Crowds and market caps

by John Q on July 6, 2005

Following up Henry’s post, I happened to reread a passage from James Surowiecki’s “The Wisdom of Crowds in which he discusses the stock market’s reaction to the Challenger disaster, the crucial point being

Did you know that within minutes of the January 28, 1986 space shuttle, Challenger, disaster, investors started dumping the stocks of four major contractors, Rockwell International, Lockheed, Martin Marietta, and Morton Thiokol, who had participated in its launch? Morton Thiokol’s stock was hit hardest of all … the market was right. Six months after the explosion, the Presidential Commission on the Challenger revealed that the O-ring seals on the booster rockets made by Thiokol became less resilient in the cold weather, creating gaps that allowed the gases to leak out.”

It struck me reading this, that I’d heard of Rockwell, Lockheed and MM in many contexts, but I’d never heard of Morton Thiokol. It turns out that they are a specialist builder of booster rockets and similar items (they’re now a division of ATK).

This seems to suggest a prosaic explanation of the market reaction. Whatever the cause, the space shuttle program was going to be shut for a long time. This would do a bit of damage to everyone involved, but much more to the rocket specialist Thiokol than to the other three big diversified companies.

The ATK website indicates that they still have plenty of shuttle contracts, so it seems as if the faulty O-rings didn’t do them much long-term damage over and above the effect on the shuttle program.

I haven’t got the full book to hand and I haven’t read the study cited there , so it may be that this explanation has already been ruled out in some way, but I thought the easiest way to find it was to post and see what response I got.

{ 15 comments }

1

JR 07.06.05 at 11:51 pm

Daniel Gross covered this in his Slate column “Moneybox” in 2003, including your suggestion. You should know that the Thiokol engineers were well aware that the rubber O-ring gaskets between the booster sections might fail in cold weather and strenuously argued that the January 28 launch should be delayed. They were overruled by managers after a high-level conference call on the evening before take-off(perhaps NASA was feeling political pressure – the Challenger was carrying school-teacher Christa McAuliffe, and she was a high public relations priority for the administration’s education initiative. President Reagan was scheduled to give his State of the Union address on the evening of the day of the launch – January 28, 1986 — and intended to highlight McAuliffe in it.)

The key whistle blower, Thiokol engineer Roger Boisjoly, later said that he knew immediately what had caused the accident. He was not alone -many engineers within Thiokol and NASA were aware of the risk of O-ring failure and were unhappy with the decision to launch. So, as soon as the accident occurred, there was a substantial community of disaffected and guilt-ridden engineers with knowledge of Thiokol’s culpability.

A poster calling himself “pollymath” on Slate’s “Fray” wrote in response to Gross’s article on August 8, 2003 (still available):

“There was information publicly available about the probable causes of the Challenger explosion almost immediately – it was on the internet.

“I was working for a computer company at the time (college leave term job) and had access to the bulletin boards that were the ancestors of places like the Fray. After a few shocked minutes watching the news on a TV that was set up in the cafeteria, I started reading the space science and NASA interest boards. Within a very few minutes (I recall it was about 20) there were postings from people with legitimate expertise saying that given the weather conditions that morning and other factors, the rubber O-rings were most likely to blame. Please remember, the content of these boards was provided almost entirely by scientists and engineers since posts were identified by real e-mail addresses and those who posted when they had nothing to say were quickly abused into silence. I read the stuff, but I didn’t dare contribute.”

Thus there appears to have been a clear route for insider information to make its way to investors immediately.

The Presidential Commission (the “Rogers Commission”) took six months because it did not want to find the truth. It valiantly attempted to obfuscate the obvious and relieve all parties of responsibility. It was unable to do so because of the work of Commission member Richard Feynman. His best remembered but by no means only contribution was a brilliantly theatrical ploy that permanently focused media attention on the O-rings. In a televised hearing, he put a bit of flexible O-ring material in a pitcher of iced drinking water set out for the commission members. Then, casually interrupting a bit of scripted testimony from a corporate witness, he fished it out and showed that when chilled it snapped like a dry twig. Feynman was on the commission as window dressing- a Nobel-winning professor with a reputation as an eccentric. Rogers and the other lawyers and fixers on the commmission thought they could control him. They were wrong. Although some members of the commission continued to blind themselves to the obvious, Feynman’s maneouver made it impossible for the Commission to ignore the truth.

I strongly recommend that you read the Challenger section in Feynman’s amusing memoir, “What Do You Care What Other People Think?”

As for Surowieki, his theory about crowds and markets comes down to a belief in magic. Yes, markets can transmit information quickly, but where there is no information markets cannot create it – and they can spread false information just as speedily as true.

2

cornellian 07.07.05 at 9:10 am

Well after the above comment I don’t have much to say (I second his reading recomendation though).

In terms of still having contracts, how many specilty rocket booster companies are there? Aside from this one failure, which was caused by a maliceless error that their engineers made efforts to correct, they are presumably good at what they do. Hence the logic of everyone changing suppliers over this seems flawed to me. I would be willing to bet that they won’t have another o-ring issues. I’m also assuming that rocket buying is done by the more rational types not a generic easily spookable market.

