Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them. - Laurence J. Peter

Tuesday, July 20, 2010

Chapter Summaries of The Moral Animal by Robert Wright

I'm writing chapter summaries of The Moral Animal by Robert Wright here. So far, two out of four parts are online. I recommend the book, as do Eliezer Yudkowsky and Patri Friedman.

Sunday, March 22, 2009

Motivated Skepticism and Motivated Credulity

Two biases that lead to "stickiness" of political beliefs are motivated skepticism and motivated credulity. When I encounter an argument opposing some strongly-held belief of mine, I am prone to examine the argument extremely skeptically, never giving the benefit of the doubt. If I find the slightest flaw - a minor factual mistake, a questionable statistic, an improbable leap of logic - I am tempted to dismiss the whole as clearly the work of a dishonest or incompetent mind. This doesn't feel wrong. It feels like being logical, and upholding the highest standards of debate. It feels righteous. This is motivated skepticism, also called disconfirmation bias.

Conversely, motivated credulity - also called the prior attitude effect - is to give a free pass to weak arguments whose conclusions I agree with. Nobody's perfect, right? This writer is clearly a paladin of a just and noble cause. If he accidentally misquoted someone, or relied on outdated statistics, or his work lacks some key citations - big deal. I won't quibble over the details. That would be petty.

I hope I make these errors less frequently than I used to, but I suspect it's impossible to be sure I'm free of it altogether. I recall some particularly egregious examples from my past, particularly when arguing about the Iraq war in 2002/2003. I was 18-19 then, and new to politics. I might write about my errors in more detail sometime.

Rule of thumb: Be skeptical of things you learned before you could read. E.g. religion. -- Ben Casnocha
Ben is right: lacking the ability to read will severely hamper your ability to form sound beliefs. Lacking an understanding of cognitive biases - such as motivated skepticism - is also a risk factor for dubious beliefs. I remember dismissing the arguments of those I thought fools based on a few tangential errors, back before I'd heard the term "motivated skepticism". That was my mistake. So I'd go further, and propose this rule of thumb: be skeptical of beliefs you formed and defended before you understood motivated skepticism.

This implies that everyone in the world should drastically lower their confidence in their cherished beliefs. Let's all do that.

I came across an example of motivated skepticism recently. Here's Andi Hazelwood and Albert Bartlett talking about peak oil and the limits of natural resources. They're discussing Malthus, saying he was basically right.
AB: I read Malthus three times. Malthus understood the problems of limits, and he understood these things. And his timetable--he couldn't have anticipated the mechanization of agriculture, which has greatly increased agricultural production worldwide...And I think with these limits showing up in terms of peak oil, peak natural gas--I think with these limits showing up people will have to reassess their proud claim that "we've proven Malthus wrong." Malthus, I think, will turn out to be... "well, yeah, he understood the problem, he was right." And when some learned scholar tells me "we've proven Malthus wrong," that scholar is telling me about himself; he's not telling me about Malthus."
The discussion then moved, as it should, to the economist Julian Simon, who would have disagreed with that. Here is how I wish the conversation had gone:
A: Julian Simon would say that agricultural innovations weren't accidental or coincidental, because whenever food became scarce, food prices increased, which increased the reward to those who found ways to increase food production. I guess we should address that.

B: Well, Norman Borlaug is considered the father of the Green Revolution, developing high-yield, disease-resistant wheat varieties. It's said that he saved millions of lives. So what motivated him?

A: According to this interview, he says he was motivated by seeing so much human misery. He was a teenager during the Depression.

B: So Simon was wrong! Borlaug wasn't motivated by money!

A: Well, maybe. Simon said that food shortages would lead to innovation through the mechanism of increased prices, and therefore growth could continue. In this particular case the mechanism seems to have been altruism rather than personal gain, but the process still worked. A resource shortage led to innovation and ensured continued growth. Ideally we would examine more examples of the motivations of people who improved agricultural yields.

B: Ok...so what does this mean for peak oil and peak natural gas, which were talking about earlier?

A: Simon would probably say that we can expect innovation to solve resource shortages for those problems too.
My fantasy pundits sound so reasonable. The actual conversation was an exercise in motivated skepticism. Instead of addressing the criticism Julian Simon would have made to them, they found an irrelevant quote to attack:
AH: And I think we should point out that Julian Simon infamously said that we had enough technology in our minds and in our libraries to allow population to continue growing for the next 7 [million] years.

