10 ms·

Predictable...but are those aholes wrong...? :-) https://news.ycombinator.com/item?id=44013692 https://news.ycombinator.com/item?id=44013692 Economics violates

by belter 1y ago

Predictable...but are those aholes wrong...? :-) https://news.ycombinator.com/item?id=44013692 https://news.ycombinator.com/item?id=44013692 Economics violates Popper demarcation criterion. Economic theories can't be falsified because you can't run controlled experiments on economies, rewind history, or isolate variables. When models fail, economists adjust assumptions ... Unfalsifiable = Unscientific.

rtsil 1y ago
Neither are litterature and peace.
- belter 1y ago
  They are not pretending to be sciences.
- immibis 1y ago
  Or astrophysics (not a Nobel prize but obviously a real science despite being non-replicable).
atwrk 1y ago
Do you think Poppers demarcation criterion is falsifiable?
- AnimalMuppet 1y ago
  Nice use of recursion...
- belter 1y ago
  It is not falsifiable because it is not a hypothesis but a framework for recognizing science. Rejecting Popper for that is like rejecting reasoning itself because you can’t run a control experiment on it...but then one turns into an economist...
  - atwrk 1y ago
    > It is not falsifiable because it is not a hypothesis but a framework for recognizing science. And if the source of that framework is not (part of your definition of) science, how do we know whether that framework is true? That was sort of my point: Poppers criterion is nice, but only works for a small subset of (natural) science - and even there has to face criticism because it is simply too naive for many edge cases.
    - belter 1y ago
      Popper’s rule is not falsifiable for the same reason a definition can’t define itself. Falsifiability judges claims, not the rules for judging claims. Mixing those levels is like asking whether the rules of chess can checkmate themselves. In category theory terms, it’s a type error, applying a rule defined within a system to the meta-level that defines the system itself. https://www.lesswrong.com/w/object-level-and-meta-level https://www.lesswrong.com/w/object-level-and-meta-level https://en.wikipedia.org/wiki/Metatheory https://en.wikipedia.org/wiki/Metatheory
IAmBroom 1y ago
> When models fail, economists adjust assumptions ... When models fail, physicist adjust hypotheses ...
- belter 1y ago
  Physics can test those hypotheses under controlled conditions...
  - RandomLensman 1y ago
    Not always/limited in some areas (astrophysics come to mind where we, e.g., cannot (yet) create a star under controlled conditions). Testing hypotheses or predictions vs observations is also a valid method.
    - belter 1y ago
      Science is not about where you test hypotheses, but how. Astrophysics builds falsifiable models and checks them against reality like spectra, gravitational waves, etc... When predictions fail, theories change. If these were economists, they would check if their equations match the economic universe they live in. :-) Instead, they conclude the agents just “did not behave rationally enough”.
      - RandomLensman 1y ago
        Not about controlled conditions then necessarily. Not really my experience with economics - a lot of awareness of reality vs model (regardless of how "beautiful" they are).
      - genericresponse 1y ago
        That's just not true. The models have real predictive power, they just have limitations. Behavioral economics, which tackles this frontier is still a growing field. Thaler, Kahneman, and Taversky won the prize in 2017 for building the bridge between economic theory and individual decision-making. At the risk of being inflammatory-- These arguments are the equivalent of saying that Newton didn't really do physics because his models of mechanics break down at high enough speeds and small enough scales.
lanfeust6 1y ago
Argentina managed to save itself from hyperinflation by drastically cutting spending. By your logic, this could not be in the least bit predictable and inflation as a phenomenon doesn't even matter. Supposing it did was fairly predictable that not setting money on fire would help recovery, what does it matter that there is no controlled scientific experiment involved? Or to put it another way, are there no facts to be gleaned from data?
- thrance 1y ago
  Argentina didn't manage shit, else they wouldn't have to be bailed out by the Trump admin. This should come as a surprise to no one, "General AnCap" may have artifically and temporarily given the impression of good numbers by destroying his country's public infrastructure, but this is at the direct cost of the future. Now they're starting to pay the price.
  - lanfeust6 1y ago
    > Argentina didn't manage shit You're deflecting. Curtailing inflation was an outcome of policy. The U.S. had fuck-all to do with it. Unless you're willing to acknowledge something so basic there's nothing else to say to someone disinterested in good faith discussion
    - genericresponse 1y ago
      Ehh... you're both wrong. Argentina ended 2024 with an annual inflation rate close to 100%. That's significantly lower, but still hyperinflation. It should be lower this year, but how much is the question. Whether that's maintainable long term is probably the decisive judgement. Argentina has had many cycles of hyperinflation, reset, hyperinflation again. The current bailout is not exactly a positive indicator.
      - lanfeust6 1y ago
        It's all relative, but yes, they're still in trouble. Currency manipulation is also unsustainable. The markets are reacting poorly now in anticipation of the Peronists winning the next govt as well.
yard2010 1y ago
These are all people using made up words to throw shit on other people.
astrange 1y ago
> Economic theories can't be falsified because you can't run controlled experiments on economies, rewind history, or isolate variables. https://en.wikipedia.org/wiki/Econometrics https://en.wikipedia.org/wiki/Econometrics https://en.wikipedia.org/wiki/Credibility_revolution https://en.wikipedia.org/wiki/Credibility_revolution Economists are practically the only social scientists capable of doing this (even if it involves a lot of rainfall IVs). Everyone else p-hacks and calls it a day.

by belter 1y ago

rtsil 1y ago

Neither are litterature and peace.

belter 1y ago
They are not pretending to be sciences.
immibis 1y ago
Or astrophysics (not a Nobel prize but obviously a real science despite being non-replicable).

atwrk 1y ago

Do you think Poppers demarcation criterion is falsifiable?

