Publication detail

Barunikova, M. Barunik, J.: Neural Networks as Semiparametric Option Pricing Tool

Author(s): doc. PhDr. Jozef Baruník Ph.D.,
Mgr. Michaela Vlasáková-Baruníková (Hlínková) ,
Type: Articles in refereed journals
Year: 2011
Number: 0
ISSN / ISBN:
Published in: Bulletin of the Czech Econometric Society, Czech Econometric Society, 18(28), pp. 66-83 PDF
Publishing place:
Keywords: option valuation, neural network, S&P 500 index options
JEL codes:
Suggested Citation: Barunikova, M. Barunik, J. (2011): Neural Networks as Semiparametric Option Pricing Tool, Bulletin of the Czech Econometric Society, 18(28), pp. 66-83
Grants: 402/09/0965: New Approaches for monitoring and prediction of capital markets 402/09/H045 - Nelineární dynamika v peněžní ekonomii a financích. Teorie a empirické modely
Abstract: We study the ability of arti cial neural networks to price the European style call and put options on the S&P 500 index covering the daily data for the period from June 2004 to June 2007. The greatest advantage of option pricing with neural networks is that we do not need to make any assumptions about the volatility of the underlying asset. We divide the data set into several categories according to moneyness and time to maturity. Then, we price all options through the categories. The results show that neural networks outperform Black Scholes model with signi cantly lower pricing error across all categories for both call and put options. Moreover, the di erence between Black Scholes and neural network errors signi cantly widens with deepness of moneyness or expiration. The deeper the option is in/out of the money and/or the longer the option
has expiration, the greater is the di erence between neural networks and Black Scholes errors. We show that neural networks can correct for the Black Scholes maturity and moneyness bias.

Partners

Deloitte

Sponsors

CRIF
McKinsey
Patria Finance