Arnold Kling  

Bruce Schneier's Latest Book

PRINT
Fund-Raising and the Independe... Allowing A Positive Price for ...

It's Liars and Outliers, and I would rate it the best economics book of the year thus far. He writes about his book here and here.

Schneier views our lives from the perspective of game theory. Every day, we must decide whether to cooperate or to defect. Do I try to arrive at work on time, or do I show up late? Do I drive safely or aggressively? Do I support the goals of my department, or do I work for myself? Does my department support the goals of the larger organization, or does it pursue its own interests? Does the larger organization work to support the goals of the society to which it belongs, or does it pursue its own goals?

He says that there are four "societal pressures" that induce cooperation: moral pressures (internalized desires to cooperate); the value of reputation; institutional and legal incentives; and security systems. He points out that in small groups (think of a band of hunter-gatherers) the moral pressures and reputation pressures are often sufficient. Larger societies need institutional and legal incentives. Security systems are in some sense a last resort.

Schneier's view is that there will always be defectors, and there will always be incentives to defect. I think this goes against a number of utopian visions. The utopian vision of the left is that human nature can be changed to get rid of defectors. The utopian vision of libertarians is that reputation systems will stamp out defectors. Schneier would say instead that reputation systems will come to be gamed by defectors.

Schneier clearly obtained many prejudices from the left side of the political spectrum. Yet I agree with a huge proportion of what he has to say, probably because he strongly rejects utopianism. More below.

I will offer my excerpts in reverse, order, from the end of the book to the beginning.


rules could be good, like a respect for human rights or a system for enforcing contracts...rules could be bad, like slavery...rules could be perceived as good by some societies and bad by others: arranged marriages; heavy taxation...Societal pressures simply enforce cooperation, without much consideration as to why the defector chose some competing interest.

Much of the book is written from the vantage point of "the interest of society." As a result, individual liberty tends to come across as only an instrumental good. We need the freedom to defect because sometimes defectors are better for society.

Sometimes assistants need to sign documents for their bosses, and sometimes attorneys and accountants need to innocently backdate documents. Sometimes defecting is a form of social lubricant: small social dishonesties that make life easier for everyone.

There are good defectors and there are bad defectors, and we can't always tell the difference--even though we think we can.

Reading the above, I was reminded of the arguments over "Robo-signing" of foreclosure documents.

we need to concentrate on the broad motivations for defection, rather than on blocking specific tactics...One example is counterterrorism, where society is much better off spending money on intelligence, investigation, and emergency response than on preventing specific terrorist threats

I would argue that one problem with financial regulation is that we are tempted to focus on "specific threats" (such as derivatives) without looking at the big picture.

Consider all four societal pressures. It's common to believe that one is enough: that reputation obviates the need for laws, or that a good security system is sufficient to enforce compliance.

This is something that libertarians should think long and hard about. In David Friedman's The Machinery of Freedom and other libertarian works, is there too much reliance on reputation, while taking moral pressure too much for granted (or ignoring its value)?

defectors are quicker to use technological innovations. Society has to implement any new security technology as a group, which implies agreement and coordination...Unfamiliarity is also an issue...Defectors are more agile and more adaptable

On the other hand, I would say that defectors have to be secretive, which makes it more difficult to use technologies that require complex cooperation (developing nuclear weapons, for example).

When you start measuring something and then judge people based on that measurement, you encourage people to game the measurement instead of doing whatever it is you wanted in the first place.

That, in a nutshell, is why principal-agent problems are difficult, why organizations tend to be dysfunctional, and why regulations work less well than intended.

In the real world, the knobs are poorly marked and badly calibrated, there's a delay after you turn one of them before you notice any effects, and there's so much else going on that it's hard to figure out what the effect actually is.

He is talking about applying different societal pressures, but these comments are applicable to just about any economic policy.

If we're going to keep government from overstepping its bounds, it will be through separation of powers: checks and balances. But it's not just government that needs to be watched; it's corporations, non-governmental institutions, and individuals. It's everyone's responsibility to keep everyone else in check.

Human society is saturated with principal-agent problems, and here he is more or less throwing up his hands and saying that we have to try to deal with them as best we can.

Facebook...because it is a for-profit corporation and not a true agent for its users

Schneier sprinkles a fair amount of this anti-corporate rhetoric into the book. The classic pro-market defense would be that Facebook is induced by market forces (consumer demand, competition) to act as its users' agent. However, I could concede that those market forces work less than perfectly. Still, I would suggest that government acts quite imperfectly as an agent for consumers, also. To be sure, Schneier does not turn to government as the all-purpose solution for market imperfections, and in that regard he shows more wisdom than most mainstream economists. In fact, he quite explicitly talks about the principal-agent problem in the context of the Transportation Security Administration, in which citizens are the principals and TSA as a less-than-true agent.

Competing societal dilemmas represent the normal state of affairs. Rarely is the real world so tidy as to isolate a single societal dilemma from everything else. Group interests are often in conflict, and cooperating in one necessitates defecting in another.

Life is messy.

