The Prisoner’s Dilemma is the obverse of Adam Smith’s theory of the invisible hand, in which the free market coördinates the behavior of self-seeking individuals to the benefit of all. Each businessman “intends only his own gain,” Smith wrote in “The Wealth of Nations,” “and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention.” But in a market environment the individual pursuit of self-interest, however rational, can give way to collective disaster. The invisible hand becomes a fist.
and
Limiting the development of [complex] securities would stifle innovation, the financial industry contends. But that’s precisely the point. “The goal is not to have the most advanced financial system, but a financial system that is reasonably advanced but robust,” Viral V. Acharya and Matthew Richardson, two economists at N.Y.U.’s Stern School of Business, wrote in a recent paper. “That’s no different from what we seek in other areas of human activity. We don’t use the most advanced aircraft to move millions of people around the world. We use reasonably advanced aircrafts whose designs have proved to be reliable.”
The second seems quite applicable to software engineering as well.
It's unfortunate that the article uses the Prisoner's Dilemma as its example, because the Prisoner's Dilemma is actually not the best game model to set up as the opposite of the Invisible Hand. A more appropriate model would be the Tragedy of the Commons, which was proposed by a fellow named Garrett Hardin in an article in a 1968 edition of Science magazine.
If you are interested in formal modeling of coordination problems, an interesting overview can be found in Elinor Ostrom's book Governing the Commons. In chapter one she gives an overview of the Prisoner's Dilemma, the Tragedy of the Commons, and the Logic of Collective Action (Mancur Olson's theory). After reviewing these abstract formal models, she looks at other possible institutional solutions for coordination problems.
Any quick one-liners or easy-to-understand wisdom that you've gained from the material? The entire subject area looks immensely fascinating, yet very dense at times.
Yes, game theory can get inscrutably complex, even through the initial models are pretty simply to set up and understand.
The basic components of any coordination/cooperation problem are the same: you have actors with preferences, you have a situation with rules that they can't change, and you have outcomes with payoffs/penalties depending on whether the actors choose to cooperate or not.
Most of the problems revolve around how to get people to cooperate, when defection is attractive for some reason. Most of the solutions involve making sure situations are not one-off interactions, that all actors have information about what each other is doing, and that in the case of multiple actors using common resources, you need to have some form of monitoring and enforcement.
If actions are one-off, the temptation to act selfishly is higher than if actors know they are in a long-term situation in which they must choose to cooperate or defect time and time again. In this way information such as reputation can then function as a signaling device to other actors.
Games such as the Prisoner's Dilemma work only when presented as a one-off situation in which both actors have no information about what they other is going to do. So the solution is 1) make sure you're not setting up one-off interactions, and 2) make such actors are in an information rich environment wherein they can make appropriate signals about their intentions and past behavior.
Most of the other games, such as Tragedy of the Commons or Battle of the Sexes, have stable solutions if you repeat them and increase the information available to the actors. The equilibrium solutions are called Nash Equilibriums, after John Nash, the paranoid schizophrenic mathematician featured in the movie A Beautiful Mind.
The tragedy of the commons works like this: If everyone follows their self interest and goes fishing in the lake, pretty soon all of the fish will be gone and it will take decades for them to repopulate. If you somehow (how to do it is a coordination problem) arrange for only a small number of people to fish, they can bring in reasonable number of fish every day for many years.
The upshot: In some markets, like consumer electronics, supply and demand regulates things very efficiently. In others, unconstrained individual interest can ruin everything.
In the financial markets, individuals acting rationally for themselves can cause damage to the overall economy: crashes, bubbles, bank runs, etc.
" If everyone follows their self interest and goes fishing in the lake, pretty soon all of the fish will be gone and it will take decades for them to repopulate. "
Or, everyone follows their self interest and realizes that they will need to work out some agreements to moderate fishing or there will be no fish.
> Or, everyone follows their self interest and realizes that they will need to work out some agreements to moderate fishing or there will be no fish.
This can have downsides, namely the sort of "credentialism" that's often derided on this site (licensing for interior decorators, etc.) All the players in some industry decide that it's in their best interest to appeal for regulation which will prevent others from entering this industry.
