[skybrian]

The problem is that most incentives can be gamed. There are good and bad ways to get customers to buy more or make the stock price go up.

People are smarter than metrics. Increase the incentives and you also increase the incentive to cheat. Sometimes the best you can do is insulate people from incentives, so people have the leeway to do the right thing without acting against their own best interest.

[kbenson]

Well, that's why there's generally a difference between market incentives and regulatory incentives. In this case, I believe we are referring to market based incentives, where gaming them is of limited use because the market should respond to that gaming.

That said, regulatory incentives and punishments have their place too, IMO generally where the market isn't responding well, or information isn't available enough to allow for an efficient market, or as a response to some other regulatory market effects. But, as you noted, you'll often get some interesting behavior right at the edge of where the regulation kicks in because there's often a hard change instead of a gradual shift as an efficient market would allow.

[skybrian]

If gaming the markets is of limited use, I wonder why market participants put so much effort into it?

I mean, consider Wall Street, Las Vegas, used car sales - the list goes on and on. It's a rich area of storytelling that goes back to the dawn of recorded history.

[kbenson]

If you say what you mean by gaming the system with any of those, Wall Street, Las Vegas, used car sales, maybe I can respond intelligently. As it is, I'm not sure what you're referring to. Most things I think of that would be considered gaming the system to do with those have to do with human rules applied to a system, instead of emergent market economics, and that's what I was referring to.

There is gaming of markets, but I think generally if it's not based on some regulation, it's because of information asymmetry (which is a market inefficiency).

All I was trying to point out in the prior comment is that there are different kinds of incentives. There are incentives that are constructed, and there are incentives that are natural. Constructed incentives are much easier to game. Natural incentives are emergent. Microsoft is incentivized to have good security for their OS now by the market in general, because it hurts them to not have good security (compared the the bast, where they could get away with lax security until it became a problem). That's emergent from the market and people deciding to use or not use their product. I wouldn't consider that "gaming the system", and if they did game it by talking a lot about security but not actually doing much, eventually the market should note that and respond appropriately.

Alternatively, Microsoft can reduce their tax burden by shifting business entities to different countries and shuffling how it appears their profit is created, so it's registered in a country with very little taxes, leading them to pay fewer taxes (not that they do, I don't know. I believe Apple and Google are reputed to do this). That's based on rules set by people, such as country boundaries and tax rates. Doing this could be considered "gaming the tax system". It requires specific changes to the rules to fix, it won't just shift naturally.

To me it appeared the comment you were responding to originally was using "incentive" in the pure form, meaning "benefit for doing so", and it appeared you were referring to incentive in the regulatory sense, where it's a human construction to influence behavior, but that's only a subset of the meaning.

[skybrian]

At a high level, the market is supposed to provide an incentive to serve customers - "solving a problem" or "help people" as the original post put it.

My point is that real-world incentives are never perfectly aligned with such lofty, nebulous goals. They are about things you can measure such as how much money you can make. Making money is not the same as helping people and no incentive scheme is clever enough to make it so. Customers are often smarter than rules but even then, customers can be fooled. So there will always a way to make money without helping people and when you increase incentives, it also increases incentive to do things that aren't actually the goal.

This means that to some extent we rely on people to follow the spirit of the incentives and not to simply be amoral incentive-maximizers.

(This is closely related to the principal-agent problem, except the principal here is society in general.)

[kbenson]

> At a high level, the market is supposed to provide an incentive to serve customers

That's not what markets do. Markets are a means of efficiently and accurately pricing things in a responsive way. Some markets don't even have customers. That price may or may not be money, depending on the market in question.

> My point is that real-world incentives are never perfectly aligned with such lofty, nebulous goals.

For markets, definitely not, since that's not really what they are for, and any created incentive will at best attempt to move a market towards that.

> Making money is not the same as helping people and no incentive scheme is clever enough to make it so. Customers are often smarter than rules but even then, customers can be fooled. So there will always a way to make money without helping people and when you increase incentives, it also increases incentive to do things that aren't actually the goal.

Nobody here has said it is. The original comment noted "People acting in their own interest are reliable." I interpreted that to mean "when there are forces urging a person or group to act a certain way for their own self interest, it's easier to rely on them to continue acting that way". If Microsoft benefits from doing something that benefits others, it's easier to rely on them to continue doing that. I'm still not sure what point you were trying to make from that, since I'm not following how your latest comment relates to that or to my call for clarification, since I thought maybe you were interpreting the statement somewhat differently than I was.

[skybrian]

My point is that "Microsoft benefits from doing something that benefits others" is never going to be entirely true, or at least not for long.