The idea is that easy money leads to overinvestment in certain sectors, which end up consuming more resources than their stable long-term share of the economy. Housing being the key example in the 2001-07 boom, or Internet technology in the previous one.

This is a reasonable description of the Austrian story.  So, no problems here.

 

 

Here, the first problem.  It's not just "something going wrong".  Rather, it's market participants recognizing that some sectors are "consuming more resources than their stable long-term share of the economy".  How does this happen?  Usually, it's because easy money policies eventually breed price inflation which, in turn, breeds tight money policies.  Since interest rates have gone up, the value of capital assets will fall - showing that many investments are no longer worth it.

In this story, government stimulus just delays the inevitable transition of people and other resources from sectors which no longer add much value to those where they can be more productive. Therefore, not only is stimulus expensive but it acts as a brake on recovery and future economic growth. So do labour market regulations such as the minimum wage.

 

This is an accurate portrayal of the story.  The problem is that the economy has an optimal use for resources based on individual preferences.  Government stimulus (generally) diverts resources away from such uses.

 

 

Right.  Similarly, I think few people would say that all the resources going into Internet companies were sensible, either.

But it does leave some key gaps.

 

Actually, we don't have to appeal to behavioral tendencies.  It's actually a direct result of very standard capital asset pricing models.  When money is cheap, interest rates drop.  When interest rates drop, assets that provide payoffs further in the future gain value relative to assets that provide payoffs closer to the present.  Therefore, sectors that are expected to offer high returns in the far future will see resources diverted toward them.  Some sectors that fit:  housing and other construction (because houses provide a consumption stream for decades), internet startups (because profits were expected to occur in the far future).  Cars are a reasonable candidate, but tend to be shorter lived than houses or successful internet startups.  Holidays are less likely, as the consumption stream from a holiday is short lived.  Gold is also less likely, because it provides no real "consumption stream".  Its value is primarily as a hedge against inflation.

Second and more importantly, this reallocation process is going on all the time, even when there is a boom. During the last 15-20 years the number of Internet software developers has consistently increased, even with a small slowdown in 2001-2. Fewer people work as secretaries and more work as reiki trainers (many of them are even the same people). So what is the difference now, and why will a stimulus package stop it from happening?

 

Here's the problem.  One fundamental misunderstanding of the Austrian theory is that it's just a "sectoral reallocation" story.  That's not quite true.  Fundamentally, the Austrian theory is really a story about error.  Easy money led to artificially low interest rates that led to erroneous calculations about the value of particular investments.  People arranged their lives based on these calculations.  Then, tighter money leads to higher interest rates, and people's hopes are dashed, and readjustment is necessary.  The "difference" is that much of the time resource reallocation is not the result of error, but during a recession it is the result of error - or more properly, the reallocation is to correct the errors that occurred during the boom.
 

Another way of thinking of the issue:  the story isn't so much a "reallocation" story as a "recalculation" story.  During the boom, there are errors in the calculations for optimal resource allocation.  In the crisis, these errors are revealed.  But, it takes recalculation to figure out where resources should be allocated.

Thinking of it this way, two things should become clear:

(1)  A Krugman-style argument of "why no mass unemployment during a boom?" doesn't work.  Starting from a stable economy, a "boom" involves the addition of new uses for a given set of resources.  Since the new uses are considered more valuable than some of the current uses, resources get reallocated without any need for them to sit idle.  But, in a bust, we're eliminating desired uses of resources.  So, resources go idle until we figure out where we want them to go.

One common assumption that economists often make is that there's an infinitely long "wish list" for society, and that society knows its wish list very well.  So, when one item is discovered to be a "bad deal", we simple reallocate resources toward the next best idea.  However, casual observation proves that this is false.  It's not unusual for us to - quite literally - not know what to do with new resources.  So, once it's discovered that we need fewer resources in housing, it's not necesssarily immediately clear where these resources should go.  Calculation time is required.

(2)  Stimulus doesn't help because it interferes with the calculations by sending bad signals.  It's as if the economy is trying to think about how to use all these resources that are suddenly idle, but the government is yelling in its ear.  This doesn't aid the "thinking" process, and, in fact, hinders it.

In short: what are the numbers?

Quibble:  I'm not sure that "what are the numbers?" is the "short" version of anything that came before.  In fact, none of what came before actually suggested that a lack of data was the problem with the Austrian story.  The argument was that the Austrian "story" applies all the time because reallocation happens all the time.  I agree with the premise (reallocation happens all the time), but that misses a significant part of the Austrian "story" - the recalculation process.  So, the conclusion does not follow.

What is the natural, sustainable rate of reallocation of people between sectors; how fast was it happening over the last six years, how fast is it now, and how fast should it be for optimal recovery and growth?

This question is seriously flawed.  First:  it's not clear that there is a single "natural, sustainable rate of reallocation".  Such a rate (if one can even calculate such a number!) would naturally depend on the structure of the economy and how one defined the sectors between which reallocation had to happen.  Second:  why should we only be concerned with the reallocation of "people" between sectors?  There are other resources - and Austrians often argue that the problem is far more about capital than labor.  Third, the rate of reallocation isn't even the real problem.  The problem is the structure of reallocation.  It's not enough that resources move from one sector to another.  They have to move out of the "wrong" sectors, and into the "right" ones.  A higher speed of reallocation only helps if we're guaranteed that reallocation is moving in the right direction.

Part of the reason that Austrians generally don't provide "the numbers" is because they are keenly aware:  (1)  that relevant economic data is highly dispersed, (2) the economy is a complex structure and (3) relevant economic laws are all counterfactual.  (1) and (2) make it impossible for an Austrian to calculate something like an "optimal rate of reallocation", because the optimal rate depends, for example, on the present and future individual preferences of the 6(ish) billion people on the planet - and such preferences are not observable to us as economists.  Now, we can make some assumptions about preferences and give you an "optimal rate", but such a rate is only actually optimal if our assumptions are true - and they're obviously not going to be, as we are making assumptions precisely because the real world is too complex.

 

Keynesian and monetary macroeconomics, whatever their weaknesses, at least provide empirical data to support their hypotheses and calibrate the theories.

I find this an interesting statement.  Keynes's business cycle theory was founded on the notion of "animal spirits" driving investors into frenzies.  I'm curious how one tests for these animal spirits without inferring them from business cycle movements.  The most updated monetarist view that I'm aware of is Friedman's "plucking" model - which basically claims that sometimes "something" goes wrong, and that creates a recession as the economy is "plucked" below trend..  Then - somehow - the economy recovers.  Observationally, this seems true - but it also seem like a nonexplanation, and it certainly doesn't seem "testable", as its only substantive prediction (eventual recovery to trend) is virtually tautological.

Austrian economics, despite its useful insights, doesn't seem to bring any testability with it.

Austrain economics is founded on a deductive epistemology.  If one has true premises and reasons deductively, then the conclusion absolutely must be true.  "Tests" are superfluous.  (Similarly, the Pythagorean theorem is true because of the logic underlying it, not because we've failed to find a right triangle for which it wasn't true.)

One only needs "tests" if one buys Friedman's positivist program for economics- and rushes into unreasonable assumptions on the basis of "parsimony".  Then, tests are necessary to show that the false assumptions don't actually hurt the model's predictions.  But, if you start with true assumptions in the first place, it's not clear what gain is to be had by testing the conclusion - especially given the "dirtiness" of economic data.  (For example, microlevel data tends to have high measurement error.  Macrolevel data relies on aggregation through questionable weighting methods.  One can also easily claim that omitted variable bias is pervasive in statistical work.)

Austrians - please give me some data and prove me wrong.