Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: framsey 04 June 2015 01:53:51PM 0 points [-]

I just wanted to mention that your last few posts have been very nice. You have clearly worked on exposition, and it has paid off.

I became interested in this sequence after your first post on mathematical ability, and I'm glad to see you factoring your ideas into digestible pieces and writing them up.

Comment author: framsey 27 February 2014 05:11:24PM 1 point [-]

What do you think of the health effects of too much sitting? That seems to be a hot topic recently. http://www.mayoclinic.org/sitting/expert-answers/faq-20058005

Comment author: gwern 21 July 2013 01:42:57AM *  4 points [-]

Are you really claiming that a theory that restricts the possible causes of an event to three has not explained the event "at all"?

This isn't a case of 1 theory which predict the cause as being 3 different possible causes. We're talking 3 entire competing paradigms, and then yeah, pretty much. "We think you're sick either because people get cancer, or they have 4 bodily humors which get imbalanced, or it's all due to malignant airs. We doctors haven't figured out which is the right paradigm, but rest assured as you die: probably one of the 3 paradigms is right!" I sure as heck wouldn't go around saying "medicine has it all figured out".

So. As I said originally: macro cannot agree on an explanation for any of the events I listed, and my examples fit your demand for examples.

Comment author: framsey 21 July 2013 02:14:58AM -2 points [-]

We think you're sick either because people get cancer, or they have 4 bodily humors which get imbalanced, or it's all due to malignant airs. We doctors haven't figured out which is the right paradigm, but rest assured as you die: probably one of the 3 paradigms is right!

There isn't this much disagreement over macro. Especially undergrad macro.

As I said originally: macro cannot agree on an explanation for any of the events I listed, and my examples fit your demand for examples.

And I explained how one could use the models taught in macro principles to think about each example.

Comment author: gwern 21 July 2013 12:45:22AM *  3 points [-]

This does not mean the theory has "huge problems" dealing with real world events! It just lacks the power to distinguish between causes from within the set Y.

Then it's not a theory capable of explaining the event at all. If all you can do is throw together a variety of heterogenous contradictory theories, you haven't explained an event.

I'm reminded of Wittgenstein, now, on family resemblances - of course one could define "a game" as an indefinitely long disjunction of possibilities "a game is anything which is either Backgammon OR chess OR go OR Halo OR tossing a ball OR soccer..." but having giving this disjunction, in what sense has one actually conveyed what a game is? Same thing here.

In the same way, economists argue about the causes of things like the Great Depression, Lost Decade, and Great Recession, but they mostly agree about the sort of causes they need to consider, and they have a common framework to think about them in (or at worst a few competing frameworks).

I believe that however we explain the Lost Decades, it will involve materialism and not souls. Do I have a perfectly satisfactory theory capable of explaining the Lost Decades? No, of course not.

Comment author: framsey 21 July 2013 01:38:14AM 0 points [-]

Then it's not a theory capable of explaining the event at all.

Are you really claiming that a theory that restricts the possible causes of an event to three has not explained the event "at all"?

Under this definition, I agree that undergraduate macroeconomics cannot explain the real world. But surely this is a rather restrictive standard for "explanation"?

I would rather say that a theory that allows us to concentrate a lot of probability mass on Y upon observing an element in X, or to concentrate a lot of probability mass on X after observing an element in Y, is doing quite a bit of explanatory work!

Comment author: gwern 20 July 2013 11:43:28PM *  5 points [-]

You've given several examples where our knowledge is incomplete. I agree! And I hope that any economist would explicitly say that there's no settled explanation for the Great Depression, or the Great Recession, or Japan's Lost Decade.

If the state of the art cutting edge economic research is that "there's no settled explanation" for Japan's Lost Decades (plural, since we're well into multiple decades, I think), then how on earth could the simple models taught to undergraduates not have "huge problems" dealing with these "real world" events?

Comment author: framsey 21 July 2013 12:13:19AM -1 points [-]

If the state of the art...is that "there's no settled explanation" for Japan's Lost Decades...then how on earth could the simple models taught to undergraduates not have "huge problems" dealing with these "real world" events?

I don't see how this follows.

Suppose we teach a theory that says things in set X can be caused by things in set Y={A,B,C}. Then I say "there's substantial argument about whether major event x in X was caused by A or by B."

This does not mean the theory has "huge problems" dealing with real world events! It just lacks the power to distinguish between causes from within the set Y.

In the same way, economists argue about the causes of things like the Great Depression, Lost Decade, and Great Recession, but they mostly agree about the sort of causes they need to consider, and they have a common framework to think about them in (or at worst a few competing frameworks).

Comment author: gwern 20 July 2013 08:50:47PM 4 points [-]

These are extreme claims, and I think I was justified in calling him on them.

You demanded examples where macro does not work well in the real world. I named 4: Japan remains a mystery to macro, and 'Abenomics' hasn't helped eluicidate it; the US crash was unexpected and falsified the 'Great Moderation', casting doubt on the entire structure of macro that didn't predict it; the Great Depression is only partially understood even if a lot of people think Friedman cracked it; and business cycles remain mysterious with panics and bubbles being not explanations but excuses. The growing influence of the macro views of Scott Sumner, who was a marginal figure at best a few years ago, offers another practical bit of evidence that macro is far from a settled science when it comes to contemporary economies.

