Replaying 1929!

From January 8, 2001

Serious Stuff to Ponder: The Second Dynamic of the Long Wave

It Maybe Wasn't Nixon

As a long-time reader of this site, you are probably aware that for several years, I've held to the notion that while the long wave of economics has traditionally lasted between 48 and 62 years, the recent extension of the long wave to a period of about 71 years has probably be driven my the underlying change in demographics wrought by modern medicine.

Simply stated, without going into speculation about which cohort (collection of people of a certain age or demographic group) has been responsible for the extension, if people live longer, they will play the economic game longers and the cycle will extend.

This week, I'm forced to revist this because of an exchange of emails this week with reader Ehor Mazurok in Alberta (you know that's in Canada, right?). We pretty much agree that what we've discussed comes very close to being the "Unified Field Theory" of long wave economics.

Now that I have his permission for me to share part of our collection of emails on the subjectI'm going to give you the "high level" view and (after the charts) some of the bakcground. I think our exchange of emails are worthy of some kinds of prize in economics if just read in the order of exchange because I think we both got a serious "Ah Ha!" when we looked at the data. I'll try to give you that same sense in a simplified and abbreviated form.

The Highest Level View:

For reasons I will demonstrate in a moment, a strong case may be made that prevailing borrowing and payoff rates, not simple interest rates drive the length of the economic long wave. Specifically, relatively low (by historical standards) debt loads and fast payoff rates rates cause the long wave cycle to be shorter while high (by historical standards) payoff periods and debt loads, not just the interest rates cause the cycle to lengthen.

Until the market peak and subsequent Depression in the 1929-1937 era, debt load, payoff rates, and interest rates had been somewhat stable over time. One reason for this may have been the existence of some linkage between hard assets, such as gold and other monetary metals and perhaps real estate. This linkage began to deteriorate following World War Two, and was formalized when the link between the U.S. Dollar and Gold was broken by Nixon in 1972.

Ehor and I have not yet had time to explore whether this relationship between payoff rates and interest rates and cycle length may be extended to shorter economics cycles such as the Juglar or Kitchen cycle, nor have we had time to assess cycle length in countries where there have been much higher prevailing interest rates or lower prevailing debt loads with faster payoff rates.

However, this theory (which I'm planning to call the "Mazurok-Ure Cycle Correlation") suggests that we would see a longer set of business cycles in countries with higher prevailing interest rates (and long debt payoff periods) and a shorter business cycle in countries with low interest rates and short payoff periods.

Consider, for a moment, Brazil. Without having the time to look at business cycles in Brazil (hey, I've got a life, right?) I would expect the business cycles to have been long while Brazilian rates were high as in the 1980's in part because of the long payoff time associated with high rates, while today we should see their cycles stablizing or shortening as Brazilian rates have been coming down and with this, the pay off rate may have started to fall..

The Domestic U.S. Relationship: Empirical Data

Fortunately, the notion of a correlative relationship between economic cycle length, debt pay off rates, and interest rates is pretty obvious, once you know what you're looking for. Like Tobin's Q (the ratio of financial to non-financial assets in an economy) the correlation between lower rates and higher frequency (e.g. shorter periods peak to peak or trough to trough with lower rates) is easy to demonstrate.

Let's look at the trough to trough periodicity of several economic cycles in the past 100 years. You will see that cycles have extended when rates have been higher, but it's not a cause and effect. The source data I'm using comes from the National Bureau of Economic Research page that defines U.S. business cycle dates at URL http://www.nber.org/cycles.html .

The NBER data shows the trough to trough length of five cycles between March of 1879 and December of 1895 adds up to 282 months. During this period, there was a lot of borrowing going on, as the railroad barons pushed west. We would expect in a review of long term interest rates to see "moderate" rates in this period.

The next group of NBER data covering June 1897 to January of 1913 adds up to 211 months. This would lead us to expect rates to have been lower in this period.

December 1914 to August of 1929, a period of low rates, that adds up to 190 months.

March of 1933 to August of 1957, a period of quickly rising rates rates, adds up to 318 months. Got this? The cycles extended dramatically when the country began to deviate from historical norms of interest rates. Rates went up and the cycle length got longer. The 1957 prime rate was 4.2%.