On the topic of the commision, I never thought about it before, but what on earth were lawyers doing on it? I don’t think that there could have been anyway that a legal issue would have caused the accident

3

James Surowiecki 07.07.05 at 9:22 am

John, I mention this possibility as well in the book — “perhaps the company’s business seemed especially susceptible to a downturn in the space program” — although Maloney and Mulherin don’t raise it (at least not in the first version of their paper). The greater dependence on the space-shuttle business may explain part of the gap, but Morton Thiokol was not, in 1986, solely a “specialist builder of booster rockets.” They were actually a small conglomerate, which also made auto airbags, salt, and chemicals. (After the shuttle disaster, the company broke up into two parts.) In fact, even after it had become increasingly clear that the company was responsible, Wall Street analysts were saying that the effect on the company’s bottom line would not be all that important, and arguing that the market had overreacted, when in fact it had nailed the right answer. So I don’t think the thesis that the market’s reaction was primarily a function of Thiokol’s unique dependence on the shuttle holds.

4

Ophelia Benson 07.07.05 at 12:42 pm

What happened with Morton-Thiokol is a really fascinating case study of hierarchy trumping expertise. I was particularly fascinated by it at the time because I’d recently seen hierarchy trumping expertise where I worked, so I paid a lot of attention. I especially remember a manager who was also an engineer being told by someone a bit farther up the ladder that it was time for him to ‘take off his engineer hat and put on his manager hat.’ In other words – take off the ‘hat’ that tells you the O rings will fail and the shuttle will explode, even though you’re right about that (and you know you’re right, and you know why; you know the O rings really will fail), and put on the ‘hat’ that tells you to launch anyway for reasons that have nothing to do with the O rings. In other, other words, act as if changing ‘hats’ somehow makes what you know about the O rings beside the point. How would it do that?

5

Jake McGuire 07.07.05 at 12:58 pm

What was the eventual effect on Morton Thiokol’s bottom line? It’s a pretty well-known phenomenon in government contracting that if you foul something up really badly you get the boot, but if you only sort of mess it up you can pretty reliably get more money to fix it.

6

James Surowiecki 07.07.05 at 1:44 pm

Maloney and Mulherin (authors of the original study) estimate that the effect on the company’s bottom line was roughly close to the decline in the company’s market cap on the day of the crash ($200 million). The company paid a few million in settlements, forfeited $10 million to NASA, had to do $400 million in repair work for no profit, and didn’t enter the bidding for the next round of booster rockets (which it presumably would have had an inside track on), losing around $150 million in foregone profits.

7

Blar 07.07.05 at 3:17 pm

I went looking for that slate article as soon as I read the post. Then I opened the comments and saw that the first commenter was already on top of things. I’ll just add this link to the slate article, agree that the market can’t take information into account unless some of the people trading in the market have that information, and leave the discussion to people who know more than me.

8

novalis 07.07.05 at 4:08 pm

jr, why wasn’t it on Usenet? I looked up the archives on Google for Jan 27-30, and nobody mentioned O-rings or Thiokol with Challenger.

9

JR 07.07.05 at 6:57 pm

Novalis, I don’t know anything about how Usenet works so I don’t know the answer. Maybe someone else does.

Re Surowieki’s comment- it happens that there have been not one but two shuttle disasters. They had remarkably similar causes, rooted in managerial failure to recognize risk rising from failing systems. The differences were (1) in Challenger, engineers had argued against launch for fear of the precise accident that occurred, while in Columbia this did not happen, and (2) in Challenger the accident was caught in detail on videotape, while in Columbia the accident went unseen. In the case of Columbia the loss of foam insulation had been caught on the video of the launch, but the engineers discounted it. Therefore, there was no ready source of information for investors and traders immediately after the accident.

In the case of Challenger, the market appears to have had immediate and accurate information and discounted the share price of Thiokol accordingly. In the case of Columbia, in the absence of clear information, the market acted irrationally and discounted the wrong stock. This is from the website of the US Navy’s Center for Contemporary Conflict, a unit of the Naval Postgraduate School:

http://www.ccc.nps.navy.mil/rsepResources/si/sept03/terrorism.asp

“This concern about the usefulness of markets in providing national security related intelligence was born out by the recent Columbia shuttle disaster, where the market’s response was not so accurate. Among the publicly traded NASA contractors, the biggest loser was Alliant Techsystems Inc, the current owner of Thiokol, which made the shuttle’s booster rockets. Alliant’s stock fell almost exactly the same amount that Morton Thiokol did after the earlier Challenger crash—about 11.66 percent. Boeing, which now owns Rockwell International, a major NASA contractor, fell 1.5 percent and Lockheed Martin fell about 3 percent (Gross, August 8, 2003). Gross (August 8, 2003) notes that, ‘The market —perhaps remembering Thiokol’s implication in the prior disaster—swiftly punished Alliant. Wrongly, it seems. Thus far, attention has focused on the performance of foam insulation lining the external fuel tanks, which were made by the Michoud unit of Lockheed Martin. The Market may be efficient. But it can also be emotional. Did traders with long memories rush to sell Alliant disproportionately because Morton Thiokol was deemed responsible for the Challenger disaster? Almost certainly.'”

