More on derivatives
Last week I wrote A Simple Explanation of How The Use of Derivatives Created The Great Recession.
I want to clarify things with an analogy that at first glance may seem unrelated. Let’s talk about poker. Its a game that seems to have grown in popularity in the last decade, especially with the advent of on-line poker, even if the rate of increase seems to me (and I haven’t had time to pay much attention lately so take it for what it is) to have slowed in recent years.
Poker is different from most casino games in that you only play against other players, not against the house. The casino – online or glitz and mortar takes a rake, which is a piece of each pot. In a tournament, the casino’s rake comes off the sum total of the entry fees (“buy-ins”); what is left after the rake is the prize pool.
When the rakes are small enough, and they usually are, a good player can have a positive expected return. So… investors willing to risks can finance players by covering the buy-ins of the players in exchange for a percentage of their winnings. Though poker is a game of skill, luck still matters a great deal, so it probably makes sense to spread one’s investment either across a large number of tournaments by the same player, or across a large number of entrants in the same tournament. I know of examples where this is being done informally, and I have seen business plans for web-based versions of these approaches on a wider scale, but the market, as such, is currently very unsophisticated. Its good for the players – they can enter more tournaments, at lower risk. Despite giving up a share of their winnings, they stand to make more money by being able to play more tournaments and tournaments with more expensive buy-ins. I understand that one well-known professional player, Chris Moneymaker (and yes, that is his real name) produced a monster return both for himself and his backer by qualifying for and winning the World Series of Poker with this sort of financing scheme.
Its not difficult to imagine how such a market could evolve… if the money is there. Over time, buying shares of a player’s winnings in a package of tournaments (say ten) in exchange for a share of the player’s winnings might become common for good players. Over time, some enterprising individuals who know and are trusted by many of the better players might find themselves a niche aggregating players together. The result – instead of being able to buy into the earnings of a single individual in a ten tournament package, you could buy into the earnings of ten or fifteen players over ten (or more) tournaments. It diversifies the risk a bit for all involved, especially if there was some sort of limited prize sharing among the players in the package. Of course, that’s a lot more expensive than financing one player… so that would lead to two additional developments.
Development 1 – exchange markets would develop. That would allow for liquidity (investors could more easily get into and out of investments). It would also allow securities to change value over their lifetime. (If none of the players cashed in the first tournament or two, the value of the security would drop, for instance.)
Development 2 – margins. Some deep pocket entities would be willing to loan investors money in order to buy into such packages.
If the market got to this point, perhaps some really sophisticated investors would jump in. They might find – using sophisticated math – that it pays to pepper each package of players with a long shot or two. They might notice other thing – perhaps their models would show that a given security contains too many players with a similar style, or who likely to knock each other out of a given tournament. Short it.
Meanwhile, the aggregators might notice the growth in demand, and start packaging more marginal players together. Its all good though – someone with a Ph.D. in econ (or even better, Math or Physics) could demonstrate that a bunch of marginal players put together do, on average, do almost as well as the top players, and hey, its Poker, all you need is a chip and chair. And a bunch of MBAs could put that proof on 17 powerpoint slides and sell it to the county treasurer in any number of locations in this country (and elsewhere), and we’re not just talking the sticks either.
If the money keeps flowing in, and the story resonates like tulips, Florida swampland or Pets.com, eventually you have rubes all over the country who think they’re going to finance a comfy retirement by plonking down for the buy-ins for some slicks who will draw to an inside straight every time. And you can even get insurance on your expected winnings, so what could possibly go wrong? Especially when every taxi driver and waiter in the country, not to mention every editor at the National Review, will happily give you advice on how best to play the derivatives market for poker. Heck, the guy in the Oval Office starts suggesting that this is probably what the country should be doing with its Social Security – everybody could direct their own account.
That’s about when the doo-doo will hit the fan. And here’s how it starts… a big investor loses big time. His package of “Poker Geniuses from Blowing Rock, North Carolina, Class B-7” comes up a dud. Unfortunately, the dude is leveraged up 50 to 1. Which makes sense, because he couldn’t lose so it was guaranteed money. Unfortunately, he did lose. So he’s gotta pay his creditors, and he does so by liquidating some other positions. Due to leverage, a lot of positions. That pushes down the value of those positions, and similar securities. But it does something worse… for some reason, it makes a few people re-evaluate the likely pay-offs of the securities that they themselves are holding. And if “Poker Geniuses from Blowing Rock, North Carolina, Class B-7” can come up empty, what about the “Potted Plant Poker Playas of Playa del Rey, L Team?” The folks holding the last tranche to pay off on something they just realized is a dog, and which has already lost value on the market when the big guy started selling left and right to meet his obligations – what are they going to do? Bear in mind, Goldman beat them out the door by six months, and they’re leveraged up. 50 to 1. And a payment is due at the end of the week and nobody is answering their calls.
