Measuring Trading Costs and Noise on Bucharest Stock Exchange. A Bayesian Approach

Abstract

Abstract In order to understand the market behavior and the market microstructure, after trade-by trade data became available, were developed and proposed many micro movement models. These models described the transactional price behavior in terms of market quality and market regulation. It is important that the academic community and practitioners disposes a good set of instruments that measures trading cost and noise because they reveal the market efficiency and the degree of competition. Since the models commonly used are time series models, our model is a computationally intensive measure for trading costs and trading noise including clustering noise and non-clustering noise, but is a reasonable instrument. We propose a movement model of intraday prices from Bucharest Stock Exchange based on Thomson Reuters Quotes and Volumes and apply it to analyze trade-by-trade stock price data. To model asset prices movements on a micro scale (high frequency trading) a stochastic counting process will be used. In order to use this model, we impose a difference between the price and intrinsic value of the asset. To find the value of the parameters we apply Bayesian estimation via filtering, which has the advantage that can incorporate all complete information about prices and trading time. Our study comprises stocks from Bucharest Stock Exchange over the last two years (2012-2013), and our question that we want to give an answer is if the trading costs are much bigger for equities that are less liquid comparing with the liquid one.