Seeking Alpha

Todd Feldman's  Instablog

Todd Feldman
Send Message
Dr. Feldman is the founder of Behavioral Finance Investment Advisors. Dr. Feldman is also an assistant professor of finance at San Francisco State University. Dr. Feldman has published scholarly journals in the area of behavioral finance and agent-based modeling in various leading finance and... More
My company:
  • Evaluating Technical Analysis Again
    Practitioners and Academics have long debated whether technical analysis works. Many traders, investors, hedge funds etc. use technical analysis to some degree.

    Currently, there is a new field being pursued in the halls of academia called agent-based modeling. Academics are not only applying these techniques to finance but to a variety of different fields. The basis of the theory is quite simple, but the implementation is difficult. The first idea is that asset prices are not random. Individuals, called agents, make up the markets and as a group affect asset prices. When the group is large and heterogeneous the interaction of these individuals will create macro phenomena that appears random. Agent-based modeling breaks down the market into individuals first to better understand these seemingly random dynamics.

    For example, suppose ten people make up the market for a stock. In order to understand the stock price dynamics better agent-based modeling assumes that it is necessary to model the behavior of each trader first in addition to the interaction of these traders before aggregating the data. Agent-based modeling also assumes that if one were to looking at the stock price alone, such as in technical analysis, it will on average be of no help in determining accurate entry and exit points. 

    Lets us take the S&P 500 index. Agent-based modeling would suggest that examining the S&P 500 versus its moving average or tops and bottoms would not provide much predictive power if any at all because the data is already aggregated. If predictive power is extracted from this process it is because some of the individual behavior is being picked up. To have a clear understanding it is vital to manipulate the individual data as well as the interaction between the individuals before aggregating. Below in Figure 1 is one of our agent-based indicators for the United States stock market. We believe this chart has more predictive power than analyzing a chart of the S&P 500. If you take a look at the magnitude of the market declines, they appear to be of similar magnitude which signals accurate entry and exit point. In addition, the market corrections are all of a similar magnitude.

    Figure 1: US Agent-Based Indicator

    Figure 2: US Agent-Based Indicator Overlay

    The overlay shows three periods, 1973-1980, 1998-2004, and 2005-2011, periods with large recessions. It can been that the 2005-2011 period resembles 1973-1980 period the most. If you look at the graph we are currently at a point where the stock market will continue to decline before moving up. This means it may be wise to take some cash off the table before investing again.

    Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
    Mar 04 3:59 PM | Link | Comment!
Full index of posts »
Latest Followers


More »
Posts by Themes
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.