So the market collapses. In fact, a few other markets collapse – who knows who has lost so money on Poker derivatives that they are going to go under, and that big multinational that wants a bridge loan does have a division that plays in this kind of thing, right? No bridge loan for you. It turns out that some of the big guys – the folks who aggregated the players in the first place, and thus understood what sort of dogs were being financed in the first place – bought into their own sales pitch and are holding piles of this crud. Inevitably, there’s a bail-out for some of the big boys. The Fed or the Treasury takes the crud off their hands, and gives them top dollar to boot. After all, they’ve convinced the government that without that market, nobody is going to finance a company that manufactures patio furniture in Indiana. Of course, none of them does get around to loaning money to that company that manufactures patio furniture in Indiana, but that’s a trivial detail. Enough of the big boys are made whole that everyone who counts is happy. And there is little thing called blackjack…
So, what did I miss?
Then there is always insurance.
Someone totally convinced that the system is practically foolproof in the long run offers insurance on investments. The insurance makes the investor feel even more secure–and the guy offering insurance thinks he has a perpetual moneymaker.
What am I missing? Where’s the derivatives? You’ve just described the classic process of bubble and bust with a particular technique — securitization — being used to create dodgy overvalued assets. But there’s always some sort of technique for dodgy overvalued assets to be produced using some form of credit during the bubble — see e.g. recent internet bubble or the 1720 South Seas Bubble and all the projector joint-stock companies invented out of thin air. But the problem isn’t with securitization, or internet stocks, or joint-stock companies. It’s with the availability of credit, based on rising asset prices, that fuels the bubble. At some point, when reality strikes, as you describe, the size of the bust is then exacerbated by the mutually reinforcing calls on credit and declining asset values.
Now, most recently we also had a market in side-bets, which clearly added a ton of fuel to the fire, because the side-bet market financing and players weren’t segregated from the main bubble market. And as Stormy indicates, the insurance aspect — when in this case the insurer didn’t actually have to have any reserves to back up his potential liabilities, just tried to cost-free monitize his tripple A rating — created a house of cards, with contagion spreading across the global financial system at a rate never before imagined. But the mechanism you outline seems to me to be just basic bubble and bust stuff.
Kind of an obvious miss there. I feel silly not including insurance, what with AIG and the monolines. Thanks.
“Where’s the derivatives?”
Writing these posts helps me clarify my thoughts, and sometimes I leave out something. Clearly I had to be more clear…
The derivatives are the packages made up of the payoffs of many players on many tournaments. What I seem to have not hammered in the story was this: the existence of these derivatives, and the fact that they could be shown, mathematically, to be safe, is what allows the market to expand exponentially. Its kosher to spend the pension fund’s money on something safe, but they’d be leary as all get-out about handing the money to some fast talking dude in Vegas. But underneath it all, its the same thing.
Additionally, the fact that these instruments can be shown to be mathematically safe is what allows the insane leverage.
Those two features – that the product is respectable and that there’s insane leverage means you have more than a bubble. They allow the bubble to grow much larger than, say, the tulip bubble or the sale of Florida swampland or even the dot coms. And the leverage adds one more feature – when the losses come in, the losers lose much more than they have. A loss doesn’t just drive them bankrupt, as I pointed out last week, it drives the winners of the bets bankrupt too, because they aren’t collecting, and they have their own obligations. Think of it this way… if you win a $10 bet, and the guy can’t pay, you’re OK. If you win a $1 million bet and the guy can’t pay, you are in trouble, because between the time you won it and you found out you aren’t getting it, you probably bought yourself something expensive.
***and the guy offering insurance thinks he has a perpetual moneymaker.***
Hey, Li’s Gaussian Cupola guarantees that the insurance can’t fail. You sayin that the math is no good?
