Eastern Europe has been undergoing rapid structural change over the last 10 years and one of the most interesting and dramatic of these changes has involved the foundation of a number of new financial markets and the creation of a system of private ownership for industry almost from scratch. Analysing this data poses a number of new and interesting problems; There is often thin trading so that a share price may be unchanged for weeks as no trades occur and then it may rapidly change to a quite different value as a trade occurs. The movement in share prices is often also truncated artificially by the market regulators so that movements greater than 10-15% may be stopped on any one day. Finally in the early periods of trading market participants may be very ill informed as to both the particular shares being traded and the general process of operating a market. So conventional finance models may prove a very poor approximation to the behaviour of these markets as those models are typically founded on the twin assumptions of rational expectations and market efficiency. This paper will address these problems, a Kalman filter process will be proposed for filtering a continuos series of trade prices from the observed thin data. This will then be extended to allow for the truncation problem. The filtered series will then be used to build time series models of market behaviour which mix GARCH-M behaviour with time varying parameters to mimic the process of learning which takes place as the infant market becomes increasingly efficient and approaches the behaviour of standard western markets.
More papers in Computing in Economics and Finance 1996 from Society for Computational Economics Address: Department of Econometrics, University of Geneva, 102 Bd Carl-Vogt, 1211 Geneva 4, Switzerland Contact information at EDIRC. Series data maintained by Christopher F. Baum ().