5 ms·
GDP might be the ultimate example of Goodhart's Law: https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law
by xnx 4y ago
GDP might be the ultimate example of Goodhart's Law: https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law
- tabtab 4y agoA common example is "lines of code". If you pay by lines of code, then coders will bloat up code to pad their metrics, making it verbose and redundant.
- wildrhythms 4y agoMore common is some mostly pointless (or annoying) new feature that an overzealous project manager wants to get users to 'use' (thus boosting some metric they can point at to secure a promotion), so they have the engineers implement a gigantic popup pointing at it. Or worse, they put the entrypoint to the feature in some high touch area so all of the power users accidentally click on it, thus logging an interaction and making the line go up. https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law
- tootie 4y agoYup. There is simply no way we can replace GDP with another metric. There's always a thousand smartasses who say the unemployment rate is fake. It's not fake, it's an indicator and can never tell a complete story. Employment, economic activity, inflation, trade are all drowning in data points that inform some specific minutiae. The solution is to not base policy on indicators but politics needs to keep things simple.
- Ambolia 4y agoUnemployment would be less misleading if it was just "total people working" / "total people in the country".
- strangeattractr 4y agoNo it wouldn’t. It would be bloated by the number of people with serious disabilities, health conditions and retirees. As your population ages you would get a bump in unemployment that tells you nothing about the demand for labour. In my first Economics subject in high school they told us to never look at unemployment without looking at the participation rate with it.
- Ambolia 4y agoI think it would be useful to know if a country has a much higher dependent population than another country. It could highlight dynamics that get ofuscated when some countries put large parts of their populations under the "unemployable" label or quietly pushed into black markets, making official unemployment numbers impossible to compare with other populations.
- tootie 4y agoThat's called employment-population ratio and it is tracked by the BLS. It's actually a bit more valuable to track prime age employment-population ratio to factor out students and retirees. But it's still a broad indicator.
- nicoburns 4y agoI'd argue that "free market" pricing in general is. The attitude that "something is valuable if people are willing to pay for it" is being increasingly overfitted.
- Phiwise_ 4y agoOverfitted to what? If the whole point of economics in a moral sense is to get people more value, than what does the question of if person A will buy item 1 at a higher price than person B for item 2 for better or worse reasons get the average person/company except for a distraction from their goal? Make item 1, sell it to person A, and go enjoy your more prosperity than you would have gotten for selling item 2 to person B. This proposal needs to be much more concrete to mean anything.
- nicoburns 4y agoOverfitted to the current legal context, market conditions, regulations, etc. A classic example would be a company which outcompetes its rivals by working around regulation that companies in that sector are subject to. That enables it to gain a competitive advantage within the system it's operating in, but is pessimisation rather than an optimisation when looking at the bigger picture and longer timescales.
- BurningFrog 4y agoBy definition, the monetary value of something is what people are willing to pay for it.
- ummonk 4y agoYes but that's not the same as the more useful (no pun intended) definition of value - namely, utility.
- BurningFrog 4y agoThey're both useful, for different purposes. I recommend being clear about which definition is used, rather than trying to enforce one meaning should always be used.
- sbelskie 4y agoBy this do you mean that GDP is being gamed?
- gumby 4y agoBasically yes, to the Bastiat point.
- cbsmith 4y agoI don't know about "ultimate", but it sure is a good example of one. In fairness though, if a politician stands at a stump and says, "see, I'm doing great with GDP", but it's because they've been serving the metric rather than the underlying principles, they're likely not going to be well received.
- club_tropical 4y agoI hate this law. There is no example of a good measure that does not become a target, nor should there be. The only question is who does the evaluating. The typical example is: hospital emergency room wait times were made a "target" and therefore hospitals stopped admitting ambulances (literally having them circling around the block) before they had the staff freed up. But there's nothing wrong with emergency room wait times being kept low. The obvious locus of evaluation, however, was moved from the patient (who is evaluating his whole end-to-end emergency care on many metrics, including but not limited to emergency room wait time) to some bureaucrat who assesses the hospital only on the wait time.