Decimalization: A Review of the Arguments and Evidence
Lawrence Harris, USC Working Paper, April 3, 1997.
Executive Summary
The minimum price increment (tick) determines what prices traders use. The tick is a decimal fraction in some markets and a fraction based on powers of two in other markets. The US Congress is presently considering a bill that would require the US stock markets to convert to decimal pricing. The likely outcome of decimalization would be a smaller price increment.
Proponents argue that decimal prices are easier to use than fractional prices and that decimalization is necessary to modernize US markets. More substantially, they argue that a smaller price increment would encourage price competition, narrow bid-ask spreads and decrease payment for order flow. Narrower spreads would ensure better prices for retail market orders, but less payment for order flow would raise retail brokerage commissions. Proponents also argue that decimalization is necessary for US markets to compete effectively against other markets.
Opponents argue that a smaller price increment would shift power from public traders to professional traders by making it easier for professionals to step in front of public limit orders. As a result, public traders will display their orders less and switch from limit order strategies to market order strategies. Quotation sizes will decline, transaction costs will rise and the markets will become less transparent.
The Canadian stock exchanges switched to decimal prices in April 1996. In the process, they decreased the tick from 12.5 cents to 5 cents for stocks priced above 5 dollars. Quoted spreads narrowed by about 4 cents but quotation sizes also declined substantially. Volumes did not change nor did the US market share of trading in cross-listed Canadian stocks. Several other international exchanges recently decreased their ticks with similar results.
The effect of tick size on market quality has also been studied by examining how stocks trade on different tick sizes. Small tick stocks have narrower spreads and less quotation size displayed than do large tick stocks. In markets that allow traders to submit limit orders that are not fully displayed, traders choose to display less size when the tick is small.
Some analysts have estimated the potential annual cost savings to the public of decimalization by multiplying half the projected decrease in spreads by annual trading volume. This method produces estimates of about 2 billion dollars per year for NYSE stocks. This is about 10 times larger than 1996 NYSE specialist profits, and about 2.5 times larger than their total revenues. This method grossly overestimates the potential cost savings because it implicitly assumes that the public trades only with dealers. Most trading, however, is between public traders. A cost savings to one public trader will represent an increased cost to another public trader.
All empirical results concerning tick size come from studies of exchanges in which a time-precedence rule encourages traders to improve price. Tick size is important in these markets because it determines the cost of obtaining precedence through price. The effect of tick size on trading in dealer networks is probably much smaller than in exchange markets because dealer markets do not enforce time precedence among dealer quotes. Decimalization is therefore unlikely to have much effect on payment for order flow or order preferencing in dealer markets.
The empirical results suggest that small market order traders and professional traders benefit from a smaller tick. A decrease in tick size hurts large traders and all public traders who use limit orders to compete with dealers.
Download the Acrobat File
Return to Larry Harris's home page
Last revised 4/10/97.