"Stick to what you know"?

In management strategy discussions a familiar command is for firms to do what they have a core competency in. And it sounds sensible enough. Yet I've recently come across examples of firms that departed from "what they were good at" and became world famous as a result.

• Would we have Honda stop making cars and go back to just making motorcycles?
• Should Vestas get out of the wind turbine business and return to home appliances?
• And Nokia? Back to running a wood pulp mill?

Beauty is the eye of the ...group?

Science Now ran a very interesting story this week that I heard on the Science Podcast. It is based on an article that appeared this month in Neuron.

The researchers asked 21 young women to rate the attractiveness of 222 female faces while lying in an fMRI scanner. To avoid any confounding effects of race or sexual desire, all the test subjects and faces were white, and all the subjects were heterosexual. The volunteers saw each face for about 2 seconds and assigned it a score from 1 (least attractive) to 8 (most attractive). They were then told how another group of women had rated the face. These "group ratings" were actually generated by a computer that gave each face a different rating than the test subjects about two-thirds of the time. As the researchers suspected, the rostral cingulate zone and nucleus accumbens [areas of the brain believed to be part of that learning system]  fired up when the group ratings did not match the subject's score. Moreover, brain areas associated with reward were less active when a subject’s predictions deviated from the group's score...

What does this imply? It seems that when the group disagrees with us we modify our beliefs, which is what social learning would imply.  I think economists would be comfortable in fitting this into a framework like the "herding" or "informational cascade" models where we update our beliefs based on the revealed beliefs of others. But there also seems to be a direct "utility" (reward) we get from agreeing with others.  This seems like a pure payoff to conformity. The question is why would evolution instill both mechanisms?  Perhaps conformity is just a good strategy most of the time. So instead of learning from others we have some tendency to just mindlessly copy them. But it suggests that someone with social disorder might indeed be more creative because maybe they don't feel that disutility from having divergent beliefs. This would in turn create benefits for the conformers because every now and then the non-conformist comes up with something so demonstrably better that some of the conformists start doing it despite the psychic penalty from deviating. Can there be an equilibrium where the majority have conformist tendencies but a small minority do not?

Bogus Boasts of Bubble Cassandras

As someone who put a small chunk of money into the Korean Investment Fund and Greater China Fund in 1997 (just before the Asia Crisis) and then a somewhat larger sum into SPY (the largest S&P 500 ETF) last autumn and this spring, I get just a little annoyed when I hear about people who claimed they saw the whole thing coming.

The problem with bubbles is not so much that you can't spot them. In 1998/1999 it seemed like everyone I spoke to recognized that the there was a dot-com bubble. The real quandary--it seems to me--is that some people continue to make a ton of money as the bubble continues to expand. If you listen to the prophets at the time they actually issue their forecasts of Doom, you may end up missing the ride up.

Here's an example. In a New York Times book review today, Daniel Gross (who writes very good articles for Slate) argues "The bailouts are especially galling given the ample warnings, like those sounded by John Cassidy of The New Yorker, who warned in November 2002 that housing would be the next crash." Wow, this makes John Cassidy look really prescient. If only we had listened to him, we wouldn't be in this fix. I might still have my nest egg!

But wait, look at the data for the Case-Shiller Index.



If you had bought the average house in 2002, the year Cassidy issued his warning (which you can read here in full but I think the title, "The Next Crash" and the subtitle, "Is the housing market a bubble that’s about to burst?" conveys the main idea), the price was about 115 (with 100 being the 1988--2008 average). Even if you sold today at 160, you would have earned nearly 40%. And if you got out at the peak (and we'd all like to think we could do that), your return is over 60%. The run up from 1997 to 2002 was just 15%. So by heeding Cassidy's warning in 2002, you miss out on something like 4/5 of the potential gains.

So the problem with bubbles is not that they're impossible to recognize, it's just very very hard to figure out how much more the bubble will expand before it bursts. One other problem. The bubble analogy is of dubious value because many of the so-called bubbles don't actually burst in the sense of a sudden collapse in prices. As you can see in the figure above, the rate of descent since the peak in housing prices is if anything less steep than the rate of ascent before the peak. How much further will housing prices fall and how long will it take? If Cassidy was right about the bubble in 2002, then we still have a long way down to go.

Cassidy also got the timing wrong for a previous "bubble," the S&P 500 in the late 1990s. On August 17, 1998, when his article "Pricking the Bubble" appeared in the New Yorker, the S&P500 closed at 1084. It would rise a further 43% over the next two years reaching a high of 1553 on March 24, 2000. Thereafter, the bubble did not burst, but deflated gradually, briefly reinflating to 1530 on September 1st, and not falling below 1084 until April 2002. (All data from Yahoo!). So if Cassidy had believed that the S&P500 was a bubble in mid-1998 and sold stocks short, he would have had to wait nearly 4 years to close his position with a profit--not too mention some very unpleasant margin calls in mid-2000.

The Five Pillars of Belief

Ok, this my second go at this topic. In "What should we believe?" I was more trying to say normatively how to evaluate a claim. But here I want to give general account of the forces that cause certain beliefs to be held.

1. Expertise: does the source of the claim have specialized knowledge on the topic that makes her view more likely to be true than a random person?
2. Disinterest: Few people give much credence to claims by scientists hired by tobacco companies...
   