10

James Surowiecki 07.07.05 at 8:14 pm

JR’s right that the market made the absolutely wrong call in the case of the Columbia, but I don’t think it was because of a lack of information. I think the CCC’s conclusion (which echoes the one I suggest in my book) is more accurate: traders relied on past experience and herded toward the most obvious answer. There was, in fact, quite a bit of information (at least within the NASA/contractor community) suggesting that the fuel-tank foam was responsible for the disaster. While the NASA Mission Management Team discounted this idea, plenty of lower-level engineers were concerned about the possibility.

11

David 07.07.05 at 9:30 pm

I don’t think it should be so surprising that a sell off of these firms would take place just after such a disaster, if there were any serious investors with sizeable holdings in the firms, nor is it obviously an example of a wise crowd. Anyone with a large position in such a firm would be aware that a space launch is a very risky undertaking, and that a major failure would be bad news for the supplier firms. Anyone with such concerns would surely monitor a brief event like a launch, and would understand that once the disaster took place, it was time to dump some stock, before someone else did. My question is, why would anybody buy the stock at such a time?

12

JR 07.07.05 at 10:20 pm

Mr Surowiecki is making my point for me. With respect to Columbia, I didn’t say there was no information; I said that the damage to the foam was seen on a video but the engineers discounted it, so that there was no “ready” source of “clear” information. Because the information was confused – concerns among engineers, rather than certainty among engineers as in the Challenger disaster — the market got it wrong. Faced with uncertainty, traders applied a crude heuristic — “the present resembles the past” — which had no value in discerning the correct cause of the accident. Surowiecki calls this “herd behavior.”

One might ask, if markets respond to uncertainty with herd behavior, what good are they? The answer is that markets do have one very important advantage over hierarchical decision-making structures: they generally do not have a long-term preference for false information. In a market, reliance on false information generally is punished and reliance on truth is eventually rewarded. The contrary is often true in hierarchical organizations. Top management often affirmatively does not want to know the truth, because decisions are driven by hidden factors that cannot be revealed (for legal, political, or other reasons). Therefore decisions must be justified by appeals to other factors which can be supported only by falsified information. Feynman made this clear in the case of Challenger, when he pointed out that engineers estimated the odds of a shuttle disaster at 1 – 100, while top management asserted that the odds were 1 – 100,000 — a fantastic figure that had no engineering basis but was necessary to defend various contradictory but politically mandated goals such as cutting safety budgets while putting civilians into space.

Being a good subordinate in a hierarchical organization means understanding what false information is desired at the top and knowing how to present it in such a way that it still conveys a subtext of enough of the truth that the top management remains marginally in touch with reality. We have seen these considerations at work in the recent intelligence “failures” over WMDs, which were not failures at all but the spectacularly successful dissemination of the desired false information.

Markets composed of independent, self-interested traders generally do not suffer from this preference for self-delusion over truth. However, markets cannot create information that is not known to at least some participants, and to believe otherwise is to believe in magic.

Further, crowds other than markets may rationally prefer falsehood if falsehood is better rewarded or has other benefits, such as making it possible to avoid unpleasant decisions or justifying one’s own prejudices or privileges. Americans believe that we spend about 15% of the federal budget on foreign aid; the correct figure is less than 1%. Britons believe that about 21% of the population is composed of immigrants; the true figure is 8%. These false beliefs persist in the face of ready availability of accurate information because they are useful to those who believe them.

13

James Surowiecki 07.08.05 at 12:03 am

JR’s point about the tendency of hierarchical organizations to reward self-delusory or self-serving forecasts and conclusions rather than true forecasts is well-taken, and is one of the main reasons why I think internal markets or other collective decision-making mechanisms can be so valuable inside organizations.

I’m not sure, though, who he’s arguing with when he says “markets cannot create information that is not known to at least some participants,” since I don’t think anyone believes otherwise. In TWOC, I write: “For the group to be smart, there has to be at least some information in the ‘information’ part of the ‘information minus error’ equation.” There’s no magic being invoked here, just the ability of markets (or other similar mechanisms) to aggregate distributed information quickly and accurately.

14

Jon H 07.08.05 at 12:44 am

“Wall Street analysts were saying that the effect on the company’s bottom line would not be all that important, and arguing that the market had overreacted, when in fact it had nailed the right answer”

Given a big firey explosion, a good heuristic is to blame the biggest, most firey things involved. Which would be the big firey solid rocket boosters.

It would be far more impressive if the market identified a cause that wasn’t quite so glaringly obvious.

15

dsquared 07.11.05 at 5:43 am

Missed this one in all the noise, but the betting crowd dropped a bollock on predicting the 2012 Olympics; Paris was odds on favourite all the way in.

Comments on this entry are closed.