AB: That's an interesting story...my correspondent asked me, "What would the world population be if it just grew at the present rate of 1% per year"--which sounds terribly small as a growth rate--"if it grew at 1% per year for 7 million years?" Well...[it's] a number that you get by writing 1 followed by 30,000 zeroes...1 followed by 85 is the number of atoms estimated in the known universe, and Julian's population size would be 1 followed by 30,000 zeroes.

AH: So, no amount of technology is ever going to make it possible for those people to fit on earth.
...
AB: There aren't enough atoms to make that many people. But Julian was worshiped by the people in Washington who wanted to hear is message. And these people were often politicians who had no scientific judgment. But Simon had a Ph.D. And he was reasonably bright in the sense that he was bright enough to know what it was that important clients wanted to hear. And so he composed things they wanted to hear. And so they were very much enamored of him.
This could be my motivated credulity talking, but I think that when you find an assertion, and you add an assumption (a 1% growth rate) to that assertion and then attempt a reductio ad absurdum, then you've only disproved your assumption, not the assertion. To be fair, Julian's assertion of seven million years of growth on Earth seems a silly prediction in several ways. But regardless of that line's worth, it was irrelevant to their argument.

So watch out for motivated skepticism and motivated credulity. And resist, if you can, the temptation to do what I just did - using my knowledge of cognitive biases as a rhetorical weapon to bludgeon my enemies. First, know yourself.

Tuesday, February 17, 2009

If you were a better-than-average driver, how would you know?

Are you good at maths? Are you good at drawing?

Your answer to these questions is probably reasonably accurate. Unless you're a mathematician or an artist, you probably don't have much pride or identity tied up in your skill in these things, so that won't cloud your judgment. (And if you are a mathematician or artist, then - relative to average people - you're almost certainly good at maths or drawing.)

Are you good at driving? Are you a good judge of character? Are you (or would you be) a good parent? Are you a good voter?

Everyone thinks they are a better than average driver. The appropriate conclusion to draw from that is that, if you need to know whether you're a good driver, the way to discover this information is not to simply introspect on the question for a few seconds. I suspect the same holds for whether you're a good parent, voter, or judge of character.

That's a serious problem! It probably doesn't matter whether you're a good driver. Even if you're worse than average, you're probably not so bad that you shouldn't be driving at all. But bad parents and aggregated bad voters do a lot of harm. If you're likely to be a bad parent, then that is actionable information: don't have children! If you're likely to be a bad voter, don't vote! If you're a poor judge of character, get a prenuptial agreement.

If you really wanted to know whether you're a good driver, you could probably find some reasonable metric and compare yourself to the average. But there are no controlled experiments in parenting or politics, which makes it very difficult to be sure you have reliable knowledge on these questions.

Friday, January 30, 2009

Destroying wealth to improve the economy

In the past, people sometimes tried to improve the economy by destroying wealth:
President Roosevelt came into office proposing a New Deal for Americans, but his advisers believed, mistakenly, that excessive competition had led to overproduction, causing the depression. The centerpieces of the New Deal were the Agricultural Adjustment Act (AAA) and the National Recovery Administration (NRA), both of which were aimed at reducing production and raising wages and prices. Reduced production, of course, is what happens in depressions, and it never made sense to try to get the country out of depression by reducing production further. In its zeal, the administration apparently did not consider the elementary impossibility of raising all real wage rates and all real prices.

The AAA immediately set out to slaughter six million baby pigs and reduce breeding sows to reduce pork production and raise prices. Since cotton plantings were thought to be excessive, cotton farmers were paid to plow under one-quarter of the forty million acres of cotton to reduce marketed production to boost prices.

History repeats:
The second plan is more politely known as "fleet modernization." It combines economic as well as environmental goals in one package.

Under a bill introduced by Sen. Dianne Feinstein, D.-Calif., owners of older cars would get vouchers worth thousands of dollars toward the purchase of newer, more fuel-efficient vehicle. For the customer to get that cash, the car dealer would have to certify that the trade-in was getting scrapped and not resold. The car's vehicle identification number (VIN) would be tracked to make sure it never shows up on a vehicle registration again.

Added: from the Wall Street Journal on January 28:
[In discussions of the stimulus bill,] dairy and beef cattle producers butted heads over talk that the government might buy up dairy cattle for slaughter to drive up depressed milk prices.

