5 ms·

> Most economic theories are tested by the scientific method I don't think they are. Can you give me an example of a prominent theory developed in the last thi

> Most economic theories are tested by the scientific method I don't think they are. Can you give me an example of a prominent theory developed in the last thirty years that is considered by the mainstream to be debunked by data?

kurthr 9y ago
I even agree with you that economics is pretty soft as science goes, more math but also a lot of p-hacking, but... they are argued, provide data, and are evaluated just like the other Social Sciences. See Picketty as a pop-econ example with plenty of research backing. The idea that in order for something to be a science a prominent theory developed in the last 30 years has to be debunked by data is silly. Most of the questionable basis for economics is much more than 30 years old (1988?). I might as well ask you the same question in Physics... is Physics no longer a science?
- soVeryTired 9y ago
  I'm not a physicist but the discovery of the Higgs boson probably put paid to a number of competing theories [0]. Some papers probably are well-reviewed and evaluated. But as a small experiment, I picked a paper at random (Artificial Intelligence, Automation and Work [1]) from this year's AEA conference and had a look through it. Despite having a promising title, there is (to my mind) no content in there. Nothing but unjustified assumptions about 'production functions' and 'productivity'. I don't see how a paper like that tells us anything about artificial intelligence or automation or work. [0] https://en.wikipedia.org/wiki/Alternatives_to_the_Standard_Higgs_Model https://en.wikipedia.org/wiki/Alternatives_to_the_Standard_H... [1] https://www.aeaweb.org/conference/2018/preliminary/paper/HZrKTnsf https://www.aeaweb.org/conference/2018/preliminary/paper/HZr...
  - kurthr 9y ago
    Well, kinda. The Higgs and Higgsless proposals mostly go back to the 60s and mid 70s, even if some papers are later. The actual discovery was on 2012 40-50 years later and the Nobel handed out. A random short paper from even a good journal like Phys Rev Letters is unlikely to be Nobel earthbreaking either, much less have effort to debunk it. As an undergrad I had to discover that the 20 year olf PRL used as a basis for the 2 week Lab (Thorium C x-section) actually had an incorrect analysis. All data taken was at the wrong energies so we had to spot the error, rederive the equation and values to measure it correctly before the halflife brought our counts down too low. That PRL was never corrected and the TAs comment was, "that'll teach you to trust some random derivation you found in an article somewhere". Similar story about the incorrect Hydrogen radius expectation value integrating factor in Baym. Even text books aren't always correct.
UncleEntity 9y ago
I'd be happy to see a mainstream economic theory that hasn't been debunked by data...

kurthr 9y ago

I even agree with you that economics is pretty soft as science goes, more math but also a lot of p-hacking, but... they are argued, provide data, and are evaluated just like the other Social Sciences. See Picketty as a pop-econ example with plenty of research backing. The idea that in order for something to be a science a prominent theory developed in the last 30 years has to be debunked by data is silly. Most of the questionable basis for economics is much more than 30 years old (1988?). I might as well ask you the same question in Physics... is Physics no longer a science?

soVeryTired 9y ago
I'm not a physicist but the discovery of the Higgs boson probably put paid to a number of competing theories [0]. Some papers probably are well-reviewed and evaluated. But as a small experiment, I picked a paper at random (Artificial Intelligence, Automation and Work [1]) from this year's AEA conference and had a look through it. Despite having a promising title, there is (to my mind) no content in there. Nothing but unjustified assumptions about 'production functions' and 'productivity'. I don't see how a paper like that tells us anything about artificial intelligence or automation or work. [0] https://en.wikipedia.org/wiki/Alternatives_to_the_Standard_Higgs_Model https://en.wikipedia.org/wiki/Alternatives_to_the_Standard_H... [1] https://www.aeaweb.org/conference/2018/preliminary/paper/HZrKTnsf https://www.aeaweb.org/conference/2018/preliminary/paper/HZr...
- kurthr 9y ago
  Well, kinda. The Higgs and Higgsless proposals mostly go back to the 60s and mid 70s, even if some papers are later. The actual discovery was on 2012 40-50 years later and the Nobel handed out. A random short paper from even a good journal like Phys Rev Letters is unlikely to be Nobel earthbreaking either, much less have effort to debunk it. As an undergrad I had to discover that the 20 year olf PRL used as a basis for the 2 week Lab (Thorium C x-section) actually had an incorrect analysis. All data taken was at the wrong energies so we had to spot the error, rederive the equation and values to measure it correctly before the halflife brought our counts down too low. That PRL was never corrected and the TAs comment was, "that'll teach you to trust some random derivation you found in an article somewhere". Similar story about the incorrect Hydrogen radius expectation value integrating factor in Baym. Even text books aren't always correct.

UncleEntity 9y ago

I'd be happy to see a mainstream economic theory that hasn't been debunked by data...