AnimalMuppet 1y ago
Nice use of recursion...
belter 1y ago
It is not falsifiable because it is not a hypothesis but a framework for recognizing science. Rejecting Popper for that is like rejecting reasoning itself because you can’t run a control experiment on it...but then one turns into an economist...
- atwrk 1y ago
  > It is not falsifiable because it is not a hypothesis but a framework for recognizing science. And if the source of that framework is not (part of your definition of) science, how do we know whether that framework is true? That was sort of my point: Poppers criterion is nice, but only works for a small subset of (natural) science - and even there has to face criticism because it is simply too naive for many edge cases.
  - belter 1y ago
    Popper’s rule is not falsifiable for the same reason a definition can’t define itself. Falsifiability judges claims, not the rules for judging claims. Mixing those levels is like asking whether the rules of chess can checkmate themselves. In category theory terms, it’s a type error, applying a rule defined within a system to the meta-level that defines the system itself. https://www.lesswrong.com/w/object-level-and-meta-level https://www.lesswrong.com/w/object-level-and-meta-level https://en.wikipedia.org/wiki/Metatheory https://en.wikipedia.org/wiki/Metatheory

IAmBroom 1y ago

> When models fail, economists adjust assumptions ... When models fail, physicist adjust hypotheses ...

belter 1y ago
Physics can test those hypotheses under controlled conditions...
- RandomLensman 1y ago
  Not always/limited in some areas (astrophysics come to mind where we, e.g., cannot (yet) create a star under controlled conditions). Testing hypotheses or predictions vs observations is also a valid method.
  - belter 1y ago
    Science is not about where you test hypotheses, but how. Astrophysics builds falsifiable models and checks them against reality like spectra, gravitational waves, etc... When predictions fail, theories change. If these were economists, they would check if their equations match the economic universe they live in. :-) Instead, they conclude the agents just “did not behave rationally enough”.
    - RandomLensman 1y ago
      Not about controlled conditions then necessarily. Not really my experience with economics - a lot of awareness of reality vs model (regardless of how "beautiful" they are).
    - genericresponse 1y ago
      That's just not true. The models have real predictive power, they just have limitations. Behavioral economics, which tackles this frontier is still a growing field. Thaler, Kahneman, and Taversky won the prize in 2017 for building the bridge between economic theory and individual decision-making. At the risk of being inflammatory-- These arguments are the equivalent of saying that Newton didn't really do physics because his models of mechanics break down at high enough speeds and small enough scales.

lanfeust6 1y ago

Argentina managed to save itself from hyperinflation by drastically cutting spending. By your logic, this could not be in the least bit predictable and inflation as a phenomenon doesn't even matter. Supposing it did was fairly predictable that not setting money on fire would help recovery, what does it matter that there is no controlled scientific experiment involved? Or to put it another way, are there no facts to be gleaned from data?

thrance 1y ago
Argentina didn't manage shit, else they wouldn't have to be bailed out by the Trump admin. This should come as a surprise to no one, "General AnCap" may have artifically and temporarily given the impression of good numbers by destroying his country's public infrastructure, but this is at the direct cost of the future. Now they're starting to pay the price.
- lanfeust6 1y ago
  > Argentina didn't manage shit You're deflecting. Curtailing inflation was an outcome of policy. The U.S. had fuck-all to do with it. Unless you're willing to acknowledge something so basic there's nothing else to say to someone disinterested in good faith discussion
  - genericresponse 1y ago
    Ehh... you're both wrong. Argentina ended 2024 with an annual inflation rate close to 100%. That's significantly lower, but still hyperinflation. It should be lower this year, but how much is the question. Whether that's maintainable long term is probably the decisive judgement. Argentina has had many cycles of hyperinflation, reset, hyperinflation again. The current bailout is not exactly a positive indicator.
    - lanfeust6 1y ago
      It's all relative, but yes, they're still in trouble. Currency manipulation is also unsustainable. The markets are reacting poorly now in anticipation of the Peronists winning the next govt as well.

yard2010 1y ago

These are all people using made up words to throw shit on other people.

astrange 1y ago

> Economic theories can't be falsified because you can't run controlled experiments on economies, rewind history, or isolate variables. https://en.wikipedia.org/wiki/Econometrics https://en.wikipedia.org/wiki/Econometrics https://en.wikipedia.org/wiki/Credibility_revolution https://en.wikipedia.org/wiki/Credibility_revolution Economists are practically the only social scientists capable of doing this (even if it involves a lot of rainfall IVs). Everyone else p-hacks and calls it a day.