Scott Sumner often points out that the market responds instantly to changes in monetary policy, and it responds based on its expectations of the future path of monetary policy. During the relevant period for the Bank of Japan, looking at market reactions to its actions and announcements it is utterly obvious that the market expects Japan to do better when the bank prints more money or people think it will print more, and do worse when the bank prints less money or people think it will print less.
The modesty argument should (and by Scott frequenly has been) actually be made against the Bank of Japan. The market is screaming that the bank should print more money, so what right does a committee have to decide it knows better? The counter-argument is that you could read the market as saying that given BoJ has made a decision, decisions to print more money are good news, but that BoJ has other considerations slash hidden information, so it can be wrong to print more money but right for the decision to not print to be interpreted as bad news. This in turn requires BoJ to have hidden information, which could be political rather than economic. I would second Eliezer in recommending The Midas Paradox if you want to know more about such things.
How did the Bank make such a huge mistake? I can think of a number of good reasons, all of which come down to politics, perception and the interests of the individuals making the decision, including their intellectual commitments. Their utility function is not the RGDP of Japan, or human flourishing, or anything remotely like either of these things. There’s also the bad reason of they thought they were right. A mix of both seems plausible—some members thought they were right, others weren’t sure, and their incentives were bad.
I assume it is a major point of the full book that in many or most situations, the academic field / governmental institution / major company is not optimizing (or at least, only partially optimizing) for the right answer, so there shouldn’t be much presumption that they will get the right answer.
The modesty argument should (and by Scott frequenly has been) actually be made against the Bank of Japan.
This is true, but I was actually trying to zoom into the particular encounter between Jack and Eliezer, where Jack didn’t know the other evidence, and the question was one of whether it was accurate to go “Given my current state of knowledge, I know that Eliezer is being overconfident” or whether it was accurate to go “Given my current state of knowledge, I don’t know that Eliezer is being overconfident”, and that the disagreement is an empirical one about the background state of institutions in the world, as opposed to one of discussing ideal bayesian agents.
I am talking about group rationality a bit, but I’m realising more and more how much modesty is a strong empirical claim as opposed to a theoretical one.
I think this comes down to many things, including Eliezer’s history of calibration in such situations, and the state of such institutions in general. In this case I think Jack was being quite fair to think that given what he knew about Eliezer, and his knowledge of Japan at the time, on average Eliezer was being overconfident here.
But that’s a calibration question as opposed to a question of whether it is generally reasonable for a person such as Eliezer to think that BoJ could be doing something insane, or what evidence Eliezer would need before being able to claim that.
To that, I would strongly answer that it is very reasonable to think BoJ could be doing something insane in this spot, it’s just a question of how much evidence you have and how confident you should be, even before we learn that actually it is consensus reality among economists and traders not in BoJ that BoJ is acting insane.
I’d also make the argument that the very fact that everyone thinks BoJ was doing something insane, and it turned out to have done something insane, is evidence that institutions like the BoJ do insane things. Not only do they do insane things, they often don’t fix them even when everyone is telling them their actions are insane.
Obviously this is one non-random example, so it isn’t that strong as evidence on its own, but the class of thing “seemingly insane thing being done by major institution that people told them was insane, they eventually fixed, and that then turned out to be insane” is reasonably large.
At a minimum such institutions strongly disagree with the argument from modesty, in the sense that they seem to believe things that the outside world tells them are wrong, quite often. Using the argument from modesty, to believe decisions made by groups and institutions that are disrespecting the argument from modesty, is at least highly weird. Why trust a group to make better decisions than you, when they’re using a worse rule to make their decisions?
Scott Sumner often points out that the market responds instantly to changes in monetary policy, and it responds based on its expectations of the future path of monetary policy. During the relevant period for the Bank of Japan, looking at market reactions to its actions and announcements it is utterly obvious that the market expects Japan to do better when the bank prints more money or people think it will print more, and do worse when the bank prints less money or people think it will print less.
The modesty argument should (and by Scott frequenly has been) actually be made against the Bank of Japan. The market is screaming that the bank should print more money, so what right does a committee have to decide it knows better? The counter-argument is that you could read the market as saying that given BoJ has made a decision, decisions to print more money are good news, but that BoJ has other considerations slash hidden information, so it can be wrong to print more money but right for the decision to not print to be interpreted as bad news. This in turn requires BoJ to have hidden information, which could be political rather than economic. I would second Eliezer in recommending The Midas Paradox if you want to know more about such things.
How did the Bank make such a huge mistake? I can think of a number of good reasons, all of which come down to politics, perception and the interests of the individuals making the decision, including their intellectual commitments. Their utility function is not the RGDP of Japan, or human flourishing, or anything remotely like either of these things. There’s also the bad reason of they thought they were right. A mix of both seems plausible—some members thought they were right, others weren’t sure, and their incentives were bad.
I assume it is a major point of the full book that in many or most situations, the academic field / governmental institution / major company is not optimizing (or at least, only partially optimizing) for the right answer, so there shouldn’t be much presumption that they will get the right answer.
This is true, but I was actually trying to zoom into the particular encounter between Jack and Eliezer, where Jack didn’t know the other evidence, and the question was one of whether it was accurate to go “Given my current state of knowledge, I know that Eliezer is being overconfident” or whether it was accurate to go “Given my current state of knowledge, I don’t know that Eliezer is being overconfident”, and that the disagreement is an empirical one about the background state of institutions in the world, as opposed to one of discussing ideal bayesian agents.
I am talking about group rationality a bit, but I’m realising more and more how much modesty is a strong empirical claim as opposed to a theoretical one.
I think this comes down to many things, including Eliezer’s history of calibration in such situations, and the state of such institutions in general. In this case I think Jack was being quite fair to think that given what he knew about Eliezer, and his knowledge of Japan at the time, on average Eliezer was being overconfident here.
But that’s a calibration question as opposed to a question of whether it is generally reasonable for a person such as Eliezer to think that BoJ could be doing something insane, or what evidence Eliezer would need before being able to claim that.
To that, I would strongly answer that it is very reasonable to think BoJ could be doing something insane in this spot, it’s just a question of how much evidence you have and how confident you should be, even before we learn that actually it is consensus reality among economists and traders not in BoJ that BoJ is acting insane.
I’d also make the argument that the very fact that everyone thinks BoJ was doing something insane, and it turned out to have done something insane, is evidence that institutions like the BoJ do insane things. Not only do they do insane things, they often don’t fix them even when everyone is telling them their actions are insane.
Obviously this is one non-random example, so it isn’t that strong as evidence on its own, but the class of thing “seemingly insane thing being done by major institution that people told them was insane, they eventually fixed, and that then turned out to be insane” is reasonably large.
At a minimum such institutions strongly disagree with the argument from modesty, in the sense that they seem to believe things that the outside world tells them are wrong, quite often. Using the argument from modesty, to believe decisions made by groups and institutions that are disrespecting the argument from modesty, is at least highly weird. Why trust a group to make better decisions than you, when they’re using a worse rule to make their decisions?
John?