3. Consensus: Schopenhauer apparently said, "There is no opinion, however absurd, which men will not readily embrace as soon as they can be brought to the conviction that it is generally adopted." But consensus isn't always mindless. Take for example the Intergovernmental Panel on Climate Change. They may be wrong but the fact that 1000s of scientists have signed on makes me more inclined to believe their claims. But those are all experts so I'm combining two pillars. How about if you come upon two restaurants side by side in a foreign city. I would choose the one with more diners (if there were space for me).
4. Observation: People say "seeing is believing" for a reason.
5. Comprehensibility: I think we need a story, a way to make sense of a claim, in order to find it fully believable. Thus, I don't think it's any accident that religion tries to counter the need for comprehensibility by stating "God moves in mysterious ways." We can't comprehend how a loving God could inflict so much suffering on humanity so we need to be told it's OK that we don't get it.
   

Niall Ferguson's "Chimerica"

My good friend from Arpoador sent me a link to read, Niall Fergusson's article, the "End of Chimerica." The article covers a lot of topics but the main point seems to be that the US and China had a special economic relationship (a savings glut in China financed a housing boom in the US) and now it's over. It's not that different from a point Krugman has made in the past about the US being a country where people made their livings selling houses to each other financed by debt from China. But housing booms haven't been specific to the US (Canada, UK, etc.) and lending to the US hasn't been specific to China (ever heard of Japan?). And I couldn't really find the place in the article where he showed that China has now or will anytime soon stop lending to the US. And what did China really have to do with sub-prime mortgage originators and the absence of oversight for CDS and related derivatives and all the rest of the things most people put at the origins of the crisis?

But the things above don't really bother me. It's things like going on and on about oil prices driven ever up (by China's insistence on growing at 10% a year--it certainly would be nice to be able to grow at 10% a year because a country "feels it must"). But the article was published in September 2008 and oil prices had been fallen sharply since mid-July (about 25% below the price mentioned in the article). Ferguson also goes on about US dollar depreciation when the US Dollar had also been appreciating since mid-July against most currencies. Books necessarily become out of date quickly but there's nothing to stop someone from updating an article.

This the quote that bugged me most:

History’s lesson is clear: the combination of high indebtedness and low growth is lethal to reserve currency status, provided there is an alternative.

Where to start? First how on earth could history have a clear lesson to offer on the economic conditions that lead to the decline in a country as a reserve currency? So far as I know, since we've had good economic data, there have been just 2 main reserve currencies (the pound and the dollar). And Ferguson himself admits there is no clear date for when the dollar became number one (somewhere between 1915 and the 1950s). So how on earth could we say with certainty that it is the combination of high debt and low growth that is fatal. what about inflation say, or the absolute size of your economy, or its share of world financial transactions? I would have started with those in my model. But I couldn't estimate the model because there's only one change! The worst thing is the last clause, "provided there is an alternative." What started out as a strong (probably false) declaration has now been reduced to something that can't be disproven. Because we don't know what an "alternative" is. He tells the Euro is one. But of there were many alternatives in the sense that that there are plenty of currencies countries might have chosen. Ferguson would protest that these weren't suitable alternatives... OK, so now what's history's great lesson? If you have world reserve currency and its economy is lousy and there's another currency that's a great alternative with a good economy people might switch? That's not a lesson from history. It's just common sense. And for all I know, it might not even be right (multiple equilibria are probably important for currency reserve choice).

My last complaint is the cheapshot against unnamed economists:

As this tremendous expansion in borrowing was taking place, many Panglossian economists tried to rationalise what was going on. Some argued that this was “Bretton Woods II”, a kind of system of international exchange rate management. Others called it a “stable disequilibrium”, something that could be counted on to continue for some considerable time.

So which economists said this was a stable disequilibrium? That doesn't even make any sense to this mainstream economist (although my google search did turn up an article on the topic in Economics Letters). It's not like Ferguson fails to reference other names in the article--he mentions the author of the book the Black Swan later in the same paragraph. And how about economists like Paul Krugman who had been raising alarm bells for a long time?

Full disclosure: when I read this article I was already annoyed with something Ferguson had said on BBC Forum. Asked to justify his thesis that financial events an important driver in history, he chose to pick the victory of Barack Obama:

"[The] Presidential election ... would almost certainly not have gone the way it did if it hadn't been for a huge financial crisis that reached its climax two months before people went to the polls."... "It's clear that things would have been very different if the financial crisis hadn't struck when it did" (my transcription)

Wait! it's not clear at all. There was only the shortest window where McCain was ahead in the average of the polls. And prediction markets gave McCain only a 30-40% chance of winning back in June. Sure Obama's position in the polls rose around the time that crisis worsened. But even if things hadn't gone crazy what certainty is there that McCain would have won? None at all. More likely in my view is that Obama would have just won a tighter victory. But who knows?

When we don't really know something we need to try to be a little more humble and say "clearly" a little less often.

Election 2008: obsessing over the numbers

County level results for Virginia and Indiana for 2004 election from Wikipedia

538 Nate Silver

Pollster.com

County red-blue map from 2004

What should we believe?

Elements that (should) contribute to a believability.

1. Mechanism: I don't believe that astrological sign can affect personality because I cannot think of any possible mechanism. On the other hand I have no trouble with the idea that iboprufen might reduce muscle pain because it is an anti-inflammatory.
2. Independent confirmation: If multiple studies come to the same conclusion this should add a lot to believability unless there is reason to believe that the results of some studies influenced the results of the others.
3. Principle of disinterest or counter-interest: There are many instances where the reporter of some information has an interest in pushing a particular view. Other things equal, we should put less weight on those reports than those provided by neutrals or those who an interest in the contrary position.
4. Understanding the opposition: When two beliefs are widespread, if you undertand (a) why it is that intelligent people can hold the opposing view and (b) how they made their mistake (which error of fact or deduction is at fault?), then you can be more confident in your own belief.