Naslov
Revisiting CAPM model with quantile regression: creating investment strategies on Zagreb Stock Exchange

Autori
Škrinjarić, Tihana ; Slišković, Marina

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Applied statistics 2018 International Conference Abstracts and Program / Lusa, Lara ; Kastrin, Andrej ; Blejec, Andrej - Ribno : Statistical Society of Slovenia, 2018, 41-41

ISBN
978-961-94283-1-3

Skup
Applied Statistics 2018 (International Conference)

Mjesto i datum
Ribno, Slovenija, 23.09.2018. - 26.09.2018

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
CAPM, quantile regression, stock market, investment strategies

Sažetak
Capital asset pricing model (CAPM) has been thoroughly investigated on stock markets over the last couple of decades. Many variations and extensions have been made of this model. In this research we explore whether conditional CAPM holds at different points of the return distribution by focusing on data from Zagreb Stock Exchange and quantile regression methodology. There are several reasons for this specification of modelling: quantile regression does not require strong assumptions on return distributions and handles heteroskedasticity of data. Moreover, CAPM model has not yet been observed by using quantile regression on Croatian and several similar CEE markets as well. In that way, we can observe if this methodology is useful to estimate systematic risk on Croatian stock market conditioned on different quantiles of return distribution. Weekly data on 5 sector indices, market return on CROBEX and return on Treasury bills (91 day) for period January 2012 – April 2018 will be collected in order to empirically evaluate CAPM model via quantile regression. Economic interpretations of results will be given as guidance for investors. Moreover, contribution of this research will be given in the simulation part, where we discuss several specifications of investment strategies based upon estimation results. Previous literature does not focus on utilizing estimation results as guidance for dynamic investment strategies. Based upon simulations of several strategies, several return and risk measures will be calculated in order to compare and rank those strategies. Since this study is one of the few which try to link statistical aspects of estimating finance model with investment strategies, there is hope that this research contributes to existing literature on aforementioned matters.

Izvorni jezik
Engleski

Znanstvena područja
Matematika, Ekonomija

POVEZANOST RADA