From April 1958 to July of 1981, interest rates went through the roof. Remember the double digit prime rate? This would suggest a a cycle about that would be grossly extended. Sure enough, the NBER data adds up to 314. The prime rate in 1981 was 18.87% (Source Federal Reserve data) .

Hmmm...what gives?

It Maybe Wasn't Nixon

The data seem to say that it's not all strictly the interest rate itself that drives the cycle length, but something more subtle: Perhaps it's the coefficient of the interest rate change, or maybe a relationship to the Fed's discount rate?

Unfortunately, a simply look at interest rates doesn't seem to fit neatly, as the prime rate data above shows. So what about the discount rate? Let's construct a table and see where this gets us.

* Source: Federal Reserve FRED system Note: because of wide swings, the average rate used for 1929 was 5.75, 1930 4.75, and 2.75% for 1933. Although some error is likely (the impact of the rates was not symmetrical) the distortion doesn't skew the conclusion. Data for all of 1981 is used and no consideration is given to the 1973 dual prime rate event.

**Source: Federal Reserve FRED system. No consideration is given to the effect of discount rate surcharges in 1980-1981.

Clearly, the data seem to show that there's no "cause and effect" relationship between the length of the underlying economic cycles (NBER data) and the prime or discount rates (Fed Reserve data).

The fact is that on a lot of places around the web, you can see a lot of critism of President Nixon's decoupling of gold from the U.S. currency (making it a true fiat currency with no hard assets behind it). But, is this really fair? The data seem to suggest that interest and discount rates aren't the drivers of the phenomena of cycle length. But if they aren't, what is?

The Hidden Dynamic

Now we get down to the guts of the discussion between my reader in Canada and the issue at hand. What we came to in our exchange is that this probably has something to do with debt loads and how quickly people are accustomed to paying off their debts.

Interestingly, there is a key role played by interest rates, but the role is a tangent from the main cycle drivers: life span and payoff times.

Payoff times are an interesting notion.but very difficult to quantify because in the land of conventional economic thought, there is no single number for "pay off". Nor, have I been able to find a single number that encompasses the net effect on the general economy of the present ratio of consumer debt to consumer spending, although you can derive some interesting notions from looking at consumer spending, consumer debt and consumer confidence going forward, as people will structure their future spending in light of a set of expectations ("confidence").

What's the biggest change that has happened since 1930 that didn't exist previously? Simply this: People no longer plan to pay off their major or primary debt, their home. The "average length of time a person plans to take to pay off their home, and at what interest rate" is not a simple number to get at. We haven't yet succeeded on this front. But the dynamic is there and it goes something like this:

In 1900, when you bought a house or a horse, you planned to pay them off quickly, and often without a mortgage. Buying a house or horse was, in fact, the way business was done. Someone would work for five to ten years, save up a pile of money, and then present typically 50% down and carry a mortgage at relatively low rates. Most working folks held the notion that it was not a good idea to have debt, and so every effort was made to pay off the debt as quickly as possible.

In 1990, if you bought a house, and that horse replacement the car, you probably planned to pay off the car, but whether you paid off the house had become a "who cares?" proposition. People buying houses today are not generally worried about payoff: they expect house prices to behave like stock pricess. Over the long haul they will always go up in value. Of course, that isn't a "fact", but the perception is stso strong that people don't really care about paying off their houses.

I heard an interesting stastic this week that underscores this point. One homeowner out of 25 in the U.S. today is delinquent on house payments. Another anectodal bit came from a colleague who does taxes on the side. He reports someone with a fairly low net worth (under $2 million) asked him how he could get the use of a hot executive jet. He was wondering how to get the use of a $5 million dollar jet.

The point is that this fundamental change boils down to people becoming "renters" of everything instead of "owners" of things. This mindset clearly boiled over into the stock market. The price of stocks always goes up, over time, goes the mantra. The craziness of the markets today may be distilled down to this simple dynamic of a change from owning an income producing (dividend paying) stock to "making payments on a stock" and then "selling it for a gain". When you think about it, paying a dividend means nothing because people are banking on stocks to go up just like they have (recently) been able to do successfully with their homes.

