Options Trading Costs Are Lower than You Think¶
Bibliographic record. Follow the original-source link for the publication.
| Field | Value |
|---|---|
| Primary domain | Volatility |
| Other domains | Microstructure, Execution Costs |
| Methods | Research Methods |
| Facets | Instrument Single Stock Options, Structure Straddle |
| Authors | Dmitriy Muravyev, Neil D Pearson |
| Published | 2020-02-10 |
| Source | The Review of Financial Studies |
| Identifiers | doi:10.1093/rfs/hhaa010 |
| URL | Open original source |
Editorial synthesis¶
Why it matters¶
A trader's chosen execution time can make the midpoint a misleading valuation benchmark. The study informs execution-quality measurement, not a blanket reduction in simulated slippage, fees or fill prices. (full_text:S44, full_text:S48, full_text:S49, full_text:S54, full_text:S629)
Main author claims¶
- The authors estimate midpoint-normalized half-spreads of 6.8% conventionally, 5.0% after the overall adjustment and 2.5% for inferred execution timers. These stock-day averages measure costs relative to option value, not margin capital. (
full_text:S365,full_text:S400,full_text:S414,full_text:S419,full_text:S437) - The authors obtain average OOS R-squared of 5.5% and 7.0% when forecasting the next month with coefficients aggregated from the previous month. Main cost tables instead rely on contract-day quote fits; the forecast test does not validate an executable OOS strategy. (
full_text:S219,full_text:S253,full_text:S272,full_text:S279,full_text:S281,full_text:S424) - The authors' Model 2 adjustment changes average ten-minute price impact from 1.526% to 0.644%. Removing forecastable movement yields a model-dependent measure, not experimental identification of an order's causal impact. (
full_text:S507,full_text:S509,full_text:S514,full_text:S521)
Data, method, or discussion scope¶
Equity options on S&P 500 constituents, January 2004–December 2015: 620,916,282 filtered trades and 1,242,893 stock-days from Nanex OPRA/SIP and CRSP. Quotes are minute snapshots; costs average stock-day statistics. The separate OptionMetrics straddle illustration imputes entry costs from comparable options. (full_text:S103, full_text:S105, full_text:S117, full_text:S119, full_text:S143, full_text:S400, full_text:S571, full_text:S592, full_text:S594)
Main limitations¶
Timing share and algo costs infer latent types under mean, median and group-cost assumptions; accounts are not identified. Historical daily fits do not guarantee attainable package execution or all-in costs. Table 10 panel A has unresolved arithmetic differences between cost statistics and return deductions. No independent replication. (full_text:S105, full_text:S219, full_text:S368, full_text:S374, full_text:S394, full_text:S395, full_text:S437, full_text:S555, full_text:S576, full_text:S592, full_text:S629)