3 ms·

People are rational actors at optimizing their own well-being, as they see it. That holds in all cases but it isn't very helpful to the politicians. What break

People are rational actors at optimizing their own well-being, as they see it. That holds in all cases but it isn't very helpful to the politicians. What breaks is trying to get people to optimize for some other factor by manipulating economic signals (e.g. with inflation). Once they know about the manipulation they'll compensate for it. Not 100%, not always even in the right direction, but enough to make the eventual outcome chaotic.

geysersam 4y ago
> People are rational actors at optimizing their own well-being What experiment can we make to empirically verify this claim? If no experiment is needed then the statement is a tautology.
- nybble41 4y ago
  I'm fine with considering it a tautology, as that implies you're using the correct definition for "rational actor" and the correct perspective (i.e., theirs) for evaluating the expected effect on their well-being. The problem is the recurring claim that people are not rational actors, either because of imperfect information (sometimes things don't turn out as expected) or because what people choose to do isn't always what someone else thinks of as their own best interest. That isn't what is meant when we say that people are rational actors. Anything can be a rational action within some system of preferences; the point of the "rational actor" model is that if you assume people are potentially irrational actors, deliberately making choices which they expect a priori to satisfy their preferences worse than other options they are aware of, then the system is underconstrained and you can't say anything about people's preferences based on observations about their choices. Assuming rationality doesn't constrain people's behavior at all, but it is a more useful way of interpreting it.
  - geysersam 4y ago
    That sounds like an unnecessarily complicated way of going about it. If someone does something really strange, we have to interpret it as a rational action in a strange value system? Instead of stating the obvious: sometimes people do strange things. Would you also consider animal behavior rational by definition? Why not? I think this way of framing things, while not imposing any constraints, as you point out, is just quite confusing. We know people are not perfect logicians, why pretend that they are? (except of course in the situations where this is a good approximation).
    - nybble41 4y ago
      > If someone does something really strange, we have to interpret it as a rational action in a strange value system? Pretty much, yes. That's the only way you get any useful information. Look, everyone knows that people sometimes make choices they later regret. That does not refute the rational actor model, because the rational actor model is about what they knew and what their hidden preferences were in the moment, at the time they made the choice. That includes preferences they may not even understand themselves. To refute the rational actor model you would need to show that the person knew there was a better choice, one whose a priori expected outcome was superior in all respects from their own point of view, and deliberately decided not to take it. Since you can't know how they valued all the possible outcomes from their point of view—they don't even know that—that means it can't be refuted at all. It's a framework for understanding people's revealed preferences, not a means of empirically predicting their behavior. At best you can use it to say that if people prefer A over B then they will choose A and not B, but it doesn't tell you whether a person will actually prefer A over B at any given point in the future, even if you can infer from their actions that they did in the past. Preferences change; circumstances change. No two choices are ever made in exactly the same context. > We know people are not perfect logicians, why pretend that they are? It's not a matter of perfect logic. It's saying that people (are assumed to) pick the course of action that seems to lead to be best outcome from their own point of view. That can, of course, include others' well-being, if they value such things. There is no requirement that their expectations be perfectly logical, or grounded in perfect knowledge. There is a separate principle, not part of the rational actor model, which basically says that people either learn to hold expectations which closely align with reality over time or tend to lose wealth, and thus influence in the economy, over time. This implies that most of the economically significant actions are taken by those who make decisions they are not likely to later regret. Of course this breaks down a bit when poor decisions get subsidized, propping up the economic influence of poor decision-makers.
      - geysersam 4y ago
        > Pretty much, yes. That's the only way you get any useful information. Nah I don't agree. Sometimes people act irrationally. A otherwise friendly person might seem unfriendly because they've had too little to eat. Or because they had a bad day at work. Typically they would agree that their temporary unfriendlyness is not in accordance with their eithical principles, nor generally in theirs or anyone else's best interest. They (we) are just limited as beings. Unable to do what they know to be best. Another way to look at it could be to say that people are not a single actor. What is the self? What we perceive as a unified I could very well be a much more shattered / distributed fenomena. Relying on the "rational actor model" limits the perspectives we can take on what a conscious being really is. And I do think logic is a *requirement* for rationality (agree knowledge is not though). Maybe perfect logic is not necessary, behavior of sufficiently rational actors is perhaps well approximated by the rational actor model. But maybe not! Question is - will increasingly rational actors eventually converge to the same behavior? If not - then rational actor theory cannot even approximate human behavior, it intrinsically has to be a theory about "human logic".