Putting Numbers To It:

To see the dynamic, let's put together a "snapshot" of a comparison between 1900 and 1990.

In 1900, a person might have made $5,000 per year income and purchased a house with a $5,000 value. The ratio of annual income to home price was 1:1 but the typical mortgage rate was 50% of value ($2,500) so the annual income to mortgage ratio was 1:0.25. The upper bound for mortgages was around 15 years. Owners were hesitant to carry paper any longer than that and bankers were pretty rigorous about loan criteria. A seasoned mortgage had good value as default rates were low.

Faster forward to 1990, and you'll see a person might have made $50,000 per year income, and purchased a home with a higher valuation, say $100,000. This throws the ratio of annual income to house price at 1:2. Thanks to more lenient loan practices, 80% loan values were typical of the period, thus the income to mortgage debt ratio was 1: 1.6 (50k income/80 k debt). The duration of the typical mortgage was 30 years and there was talk of 35 years.

So what has happened? A debt which would have been paid off in 15 years in 1900 is now a debt being paid off in 30 years, or possibily never. The underlying dynamic that comes into play looks like this:

In 1900, a person with income (I) expected to pay off debt (D) in period (P) given prevailing rates that were low (R_L). The low rates made a payoff in 15 years almost certain.

In 1990, a person with income (I) expected to pay off debt (D) in period (P) given prevailing rates that were high (R_H). The low rates make a payoff in 30 years somewhat questionable. House debt keeps recirculating and being restructured. Although it's a false sense of wealth, loand-to-value (LTV) ratios have been climbing compared with historical norms.

Put another way, if you paid off a house in 1900 with 25% of your annual income, it was done in 15 years. But today, if you pay off a house with 25% of your annual income you might never pay off the house! Interest rates therefore include expectations, that in turn drive how people structure long term debt.

Impact on the Economic Long Wave

Let's suppose, for a moment, that there's a cause and effect relationship between how quickly people pay off their debts (and remember, there's no simple number for this yet) and how long an economic cycle lasts. If this is the case, and we apply the difference in mortgage payoff rates, we would expect the length of this cycle to extend to twice the previous cycle length because payoffs of debt are now twice as long (15 years versus 30 years).

Now, given that the long wave had been running around 56 years, doubling the length of the wave would cause us to anticipate a crash and Depression of the 1930's magnitude about 112 years lafter 1929. That would put us at 2041, and Harry Dent with his "Roaring 2000's" would look like a genius.

Unfortunately, I doubt things will work out that way, because it's clear to any thoughtful observer that the 1929 was not the "break date" where suddenly all long-term mortgages were rewritten to a 30 year basis. This is something that has happened over time, and it's like writing a formula to quantify the movement of a glacier. The lengthen began to exceed old rates sometime in the 1960's or maybe the late 1950's. Remember from my report of last week that this was about the period when the ability to live as a single-income family began to dissapear in the U.S.?

Let's suppose, for a moment, as my reader in Alberta suggested, that the change may have coincided with the closing of the gold window by Nixon. If this were the case, then we would expect to see part of the present cycle (the part from 1929 to 1972) operating on the 56 year "normal long wave" rate, and the part from 1972 to 2000 operating roughly the "double speed" rate implied by the extension of mortgage payoffs.

This means 43 year of the 56 year cycle has been completed by 1972. That's 76.8% of the long wave cycle we had expected. 100% minus 76.8%, or 23.2% of the cycle had yet to be run. This would have run at roughly the 112 year rate.

23.2% of 112 years is 25.98 years. That 26 years from 72 would have put us at 1998. But remember, the cycle length has continued to lengthen as the amount of long term debt.

A "side proof" comes from placing the market crack in 1920-21 next to the 1987 break. As I've stated perviously, there's a high correspondence. But now with this statement of dynamics as a model, let's plug in some numbers. The peak to peak was 56 years in the previous cycle from 1873 to 1929. The break happened in 1920. At that point, 84% of the cycle had been completed. Looking at 84%, we would expect that 1929+(84% of 56) + 1972+16% of 112) should give us a market break, and sure enough, it does: 1990-1991's break.