Oh, absolutely. I just don't buy the idea that the default expected behavior from people is to shot themselves in the foot so blindly (e.g., a race to see who can empty the lake first).
That people may slowly stab themselves in the foot repeatedly is more likely.
The problem is that even if 90% of the fishers in an area are responsible, the remaining 10% can ruin it for everyone. And no one has any incentive, other than their conscience, to cooperate. And that's why it's a tricky problem.
One difference between designing airplanes and govt regulation is that designing airplanes isn't prone to folks making design decisions to benefit to friends and family.
Political corruption is pretty much a fact of life whenever govt gets involved in money.
I don't see why folks who complain about corruption involving govt contracting believe that financial regulation will be any better.
I think this is a better way of thinking about it than "greed". The next natural conclusion that follows is that nobody actually knows how to fix this. It may even be an intrinsic attribute of a growing economy; it may literally be impossible to avoid, short of not having a growing economy.
Centralized management makes the lockstep-motion problem radically worse, solving the positive feedback problem by nearly eliminating feedback and cutting straight to the failure case. ("But the central planners can use feedback!" - not on even remotely the same level of granularity, by definition of "central".) Cutting down the positive feedback loop powering the bubble pretty much inevitable involves cutting out the positive feedback loop driving the economy in general. What's left, really?
Rather than preventing crashes, we probably ought to focus on having the wisdom to be prepared for them when they come.
Growth does not imply the use of more resources. Economic wealth comes from knowledge and command of atoms. Software is wealth, books are wealth, and so on. Such wealth clearly requires a baseline of control over resources and energy, but we get more efficient with it over time so it's not the linear relationship that so many seem to think.
There's a lot of growth left to do, and all precluding our growth now will do is make damn sure we never actually build a sustainable society, because it's not possible with our current technology.
We'll worry about how to manage a non-growing economy when we get there. That situation is so unlike the current situation that even speculation has rather limited value.
(See also: http://www.jerf.org/iri/post/2417 , which is too long to put here and not entirely on topic, but is fleshed out a bit more.)
Or, perhapse "Why is that obvious / why is it a necessity?"
If humanity expands into space, there is absolutely no reason for the population to ever stop growing. Space contains more energy, more living space, and more resources than we could possibly use.(1)
In fact, give the release of resource constraints, there are good arguments to keep growing--"given enough eyes, all bugs are shallow" applies to more than just software. More people is a good thing.
I would have responded earlier if I had noticed this.
You're right that it's not quite a necessity. It's concievable that population growth could continue, albeit at a very slow rate, until people are ready to colonize space. I don't doubt that humans will figure out how to do that-- if they survive long enough.
But it seems likely to me, maybe even obvious, that at the current rate of growth people are going to commit evolutionary suicide long before that becomes possible. This because the earth is very limited and recently economic growth has been really fast. In addition to the Club of Rome stuff recommended above, check out: Capitalism Nature Socialism, a pretty good journal which often discusses these issues, and Beyond Growth by Herman Daly.
Edit: note that it was I who upvoted your comment.
Thanks for the upvote. I had actually wondered--is there any way to know who has up- / down-voted a comment?
As to the growth issue, I'm not quite sure what you mean by "evolutionary suicide." My first guess would be that this term applies to a species that fails to adapt to a changing environment or, in the case of a sentient species, chooses to consciously direct its evolution in a direction that renders it unfit for its environment. That doesn't seem like a good fit for your argument though, so I think what you likely meant was that the population will outgrow the available resources, and there will not be enough food / energy / medicine / etc for everyone. If I'm wrong, please correct me.
Assuming that this is what you meant, I see several implicit assumptions that are worth examining. They are:
1) Humanity does not currently know how to colonize space.
2) It will be S years before we colonize spaces, and S is a large number ("a long time"). This probably means decades, possibly centuries, although you don't specify.
3) The current rate of worldwide population growth will be sustained into the indefinite future.
4) The resources of the Earth are too limited to support the world population that is projected to exist in X years, where X is a small number of decades.
5) X years < S years.
6) Our current knowledge of the Earth's resources is complete and correct.
If any of these points prove incorrect, I would be hard pressed to agree with your argument.
