Arnold Kling  

I, Super-Senior CDO Tranche?

Rod Replies... Alan Blinder Testifies...

We know Leonard Read's famous "I, Pencil" essay, where he says that nobody knows how to make a pencil. Somehow, the market arranges it so that wood, graphite, dyes, cutting tools, and other inputs come together to make a pencil.

It feels very good to say, "Nobody knows how to make a pencil." Now, say, "Nobody knows how to make a super-senior tranche of a mortgage-backed security." Does it feel good? Or does it feel like there is a swindle in there, somewhere?

I don't mind that the folks who manufacture dyes that are used on the outside of pencils know nothing at all about graphite. But somehow I mind that the folks creating complex mortgage-backed securities know nothing at all about what goes on in putting together the loans. Why is that?

1. Maybe it's my problem. Maybe finance is no different from pencil-making, and I'm just being neurotic about it.

2. Maybe the financial process is more tightly integrated. The various steps are more highly interdependent, so it's harder to partition knowledge safely.

3. Maybe the financial process is more prone to catastrophic sorts of failures. If somebody makes a mistake somewhere along the line in the pencil-production process, then the error is bound to be caught before anything really terrible happens. In the financial process, errors can be disguised by good luck or misleading accounting, until it finally catches up with you.

Comments and Sharing

COMMENTS (23 to date)
Les writes:

It seems to me that there is a world of difference between "I, Pencil" and creating complex mortgage-backed securities.

In "I, Pencil" its all about the spontaneous order of free markets, and the cooperation between strangers without the slightest trace of government interference.

But in creating complex mortgage-backed securities its all about political "affordable housing" (NINJA mortgage loans), the government oligopoly of bond-rating agencies, the bank regulators, Fannie and Freddie, class warfare and other dysfunctional, disruptive political moves.

In brief, "I, Pencil" is all about harmony, whereas creating complex mortgage-backed securities is all about taking advantage of others and beggar-my-neighbor trickery.

Helder Ferreira writes:

An error somewhere along the line, compounds. With pencils or paper, consequences are easily manageable, with medicines or securities, consequences might be nasty. That's the only difference

Outer Life writes:

When it comes to the creation of super-senior CDO tranches, I think "I, Nuclear Bomb" is a closer analogy than "I, Pencil."

charles warren writes:

...nobody knew what went into a CDO, because the buyers didn't care. as long as it was AAA rated they were fine with anything...
(Goldman "cheated" to a small extent. they actually learned something about the product and made out big time.)

James writes:

I'll go with neurotic, or just mistaken. The meltdown occurred because of risks that people took. All the associated financial instruments just redistributed that risk, but the problem was taking those bad risks in the first place.

You can believe the counterfactual that if it were not for securitization, lenders would not have offered loans to unsuitable borrowers, but 1) that doesn't account for why the meltdown happened decades after securitization started and 2) that gives no account of the origin of securitization. It makes more sense to believe that people already wanted to extend high risk loans, and so they created all of the associated financial instruments.

ThomasL writes:

I'd toss in skepticism about the existence of realistic alternate uses of the component parts.

If any of the component parts have no (or extremely limited) alternate uses, and so presumably exist solely or primarily for the purpose of constructing whole CDOs, it becomes more worrisome that the whole it not fully understood.

Metal, wood pulp, graphite, gum, &c. all have so many alternate uses--a kind of self-certification of their merit as commodities--it is not at all worrisome that no one understands all the details about where they come from.

Pietro Poggi-Corradini writes:

The difference might be in the price signals. If a price along the line is hugely distorted that creates problems and maybe pricing financial instruments is harder than pricing dyes.

Felix writes:

But somehow I mind ... Why is that?

Because you just got burned by mortgage-backed securities and there hasn't been a recent case of millions of people being poisoned by chewing on bad yellow paint from pencils.

Now, as to whether you should be more worried about a 20 cent pencil or a 20 cent investment in a senior tranche - is another question. :)

Salem writes:

No-one knows how to make a pencil, but many people know how to assess the risks involved in buying one. It does not matter if no-one knows how to make a super-senior CDO tranche. It does matter if no-one knows how to price the risk.

