The reason you can’t rigidly separate positive from normative economics is that you can’t rigidly separate claims of fact from claims of value in general. Human language is too laden with thick concepts that mix the two. The claim that someone is a “slut” or a “bitch”, for example, melds together factual claims about a woman’s behavior with a lot of deeply embedded normative concepts about what constitutes appropriate behavior for a woman. The claim that financial markets are “efficient” is both an effort to describe their operation and a way of valorizing them. The idea of a “recession” or “full employment” or “potential output” all embed certain ideas about what would constitute a normal arrangement of human economic activity (...) You could try to rigorously purge your descriptions of the economy of anything that vaguely smells of a thick moral concept, but you’d find yourself operating with an impoverished vocubulary unable to describe human affairs in any kind of reasonable way.
I found that very poignant, but I’m not sure I agree with his final claim. I think he’s committing the usual mistake of claiming impossible what seems hard.
Is it even hard?
JFDI,
or as we might say here, shut up and do the impossible. Is “efficient” a tendentious word? Taboo it. Is discussion being confused by mixing normative and positive concepts? DDTT.
Yeah, agreed. It’s entirely possible to describe a system of economic agents without using such value-laden terns (though in some cases we may have to make up new terms). We don’t do it, mostly because we don’t want to. Which IMHO is fine; there’s no particular reason why we should.
Yglesias seems to be committing an error here by confusing technical jargon with common English. Efficient has a very specific meaning in economics (well, two specific meanings, depending on what kind market you’re talking about). The word efficient is not meant to refer to universal goodness and it’s a mistake to treat it as if it were.
The three you mention are all subtypes of the same efficiency—informational efficiency. Informational efficiency is used in finance and refers to how well a financial market incorporates information into prices. Basically a market is informationally efficient if you can’t out-predict without using information it doesn’t have. The weak / semi-strong / strong distinction merely indicates how much information it is incorporating into prices: weak means it’s incorporating it’s own past prices, semi-strong includes all public information, and strong includes all information held in private as well.
The other type of efficiency is allocative efficiency, a concept used in microeconomics. An allocatively efficient market is one that assigns goods to the people who place the highest value on them (subject to the constraints of each person’s endowments). It is effectively a utility-maximising condition. The whole concept of market failure in economics is built around situations where markets are failing to be allocatively efficient.
The first thought that I have when considering how to describe the economy without using normative language is that all of the values that are commonly measured (i.e. GDP, unemployment, etc.) are chosen to be measured because they are proxies for things that people value.
In fact, the whole study of economics seems to me like the study of things people value and how they are distributed. If you choose proxies for value you’re having a profound effect on what gets measured (consider the recent discussions of statistical significance as a proxy for evidence) and if you try to list everything that everyone values you end up butting up against unsolved problems.
--Matt Yglesias
I found that very poignant, but I’m not sure I agree with his final claim. I think he’s committing the usual mistake of claiming impossible what seems hard.
Is it even hard? JFDI, or as we might say here, shut up and do the impossible. Is “efficient” a tendentious word? Taboo it. Is discussion being confused by mixing normative and positive concepts? DDTT.
The quote smells like rationalising to me.
Yeah, agreed. It’s entirely possible to describe a system of economic agents without using such value-laden terns (though in some cases we may have to make up new terms). We don’t do it, mostly because we don’t want to. Which IMHO is fine; there’s no particular reason why we should.
Yglesias seems to be committing an error here by confusing technical jargon with common English. Efficient has a very specific meaning in economics (well, two specific meanings, depending on what kind market you’re talking about). The word efficient is not meant to refer to universal goodness and it’s a mistake to treat it as if it were.
I know of three, although it is a matter of parametrization (weak, strong, semi-strong). What two meanings do you have in mind?
The three you mention are all subtypes of the same efficiency—informational efficiency. Informational efficiency is used in finance and refers to how well a financial market incorporates information into prices. Basically a market is informationally efficient if you can’t out-predict without using information it doesn’t have. The weak / semi-strong / strong distinction merely indicates how much information it is incorporating into prices: weak means it’s incorporating it’s own past prices, semi-strong includes all public information, and strong includes all information held in private as well.
The other type of efficiency is allocative efficiency, a concept used in microeconomics. An allocatively efficient market is one that assigns goods to the people who place the highest value on them (subject to the constraints of each person’s endowments). It is effectively a utility-maximising condition. The whole concept of market failure in economics is built around situations where markets are failing to be allocatively efficient.
The first thought that I have when considering how to describe the economy without using normative language is that all of the values that are commonly measured (i.e. GDP, unemployment, etc.) are chosen to be measured because they are proxies for things that people value.
In fact, the whole study of economics seems to me like the study of things people value and how they are distributed. If you choose proxies for value you’re having a profound effect on what gets measured (consider the recent discussions of statistical significance as a proxy for evidence) and if you try to list everything that everyone values you end up butting up against unsolved problems.