Observations and Updates – USA (1790 – Present)

lpplmarketwatch Leave a comment

Looking at the 10080 day window of The Bubble Index: DJIA, there is a low in 1948 and a peak in 1968, which is followed by another low in 1988; the final peak displayed is reached in November 2007. It is clear that there is a 20 year upswing followed by a 20 year down-swing. Extrapolating this trend, the end of the current down-swing will be in 2028, give or take a year or two. This will be followed by a peak around 2048. What this means, I am not sure; however, in the time period between 1968 and 1988, the DJIA was relatively flat. I could speculate and say that the same will be the case for the DJIA for the period from 2008 to 2028. Then, in the period 2028 – 2048, a new large scale bull market will emerge. (UPDATE: the data from 1790 – Present show that there is not a predictable pattern)

Currently, I am working on extending the DJIA Index from 1896-Present to 1790-Present using economic indicators and stochastic simulation. This will allow The Bubble Index to estimate daily levels in the time range from 1790-1896. Although the data will be simulated and artificial at the daily level, it will be made in such a way to fit the industrial production data going back to 1790. This will allow the creation of a reliable extension of the longer day window indices (1260, 1764, 2520, 5040, 10080 days) back to the 19th century.

UPDATE:

Download PDF

The Bubble Index: DJIA (10080 days)

lpplmarketwatch Leave a comment

I think this is a very interesting graph.

Here is the output of The Bubble Index of the Dow Jones Industrial Average with a 10,080 day window. Note the peak in late 1960’s and the peak in late 2007. NOTE: Value of 100 corresponds to the maximum level of the index, reached on November 16, 2007. Also, I have included a similar graph of the S&P 500 located below.

Download PDF

Hierarchical Edge Bundles: Visualization of Adjacency Relations in Hierarchical Data by Danny Holten

lpplmarketwatch Leave a comment

A thought inspiring article, entitled Hierarchical Edge Bundles: Visualization of Adjacency Relations in Hierarchical Data, by Danny Holten, would provide a possible way to visualize the flow of capital via hierarchical financial networks. The hierarchy would consist of governments, banks, exchanges, and traders. With such a visual tool, one may be able to identify connections in the network which are vital to the entire system. The more densely connected regions of the network would represent areas for officials to focus their efforts in either preventing a crash or ensuring a “safe landing” for an inflated economy.

For some reason, I found that the following picture would be perfect for visualizing the hierarchy of big institutions. The bigger bubbles would represent the larger institutions.

Potential New Site

lpplmarketwatch Leave a comment

In previous posts I have mentioned, with few details, a potential way to display the state of the trading network. Over the past few weeks I have been playing with the idea and progress is being made. The site, Particle Markets, will be the center of these new ideas to publicly and freely display the state of trading networks.

Currently the site displays a simple, 2D Ising type model, progressing through time. There are 176,400 total traders, and they all start with Long positions. As time progresses many of the traders sell and become either neutral or short. All traders are long/short 1 share. The summation of their numbers after each time step determines the change in price of the traded security.

Short positions: RED
Long positions: GREEN
Neutral: BLUE

This is a simple model. Future models will incorporate more details. I hope to be able to use actual data one day and display daily changes in the actual human network; in the same way a meteorologist shows viewers his Doppler radar to warn of storms, Particle Markets will be able to show investors the “market weather.”

The concept applies mostly to the stock market, since a trader holds a contract which has no “delivery date.” I hope to apply the same ideas to commodity markets.

Future Update Ideas:

The Bubble Index with Fake Data

lpplmarketwatch Leave a comment
Here are two examples of random data ran with the algorithm of The Bubble Index. The first is a Geometric Brownian Motion and the second is a Weiner Process. It’s interesting to note the the shorter time frames (52, 104, and 153 days) never reach more than 20 and most of the time lie below 10. The longer time frames (256, 512, 1260, and 1764 days) can sometimes spike to large values. Comparing the output of the Geometric vs. non-Geometric Brownian motion, one can note the effect of exponential growth in the corresponding larger values of The Bubble Index. (Values are standardized)