Second Great Depression?

[RebbePete]

Re: Second Great Depression?

It can be quasi-periodic as I believe Kondratiev waves are supposed to be. This does not mean there is no predictive value. I think Kondratiev waves explain the credit cycle quite well and I wouldn't call it magic. Although there are many magicians who incant it as there are many magician who incant science.

ASH, as you're an engineer, I won't hold back. If you do an FFT of log daily price series, and plot the power spectrum on log-log plot, a purely lognormal returns time series would have a slope of -2. I have done this. Large cap indices like the S&P or FTSE conform pretty well except for the very low-frequency components, corresponding to long economic cycles. Smaller cap co.s, emerging mkt indices and commodities all have different signatures indicating long-range order.

This is not particularly new - Mandelbrot (who as a sideshow invented fractals) found the same thing for cotton prices for his PhD thesis.

Steve Keen's models all have quasi-periodic behavioural regimes (which apparently we are leaving into a debt-deflationary crash absent money printing).

Turbulent fluid flow displays cyclic behaviour but is not harmonic.

So it doesn't have to be harmonic to be quasi-periodic or cyclic. I would propose that the economy comes under such a heading. Ie. history rhymes but does not repeat.

I have a Masters in EE (as well as being a part-time rabbi, long story). I only say that to indicate some expertise in the subject matter.

The problem with Fourier analysis is that it only looks at the behavior of the system under the current set of inputs. To really understand the system, you need to discover the transfer function. Because of the quasi-periodic effects some people have reported, there probably are some feedback paths with delays in the world economic system. Unfortunately, there is no way of knowing whether there are other feedback paths that become more dominant when the system is stimulated by an impulse or step function (like is happening now).

Here's an example for the non-techies. Take an empty plastic water bottle with the lid off. You can press gently in the side and release it, and the bottle returns to its shape. You can set up an experiment where the bottle oscillates - bounces back and forth by being periodically poked - based on the elastic characteristics of the bottle. You can even fill it with various substances and generate graphs of how they affect its resonance ("bouncing") characteristics.

Now push it further until the side of the water bottle deforms inward. That wasn't predicted by the model, and now, the characteristics of the bottle that you measured have been permanently altered. You took the system beyond the limits of the oscillating model you had developed.

Or, picture a thermostat in a house. If you measure the temperature in a house on a cold day, you will get an oscillation up and down as the thermostat reacts with lag to heat loss through the walls. You can measure it for a long time and find cyclic variations, short term and long term (e.g., day night as the sun illuminates different parts of the house). Now add a malfunction to the heating system so that, whenever it goes on, the basement catches fire. You've taken the system past the limits of your model.

Nowadays, physicists and engineers can model things like these - the crushing of the bottle or the fire in the basement - but the model was developed by a meta-understanding of the system and not by watching its behavior over time. That meta-understanding is akin to macroeconomics, although the world economic system is so horrendously complicated that we have far less understanding of its operation than we do of the physics of materials.

People tried an approach called "deterministic econometric models" that were all the rage back in the stone age when I was in college. They attempted to model the feedback cycles in an economy, but these had limited accuracy. They just weren't complicated enough to be of use outside of some very simple markets (I believe there was a famous model for hog futures that didn't do too badly on a small time interval).

[sunskyfan]

Re: Second Great Depression?

"I have a Masters in EE (as well as being a part-time rabbi, long story). I only say that to indicate some expertise in the subject matter."

This is why I love this site ... great points and cool people thinking!

[powersown]

Re: Second Great Depression?

Hi powersown. In my opinion, reading America's Bubble Economy will not be of much benefit to you. Most of it was not written by Erik Janszen, and it is written at a much lower level than what EJ publishes on iTulip. In other words, it seems "dumbed-down" to me, and very light on analysis -- not characterstic at all of EJ's usual work.

Hi Ash,
Thanks for the tip.
As soon as I get this months check I will be subscribing. Then I can get up to speed ;)

[ASH]

Re: Second Great Depression?

Hi Ash,
Thanks for the tip.
As soon as I get this months check I will be subscribing. Then I can get up to speed ;)

I think a lot of the useful reading is probably in the free sections of the site, although I wouldn't discourage you from subscribing. Try reading the articles which are linked in the lower-right corner of the main iTulip web page (in sections such as "Economics" and "Financial Markets", etc.). I think those are free, and they span a lot of the analysis behind EJ's interpretation of the present economic events.

[rj1]

Re: Second Great Depression?