An obvious societal pressure system...would be to require advisors to disclose any conflicts of interest; but this can have the reverse effect of increasing the number of defectors. By disclosing their conflicts, financial advisors may feel they have been granted a moral license to pursue their own self interest

If specific regulations tend to crowd out fundamental principles (such as fiduciary responsibility), financial regulation will be dysfunctional. I think this is a big deal.

The emergence of long-distance commerce in the Western world was aided in part by the involvement of European Quakers, who earned a reputation for dealing honorably with their business partners.

He argues that the threat of being expelled from the Quaker community for shady dealing was helpful in this regard.

From the perspective of trust, societal dilemmas involve a Red Queen Effect. On one hand, defectors should evolve to be able to better able to fool cooperators. And on the other, cooperators should evolve to better recognize defectors. It's a race between the ability to deceive and the ability to detect deception.

Robin Hanson, check the box if you've read this far.

Societal pressure failures occur when the scope of defection is either too high or too low: either there are too many burglaries, or we're spending too much money on security to prevent burglary.

This is the non-utopianism to which I referred above.

Instead of having to trust individual merchants, people can trust the laws that regulate merchants. Instead of having to evaluate the trustworthiness of individual borrowers, banks and other lending institutions can trust the credit rating system.

I think Schneier could have written at great length about this process of substituting one layer of trust for another. It is my view that the financial industry operates this way. In some sense, the "financial crisis" was like the levels of a house of trust collapsing on one another. Much, much more could be said about this.



COMMENTS (9 to date)
Becky Hargrove writes:

I definitely need to read this book: the 'lead, follow or get out of the way' pattern has intrigued me for years.

DPG writes:

I like the cooperate/defect approach.

Applied to political systems, I think it explains why democracy is preferred. It is the most effective system at limiting defection in the pursuit and exercise of political power (although still not perfect, e.g. Churchill).

Re: Charles Murray, marriage, and male labor, I again think this cooperate/defect is useful. Young men have the choice to cooperate: stay in school, get a job, woo a girl politely, and provide for a family. Or they can defect: live on the dole, avoid serious relationships, fritter away their time on leisure.

Kevin L writes:

When it comes to reputation, I think government is a terrible source. Any benefit in being "democratic" is outweighed by the (usual) tactic of eliminating regulation by exit. That is, we are forced to accept the government's reputation-granting credentials, at least at some level. I'm thinking of the Federal Reserve, FDA, and other areas where the government either grants itself or another group a monopoly on verifying reputation. Even in this category there are more categories. If you distrust the Fed enough, you can try to arrange payment in other currencies, but legally you must accept Fed notes if they are offered. With medical licensing, doctors cannot legally do many things without the AMA's approval, even if their customers trust them. So one party or another is being forced to cooperate with a reputation source, and the feedback loops that exist in a free system are totally cut off.

Seth writes:

DPG - I'm not sure the latter is a defection any longer.

mobile writes:

Some Facebook users think the purpose of Facebook is to provide services to Facebook users. They try to interpret or predict the actions of Facebook from that perspective.

When I run into them, I repeat the quote "If you're not paying for it, you're not the customer. You're the product." (image with pigs)

Turns out that Bruce Schneier is the inspiration for that quote, if not the direct source.

Matt C writes:

When I run into them, I repeat the quote "If you're not paying for it, you're not the customer. You're the product."

Too true. Also applies to the U.S. health care system.

Bryan Willman writes:

Human minds are finite, and therefore people have to delegate evaluations. You probably can't be your own investment advisor AND do the building inspection for your house AND health inspections on all of the restaurants you eat in. You have no choice but to trust to other people's judgement.

If a society works at all, it means these mechanisms work out OK most of the time.

A real problem with entities like facebook is that they don't really make clear to people how they make money, and have all sorts of features and network effects that draw people in.

Bryan Willman writes:

It's actually worse than just facebook, I had an exchange on a bulletin board in which it took a while to persuade somebody that google does not care one whit if he found some search result useless. He (and you and I) are not the customer, and so whether any particular search result serves us is of literally zero accord to google. Only when masses of people (eyeballs to advertisers) defect to some other service do they care.

The world is full of such things, and has been for a long time (newspapers and their dual revenue selling of both subscriptions and advertising.) What amazes me is how few people seem to understand this.

Ghost of Christmas Past writes:

In a way, "reputation," though a darling of libertarians, shares a flaw with "affirmative action" (the darling of leftists).

You see, reputation schemes may protect certain groups of people (e.g., "all bidders at a given auction website") but sacrifice individuals (e.g., "the eleventh through eighty-fifth buyers, who get totally cheated by a crooked seller who made sure to earn a 5-star rating from his first ten buyers just so he could cheat a bunch of others before their negative feedback could percolate through the system").

When a leftist says "affirmative action" is okay because it only harms a small percentage of white people, a libertarian rightly replies that AA harms a few individuals a large amount-- it doesn't harm a bunch of people each a small amount-- so the propaganda about groups is just misdirection.

When a libertarian says a "reputation" scheme will be fine because defectors will only harm a small percentage of group members, we should be cautious-- defectors may inflict severe harms on individuals before their reputations erode enough to warn off other members of a group.

Comments for this entry have been closed
Return to top