Considering some of the news reports I have seen in the last year or two, I suspect that there might be more resources out there than we think, and that space colonization might be closer than we think:
- Yesterday's NY Times carried a major article about a new technique for natural gas recovery from shale.
- I am aware of at least one study that shows a space elevator could be built for about the price of Boston's Big Dig (i.e. ~10-20B). With such a construct in place, the two major hurdles in space development (launch costs and risk of catastrophic launch failure) go out the window. Even if that study was low by an order of magnitude, it's still something that any G8 government could afford, as could a conglomeration of major corporations.
- JAXA (the Japanese NASA) is currently working on a solar power station project. If powersats prove out, they will represent a major incentive to start building up space infrastructure, which will naturally lead to more people staying in space longer, which will naturally lead to colonization.
- It is pretty clear that population growth tends to level off and even decline once a nation reaches a certain standard of living and life expectancy. Europe and Japan are good examples of this: negative population growth and/or aging populations. More people aren't the problem, they are the solution: we need more minds coming up with ideas, and more hands producing wealth world-wide.
You're right. That's what I meant by evolutionary suicide, and my argument (well, it's not really mine) relies on all six assumptions, more or less.
Although:
Colonizing space and colonizing space to the point where you don't need to worry about earth's carrying capacity are two quite different things. Energy and minerals might be obtainable from space somewhat soon. But we can't eat rocks or sunlight.
I don't claim population growth will be sustained into the indefinite future. I claim that it will be sustained until it stops, and that how it will stop is a question of serious concern.
As I see it there are several possibilities:
1.Population growth continues apace and, due to reproductive time lag, there are suddenly far too many people on the earth. This has a quite catastrophic domino effect and an appalling number of people, maybe 100%, die.
2.Population slows and then stops, but people's lifestyles get more and more decadent and far too many resources are transformed into far too much waste. Something, or many things, run out and disaster ensues.
3.Both population growth and resource consumption naturally find some sustainable level. Maybe that would be somewhere below where it is now. Maybe someone would invent something fabulous, say a glass elevator like in Charlie and the Great Glass elevator, and there is room for lots more consumption and lots more people.
Obviously, I prefer the third option. Maybe it will be that. I'm just sufficiently unsure to be worried about it.
At least in principle, I do not see why it cannot. Growth implies that the GDP is going up, considering that IP/software/etc can all contribute to the GDP then as long as people continue to work creatively then it can in principle grow indefinitely.
Failure is a feature of capitalism. It is to be expected.
However, what shouldn't really happen is a cluster of failures. That's caused by misinformation in the pricing system. Real saving rate doesn't matched up with interest rate, so people are being fooled by non-existent capitals.
Absolutely. Clusters of failures are just popping bubbles built on incorrect information.
Everyone knows that everyone is after more money, it's a prime motivation. This makes it easy to manipulate if you can make a credible looking "investment".
As Marx correctly understood, capitalism is best likened to a force of nature: it doesn't have any moral basis and can have both morally positive and negative consequences. This is why we need governments to keep capitalism in check: unchecked capitalism also leads to activities we find morally unacceptable. Like evolution favors only those traits that add reproductive value, capitalism favors only those activities that increase capital. This may mean obtaining a monopoly, deceiving customers and competitors, employing slaves and generally any other activity we disapprove. In a fully rational market, the customers would not accepts those practices, but as Keynes said: the market can stay irrational longer than you can stay solvent. Moreover, capitalist principles cause an important part of the customers not to care about the immoral consequences: they don't see them because the market isn't transparent or because they do not want to see them. We therefore cannot trust the market to merely expose the positive sides of capitalism. Failures, even clusters of failures, are to be expected and can only be avoided by government intervention. As long as a government lauds the free market and has the irrational expectation that the market will straighten itself, we will continue to suffer the consequences of unchecked capitalism.
Wow, it's hard to know where to begin with this post. It seems strange to me that you're trying to use Marx to justify a liberal position. After all, Marx would be the first to disagree with you about the feasibility of using the state to check the market. Marx saw the state as little more than the 'executive committee' of the bourgeoisie, so using the state to check the market is just not going to happen until revolution leads us out of the capitalist stage of history.