Yeah, it's classified now because it "doesn't work" according to the army.

Funny how you classify something that "doesn't work" AND allow the DV's who visit in theater to wear this stuff for "testing purposes".

It's like 4x as expensive as our SAPI plate shit (the 1 hit wonder and then you're done under), and from everything I've seen, worth it's weight in gold if you are being shot at.

Read up on it. Looks good. Has any foreign government picked it up for their military?

[grizam303]

Re: Second Great Depression?

I linked my ex-infantry friend the wiki article on it, and he said "I don't even need to read the article, it's awesome." Any armor that can withstand multiple shots is amazing IMO, the fact that it stops steel-core is fantastic.

[olivegreen]

Re: Second Great Depression?

For example, we forecast 25% and higher unemployment among 18 to 25 year old men. The outcome in the US case is unpredictable, but worst case is bad. Just ask the Greeks where unemployment among the 18 - 25 age group has been above 25% for years.
[/CENTER]

This unfortunately is what the "elders" historically speaking have used as an underlying yet prime reason for war... no? Must put youthful rage somewhere productive when there are no jobs or futures...

Olive

[Master Shake]

Re: Second Great Depression?

See Are You a Doomer?

I tried to take the test, but the last two questions were missing.

[ricket]

Re: Second Great Depression?

This unfortunately is what the "elders" historically speaking have used as an underlying yet prime reason for war... no? Must put youthful rage somewhere productive when there are no jobs or futures...

Olive

The sad part is that people actually believe war is beneficial.

I say it only halts production and destroys capital and that can never be good for anyone.

[*T*]

Re: Second Great Depression?

*T*, pretty advanced stuff there. I wonder if all these models take into account all of the world CBs printing in unisson?

Keen's model had govt. spending I think, but did not include printing money. The other methods are just data analysis and not really models..

[*T*]

Re: Second Great Depression?

Thanks for the commentary, T. I agree with you. That's why I was pretty careful in my post to talk specifically about the problem of timing. All it takes to get cycles is some kind of "mass" and a "restoring force", and the last part of my post was a nod to there being economic cycles -- just not harmonic economic cycles. The whole point of my post was that although the idea of economic cycles can help with the analysis of the big picture, the cycles can't be used for timing because they aren't harmonic.

(& sunskyfan)

I think we all agree then -- there may be a point where the data says things have gone "too far" but you can't really pin it down until the saturation point occurs, at which point you know things have turned.
Which advocates a good risk management approach.
Makes me feel better about calling the top of the UK housing market every year for the last 10 years then.

[powersown]

Re: Second Great Depression?

"Both writers anticipated the onset of another big depression at that time. Here we are over 20 years later and it looks like it may have finally started. I am wondering what happened between then and now? Why the delay?"

Sorry, I haven't read the intervening thread. But I find it interesting how you're interrogating reality for its 'delay' in not conforming to a wave on a Powerpoint chart. This reminds me of the Kafka aphorism: "a cage went in search of a bird." This may be sacrilege but perhaps the arm on the printer was misaligned?

My intent was to inquire of anyone familiar with K-wave cycles as to any insight regarding what forces may have caused a deviation in a pattern that had been fairly consistent for the past 250 years.
I had been out of town for the past 15 years and thought maybe someone might have been paying better attention than me.
Does that make more sense?

[ASH]

Re: Second Great Depression?

My intent was to inquire of anyone familiar with K-wave cycles as to any insight regarding what forces may have caused a deviation in a pattern that had been fairly consistent for the past 250 years.
I had been out of town for the past 15 years and thought maybe someone might have been paying better attention than me.
Does that make more sense?

Perhaps it was the Fed, under Greenspan, blunting the economic cycle with interest rate policy. Perhaps it had something to do with the changes in reserve fraction requirements that happened in the early 90's.

Still, to return to an earlier point, the factors which affect the timing of the cycle change on a time scale much, much shorter than 250 years. It's frankly implausible that any economic pattern you might identify on that sort of time scale will be useful for timing (except the trivial case of things with a 1-year period).

[mcgurme]

Re: Second Great Depression?

At another level, which I find more interesting, the Fourier transform is the basis for how a lot of subtle stuff works. For instance, the Heisenburg uncertainty principle from quantum mechanics is really just a consequence of the Fourier transform (and its inverse). When I take an x-ray diffraction pattern in lab, I am really just taking the Fourier transform of a physical piece of material -- that's an experimental way of performing the mathematical task. The Fourier transform also explains how you can get "white" light out of a supposedly monochromatic laser, if you pulse it fast enough, and lots of other n.

