[cotsuka]

I'm saying calling something Austrian or Keynesian is inherently political. We should come up with economic models, whether they're Austrian, Keynesian, alien, or dumb, test them against data to see if they hold up, and refine until we get close. Aka, what scientists do.

Great article the other day from Krugman about this very topic. Economics might be a science, but economists are not being scientists about it. http://krugman.blogs.nytimes.com/2013/10/21/maybe-economics-...

[mistermann]

You can make models all day, and I think it should be done in the pursuit of knowledge, but to imply these models can accurately predict behavior is a bit naive IMHO.

Economics in a globally integrated economy are of similar complexity as weather, if not worse.

Keynesians have been driving the bus for quite some time now, if they have it all figured out as many seem to imply, why do so many things (deficits and debt plus extreme government interference in the bond market, as just a few examples) seem to continue to deteriorate.

I don't know if Austrians would do a better job, but to look down one's nose at them seems a bit delusional to me.

[bermanoid]

Austrians, for the most part, make no testable predictions. They essentially reject math as a tool, and insist that unless you implement their preferred systems 100%, then they won't work at all, and then explain all observed failures by claiming that it would have worked better if things had been pushed further.

Even if they're on to something, they don't give the unconvinced very much to work with. Most of their arguments push every economic dial in a single direction no matter the current state of the world, which says to me that there's no way that approach could lead to a reasonably balanced economy.

I think econ is messed up in all sorts of ways, but I don't think the Austrian school has the answers, they strike me as far too religious and unbalanced to be believable, and quite frankly I see way too many words and far too little math coming out of their camp to buy any of it. Every other science leans on math to great benefit, I refuse to believe that econ is somehow the one exception where it's not useful, especially since at its core it's about optimizing certain numbers.

[mistermann]

I'd suggest any unwillingness to make specific mathematically testable predictions is correct acknowledgement that the field is too complicated to do so. As I've said, you can make a model and tweak it by picking and choosing inputs until you can make it successfully back test with historic data, but any comprehensive correctness is likely tenuous if not specious.

Having a complex model that appears correct but is actually wrong is quite dangerous. Observe long term charts (debt, debt to gdp, etc) in various countries throughout the world. We're told by economists and the federal reserve in very serious tones using large words and complex models that everything is all figured out, but the charts continue to get worse. Even during the glorious Clinton years of supposed robust surpluses, debt increased every year. Hindsight being 20/20, we now know we were just in a bubble.

Austrians speak in broad principles rather than specific mathematical formulas, and in my opinion, their broad (vague) predictions of what will happen if you do <x> are far closer to honest reality than any other camp.

[shard972]

> They essentially reject math as a tool, and insist that unless you implement their preferred systems 100%, then they won't work at all

What? What Austrian has ever said that?

Austrian's very rarely deal with absolutes in the real world and mainly deal with relative analysis (If you go with X bad policy, these negative things can occur).

> then they won't work at all, and then explain all observed failures by claiming that it would have worked better if things had been pushed further.

What occurrences are you referencing when an Austrian got policy control?

> and quite frankly I see way too many words and far too little math coming out of their camp to buy any of it.

They reject the idea that maths is a useful tool in economics due to it being a social science.

[sien]

Economics forecasting is harder than the weather.

How do you build an economic model that can deal with the introduction of mobile phones? What about some new energy source that cut the cost of energy dramatically? What about the fall of Communism?

Also the weather doesn't have expectations that react to the weather forecast. Economic models get incorporated into people's expectations and this feedback can actually alter the economy. If you read up on The Phillip's Curve and stagflation you can see an example of this interaction.

Weather is actually decent for forecasting. You can constantly get new data to verify your model against. If you are forecasting for a week you have hundreds of tests to do per week.

Any forecasting where the predictions matter over the scale of decades is harder because the time span means you will have less data available for validation.

[jswinghammer]

The basic problem is that the models are not testable in the way you think they might be. Economics is inherently a deductive discipline and anything you can think of as coming from economics that is valuable in any way came from someone's head and not from testing models. Economists get into trouble by trying to pretend they are physicists.

[dwaltrip]

Economics is not a science. You can't test hypotheses rigorously. Which makes these debates 100x harder.

[nl]

This is true, but it is also true for other sciences (eg meteorology, mentioned elsewhere).

I'm not sure what the solution to the untestable hypothesis problem is.

[phyalow]

Krugman<-Spew