Current position in the K-wave (corrected)

I found some errors that might confuse readers,  Here is a revision:

 

A common view of the K-wave is that a downwave (or what some call the "B phase" of the wave) began in the early 1970's. The average length of the B phase is typically 25-30 years, suggesting that we should be nearing the end of the B phase and the beginning of the next A phase. Empirical work using some new tools suggests otherwise. It appears that we are about halfway through the B phase, roughly equivalent with 1929-1930 in the last K-wave as described below:

In the 1920's Kondratiev showed empirical support for his wave most dramatically by the wave-like structures shown in 19th and early 20th century prices. Here is a plot of the U.S. producer price index (commodity-based):

http://csf.colorado.edu/authors/Alexander.Mike/K-PPI.gif

K-peaks can be seen in 1814, 1864 and 1920 and troughs in 1843, 1896, and 1932. After 1932 what looks like runaway price inflation began and one can't find the wave any more. Interest rates also show peaks and troughs of Kondratiev spacing. There were "K-peaks" in interest rates in 1814, 1861 and 1920 that were close to the Kondratiev price peaks shown in the figure. There was also a major peak in interest rates in 1981.

If we use interest rates by themselves we would call 1981 the most recent K-peak and the beginning of the B phase. But other turning points, such as the shift to a lower productivity regime and the stagnation of real wages after 1973, can be used to argue for a much earlier K-peak. An early 1970's K-peak has the advantage of being approximately 54 years after the previous K-peak in 1920, that is "right on schedule".

One can deal with tendency of the post-1932 inflation to obscure price cycles using a concept I call reduced price. Reduced price (rP) is simply the actual price index (P) divided by the value (Pm) expected from a simple monetary-based model, that is:

rP = P / Pm, where Pm = aS + b, were a and b are constants and S is a variable I call monetary stimulation, defined as:

S = (D + M)/GDP where D is cumulative government deficits and M is money supply

For money supply I use M3 after 1959 and M2 before. (the two were nearly equal in 1959). I describe the approach in more detail here:

http://www.gold-eagle.com/editorials_01/alexander051401.html

A plot of rP is shown in the following figure in black:

http://csf.colorado.edu/authors/Alexander.Mike/RP-EAR.gif

K-peaks are identified in 1864, 1918 and 1981. K-troughs are identified in 1897 and 1946. These points are very close to Kondratiev peaks (1861, 1920, 1981) and troughs (1901, 1946) in interest rates.  Additional structural features are identified in the reduced price plot. After the K-peak, reduced prices fall to a region called the plateau where they are roughly constant for a while. Prices then "fall off of the plateau" and bottom after a few years at what is called the vortex bottom by Brian Berry. After the vortex they rise to a secondary peak that I call the DG peak (also after Berry). After the DG peak, prices fall to the K-trough, which is lower than the vortex bottom. We can see that 1932 is not a K-trough, but rather, a vortex bottom. 

 

Prior to 1932 these structures can be seen in the raw price plot as shown by the dates on my first figure above. We can even see the vestiges of a DG-peak in 1937 before secular inflation obliterates all traces of the K-wave.  Thus, we see that reduced prices shows the same structure as prices before the onset of secular inflation. 

That is, the transformation of raw prices into reduced prices does not introduce distortion into the pre-1932 record. After 1932, when the raw price data no longer shows any structure, we see a 1937 DG-peak (the 1942 peak is purely a war-time effect) and a 1946 K-trough that is lower than the 1932 vortex bottom (this is what makes it a trough). After 1981 we see a rapid drop to a "plateau" structure. What this tells us is that reduced prices captures the ongoing K-wave that lies "underneath" the raw price data. Reduced prices clearly show that we have been on the plateau since the mid-1980's (that is, the period from 1986 to 2000 has been "Kondratiev-analogous" to 1923-1929).

The recent stock market peak in 2000 suggests that we have arrived at the end of the plateau period and are in the early stages of descent to the vortex. My secular market trend model indicated that this peak was the end of the 1982-2000 secular bull market trend (fully analogous to the 1921-1929 secular bull market). In real time, I had identified the end in 1999 (prematurely) see:

http://csf.colorado.edu/longwave/oct99/msg00820.html

The relevant webpage is no longer at cybercities, it's at CSF:

http://csf.colorado.edu/authors/Alexander.Mike/Stanpor3a.html

The stock cycle analogy suggests the recent stock peak in 2000 is analogous to 1929. Confirmation of this will be if reduced price falls down from the plateau, which it might have started to do (I will be monitoring events in real time).

Some of you probably have noticed the plot of "ex-ante real rates" on my reduced price graph in red. These come from a working paper by James W. Kolari and Ariel M. Viale of Texas A&M. Ex-ante real interest rates can be thought of as investor beliefs about future real interest rates, that is, the future monetary environment. They can only be obtained using sophisticated econometric models. The rates I plot were obtained as an output from a three-state Markov switching model (and I have no more idea what that is than most of you). Dr. Viale kindly sent me a file of the ex-ante rates so that I could look for Kondratiev-like structures. I smoothed them with a nine-quarter centered moving average and plotted them along with my reduced prices.

The plateau period corresponds to a period of high ex-ante rates, implying "fear of inflation" amongst investors. This period then collapses to a low-rate regime right around the vortex. At present were are in a high-rate regime that began after 1980. The fact that ex-ante rates have not fallen yet suggests that the vortex is still in the future. This observation is consistent with the end of the plateau location suggested by reduced price and the stock cycle.  This is important since Berry (the originator of the vortex concept) places it in 1987.  He appears to have been premature. 

Having provided an empirically well-supported argument for placing us at a "1930-like" position within the B-phase of the K-wave, the question becomes what does it mean? This I will address in a future post.

For now, I was wondering if this view of cycle position is what most readers here expect or is their view of our position different? I would like to discuss this.

 

Thanks

Mike Alexander, author of
Stock Cycles: Why stocks won't beat money markets over the next 20 years.
http://www.net-link.net/~malexan/STOCK_CYCLES.htm