geysersam 4y ago

> People are rational actors at optimizing their own well-being What experiment can we make to empirically verify this claim? If no experiment is needed then the statement is a tautology.

nybble41 4y ago
I'm fine with considering it a tautology, as that implies you're using the correct definition for "rational actor" and the correct perspective (i.e., theirs) for evaluating the expected effect on their well-being. The problem is the recurring claim that people are not rational actors, either because of imperfect information (sometimes things don't turn out as expected) or because what people choose to do isn't always what someone else thinks of as their own best interest. That isn't what is meant when we say that people are rational actors. Anything can be a rational action within some system of preferences; the point of the "rational actor" model is that if you assume people are potentially irrational actors, deliberately making choices which they expect a priori to satisfy their preferences worse than other options they are aware of, then the system is underconstrained and you can't say anything about people's preferences based on observations about their choices. Assuming rationality doesn't constrain people's behavior at all, but it is a more useful way of interpreting it.
- geysersam 4y ago
  That sounds like an unnecessarily complicated way of going about it. If someone does something really strange, we have to interpret it as a rational action in a strange value system? Instead of stating the obvious: sometimes people do strange things. Would you also consider animal behavior rational by definition? Why not? I think this way of framing things, while not imposing any constraints, as you point out, is just quite confusing. We know people are not perfect logicians, why pretend that they are? (except of course in the situations where this is a good approximation).
  - nybble41 4y ago
    > If someone does something really strange, we have to interpret it as a rational action in a strange value system? Pretty much, yes. That's the only way you get any useful information. Look, everyone knows that people sometimes make choices they later regret. That does not refute the rational actor model, because the rational actor model is about what they knew and what their hidden preferences were in the moment, at the time they made the choice. That includes preferences they may not even understand themselves. To refute the rational actor model you would need to show that the person knew there was a better choice, one whose a priori expected outcome was superior in all respects from their own point of view, and deliberately decided not to take it. Since you can't know how they valued all the possible outcomes from their point of view—they don't even know that—that means it can't be refuted at all. It's a framework for understanding people's revealed preferences, not a means of empirically predicting their behavior. At best you can use it to say that if people prefer A over B then they will choose A and not B, but it doesn't tell you whether a person will actually prefer A over B at any given point in the future, even if you can infer from their actions that they did in the past. Preferences change; circumstances change. No two choices are ever made in exactly the same context. > We know people are not perfect logicians, why pretend that they are? It's not a matter of perfect logic. It's saying that people (are assumed to) pick the course of action that seems to lead to be best outcome from their own point of view. That can, of course, include others' well-being, if they value such things. There is no requirement that their expectations be perfectly logical, or grounded in perfect knowledge. There is a separate principle, not part of the rational actor model, which basically says that people either learn to hold expectations which closely align with reality over time or tend to lose wealth, and thus influence in the economy, over time. This implies that most of the economically significant actions are taken by those who make decisions they are not likely to later regret. Of course this breaks down a bit when poor decisions get subsidized, propping up the economic influence of poor decision-makers.
    - geysersam 4y ago
      > Pretty much, yes. That's the only way you get any useful information. Nah I don't agree. Sometimes people act irrationally. A otherwise friendly person might seem unfriendly because they've had too little to eat. Or because they had a bad day at work. Typically they would agree that their temporary unfriendlyness is not in accordance with their eithical principles, nor generally in theirs or anyone else's best interest. They (we) are just limited as beings. Unable to do what they know to be best. Another way to look at it could be to say that people are not a single actor. What is the self? What we perceive as a unified I could very well be a much more shattered / distributed fenomena. Relying on the "rational actor model" limits the perspectives we can take on what a conscious being really is. And I do think logic is a *requirement* for rationality (agree knowledge is not though). Maybe perfect logic is not necessary, behavior of sufficiently rational actors is perhaps well approximated by the rational actor model. But maybe not! Question is - will increasingly rational actors eventually converge to the same behavior? If not - then rational actor theory cannot even approximate human behavior, it intrinsically has to be a theory about "human logic".