Your equation of evolution and capitalism has a nasty heritage of Social Darwinist thinking. I think it's a misunderstanding of both evolution and of capitalism. Better not to make such an analogy.
Monopolies, deception, fraud, slaves, and so on are not at all unique to capitalism and have been with humanity throughout history. There's no compelling reason to believe that capitalism favors these practices for rational or irrational reasons.
It's not clear why the government is the only institution in society that can avoid failure. It's moreover not clear that the government doesn't play a role in setting up failures as a consequence of its intervention. Marx (or Lenin at the least) would certainly hold that it's the state that creates the very phenomena like monopolies that you decry as "irrational."
My intention is not to use Marx to justify anything; I just want to lay focus on something he and Engels got exactly right and which is being forgotten: capitalism is amoral (a term that is not be confused with immoral). As a result of this, parts of their analysis are extremely valuable and it's a damn shame the anti-communist tendencies have swept that under the rug.
By stating that evolution is an amoral process, I completely condemn any form of Social Darwinism: you cannot derive moral values from amoral principles. Evolution is completely amoral: that which reproduces itself, will continue to exist in some form, independent of our moral judgment about it. Social Darwinists are therefore guilty of the naturalistic fallacy
However, the same goes for capitalism: that which increases capital the most stays in the best position to keep increasing capital. Because of irrationality and a lack of market transparency, the best moral intentions on the part of individual humans cannot prevent immoral consequences from arising. I'm not saying any of those immoral consequences are unique to capitalism. I am saying they are unavoidable when you let capitalism go unchecked.
It's not clear why the government is the only institution in society that can avoid failure
And I cannot prove it is. However, as far as I can see, no other institution has the power nor the mandate to guide capitalist forces. This is the exact reason you have things like the Federal Trade Commission and an Antitrust Division in the DoJ, both on a government level. Who else could do that job?
If the government is trying to head off failures, then it become a quasi-political game. Politics also happens to be a far more violent affair than the play-wars of businesses.
To be threatened with a gun for the purpose of forceful extraction of your wealth is FAR DIFFERENT than peaceful voluntary exchanges championed by capitalists.
Also you misunderstood failures(as in loss). Failures is a feature of capitalism. Failure play an improtant part of resource allocation by telling companies what not to allocate to.
Likewise, success(as in profit) tell companies that this is what consumers want. It could be that consumers want lot of trashy poor quality mass produced goods or high quality expensive goods. The market help companies decide what to produce.
"What boosts a firm’s stock price, and the boss’s standing, is a rapid expansion in revenues and market share. Privately, he may harbor reservations about a particular business line, such as subprime securitization."
The problem is misaligned incentives. Cash bonuses paid out long before the consequences of decisions become clear create problems.
Financial institutions can partly remedy this by replacing annual bonuses with long-term reward schemes like that adopted by Credit Suisse in 2005. If a decision proves to have been reckless/short-sighted a few years on, bonuses can be retracted.
If a decision proves to have been reckless/short-sighted a few years on, bonuses can be retracted.
Retracting something already given can be difficult, especially if the employee has moved on to another place. There are ways, of course, with "clawback clauses" and similar things, but that seems unnecessarily clunky if I understand them correctly.
Basing bonuses on long term performance after it has had long enough to play out and (at least most of) the consequences known may very well be a good idea though.
That's an appealing theory, but the only evidence produced to date that compares the performance of folks who got long-term compensation with those who got short-term shows that the firms that used short-term compensation did better than the ones that used long-term compensation.
Because much of the perceived value in capitalism is based on assumptions, and assumptions can change in value much faster than physical assets?
For instance, Detroit has not been a valuable place to have a business or home for decades but it has faded slowly, but the assumption that housing prices would go up forever was worth trillions, but as soon as it was disproved, those trillions disappeared fast.
Financial market affects today's economy a lot. There's a great demand for assumptions and forecast involved in financial services. However the reliability of today's forecast and analysis methods leaves a lot to be desired. This situation is the cause of the current crisis.
The title of the article is "Rational Irrationality" and not letting the system to crash now, seems to me (and to many others) very irrational.
