Skip to Content

Stanford GSB News

 

Researchers Put a Price Tag on London Stock Trades

March 1995

STANFORD GRADUATE SCHOOL OF BUSINESS—Electronic trading technologies have drastically reduced the costs of financial transactions and put tremendous pressure on financial exchanges to lower their costs. Responding to this pressure, in 1986 the London Stock Exchange switched from a closed, floor-based, broker-dealer market to an open electronic quotation system dubbed SEAQ that operates much like the NASDAQ system in the United States.

The London exchange's regulatory changes emphasize liquidity over transparency, renewing the debate on whether such rules affect the costs of financial transactions.

The question of what it actually costs to conduct a trade remains elusive. Unlike a retail item such as an automobile, which comes with a sticker stating the manufacturer's price, the true value of a security is difficult to determine. Bid and ask prices are set by market makers, but individuals dealing in a market may be able to negotiate discounts.

Stanford Business School researchers Peter Reiss, MBA Class of 1963 Professor of Economics, and Ingrid Werner of Ohio State University, studied the new London system to try to determine the actual costs of trading securities. "The question is, how much am I actually paying to the market maker if I buy at the ask price? This depends on what the true (unobservable) value of the security is," says Werner.

Most studies of market transactions have assumed that the true cost of a security is midway between the best bid and the best ask price. Reiss and Werner disagree, arguing that that midpoint doesn't take into consideration all potential discounts. They analyzed the relationship between investor transaction costs and the best bid-ask spreads—also known as "the touch"—and found that dealers operating on the London Stock Exchange offer systematic discounts from posted prices.

Their study of transaction costs is unusual, say Reiss and Werner, in that "to the best of our knowledge, this is the first study to account simultaneously for differences by type of trade, trade volume, and type of security." They analyzed the spreads between the bid and ask price and the extent to which those spreads are discounted by market makers—individuals who are obliged by market rules to post prices and take trades at all times.

Theoretical models of market prices have postulated that large trades obtain less favorable transaction prices than smaller ones. Researchers have argued that large trades pose at least two problems for the dealer: the dealer has to assume the risk of changes in demand adversely affecting the value of his position and also faces the risk that the party seeking to make a large deal may be privy to information. Analyzing their data, Reiss and Werner found the opposite to be more likely.

"Our estimates reveal that medium to large trades on average received discounts from the touch spread. These discounts increase the wider the touch. Small and very large trades pay the touch (and sometimes more)," write Reiss and Werner. They also discovered that market makers and dealers (who generally trade for a specific institution) price customer trades differently. "Market makers only discount very large trades; dealers regularly discount medium and large trades. Market makers rarely discount trades with other market makers over the phone, but do so when trading anonymously using interdealer brokers," say Reiss and Werner. "The pattern we observe suggests that neither simply asymmetric information nor inventory risk models can easily explain why dealers widen spreads and then selectively discount," they say.

—by Cathy Castillo

Related Information

Anonymity, Adverse Selection, and the Sorting of Interdealer Trades, Peter Reiss and Ingrid Werner, Review of Financial Studies, August, 2005 Details