What are the known flaws and limitations of OptionMetrics data?

Several researchers are skeptical about the database quality, but their argumentation is somewhat unclear to me.

For instance, Constantinides, Jackwerth and Perrakis (2008) (link):

[...] we were conservative because of concerns regarding the quality
of the Option Metrics database.

No argumentation here. Then, Van Binsbergen, Brandt and Koijen (2012) (link):

Using closing prices from OptionMetrics for all quantities does not guarantee that the index value and option prices are recorded at the same time and induces substantial noise in our computations [...] For instance, the options exchange closes 15 minutes later than the equity exchange, which leads to wider bid-ask spreads in options markets during this period.

Here, the authors explicitly define the problem, but the WRDS OptionMetrics manual (link) states the opposite:

OptionMetrics compiles the IvyDB data from raw 3:59PM EST price information.

which could actually be a recent change to the methodology, unknown to the authors of the latter paper.

P.S. There's also a StackExchange answer questioning (no pun implied) quality of OptionMetrics dividend info (link) — again, without argumentation.

UPD #1: Found more info on the closing price issue:

Option prices used in implied volatility calculations up to March 4, 2008 are end of day prices. Starting from March 5, 2008 we have been capturing best bid and best offer as close to 4 o’clock as possible in an attempt to better synchronize the option price with the underlying close. Currently all option quotes, except for VIX option prices, are captured at 15:59 EST.

1 Answer
1

OptionMetrics has its flaws but it has been widely used in economics/finance research.

Regarding the Constantinides, Jackwerth and Perrakis (2008) paper I am unsure what their concern are. The Binsbergen et al. comment is easier to address. They basically have a confidential dataset that they use to estimate dividend strips. To do so, they need put call parity to hold and the put call parity relation has a time $t$ subscript therefore it is important measure call and put prices at the exact same time. OptionMetrics only has end of day quotes and it might indeed be the case that there is lag between observations. In any case I do believe you could establish the put call parity relations using option metrics data. Binsbergen et al. obviously had to defend their novel dataset. There is also some controversy regarding the slope of the term-structure of returns and/or equity yields. You should read Bansal, Miller and Yaron (2017) for more information on that.

Finally, regarding the option metrics data quality, I would not be concerned that much about the timing of the observations as I would be with the usage of their implied volatility surface. Although this has been used quite extensively in research, indeed OptionMetrics does an adjustment to prices of American options to account for the early exercise premium. Usually the profession ignores this problem, but if you are not using deep out of the money options then you actually do not know the exact adjustment OptionMetrics is doing. It is a bit of a blackbox.

Hope this helps.

Thanks for contributing an answer to Quantitative Finance Stack Exchange!

But avoid …

Use MathJax to format equations. MathJax reference.

To learn more, see our tips on writing great answers.

Required, but never shown

Required, but never shown

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy