The basic methodology to estimate fractions of strategies used by market participants is pointed out by the figure below. The data to estimate is on the micro level of real financial markets (see right part of the figure) and thus cannot be observed (at least not at high frequency). The collective behavior of the market participants (i.e., the demand and supply) are aggregated by the market maker. By clearing the market, she adjusts prices accordingly. This results in the observable macro level market behavior. The price time series of the real market is to be explained by an agent-based financial market model (see left side of the figure). The structure of this model on the micro level is created by researchers in finance in way, such that the resulting macro level behavior replicates the stylized facts of real financial markets (e.g. excess volatility, long tails in the price return distribution, excess kurtosis…) on a daily basis. This involves calibrating fixed model parameters accordingly. In contrast to the real financial market, the fractions of the market strategies used are known in the model by observing the modeled economic agents’ behavior. This behavior could be pursuing either a chartist or a fundamentalist strategy. These strategies are modeled in the way described on this site. Estimating the time-varying model parameters of the chartists’ and fundamentalists’ fractions to the real financial market involves choosing them, such that the macro level price time series matches the one of the real market. If this is the case, we assume, that the chartists’ and the fundamentalists’ fractions in the model replicate the real market participants’ behavior. The process of estimation has to solve the optimization problem of choosing time-varying model parameters, such that an error metric is minimized. This metric measures the difference between the market model’s and the real market’s observable macro level behavior. The problem with this basic methodology is that the agent-based financial market models have a complex mapping from the micro to the macro level behavior. Thus, the estimation is computationally very intensive and only rather simple models at low frequency (e.g. yearly strategy fractions) can be estimated. One recent example is described in the paper of [Boswijk et al. 2007].

In contrast to the above basic approach, the estimations presented on this website are based on a methodology, which allows estimating highly complex real multi-agent-based financial market models at high frequency. The methodology is the result of several years of intense research conducted at Humboldt University, Berlin, Germany and Hohenheim University, Stuttgart, Germany. The approach has been presented by the main author at the 14th International Conference on Computing in Economics and Finance 2008 at the Sorbonne University in Paris, France.
The following figure compares estimation results of our approach (red line) vs. a very recent basic approach of [Boswijk et al. 2007]. We use a more complex model and yield more detailed, daily estimation results, while Boswijk et al. only yield yearly results.

References:

The whodrivesthemarket.com approach:

• Klein, A., Urbig, D.: “Who drives the Market? Estimating a heterogeneous Agent-based Financial Market Model using a Neural Network Approach”, presented at Conference on Computing in Economics and Finance (CEF), Sorbonne University, Paris, France, 2008.

Estimations following the basic methodology lined out:

• Alfarano, S., F. Wagner and T. Lux (2005): “Estimation of Agent-Based Models: The Case of an Asymmetric Herding Model,” Computational Economics, 26, 19-49.
• Boswijk, H. P., C. H. Hommes and S. Manzan (2007): “Behavioral Heterogeneity in Stock Prices,” Journal of Economic Dynamics and Control, 31(6), June, 1938-1970.
• Vigfusson, R. (1997): “Switching Between Chartists and Fundamentalists: A Markov Regime-Switching Approach,” International Journal of Finance & Economics, 2(4), October, 291-305.
• Westerhoff, F. and S. Reitz (2003): “Nonlinearities and Cyclical Behavior: The Role of Chartists and Fundamentalists,” Studies in Nonlinear Dynamics & Econometrics, 7(4).

Agent-based financial market models, some rather simple examples:

• Brock, W. A. and C. H. Hommes (1998): “Heterogeneous Beliefs and Routes to Chaos in a Simple Asset Pricing Model,” Journal of Economic Dynamics and Control, 22(8), 1235-1274.
• Kirman, A. (1993): “Ants, Rationality, and Recruitment,” Quarterly Journal of Economics, 108, 137-156