On Inferring the Direction of Option Trades
47 Pages Posted: 27 Dec 2001
Options market microstructure research relies primarily on transactions and quote data available from the CBOE. In most implementations, the data do not include information on whether a given trade is buyer-or seller-initiated. As a result, some researchers have used trade classification rules developed for common stocks. These rules include the quote, the tick, the Lee-Ready (1991) and the Ellis-Michaely-O'Hara (2000) methods. Using a proprietary CBOE data set that reports trade direction, we find that the correct classification rate for the quote rule is 83%, and that for the Lee-Ready, Ellis-Michaely-O'Hara, and tick rules is 80%, 77%, and 59%, respectively. The main forms of option trade misclassification include outside-quote trades and reversed-quote trades (i.e., buying at the bid and selling at the ask). Other factors, such as trading frequency, volume, moneyness, and maturity have indirect effects by influencing the probability of outside-quote and reversed-quote trades. Underlying asset price changes around the time of the trade are found to enhance classification precision. On further analysis, we are able to isolate trades that are misclassified almost 50% of the time by any method. These are the components of index spreads and index combinations other than those executed on RAES. These trades comprise approximately 15% of the sample and their elimination results in higher than 87% correct classification rate for the quote rule.
Keywords: buyer-initiated, seller-initiated, tick rule, quote rule, transaction data, options
JEL Classification: G10, G13
Suggested Citation: Suggested Citation