Office of Research A New Methodology for Measuring Liquidity-Induced Transaction Costs

This paper introduces a parsimonious methodology for measuring liquidity costs by parameterizing the relationship between price and order flow. Unlike the effective spread, our liquidity measure does not assume the midpoint of the spread equals the true price. We find that the marginal price impact of order flow declines rapidly and more liquid firms do not necessarily have lower liquidity costs. In comparison to our measure, the effective spread significantly over-estimates realized liquidity costs, especially for less liquid firms. Our empirical study also reveals that NYSE specialists strive to obtain a profit in dollars rather than as a percentage of the transaction.