The fundamental difference is that there are objective, external ways to test the quality of a pencil and the risks involved in using it. The only external way to test the quality of a CDO tranche is to suck it and see.

Chris W writes:

The dye-makers don't need to make huge assumptions on what the graphite-makers are doing in order to make their contribution, and likewise for everyone else in the pencil-making process. Each group contributes something, and each group can contribute something flawed, but bad graphite doesn't mean that the wood is bad too. Hmmm, maybe this doesn't really get to the point . . .
But I think these complex financial products were less like "I, Pencil" and more like "I, Mystery Stew". Mortgage originators had their flawed methods, and so the people packaging the CDOs had bad assumptions about how to appropriately spread risk. Ratings agencies had models for rating CDOs, but had bad assumptions in their models about how housing prices could drop and also about things like vintage (i.e. too high concentration of mortgages originating in a single year). And the investors had bad assumptions about how badly each group is incentivized to provide a quality product. And that's not just about how high the credit rating is because it is true that investors had an appetite for risk. But still, these products did not work as advertised. These financial products are useful to society because they can distribute risk to those most willing to bear them, but when investors are duped into taking on significantly more risk than they had planned or desired, this weakens the financial system and the economy.

Pandaemoni writes:

I guarantee that each and every securitization had an item styled as a "Credit and Collection Policy" (or words to that effect) that did in fact try to pin down how loans were originated (including standards used to extend credit and value assets) and how deliquent amounts were collected.

The "super senior tranche" was then just a question of writing into the documents a payment waterfall, with that tranbche always being the first to be paid (after fees of the collection agent).

So, there was an effort to know what was going on. Issues then certainly abounded like, (i) companies that turned a blind eye to violations of the Credit and Collection Policy...because you could make more money ignoring it; (ii) appraisals that were overgenerous, (iii) a lack of concern (beyong lip service) about possible real estate market downturns, and many others.

Whatever the case with pencils, I am sure they did know how to make a securitization and all its tranches from scratch. That they know how to do it does not mean that they were fastidious in their oversight of it, of course, just as a pencil manufacturer could manage the employees who cut and delivery the needed trees poorly.

In this case the incentives at the loan making level heavily favored loan generation in the short term, and the negative consequences of those loans defaulting were longer term, and not bourne directly or solely by the originators responsible. Sometimes you could claim that it all amounted to fraud and breach of contract at the origination level, but good luck proving that, and good luck suing the origination team within your own they'd likely have been part of the problem.

I still do not doubt that if you asked anyone in the asset finance group at Citibank how the loans were generated, he'd point to the Credit and Collection Policy for that deal.

Charlie writes:

I think dye is the wrong part of the pencil process for you metaphor; we should use lead/graphite instead.

At some point in the pencil process, all the various parts are put together: dye, wood, graphite. The company that does that does not need to know about how dye is made, wood is cut, etc. But they do have to make sure they use graphite instead of lead.

Similarly, the people creating mortgage back securities should make sure they don’t use lead (bad loans that will fail and blow things up) but instead use graphite (good loans that won’t blow up).

Ivin Rhyne writes:

Dr. Kling,

Might I suggest that your uneasiness derives from a fundamental (but perhaps unrealized) understanding of economic first-principles. When I talk about economics with non-economists, I start by explaining that people purchase “products” and “services”. Economists tend to look at these products as a bundle of “goods” and “bads”. For example we can theoretically “slice a Twinkie” multiple ways. We can look at it’s calorie content, it’s sugar levels, it’s aestetics, fats, etc. ad nauseum.

Because we can’t actually taste and test a Twinkie before buying it, we have to trust that the bundle of goods and bads of the specific Twinkie on the shelf of the 7-11 at two o’clock in the morning is roughly equivalent to the hundreds of Twinkies we have eaten in our lives. This predictability of the bundle of goods in a product is what I call its level of quality.