Thursday, January 29, 2009

Home ownership and labor mobility: a trade-off

Governments tend to promote/subsidize home ownership. Why? Rourke O'Brien and David Newville write in the Washington Post (January 11, 2009):
Studies have shown that families who own their own homes are more likely to be involved in their communities, to report higher satisfaction with their lives, and to vote. Homeownership also has positive impacts on children, such as increased high school graduation rates, fewer behavioral problems and better job outcomes after school, and these effects have been found to be strongest among low-income homeowners.
Citation needed, but ok...
What's more, homeownership has been and will continue to be the single biggest source of wealth for low- and moderate-income families. Even if housing prices don't rise much, the forced saving that comes from paying down a mortgage can help families build equity that can then be leveraged to finance a child's education, provide financial security in retirement or pass wealth on to the next generation.
"Even if housing prices don't rise much"? What if they fall by 40%?

We should consider the costs of promoting home ownership too. Arguably, it contributed to an enormous financial crisis. (Note: don't believe that explanation - or any other explanation of the financial crisis - unless you have a good reason to think you're better at economics than Nobel Prize-winning economists, because Nobel Prize-winning economists disagree on this.)

Also, home ownership reduces labor mobility, as Will Wilkinson points out. And with a labor force more heterogenous than ever before*, America needs labor mobility more than ever.

Maybe governments should pursue policies that promote labor mobility. (Or maybe governments should leave people the hell alone.) One downside to that might be that labor might mobilize itself to some other country (say, to Taiwan - it's a great place to do business), or even to some still further frontier.

*That is, the skills of the labor force are more specialized than before. Fifty years ago America had more factory jobs than it does now. My understanding is that one high school dropout could work in one factory about as capably as another factory. But you can't easily switch from being a website developer to a biotech engineer, or whatever it is people are doing in America these days.

Sunday, January 18, 2009

How do you distinguish good experts from bad experts?

How do you distinguish good experts from bad experts? It would be an extraordinarily useful skill - a meta-skill that gives you access to so many others.

Priests and biologists both claim to know the origin of human life. How do you choose between the two? It seems obvious to me, but people disagree on this so there's something non-obvious about it.

How do you distinguish between a good mutual fund manager and a bad one? It's really hard. So you might choose to get an index fund instead, which is probably not a bad idea. In one sense, that would be judging the market to be your good expert. Patri Friedman didn't do that in 2008; he chose a bear strategy instead. This is evidence that Patri Friedman is good at choosing good investing experts or is a good investing expert himself (but not sufficient evidence - watch him for another ten years).

How do you distinguish between bad science and good science? Patri Friedman thinks he can. squid314 challenged him:
You need to prove that you can "beat the house"; that your own judgment is likely to be less fallible than the nutritional scientists' for some reason. Top nutritional scientists should know everything you do about the problems with certain kinds of studies and take that into account. So all of the pitfalls that apply to nutritional scientists equally apply to you, unless you have some specific reason to think they don't. And then the experts have the extra advantage of much greater familiarity with the subject matter.

Patri claims "that by using [rigorous scientific procedures], digging into a few papers, correcting for funding biases, one can occasionally decide to take a contrarian view, and be correct."

I suspect that that approach would yield a strong conclusion on the priest/biologist question.

How do you distinguish between good economics and bad economics? Nobel Prize-winning economists Paul Krugman and Paul Samuelson have a different view of macro-economics than Nobel Prize-winning economists Milton Friedman and Friedrich Hayek (of course, all four of them disagree, but if you're going to divide them into two groups, that's a sensible categorization). Who are the better economists? How do you know whether you are a good enough economist, or a good enough judge of economists, to decide?

Monday, December 22, 2008

Three narratives about entrepreneurial motivation

Narrative One: entrepreneurs are motivated by financial incentives. Reduced regulation and reduced taxes can increase entrepreneurship.

Have you seen this editorial in the Wall Street Journal about the decline of entrepreneurship in Silicon Valley?
According to the National Venture Capital Association, in all of 2008 there have been just six companies that have gone public. Compare that with 269 IPOs in 1999, 272 in 1996, and 365 in 1986.

Michael S. Malone says that Washington is killing entrepreneurship with capital gains taxes and regulation such as Sarbanes-Oxley. The mechanism is that Sarbanes-Oxley makes the overhead of IPOs much more expensive, which reduces the incentive for entrepreneurs to start companies because the likely gains are lower, and reduces available VC capital for the same reason.

There were some supporting anecdotes from the discussion on Hacker News. From ohhmaagawd:
I know a CEO of an Internet company that planned on doing an IPO in 2008. Of course the economy tanked so it's not happening.

