ABSTRACT
Over the three-year period from 1987 to 1989 several hundred stocks were converted from manual limit order books to the new ‘electronic book’. The implementation schedule lists the conversion date, post number, and specialist for each stock that was converted in this period. Daily trade data are available in the CRSP database and include closing prices, diurnal range, and trading volume. The mean absolute diurnal price range is used as a measure of intraday volatility and the net mean absolute daily change is used as an indicator of interday volatility. Intraday volatility is partitioned into a component that varies with interday changes and a component in excess of that which cannot be explained in terms of interday price fluctuations. The excess intraday volatility is the response variable under investigation. We use a longitudinal quasi-experimental design to perform a comparison in ‘event time’ of the excess volatility before and after the treatment which is expected to cause a step change in the response. The treatment is the implementation of the Electronic Display Book on the floor of the New York Stock Exchange. A linear regression model is used to control for other variables that may also affect excess volatility.
