Return to search

On the trading volume and time effects on the bid-ask spread components of NYSE and NASDAQ common stocks

The study of the bid-ask spreads of stocks is important since they constitute the mechanism through which trading costs are incorporated into prices and recovered by market-makers. Realized spreads are a measure of the trading costs which private and institutional investors have to cover whereas quoted spreads are important in revealing the price-generating mechanisms involved. A comprehensive trade-indicator model of the bid-ask spread of common stocks has been developed which, unlike previous research, has incorporated trading volume, the depth at the quoted prices, the waiting-time between trades, as well as the fixed-cost-of-trade into equations for the changes in the quoted-spread, the ask, the bid and the transaction prices. The parameters of the models have been estimated using intraday data of NYSE and NASDAQ stocks, split into deciles on the basis of their trading activity as measured by the number of shares traded. For both exchange mechanisms I find that a large adverse-selection cost, which depends on the trading volume and a smaller inventory-holding cost are present in the quoted spread and that the parameters of both of these vary with trading activity. These costs are asymmetric in the bid and ask sides of the quoted spread, a feature not analyzed in previous empirical work. Only a small part of these costs is recovered in the realized spread through trading. My estimates of the adverse-selection cost present in the realized spread are close to the values given by other researchers but the size of the inventory-holding cost is found to be much lower probably owing to the shorter time-horizon of the data employed. The depth at the quotes is found to be symmetric, to affect the size of the spread of both the NYSE and NASDAQ stocks and also to be present in the realized spread. The parameters of the above components, as well as the fixed-cost-of-trade, are estimated and their patterns for the two trading mechanisms examined are compared and contrasted. Weak evidence is found for the waiting-time between trades both in the quoted and the realized spread. The results of this thesis, apart from offering support for recent empirical evidence which indicates that information first enters the price-process through the depths of the quotes and not the spread, also contribute to the formation of a more theoretically sound explanation. Moreover, the finding in this thesis that the adverse-selection cost for NASDAQ is larger compared to that of NYSE stocks is in line with other recent empirical research

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:403135
Date January 2004
CreatorsSpirakos-Papastavridis, Spiridon
PublisherUniversity of Warwick
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://wrap.warwick.ac.uk/1223/

Page generated in 0.1462 seconds