"too big to fail" syntagm does not look good either to the author of the article, in the end he says:
"As memories of September, 2008, fade, many will say that the Great Crunch wasn’t so bad, after all, and skip over the vast government intervention that prevented a much, much worse outcome. Incentives for excessive risk-taking will revive, and so will the lobbying power of banks and other financial firms." ... "The next time the structure starts to lurch and sway, it could all fall down."
Bailing out big companies is not capitalism as we learned about it. And the bottom line is it hasn't crashed now, the governments did not let it crash. Whether this is good or bad, I guess we will see in the future.
Although capitalism is crash-prone to some extent it is the most reliable system mankind has ever come up with. It is arguably because it's idea of individual success reflects the nature of the animal the system is supposed to work on.
Sure, what I mean is that the pursue for individual success is a key feature of Homo sapiens species. Traditional capitalism declares it as its main trait. The fact that participants of a capitalist system are encouraged to act according to their natural behavioral patterns makes capitalism the most stable system ever known.
Capitalism is one of the few social constructs that encourages negative traits. Who is to say that selfishness is better than altruism? Altruism seems to be a human trait as well -- we do act altruistically on occassion. Is it impossible to build a financial system based on altruism? We tend to think quite highly of those who give of themselves more than take. Why don't we build a financial system that rewards those who think of others? Is such a system impossible?
Society rests on and perhaps morality itself is based on the repression of innate desires. Males have a desire to procreate, but rape is forbidden. We have a desire to promote the success of those with our genes, but discrimination is forbidden. I would say our society is better off because we shun, outlaw, and degrade the exercise of traits we would like to see less of in those who share our society with us.
Humans are at a crossroads here. We have the knowledge, the vision, and the power to evolve ourselves into the species we choose to be. Why, if there is an alternative, would we choose to create a system that rewards those for behaving in a manner that we don't like, in a manner that we would like to see less of?
Imagine saying, "You're so greedy!" Now imagine saying, "You're so giving!" It's a totally different feeling. Imagine a society that promotes the giving over promoting the greedy.
Capitalism creates an environment where both positive and negative traits can be materialized. Everyone agrees that altruism is better than selfishness in moral sense. However one must be quite civilized and have high quality of life and standard of living to exercise altruism. Those whose individual demands are not met simply cannot afford altruism. A financial system aimed at rewarding those who think of others (surprise!) exists right now. Those who have extra money, time or other resources can invest them in those who are in need through loans, grants and scholarships. I'm sure any innovative ideas in this area would be accepted by the HN community and society in general. This proves that the current society/economy is open for a change toward a more altruism-encouraging system.
Western society and morality is based on the relative tolerance to innate human desires and wishes in contrast to other cultures. That's why it has created a financial system that made the countries that chose it developed.
"Males have a desire to procreate, but rape is forbidden. We have a desire to promote the success of those with our genes, but discrimination is forbidden."
Morality restricts certain ways in which a human can fulfill their needs. However rape is not the only way to procreate and discrimination has very little to do with promoting the success of one's genes. Today's social and financial system is successful among other factors because it provides equal and legitimate ways to have sex or multiply for everyone.
We are at crossroads, we have opportunity to choose the way financial system would look like in the future and that's why it is especially important to understand why the current financial system is successful. It is, despite difficulties, which, for example led only to a 2% drop in US GDP growth rate in 2008 compared to 2005. The GDP is still increasing! You can have look at http://www.data360.org/dsg.aspx?Data_Set_Group_Id=353&co... (US GDP statistics) to see how stable economy is. These disturbances are annoying, but their impact on lifestyle is minor.
"Imagine saying, "You're so greedy!" Now imagine saying, "You're so giving!" It's a totally different feeling. Imagine a society that promotes the giving over promoting the greedy."
What does this have to do with the initial argument? Could you please point out where I disagreed with this?
> However one must be quite civilized and have high quality of life and standard of living to exercise altruism.
Not necessarily. There are plenty of people that have nothing that still devote time/money into helping others before themselves.
I hate to reference the Bible, but there was a parable in there about a woman donating 2 cents to the poor verses wealthy men donating piles of money (I think this was 'bags of silver pieces or something like that). She gave more than they did if you consider the percentage of their total assets/worth that was donated.