The theory forces me to look at the 1987 break as an endougenous event, perhaps forced by the Fed to try and avert something they saw coming. The fact that the 1990-91 break fits is very interesting, and when you line up 1920-29 as the final run-up against 1991-2000 as the final run-up this time, you see the precision of fit. 1987 was a blip on the way to 1990-91's pause.

There's a lot of other anecdotal stuff supporting the notion that personal long-term debt, especially housing debt, is what drives the long wave. The peak prime interest rates in 1929 were around 6%, and they got back into this area in 1966-67, and although they climbed to around 8.5% for a while, they declined again and didn't move "semi-permanently" above 6% until late 1972.

There's clearly a "tough to document" dynamic going on here. LTV ratios approach the dynamic, but don't answer everything. The dynamic involves rates, and how people react to them, how people structure long term debt in their own lives. LTV's cover housing, but what about credit cards, and how about all those car leases?

There's a lot of fertile ground to be tilled in this area of economics, but little incentive to do the research. The big lenders don't want to do anything other than make money and the average home buyer isn't likely to change borrowing habits until layoffs associated with this downturn, that I argue is the start of Depression Two, have taken hold. The Fed no doubt sees it, but to put the facts on the table would only cause panic and not be productive. The dynamic does explain the horrible box the Fed is in, however, because as they lower rates, they tend to curtail long term expansion of mortgage debt, and that shortens the long wave cycle. Save the economy today but speed up the severity of the pending Depression by working down debt; it's a hellavah choice.

In the meantime, I'll keep buying the odd gold coin even though as I said in the title: Maybe it wasn't Nixon (that caused the cycle to extend). But for now, stabilizing rates and continuing asset inflation game seems to be government's only way out of a very bad spot.

Theorectical Basis of the Mazurok-Ure Correlation:

This past week, Ehor Mazurok & I exchanged a series of emails that led to the view expressed in our "correlation". Here's some of the first material he sent me:

Thank you for your reply.  Was away all day today so didn't get a chance to
get back to you until now.  Regarding what you wrote;
<SNIP>
>we are really in a transition phase now between the shorter
>cycle (52?) and the longer cycle ( around 77-80?) and that perhaps in this
>"transition zone" we are getting the extension beyond what Kondratieff and
>others would expect.
>
>Does this make any sense to you?

