Derivative security Dn can be represented as a polynomial of infinite order in hn :
Dn — ao + aɪhn + a2hn + ∙ ∙ ∙ ■
Clearly, Dœ — ao and its linear approximation is:
D1 — αo+ α1h1 h2D1 — h1D2 „ „
> D, = ɪɪ--√ 2 — 2D2 — Dv
(13)
D2 = ao + a1h2 h2 — h1
As noted, D1 can be obtained as the value of the corresponding Europeamstyle option. D2 can be calcu-
lated from recursion (12). This task is sufficiently difficult per se, so, quadratic or higher order approximations
for Dœ are not feasible.
6. Empirical Application
6.1. Data
The tools developed in the preceding sections are used to predict prices of S&P 100 index options.
The S&P U.S. 100, a subset of the S&P 500, is comprised of 100 leading U.S. stocks, which together
represent almost 45 percent of the market capitalization of the U.S. equity market. As of June 2004, the
5 largest companies included were: General Electric, Exxon Mobil Corp., Microsoft Corp., Pfizer Inc., and
Citigroup Inc. The S&P 100 index was originally developed by the CBOE and later transfered to Standard
& Poor’s for management.
The CBOE offers three distinct options on S&P 100. Two of these have relevance to this paper. The
most popular one is the American-style S&P 100 index option, commonly known by its ticker OEX. As a
rule, on each trading day, OEX puts and calls have high trading volumes and open interest for a wide range
of strikes.
Less popular, but still actively traded at CBOE is the European-style S&P 100 index option, ticker
XEO. Both OEX and XEO are cash-settled, and apart from the difference in the exercise style, share same
characteristics (exercise dates, minimal strike intervals, minimum ticks, etc.).
I collected closing CBOE prices of both OEX and XEO on 7 consecutive trading days: June 30th - July
2nd, July 6th - July 9th. Only options with positive trading volume and open interest are used in estimations.
July 9th data is set aside for out-of-the-sample predictions.
As a proxy for the risk-free interest rate I use the T-bill rate. Rates for different times to maturity are
obtained by interpolation: linear OLS fit to the whole set of T-bill quotes reported in the Wall Street Journal
12