***and the guy offering insurance thinks he has a perpetual moneymaker.***
Hey, Li’s Gaussian Cupola guarantees that the insurance can’t fail. You sayin that the math is no good?
Ah, I see what you were saying. But a package of a diverse set of payoffs isn’t a derivative. It’s just securitization of a bunch of claims. The fact that the claims themselves are contingent (involve the performance of bets) doesn’t make them any less a financial asset than a railroad bond or a home mortgate, and packaging a bunch of claims is just securitizing them. The potential buyer of a piece of a securitized package of claims like your package of pay-offs is provided with (a) the benefit of diversification for a modest sum relative to the cost of assembling a diversified set of claims on his own, and/or (b) the ability to choose a risk and maturity or cash flow structure (tranche) which meets his preferences, again in ways the buyer couldn’t readily reproduce by buying on his own.
Now, in our most recent bubble, lots of high-level math was used to achieve diversification and cash flow objectives in securitized packages. I agree with you that the high-level math superficially appeared to make the instruments “safe” in the sense of a very high probability they would achieve their cash flow (risk and maturity) goals. The problem wasn’t with the math. It was with the underlying assumptions — (1) that there wasn’t a bubble in the underlying asset classes (mostly residential and commercial real estate) and/or (2) even if there was a bit of a bubble, that the assets in the securitized packages weren’t correlated, so the diversification designed by the quants would indeed be able to reduce default risk to meet rating objectives even if some real estate markets were experiencing bubbles.
I certainly agree that the bubble process was compounded — both in size and duration — by the magic dust sprinkled on new dodgy securitized assets by the quants. Althoughy ou’d be surprised how many fly-by-night “real assets” like tulips or Florida swamp acreage can be “manufactured” at the height of a bubble. But you’re right that a bubble in “financial assets” can inflate more and longer than one based on “real assets”.
Unfortunately, both the rating agencies and the sell-side folks who manufactured these new dodgy assets embraced the quant methodologies without challenging the underlying assumptions. And all too many portfolio managers took the sales spiel on faith, comforting themselves with high ratings that met their regulatory investment criteria. So in the most recent bubble, the ongoing creation of dodgy assets was facilitated by the quant magic dust since it seemed on its face so plausible and was validated by regulation-mandated ratings.
So we don’t need derivatives to “explain” the recent bubble. That’s not to say, however, that derivatives weren’t important to the size and duration of this particular bubble and the speed of contagion and magnitude of asset-price declines when the bubble burst. Here’s what I’d add to your tale. [con’t in next comment]
[comment con’t] Now to derivatives. In your example, a derivative would be a side-bet between two kibbitzers that pays out based on some aspect of performance related to the original claims/assets (either individual or securitized) but that doesn’t actually involve receiving payment from any tournament or player syndicate. In the most recent bubble, there were folks who pointed out the emperor had no clothes. Some of them even used derivative markets to make side-bets against the bubble. And they found greedy idiots like AIG and the monolines to take the other side of their bets. In the final stage of bubble madness, those bets were themselves used to create new (very, very dodgy) assets that, with quant magic dust, were deemed to be “safe”.
However, because both the asset (bubble) markets and derivative markets were linked by an orgy of credit creation (leverage) and cross-market collateralization (producing margin calls in a declining-price market), when the air started leaking out of the bubble one place, contagion spread across all markets. As you pointed out, even the folks who made the right side-bets or tried to “insure” against asset-price declines were at great risk of losing huge amounts. Since no one knew where the game of musical chairs was going to stop and the ulitmate losses realized, all credits were suspect. So we had a “bank run” in the shadow banking system (repo market) which was only stopped by the Fed temporarily substituting itself for the money-markets and plugging the biggest holes in the derivatives markets.
So as I see it, derivatives were facilitators of the size of the asset bubble and subsequent crash and they fueled cross-market contagion. They may indeed have been the ingredient that turned a nasty situation into a systemic meltdown. What they shared with securitization was that a financial innovation designed to manage and reduce risk turned out to have compounded and distributed risk in very dangerous and unpredictable ways.
So, what did I miss?
Actually, that wasn’t bad. If you were saying that financing innovations enhanced the housing bubble, I would agree. But the question is, to what degree, and what was the contribution of which innovation. You seem to blame everything on CDS (I think? or what other “derivatives” do you have in mind?). Also, remember AIG stopped writing CDS in 2005.