I totally agree with you on this point, and this is what I believe is
happening right now.  It does make sense to me, but I've ran across a lot
of people who don't buy this. Perhaps some background as to how I arrived
at all of this.
When the first Apple II computer came on the market, I immediately saw its
potential for stock market analysis.  Stock market quote lines at the time
were available only to brokerage houses, so I wrote a computer program that
allowed me to input quotes printed in financial newspapers.  Quotes were
typed in to the computer in abbreviated notation, and the program generated
and stored on floppies the data required for analysis.  The stock exchange
I picked was the VSE (Vancouver) Curb which had about 300 entries.
Other programs were written to find high performing stocks and to construct
a data base to analyse those stocks.  On the basis of this analysis, I
developed some routines for identifying potential high flyers.  To get an
idea how good those routines were, I wrote another program called
"Dartboard!"  Basically, Dartboard was a software blind random pick, that
was used to check how well the picks by my routines performed against a
random selection.
In all of this two curious patterns emerged, which after some number
crunching, I developed a model that enabled me to generate those patterns
from a limited data set. The attached file "fig1.jpg" is a diagram of that
model.  I hope you are able to read those files, if not, let me know which
format is easiest for you to read.
The two patterns shown in file "fig1.jpg", are the basic model, and the
non-linear model. Both models start at the 'Source'.  Here, for some reason
a deviation occurs from a basic trend, called Trendline 0.  That deviation
is a rapid expansion that continues toward a crisis point.  At the Crisis
point, the decline can occur in two ways.  A dramatic bust, followed by an
Abort, or along a gradual FALL line.  This all ends at the End of Fall, or
EOF, where a new trend is re-established.
The non-linear model, is essentially a basic model with more than one
crisis point.  Each crisis point is progressively higher than the last one,
and at the final crisis point, the decline can occur in the same two ways
as in the basic model.
The straight lines constructing the models, represent the statistical
average of a trend underway.  If raw data is normally distributed, then
these lines represent the "mean" of this distribution.  I chose to
construct the models in this manner because straight lines can easily be
described by straight line equations in the form of Y=m*X+b  where X,Y are
the data coordinates, m the slope, and b the intercept.  I have found that
a relationship exists between those straight lines.  This is a handy thing
to have, because once Trendline 0 is established plus 1 data point on the
expansion line, the whole pattern can be built from this information.
It was well into the middle of 1980's when one evening I was browsing
through an economics data book that graphed 100 years of the PPI (It was
called the Wholesale Price Index then) that caught my attention. What I saw
was a pattern of the PPI leading up to the Great Depression that resembled
the basic model.  A quick check showed an excellent fit.  The attached file
'fig2.jpg' shows a reconstruction of this test.  In the file, Figure 2 s 
hows the minimum data set required to construct the whole pattern. Figure
2.1 shows the actual result.  In figure 2, the equations derived for the
straight lines and their intersections are shown.  Figure 2.1 compares the
Forecast Data with Actual Data.
Using PPI data after year 1935, data was analysed to the current year at
the time which I think was around 1988.  Results showed the PPI was
following the non-linear model, and we could expect multiple crisis points,
instead of one in year 1920.  I discussed this with a friend of mine who is
a professional economist, and he encouraged me to document this.  If it
wasn't for his reminders (badgering?) it would have never been done.  But
in June of 1992, I finally put it all together.
When this work was completed, the economy was in a recession.  From the
non-linear model constructed, it was evident the economy was already
starting to recover, although none of the economic indicators have shown
this.  It was interesting to observe that as the economy started to
recover, governments began to advertise they were putting programs into
place to get the economy moving again.  What claptrap!!!  If gov't really
knew how to get the economy going, then they should know how to prevent a
recession.  Since a recession did arrive, it can only be concluded, they
don't know what to do, but were astute enough to get some mileage out of
the whole thing.
I felt some comfort in the fact that I was able to predict a turning point
for the 1992 recession, but lets face it, it was too close to the time my
work was completed, and besides it could be a lucky guess too.  In my view
the true test of the model would occur in 1997, the timing of the next and
last crisis point some five years away.  File "fig3.jpg" shows the
non-linear model constructed from data available at the time.  All of the
straight line equations and correspoinding turning points are listed on
that diagram.
As you can see from this file, the last crisis point identified as C3 was
predicted to occur in 1996.9 at PPI 1116.3, base year 19140.  The actual
turning point of the PPI was projected to occur at the same time, but at a
lower value of 1097. To convert this PPI figure from base year 19140 to
19900,  multiply PPI(1914) X .10126.  This calculation produces the
following figure 1097 X .10126 = 111.08.  Carrying on further, the work
showed that in .9 year, or 10.8 months into 1996 (near end of October), the
PPI will start to drop from a high of 111.08, base year 1990.
The forecast was in error.  The decline actually started in January of 1997
from a PPI of 111.6, base year 1990.  It came about 2.5 months later and
the turning point was actually +.52 higher than predicted.  Actually not
bad for something that was put together in 1992.
There was, however, a miscalculation on my part in the 1992 work.  I
assumed the PPI decline would proceed along bust line CR3 and abort on line
A3 (file fig3.jpg). It didn't occur to me that PPI would decline along F3.
 In retrospect, it has dawned on me that F3 is a managed decline, where a
bust along CR3-A3, is a natural hands off response!
In 1997, the PPI index fell steadily for 6 straight months in a row, where
it plateaued and the decline stalled.  When that happened, I began to
suspect, the decline was being managed, and the stock market bubble may be
the result of FED intervention.  From todays point of view, it appears Mr.
 Greenspan is succeeding, at least in the short run.  Changing seats on the
Titanic may have looked like a good idea at the time, but the end result
was just the same. It sank!
Referring back to my previous message, the Source of the basic model in
1930 occurred in year 1915.  The Source of todays non-linear pattern was
around 1972.  The interval of these sources some 57 years, is my belief,
represents the Kondratieff cycle.  To check this out, I found WPI data back
to the year 1700, and all the other sources were well within the Kondrat 
ieff interval.  However, when looking at the peaks, the last peak occurred
in year 1920, but this one occurred in 1997.  A significant difference from
the past.  Looking at End of Fall at 1935 and the projected one at 2065,
the difference is even greater, 130 years.  Conclusion being, as you
indicated, we are in a transition period where the Kondratieff is being
extended to a duration that very few people anticipate.
So where is all of this going?  I am afraid, all is not well.  Looking back
on my stock market experience, any stock that went through the non-linear
expansion was quickly delisted.  Bre-X is one such example.  In short,
there is no recovery in sight for the economy that I can see without going
into a very dangerous depression.  I can summarise my guess as to what is
going to happen in file "fig3-a.jpg".  This is a "cut" from original Figure
3, with a few probable descent lines.  Since we are in a managed decline, I
visualise an erratic decline along line F3.  A slight recession occuring
where PPI falls well below F3, then a minor rebound heralded as the end of
the recession, and another recovery in place, probably timed for the next
Presidential election.  After that, another decline or recession, followed
by another recovery.  It may look normal, but the difference being, is a
continuous overall decline in the standard of living.  There is always the
chance of an exogenous shock that could cause a precipitous drop to A3,
however, that is unlikely as long as the USA still remains as the dominant
world power.  The table to the right of this diagram is a cut from
International Financial Statistics showing when the turning point occurred.
 This is the data I used for my work after the 1930 depression.  Further
work has shown that an early date for the depression would be 2014, and a
late date of 2020.  I don't believe the End of Fall will ever be reached
this go around.  The economic dislocation and social pressures will be so
great after 2020, that our existing social systems will cease to exist and
some new economic order will be re-established.