Ash,
You did a great job explaining FFT, I'm sure readers gained a lot from that.

But I do have to disagree with your assessment of the Heisenberg uncertainty principle. A fundamental part of quantum mechanics is the of the collapse of the quantum coherent state (described by Schroedinger's eqn) into a classical one. This is not just a simple matter of a FFT of a piece of material. It is the more fundamental concept that, when you see diffraction of your X-ray, it is telling you that the photon's potential for interaction and coherence collapse is not actually localized . The FFT is a way to see this after the fact, but IMHO, it does not tell us anything fundamental about why the uncertainty exists in the first place or, why disparate parts of a system can have s p o o k y* action at a distance (as Einstein termed it), i.e., non-locality. The lower the mass of the particle, the more non-local it becomes - which is what Heisenberg's principle is about.
Morgan

* For some reason this board system is programmed to change the word s p o o k y into "friendly", so I cannot get it to reprint that word, without the spaces.

[ASH]

Re: Second Great Depression?

Ash,
You did a great job explaining FFT, I'm sure readers gained a lot from that.

But I do have to disagree with your assessment of the Heisenberg uncertainty principle. A fundamental part of quantum mechanics is the of the collapse of the quantum coherent state (described by Schroedinger's eqn) into a classical one. This is not just a simple matter of a FFT of a piece of material. It is the more fundamental concept that, when you see diffraction of your X-ray, it is telling you that the photon's potential for interaction and coherence collapse is not actually localized . The FFT is a way to see this after the fact, but IMHO, it does not tell us anything fundamental about why the uncertainty exists in the first place or, why disparate parts of a system can have s p o o k y* action at a distance (as Einstein termed it), i.e., non-locality. The lower the mass of the particle, the more non-local it becomes - which is what Heisenberg's principle is about.
Morgan

* For some reason this board system is programmed to change the word s p o o k y into "friendly", so I cannot get it to reprint that word, without the spaces.

I'm afraid I have to disagree with you, since this is my area of professional expertise.

My point regarding the Heisenburg uncertainty principle is unrelated to my point about x-ray diffraction. A particle with an exact location is in an eigenstate of position: a point in real space. A harmonic plane wave (in space) with an exact wavelength is in an eigenstate of momentum: a point in Fourier space. The incompatibility between position and momentum, and the Heisenburg uncertainty principle, simply fall out of Fourier analysis. If you apply Fourier analysis to find the spatial frequency components of a delta function in position (the point-like particle) you get a white spectrum (completey undefined momentum); conversely, the real-space image of a point in Fourier space is a wave that extends through all space (completely undefined position). The same relationship applies when the domain is time rather than space, leading to the energy-time Heisenberg principle. Fourier analysis can also be used to prove that a Gaussian wave packet is the minimum-uncertainty wave function, for which:

delta-x * delta-p >= h-bar/2

which is the quantitative form of the uncertainty principle (for space and momentum).

My comment about x-ray diffraction stands. X-rays diffract from a regular lattice of atoms if the incoming x-ray's wave vector matches the periodicity of the crystal. One is essentially testing for periodicity (taking a Fourier transform) by throwing waves at the crystal. This is why solid state physicists spend all their time working with the "reciprocal lattice", which is the Fourier transform of the crystal. If you want to compare wave phenomena to the spatial periodicity of the crystal, you move everything to Fourier space, where both the waves and the periodicity of the crystal are points.

What you wrote about quantum coherence, the Schroedinger equation, and "collapse into a classical state" isn't correct. In fact, the statement "the lower the mass of the particle, the more non-local it becomes" is not technically correct and misses the point. The deBroglie wavelength may be shorter for higher-mass particles, but with regard to nonlocality, it's really all about eigenstates versus superpositions, and measurement. If you have two conjugate variables which don't commute, you can't have simultaneous eigenstates of both. You can put the system into an eigenstate of either with the appropriate measurement, and you can have a superposition state in which both variables are indeterminate, but there's never anything "classical" about it. As far as delocalization goes, that has nothing to do with mass. That only has to do with whether you are forcing the system into an eigenstate of position or not by measurement. Photons, which have no rest mass, can be localized just as easily as electrons. All you have to do is attempt to measure their position -- for instance with a phosphor screen or a PMT. If you want to delocalize something, then measure its momentum.