I wouldn't rely on poor people's altruism as on a valuable contribution to others. Even if giving those 2 cents was an important decision for the woman considering her poorness, the piles given out by a richer man were much more valuable. It is absolute value of a contribution that matters for a receiver of good here.
That may be, but the original statement said that one needed to be 'rich' to exercise altruism, and despite the downmodding of my previous comment that point still stands.
Who is to say that selfishness is better than altruism? Altruism seems to be a human trait as well -- we do act altruistically on occassion.
It's not clear to me whether that's a bug or a feature. Most cases in which altruism would be possible in the ancestral environment are cases in which acting "altruistically" actually benefits the actor in the long run. We're now in a world where it's less clear that those actions will benefit the actor (or actor's genes), on balance. Altruism does result in arguably worse outcomes in many cases (food aid to Africa, liberating Iraq, etc) in the modern world.
Is it impossible to build a financial system based on altruism?
Well, it doesn't seem stable. A financial system based on self-interest is stable when there are a few people with altruistic motives in the system. A financial system based on altruism would rapidly fail in the presence of a few self-interested people in the system.
> We have a desire to promote the success of those with our genes, but discrimination is forbidden.
How so? Am I not allowed to be a white male that exclusively dates white females? Is there affirmative action for dating/mating now?
Discrimination is forbidden in ways that have nothing to do with genes and evolution. Do you really think that the KKK was all about 'promoting better genes?' In modern times, they may have convinced themselves of this but it is bunk science. Neo-Nazies, skinheads, KKK members, all of them discriminate against others, not because they want to 'get rid of others to promote their genes.'
And you say 'those with our genes,' but that normally means one's offspring. Even within the same 'tribes,' preference is given to one's offspring above all else.
This has more to do with human nature to want to join a community and exclude others, which is closer to our warring nature than anything to do with procreation.
Altruism require giving away resource with no expectation of benefit. So in short, it is sucide of the individual if practiced to extreme.
A more substainable form is voluntary exchanges in which individuals exchange good and services for their percevied benefit. That's cooperation based on "stratching your back, you stratch my back".
Also note that individuals are what composed of groups. They are the smallest building block of society. Focus on them, and you benefit the group. The ideal situtation is that you would never need to sacrifice an individual for the whole group.
The author of the article almost alludes to this when they write:
>Consider the freeze that started in August of 2007. Each bank was adopting a prudent course by turning away questionable borrowers and holding on to its capital. But the results were mutually ruinous: once credit stopped flowing, many financial firms—the banks included—were forced to sell off assets in order to raise cash. This round of selling caused stocks, bonds, and other assets to decline in value, which generated a new round of losses.
We might be less crash-prone if we had system that didn't crash when the growth of the money supply was interrupted. It seems like it's not just maintaining the current level of credit is necessary for stability, but the indefinite exponential expansion of credit is necessary to stave off problems.
The vast majority of economists thought that such a crash couldn't happen. They thought that it was impossible because we are so much smarter now than we were back in the 1920's.
I have been seeing a lot of these "Look at this forgotten/unknown economist who predicted/explained everything!" type of articles. Yes, they are probably almost always wrong, but how is everyone on HN supposed to know?
The Prisoner’s Dilemma is the obverse of Adam Smith’s theory of the invisible hand, in which the free market coördinates the behavior of self-seeking individuals to the benefit of all. Each businessman “intends only his own gain,” Smith wrote in “The Wealth of Nations,” “and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention.” But in a market environment the individual pursuit of self-interest, however rational, can give way to collective disaster. The invisible hand becomes a fist.
and
Limiting the development of [complex] securities would stifle innovation, the financial industry contends. But that’s precisely the point. “The goal is not to have the most advanced financial system, but a financial system that is reasonably advanced but robust,” Viral V. Acharya and Matthew Richardson, two economists at N.Y.U.’s Stern School of Business, wrote in a recent paper. “That’s no different from what we seek in other areas of human activity. We don’t use the most advanced aircraft to move millions of people around the world. We use reasonably advanced aircrafts whose designs have proved to be reliable.”
The second seems quite applicable to software engineering as well.