Ehor's chart from 1992 is shown below:

You must admit, not a pretty picture, and all because of something that
happened before 1972.  My guess is, the trigger for this thing was Nixon
closing the gold window.  It all makes sense!
What do you think?

I then sent him a note suggesting that if underlying mortgage structuring (as a major component of aggregate long-term debt was indeed the driver, the interest rates should be an indicator, but not predictor of where things are heading. he responded with some excellent comments (and a bit of math):

I like the unique way you calculated the
length of the new cycle using percent of ideal cycle length.  Agree with
you the increased cycle length did not occur overnight, but probably did
come about in a continuum of events.
In your message, you mentioned the Kondratieff long wave historically was
48-56 years. Although there are many figures quoted for the Kondratieff
span, the one that seems to be repetively used is 54 to 60 years.  I have
seen figures from 48 years, to as long as 65 years.  For my work, I just
averaged the extremes (48+65)/2 = 56.5, and used this as the nominal length
for the Kondratieff, which agrees with your figure of 56 years.
Regarding your comment on higher interest rates, no disagreement there.  In
the past I have done a lot of work analysing mortgages and how interest
rates effect them.  From the interest rate formula, increased rates
definitely prolong the repayment of debt.  Since I believe the Kondratieff
is a debt cycle, then prolonging the repayment of debt, essentially
prolongs the cycle.  A lot of people still don't buy the concept, that the
Kondratieff is essentially a debt cycle.  This cycle is remarkably
consistent as far back as year 1600, and maybe even earlier.  Think of the
sociological and technological changes that have occured since then, and
yet this cycle remains relatively unaffected.
However, if one looks at debt over that period, its management has really
not changed that much.  So the two have to be interconnected!
In my view, as long as debt exists, we will have depressions.  Its
unavoidable.  If one looks at production and debt, the mathematical
concepts
governing those factors are significantly different and work against each
other.  For example, production is linear, while debt is exponential.  If
production of food is to be doubled, then inputs have to be doubled.  A
failure of input causes a production output failure.  With debt, an input
failure of cash, does not produce an output failure, on the contrary, it
produces a greater output.  The unfortunate part in all of this, is debt
consumes productive output, and being exponential in nature, it can easily
outpace production.  So if we assume a linear production function such as,
                        P = a*t + b   a linear function based on time "t"

and debt                D = (1 + i)^t  likewise a function of "t"