The idea of quality, in the story of I, Pencil, is never used explicitly or implicitly. But I argue that we rely on the knowledge, both personal and institutional, of the final assembly plant to provide us with a writing utensil that performs as expected. In other words, that has a predictable bundle of goods and bads. In the case of the Super-Senior CDO tranche there is no reason to believe that those assembling the final product have the personal or institutional knowledge to provide a product that performs predictably. And thus your uneasiness.

Put another way, while no one person knows everything necessary to make every part of a pencil, we have confidence that SOMEBODY knows how to assemble them correctly. There is no indication that anyone knows how to assemble these derivatives correctly.

In the marketplace if your favorite pencil manufacturer suddenly started making sub-par products, you might eventually switch to another brand. In the case of exotic financial derivatives, now that you have seen the poor “quality” of the product and seen just how inept those on the assembly-line floor are, it would be quite understandable for you to beat a hasty retreat away from brokers bearing derivatives.


Ivin Rhyne

Kevin writes:

Arnold you're actually incorrect that nobody knows how to make a super-senior tranche of a mortgage-backed security. It can easily be done by one person with a calculator, or if you prefer, one person with a rock and a cave wall to scratch on. This is why your #3 is spot-on.

Bill Drissel writes:

Somewhere in here we need to mention the ratings which somehow contributed to the valuation of the CDOs, tranches etc.

It seems to me that the rating agencies utterly failed to appreciate the downside potential. I wonder why they haven't been fired like Enron's auditors were.

Hyena writes:

At a basic level there is no difference. Financial firms manufacture securities, including CDO tranches. There can be issues with suppliers which lead to product failure, sometimes catastrophic and with consequences which are difficult to contain.

It's harder to think of this occurring with pencils because we're not used to it being a danger. Imagine though a case where the paint on pencils contained a significant quantity of lead. This could easily happen and we've had examples with toys in the last 5 years. This is probably a catastrophic pencil failure and it, too, will have far-reaching consequences which will prove difficult to contain.

Whether you're being neurotic depends on how much you think you should worry about catastrophic CDO failure v. leaded paints on pencils among other scenarios.

Daniel Kuehn writes:

I think it would be a little too much to approach everything in finance this way, but I think the source of the discomfort is merited. The fact is, it's wrong to say "no one person knows how to make a pencil". They do know - they know all that they need to know from price signals. How do you make a pencil? You hire a bunch of trained professionals and suppliers to put it together at the going market rate!

The problem with MBSs is that there is a real concern about the operation of the price mechanism.

Now, if you believe that that problem originated from government action alone then that's an easy foil for you. If you believe natural bubble psychology was involved then maybe that's a reason to be wary of certain other products too from time to time.

floccina writes:

Maybe the only problem with CDOs is they have not been around long enough to enable judgments of quality.

Colin K writes:

It's not just price signals. Manufacturers require suppliers to deliver materials that conform to well-defined standards of quality. I can take a rod of graphite, test it's physical properties, and be fairly certain it will make a good pencil, even if I don't know how to manufacture graphite rods myself.

My question is, were the quality tests for the raw mortgages insufficiently precise, or were they actively ignored? Or, did the CDO stack become so complicated that it was impossible to know the quality of mortgages with sufficient precision to prevent collapse?

Noah Yetter writes:

One word: leverage.

Noah Yetter writes:

OK maybe a few more words. Fundamentally, the producers of financial securities are more like casinos than pencil factories. When you buy into a CDO tranche you are literally gambling. The difference is that while casino games are based on calculable probabilities and therefore known risk, the business (and government) world that forms the foundation of financial gambling is defined by uncertainty which cannot be calculated. Yet we treat it as if it can be. In doing so, not only are we telling a lie to ourselves we are institutionalizing that lie through the establishment of the ratings agency oligopoly, requirements to use it, and requirements to hold securities meeting a certain rating threshold.

Jeff writes:

There is no mechanism in place wherein you can design a pencil to fail and then bet that it will.

Ivan writes:

How about:
For how long have we had the expertise to make pencils? What happens if we start making deffective pencils?

Now use the same logic to your CDOs


Comments for this entry have been closed
Return to top