And remember, CDS brought down one (1) firm. Regular old mortgages brought down Wamu, Wachovia, Bear, Lehman, Countrywide, Fannie/Freddie, and scores of smaller banks. The ratio is probably 50:1. So you have to ask yourself: was it the mortgages or the CDS?
“Ah, I see what you were saying. But a package of a diverse set of payoffs isn’t a derivative. It’s just securitization of a bunch of claims. The fact that the claims themselves are contingent (involve the performance of bets) doesn’t make them any less a financial asset than a railroad bond or a home mortgate, and packaging a bunch of claims is just securitizing them.”
Strictly speaking, derivatives are not necessary for my story, and I was being sloppy with language. In practice, however, derivatives are going to show up in my points 1 and 2. A derivative of some sort can further remove some county in Nebraska from the reality that it is gambling on poker players. Make something where the payoff looks a bit more guaranteed than it would from just packaging these players together and your proven to be safe security becomes “you won’t lose money except once in 1.7 trillion years” derivative on a good Powerpoint slide. Here’s the pitch: “you lose a touch off the top of on the returns of the safe security, and get a guaranteed payoff anyway if none of your players come in the money.”
But more importantly, the leverage is often easier to accomplish with derivatives, particularly when some of them are traded in opaque markets. It seems that people keep falling for this stuff, but maybe, just maybe, someone will ask a question or two if they see what the big boys are actually trading amongs themselves.
Yes. You’ve stated it much better than I did.
Yes. You stated it much better than I did. But we both forgot to mention one thing (that was intended for the original post)… this process does not require the underlying bubble (housing, poker, whatever) to come to an end in order to collapse. It just requires big enough losses by some players. And in these zero sum games, those are guaranteed to occur at some point.
I didn’t emphasize it in the post, and meant to… one point of this example is that you don’t need the underlying bubble to come to an end for everything to fall apart.
In 2007-2008, the bubble collapse brought down the financial instruments. In my example, the underlying market remains intact but you still have a financial crisis due to bad bets. (And yes, we’ve seen it in the real world. One example – LTCM. Sure, Russia tanked on them, but it was still a bet by them that supposedly created this emergency on Wall Street though other players weren’t even involved except through LTCM).
You have noted more than once that AIG stopped writing CDS contracts in 2005. That may very well be though I have not been able to find verification of that fact. While searching around I did come across this very interesting suit filed in California against AIG. It provides scattered bits of information none of which is verified on its face, but it is interesting for waht is claimed. It reads lioke a RICO indictment. It also makes a good point regarding the losses suffered by AIG investors who bought in around 2004-2006. Too bad for them.
It is a very good story. If we substitute housing, the story is formally just about exactly the same. In reality, instead of one or a few winners against a pool of losers, housing was thought to be a pool of winners against one or a few losers. The math can tidy up that difference through adjusting expected returns. The system you’ve described means we don’t care whether there are few winners or many, formally, as long as we know the expected return.
The reality is that the housing market could become subject to an upward creep in expected return, but realistically, the market for participation in poker stakes should not. We know the expected return from the poker tournament – setting aside crime, rule changes and the like, the expected return doesn’t change. In a nice little irony, the derivative market for participation in poker stakes is less speculative than derivatives for mortgages. Your example is isn’t terrible enough for the mortgage market.
I wouldn’t read a plaintiff’s suit for info. Maybe the decision.
There has been a LOT written about AIG. This is particularly good: http://www.washingtonpost.com/wp-dyn/content/article/2008/12/30/AR2008123003431.html?sid=ST2009013000235
It’s the third of 3 parts. The first two are here
one point of this example is that you don’t need the underlying bubble to come to an end for everything to fall apart.
This I gotta see.
Yes, the math is no good, See Taleb and his black swan. The assumptions under modern econoic theory are alot more arbitrary than anyone in Econ wants to admit.
I agree and there will always be big losses with deriviatives and securitization because, by nature, they create bubbles and spur dangerous amounts of leverage because they disassociate cause from effect. Their very construction hides any fundamental understanding of their probable effect from the recipients and intermediaries of these complex amalgams of risk. IN this process, mathematical modeling is based upon faulty assumptions and are used as a sales tool to convince otherwise conservative fidicuiaries that these products are supported by sound economic and financial principles when they are decidedly not. These constructs distribute risk in ways that are unpredictable and dangerous.