as "t" increases, it doesn't take long before D > P , and there is not
enough production to settle debt, so a depression results (effectively an
economic bankruptcy) that reconciles debt with production.
I've done a lot of historical research on this subject, and it was very
interesting to find that hundreds of years ago, there were economic periods
where charging interest on debt brought about severe punishment.  Could
that have been some Regents efforts to control the exponential nature of
debt?  A tempting conclusion.
Regarding my comments on the pending decline, believe me, I really
struggled with that one!!!  In my work, the 2014 to 2020 time frame
translates to an expansion of the 1929 to 1930 time frame.  After 2020, we
are looking towards a magnified 1935 to 1937 time frame.
When I was completing my work in 1992, I wondered if there was a way to
compare the intensity of the last depression with the next one.  Decided to
use Fourier Analysis to break down the PPI into cycle components from Years
1896 to the projected year 2000.  The Fourier terms obtained were Sine and
Cosine terms in the form of,
A*Cos(x)   and  B*Sin(x)
where "x" was the frequency or length of cycle, with A and B being the
cycle
amplitudes.  The resultant amplitude was caculated using this formula,
K = SQRT(A*A + B*B)
Results were;
    Cycle in Years          A             B        K(Resultant)
         3                24.59        -83.92          87.45
        19                40.23         13.72          42.51
        38                35.02        109.59         115.05
        57               449.02     -2,242.43       2,286.95
        76            16,764.27     -8,732.14      18,902.14
        95           -67,811.92    -52,989.64      86,060.20
       114           -88,955.16     11,381.49      89,680.31
       133           198,337.34     66,558.64     209,207.44
       152           -99,623.86    141,364.13     172,941.41
       171           -55,459.89   -229,387.18     235,996.35
       190            55,834.63    166,185.44     175,314.31
       209          -135,087.48     82,976.88     158,536.40


A word of caution here when it comes to Fourier Analysis.  From my
experience, Fourier analysis has no predictive power at all!!!  It can only
reproduce a given or known pattern by an infinite series of sines and
cosines.   Fourier, however, can reproduce the cycles that make up a known
pattern with a high degree of accuracy.
One slight problem with this analysis, is that it is based on the
assumption the PPI pattern from 1896 to 2000 repeats itself exactly, which
it doesn't.  So this assumption in itself has to produce some errors in the
final result.  Nevertheless, even with this error, I believe good
information can be extracted from these results.
Looking at the resultant amplitudes, the most prominent is for 171 years,
(235,996.355) and after that for 133 years, (209,207.44).  I really can't
attach any meaning to those amplitudes, except they are just "order of
magnitude" figures.
What was interesting, was the appearance of the 57 year cycle, which is the
average length of the Kondratieff.  Its amplitude is, 2,286.95.  Assuming
this is the magnitude of the Kondratieff, and we know what devastating
effect the last depression had based on this cycle, so through this
analysis, it should be possible to estimate the order of magnitude of the
next depression.
So on what cycle will the next depression be based on, 77 or 112 years?
Assume it will be based on 77 years.  From the table, the closest cycle is
76 years with an amplitude of 18,902.14.  Comparing this with the 57 year
amplitude,
K76/K57 = 18,092.14 / 2,286.95 = 7.911
which implies that if the next depression is based on a 76 year cycle, it
will be about 7.9, or approximately 8 times worse than the last one!
If the next depression is based on a 112+ cycle, the next nearest cycle in
the table is 114 years, which by similar comparison suggest the next
depression will be 8 times greater than the last one.
Heaven forbid if we are on a 171 year cycle.  That amplitude is 235,996.35,
and when it is compared to the 57 year cycle amplitude, the next depression
would be
235,996.35 / 2,286.95 = 103.19
103 times greater than the last one?
What this really shows, is the dangers associated with an increased cycle.
>From what I have read, the last depression was bad enough without doubling
its intensity.  But, making it 8 times worse, I doubt if existing social
systems at the time will able to withstand this. This is the reason for my
pessimisim.
I have given some thought to the specific debt load that can be sustained
by an economy, and its a question without solution so far.  If financial
transactions were constrained by unyielding laws, then figuring out a
specific debt load that would exceed limits of an economy could be easily
calculated.  However, laws regarding financial transactions change, or
yield to accommodate devastating financial transgressions thus prolonging
the inevitable conclusion.
For example, the IRA's and Keogh retirement plans.  Why are they so popular
now, and not 30 years ago?  Is it governments valid concern that citizens
should have enough for their retirement, or is this a voluntary form of
deferred consumption?  Lets face it, if you have an economy awash with an
expanded money supply, you have two ways to solve the problem, rein in the
money supply, or prevent the populace from spending it.  Reining in the
money supply results in recessions, on the other hand, if the populace can
be made to defer consumption voluntarily, that is the best of both worlds.
 If the populace doesn't volunteer to defer consumption, it can be forced
by laws to do so such as currency restrictions, changing debt margin
requirements, and so on.  Under these circumstances trying to determine a
specific debt load becomes like trying to squeeze a balloon into a smaller
space.  You squeeze the top, and it pops out the sides.  Get someone else
to squeeze the sides, it just finds another place to pop out.  When you
finally think you have it, it bursts!  :-)
No problem in batting this idea around with you.  I believe today
economically we are in uncharted territory, the likes not seen in history.
 Regardless what economic path is taken, the bottom line in the end will
always be, Gold will Rule the DAY!

