I am analysing option-implied RNDs and risk preferences for my masters thesis, so forgive me if I sound like a beginner in derivatives.
I use WRDS to download my historic options data. I am looking at SPX European Friday-settlement options with τ = 5 weeks, looking at a month by month nonoverlapping data. I'll give an example:
Information Date: Dec 13th, 2017
Exercise Date: Jan 19th, 2018
Index Close: 2662.85
Average of bid and ask for this strike: 1663.50
Implied vol: 1.3931
The next available strike with volume > 0 was K = 2000, with strikes above having reasonable gaps and volumes. This is not the only dataset where I see this. Is this normal and is it okay to include it in the dataset to build the RND? I assume the answer will be related to trading fees, but just want someone with more knowledge to confirm.