Automatic Order Matching and Latency in the Dissemination of U.S. Equity Data

Abstract

This study reveals that public U.S. market data at present yields incorrect prevailing quoted prices for 29% of trades from the NYSE and 33% of trades from Nasdaq. Quote updates that accurately occur after trades at an exchange are often disseminated before the trades on the consolidated data feeds and are thus recorded in TAQ data in incorrect order. This generates substantial errors in trading cost and price impact measures. I identify the culprit behind the errors, recommend a simple approach that yields correct prevailing quoted prices, and discuss changes to market design that would prevent the issue.