As of this writing, we are still working on the math, but a couple of major Policy points have emerged. The most significant of these is that regardless of the depth of rate decrease undertaken by the Federal Reserve, there will be no appreciable impact on the depth of Depression Two (that we entered into March of 2000 with the peak of the markets measured by the Aggregate Index that corresponds to the DJIA in 1930).

I also received a fine note from a reader who said basically, "how can things be 8 times worse than the Depression?" and asking me to clarify some detail. Here is what I told him and you may find useful:

William,

Thanks for your note.  The 89% decline, or possibly more, would come in a
variety of indices.  As a result of the "complexification" of the world
since 1929's watershed event, there's no longer a single proxy for the
economy in markets.  This is why I developed the Aggregate Index.

It says that yes, the Dow was pretty much indicative of the economy then,
but today you need to look at broader measures and other sectors.  The
Aggregate recognizes the S&P and NDX as broader measures in these regards.

On your second question, the one that asks "eight times worse - or even
twice as bad" related to what?" - is a good one!

I tend to believe that in terms of the Dow, or broader markets, the work
out will be an exponential market in reverse.  In other words, what was an
89% decline in the past may be a 92% decline, but that part of the curve
could be powerfully exponential. It could be that 1000 times worse might
only cause the Dow to experience a 93% decline.  The value of the Dow
should theoretically never go to zero.

On the other hand, "twice as bad" might mean a 50% unemployment rate
instead of a 25% rate.  But again, we haven't gotten to the math on that
part.  What we need to see are what drives the dynamic (the underlying
dynamic would be an annual calculation of debt, rate, and time to pay off,
so that the time to pay off becomes clear.  Once you have that, you will
some logical bounds below which rates can't fall at one end, and above
which they won't rise (for long) at the other.

Back to the point, however, remember the "eight times as bad" could simply
be a reversion to the historical low, but start much higher up the
exponential debt build up, that Alan Newman has so eloquently captured in
"Pictures of a Mania" collection.

Clearly, once things get anywhere near equal to 1933 or 1937's lows, there
is so much social pressure that government will be forced to create a war
in order to keep order.  It can be fairly argued that the end of the
previous Depression didn't happen until we started rearming in preparation
for WWII - and there's a fair case to be made that Roosevelt forced the
Japanese into WWII by interdicting their trade routes.  Such would likely
happen again, but the consequences today are much shorter, characterized by
several brilliant flashes of white light that vaporize everything inside
the >5 PSI overpressure blast zone, if you follow my drift.

This may be why the Mayan calendar stops in 2012. It's also why I tell all
my friends who own boats that this is not the time to be selling them...

More to come next week as we develop this....the markets today didn't seem like they were listening too close to my call for one last big rally, but I still have faith and plenty of MSFT calls for January! -GAU

[January 22 Note: The calls worked out fine. The return was >75% for a 4-week trade]

Write when you get rich.

George Ure

Back to main page

All contents (c) 1998-2001 by George A. Ure, MBA, except authors as linked or noted