Economical Cycles: overview

� Timing Solution   http://www.timingsolution.com     December 1, 2008 

 

Introduction

 

 

Though the title of this article is "Economical Cycles", you will not find here some theory or thesis. My purpose is simple: just to show you that cycles in economic activity do exist. I will do the cyclical analysis for different financial data. I will do it for several selected data sets. The reason why I do the selection of financial data is very simple. For some financial data it is very difficult to perform the detrending procedure (normalization); therefore it is impossible to perform the spectral analysis of those data. 

For example analyzing GDP we can only select the recession/slowdown periods. Here they are (and it looks like the year 2008 shares the company of these pink zones): 

  

 

This is obviously related with 8-12 year Juglar cycle.  And this is all what we can tell about these data as they are. The further analysis requires the detrending of these data, and this is practically impossible. I tried to use different modifications (like GDP per capita), and still they require the detrending; the progress of the modern technologies is too fast... Let's work with what we can.

Basic Cycles: Dow Jones Industrial Index from 1789 year

 

 

I use these data to identify the most important business cycles. 

Look at the spectrogram calculated for this financial instrument (to remove the trend, the relative oscillator with the period of 70 months is used):

 

 

 

You can see that the most energy is concentrated in four cycles. Here they are in order of their importance:

 

9 year Juglar  cycle  - I observe this cycle very often with different economical data. It is interesting that the period of this cycle varies in a very wide diapason (8-12 years). Possibly it interferes with 11 years sunspot activity cycle.  They call this cycle "fixed investment cycle". The phases of this cycle are:  growth - speculative investment - bubble - burst .... 

I believe this cycle is formed by human factors, here they are:

40 months inventory Kitchin cycle 

 

There are also two minor cycles:

 

Non name #1:  57 months cycle - I observe this cycle very often for different financial data.

 

Non name #2  5.9 year cycle

 

This is is how two major cycles, Juglar and Kitchin, work:

 

 

 

     

Here you can see how another modification of Juglar cycles works,  we calculate  Juglar cycle with 6 overtones to see inter cycle details:

 

 

     

 

Producer Price Index (PPI) from 1800 year (before 1978, Whole Price Index).

 

The main problem in analysis of this index is difficulties to normalize data - eliminate trend (especially for data beyond 1970s). 

 

The most powerful cycle here is 28 years cycle:

  

 

 

 

Possibly 5 and 15 years are important for PPI as well.  I have no idea why this cycle is so important, maybe it is somehow tied with commodity cycle.

 

This is how this 28 years cycle works in time (4 overtones used):

 

 

 

Here is the superposition of these three the most powerful cycles:

 

 

Total Layoffs and Job Losers from 1967 year.

 

The most influent cycles here are 5 and 11-year cycles:

 

 

 

 

 

 

I don't known either 11.1 is current period of Juglar investment cycle or the sunspot activity cycle works here; or they may be working together. 

 

This is the forecast based on these cycles:

 

 

Actually, this last diagram is confusing for me. It contradicts to everything above. If stock market (DJI) goes down, PPI goes down, - why does unemployment go down as well. Logically, it should go up.

Though it might go down as the result of some special government projects that are discussed so much nowadays...