He told me that SOX required the company to have auditing/traceability and controls for everything. This meant they quit using any software as a service applications, they shut down IM, they have to keep track of everything on every machine, banned a whole list of applications that would interfere with SOX (iTunes, file sharing, and music app), etc.

They had to hire lawyers, more accountants, consulting firms. He told me they were looking at > 10m to comply. And that doesn't count lost productivity of having to act like a draconian big ass company where they had to have design docs for everything they do, detailed project plans, get rid of the tools they use (as listed above), etc.

In other words to go IPO now you give up agility, spend more money, and become less productive.

From russell:
I worked for a company a couple of years ago that was planning to public in a year or two. The CEO told us that complying with SOX would cost $3 million for the IPO and over $1 million per year after that. That's a huge burden for a company with revenues of $35 million at the time. I have read that it is now $5 million for an IPO and that no company with revenues under $100 million can afford to be public.



Narrative Two: entrepreneurs are not primarily motivated by financial incentives; they are driven by the joy of making a difference and creating value. Encouragement from national leaders can increase entrepreneurship.

In The Cure To Our Economic Problems, Mark Cuban (billionaire entrepreneur) says that financial incentives aren't as important as some people think:
Entrepreneurs who create something out of nothing don’t care what tax rates are. Bill Gates didn’t monitor the marginal tax rate when he dropped out of Harvard and started MicroSoft (btw, it was a ton higher than it is today). Michael Dell didn’t wonder what the capital gains tax was when he started PC’s Limited, and then grew it into Dell Computer. I doubt that any great business or invention started with a discussion or even a consideration of what the current or projected income or capital gains tax was or would be.

The impact of tax rates on productivity and development is something economists masterbate about, enterpreneurs don’t waste their time thinking about it. We have business to do.
...
Entrepreneurs live for the juice of making their dreams come true. Of having a vision and fighting to see it come true. The joy of mission accomplished and the scoreboard of the financial rewards.

The solution?
The cure to our economic problems is the Entrepreneurial Spirit of All Americans. Instead of bitching at each other, could one Presidential candidate please show even the least bit of leadership and character and stand up for and encourage the entrepreneurs in this country ?
...
What we need is our candidates to stop yelling at each other and starting looking at the American people and encouraging the best of who we are. That is who I want to get behind. That is what I would like to see for our country. That is what will energize and motivate people to create companies and invent products that will turn the economy.

That's a bit vague, but he's a billionaire and I'm not, so who are you going to listen to?



Narrative Three: Entrepreneurs are held back by concerns about health care. Universal health care can increase entrepreneurship.

Universal health care, it is argued, would encourage entrepreneurs in America, because:
-People who get health insurance through employment (which is nearly everyone) are reluctant to leave their job to start a business because they'd lose their health insurance and would pay a much higher rate for private insurance.
-People fear the cost of having to organize health care for their employees.

Morgan Wilkins argues:
Universal access to health care would not only increase individual liberty and alleviate wasteful spending, but would also increase entrepreneurial activity – another value central to conservative thinking. Many entrepreneurs are discouraged from embarking on their new business venture by fears about the high cost of providing employers with health insurance. Small business employers are then forced to make tough decisions about whether to absorb the high costs, pass them on to employees, or not offer benefits at all. A universal health care system would eliminate these concerns and encourage hesitant entrepreneurs to test their innovations and ideas.

Or Daniel Brook:
America acts as if all is well when in fact we’re one of the only developed countries with a rate of self-employment even lower than France’s. While surveys show that Americans are nearly twice as entrepreneurial as Europeans, we’re only half as likely to actually become self-employed.

What is holding Americans back? In two words: health care.
...
In other developed countries, where self-employment rates tend to be higher, taking the leap to working for yourself doesn’t affect your health care coverage or your family’s. In publicly funded health care systems, entrepreneurs pay less into the system during the few lean years that often accompany starting a business. Once you get off the ground, you pay more. That benefits the country’s health and its economy.

I did see a study about this though.
Some commentators have suggested that the absence of portable health insurance impedes people from leaving their jobs to start new firms. We investigate this belief by comparing wage-earners who become self-employed during a given period of time with their counterparts who do not. By examining the impact of variables relating to the health insurance and health status of these workers and their families, we can infer whether the lack of health insurance portability affects the probability that they become self-employed. The evidence does not support the conjecture that the current health insurance system affects the propensity to become self-employed. Hence, whatever its other merits, there is no reason to believe that the introduction of universal health insurance would significantly enhance entrepreneurial activity.

But it's just one study. You could apply motivated stopping or motivated skepticism at this point, if you want to.