A good principles class will spend quite a bit of time talking about each of your examples, and show how to think about them using the standard tools.

I'm sure it would. Doesn't mean it is right or that macro has a consensus view on each such as how to fix and prevent and predict them.

Comment author: framsey 20 July 2013 11:28:18PM *  0 points [-]

You demanded examples where macro does not work well in the real world. I named 4

I demanded examples of models taught to undergrads that have "huge problems dealing with the real world." The same poster went on to say that these models are so dangerously wrong one must be intellectually inoculated against them!

You've given several examples where our knowledge is incomplete. I agree! And I hope that any economist would explicitly say that there's no settled explanation for the Great Depression, or the Great Recession, or Japan's Lost Decade. But that is quite different from saying that the models we DO teach have "huge problems dealing with the real world" and "are NOT how the world works". I think the basic models ARE effective tools for understanding how the world works, and a good teacher will explain both their uses and their limitations.

In brief, the fact that there are things we don't know does not mean that what we do know is wrong, and a good principles class should teach both what we know, and what we don't.

Comment author: benkuhn 18 July 2013 07:51:27PM 4 points [-]

Are complex macro models how the world actually works? What are our most successful macro models, and how successful have they been? My impression (based off not that much data, admittedly) was that this applied to all macro, not just undergrad.

Comment author: framsey 20 July 2013 11:26:27PM *  0 points [-]

Are complex macro models how the world actually works?

Nope! All models are huge simplifications.

What are our most successful macro models, and how successful have they been?

A controversial question!

The conventional approach in (academic) macro is to build (relatively) simple models that can match particular stylized facts. Thus we have lots of models that can predict certain patterns in the data, but don't pretend to explain everything. Some people think we shouldn't do anything beyond this! (See Caballero, Pretense of Knowledge Syndrome)

Other people do try to build models that can match all the data. A standard cite is Christiano, Eichbenbaum and Evans (2005), and for another approach see Smets and Wouters (2007) (they even call themselves Bayesians!).

Then there are macro forecasters who try to accurately predict the future using non-theoretical statistical models. An example of this would be the work of Frank Diebold at Penn. These models can do a lot better than the above at predicting future data absent changes in the policy regime, but are presumably less effective at predicting the effects of novel policies (see the Lucas Critique).

My own opinion is that if you want to predict the next data point, use a forecasting model, but if you want to know the effects of a new policy, your best bet is to rely on simple models plus judgement. Good economists know more than any model!

Comment author: Matt_Simpson 20 July 2013 08:29:12PM 1 point [-]

What do you mean by 'content' here? The basic narrative each model tells about the economy?

I think I agree with you. The big difference between the models I learned in undergrad and the models I learned in grad school was that in undergrad, everything was static. In grad school, the models were dynamic - i.e. a sequence of equilibria over time instead of just one.

Comment author: framsey 20 July 2013 09:08:31PM 0 points [-]

What do you mean by 'content' here? The basic narrative each model tells about the economy?

Right. Plus most undergrad models have an analog in grad macro, i.e. the AD-AS model and the New Keynesian model, or Quantity theory of money and a basic cash in advance model.

The big difference between the models I learned in undergrad and the models I learned in grad school was that in undergrad, everything was static. In grad school, the models were dynamic

True in general. Some intermediate macro courses use a two-period framework to explore basic dynamics. Williamson's textbook does this.

Comment author: gwern 20 July 2013 07:45:20PM 4 points [-]

Contemporary Japan, and America; the Great Depression; business cycles in general.

(I threw in a fourth one for free because I love the 'name three' heuristic even if you chose a really bad topic to use it in.)

Comment author: framsey 20 July 2013 08:27:12PM 1 point [-]

Lumifer said (1) the state macroeconomics as a whole is bad, (2) what you learn in a principles course is not how the world actually works, and (3) macro models have huge problems dealing with the real world. These are extreme claims, and I think I was justified in calling him on them.

To your examples -- the AD-AS model (or IS-LM in older textbooks) can be used to think about business cycles in general, and the liquidity trap in particular, which covers most of your examples. The Great Depression needs discussion of monetary policy (gold standard, Friedman-Schwartz, etc), and all of your examples need some discussion of financial crises, banking panics, and asset bubbles. Japan is not that mysterious once you consider demographics and per-capita growth rates.

A good principles class will spend quite a bit of time talking about each of your examples, and show how to think about them using the standard tools.

Comment author: Lumifer 18 July 2013 03:22:10PM 15 points [-]

I would caution to be sceptical of undergrad-level economics, in particular macro. The usual models taught to students have huge problems dealing with the real world.

This is not to say that one should not study economics. However the state of this discipline is pretty bad right now (again, in particular concerning macro) and it's generally acknowledged that simple macro models are NOT how the world actually works.

Comment author: framsey 20 July 2013 06:15:16PM 2 points [-]

I would caution to be sceptical of undergrad-level economics, in particular macro. The usual models taught to students have huge problems dealing with the real world.

Name three.

View more: Next