Roll’s Model of Bid Ask Spread: Implicit Evidence from BSE

Abstract

This paper empirically studies the bid ask spread model as proposed by Richard Roll using the data from Bombay Stock Exchange (henceforth, BSE). The objective is to understand and determine whether the impact of various events like the abolition of Badla, introduction of Electronic Trading and Futures Trading in BSE, are sufficiently captured by Roll’s model or not. The order processing cost model was taken into consideration for the three period event study, namely: i. Pre and post impact when the Badla was banned, ii. Introduction of Electronic trading in BSE, and iii. Commencement of futures trading in BSE Although the empirical findings presented here are those of the stocks which comprises the Sensex, but it gives a good picture of the market response as a whole as the volume of trading in the BSE is highly skewed in favor of the Sensex scripts.The results indicate that Roll’s model does not fit the data very well. There is a large number of covariance of price change which is positive, which is contrary to the proposed model. The variation of the calculated spread is also significant on a month to month basis. Further research is needed to propose a theoretical framework for Indian stock market.