you will find that they are not so unwanted after all.

Morgage securities, with the exception of interest only tranches, will always have some value. And if they were based upon high risk loans the securities were heavily discounted to begin with. There can be additional market driven discounting but only the toxic waste tranches will be discounted to pennies on the dollar. And those tranches make up relatively small pieces of the entire issue.

The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose.

[Ginger Yellow]

"With all these valuation problems, I'm really at a loss as to why these various instruments aren't standardized in some fashion so this sort of problem is far less likely to occur. (I don't see why any of these derivitives are a bad idea in and of themselves.)"

Well it's funny you should say that. The European Securitisation Forum has just put out a draft Standardised CDO Data File for consultation. It's by no means a panacea, but it should help investors compare one deal to the next at least and get a better idea of relative pricing. I assume the American Securitization Forum is doing something similar. The problem is that CDOs are often used to create bespoke risk/reward packages for investors, so there's relatively little incentive to standardise, especially since people have been lapping them up for the last few years.

The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose.

The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose.

The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose.

The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose.

"The use of leverage by hedge funds to buy up these mortgages means they will sell them at a loss for their investors. But the pension funds and others who are waiting in line to buy these securites will be buying them at very favorable prices. Some people win and some people lose."

This is very true. There's a somewhat similar situation in the UK subprime (or non-conforming) market at the moment, which went through a rough patch a year ago. For a variety of reasons people are shorting triple-B bonds aggressively, with the result that cash spreads have nearly doubled in the last month, despite the fact that the underlying loans have stabilised. The people who are still going long on cash bonds are going to make a fortune, so long as they don't pick a crap servicer.

[Bernard Yomtov]

tzs makes good points. One of the problems is that the markets for these things are often very thin, so the valuation models can fail, in part because things get discontinuous. There are price gaps that you don't see in the trading of Microsoft, for example.

Another problem is that even simple mortgages aren't that easy to deal with. The prepayment option makes them more complicated than debt securities like bonds. Start all the slicing and dicing and things get very tricky.

[tzs]

Righto about the repaying option complicating matters. I remember going through the calculations for valuing European-style options (fixed date) and then having to deal with the calculations for U.S.-style options, where the purchaser can exercise the option at any point up to the expiration date. A world of difference in complexity.

Also--sparse trading. Meaning discontinuities, lack of feedback from the market, etc., etc., and so forth. One reason why derivatives don't make any sense for a lot of stocks--the spread between the put and call prices is such that you can tell there hasn't been enough trading to get "equilibrium prices"

Add to this the fact that CDOs are usually individually financially engineered and it's even harder to get an idea on risk and as to whether they are correctly priced.

Remember when during the whole dot-com craze a whole bunch of theories came out about how the dot-com companies should be evaluated according to completely different criteria and forget the fact that none of the companies had made any money or looked to be making money in the near future? "Eyeballs", "network economies," and all the rest?

Now we're learning Yet Once Again that all those wonderful evaluations about CDOs contain the exact same silliness.

Yah, I know. This time it will be different. Trust me.