FACTOR BASED APPROACH TO EFFICIENT PORTFOLIO DIVERSIFICATION IN THE ILLIQUID AND UNDEVELOPED STOCK MARKET

In the past decade a variety of research analyzed the possibilities of moving away from inefficient cap-weighted benchmarks by empirically testing different "smart" beta strategies which aim was to offer a more efficient harvesting of risk premium in the developed equity markets. While successful in the developed markets such strategies are not easy to implement in the undeveloped and illiquid markets in which it is hard to outperform the cap-weighted benchmark due to fewer investment opportunities and missing or unreliable data. In this paper we analyze the possibility of using fundamental weighting, a strategy criticized for exposing investors to additional and implicit risk factors, as a tool to intentionally tilt portfolio towards illiquid and undeveloped market's specific and unobservable risk factors. In the second step we perform principal component analysis to capture the underlying risk factors of the fundamentally weighted portfolio and use them to optimize portfolio's performance in the analyzed sample by reducing its volatility. Thus far fundamental-weighting has been shown to be able to outperform the cap-weighted index in such environment but no attempt regarding control for implicit factor exposure of such portfolio has been reported. On the other hand, approaches focusing purely on portfolio risk reduction by estimating minimum variance portfolios failed both from in-sample and out-of-sample perspective. Therefore, we limit our analysis to in-sample estimation only in order to assess potential for further testing. We test 15 revisions of the cap-weighted CROBEX index in the period from 2009 to 2016 and find that the proposed approach significantly improves portfolio's performance when compared to the (untilted) minimum variance portfolio estimated by applying the principal component analysis. We also find that it outperforms the cap-weighted benchmark in the illiquid and undeveloped Croatian stock market, albeit on the in-sample basis. However, research results do not show the possibility of increasing the portfolio efficiency relative to the cap-weighted index by reducing its volatility which can be explained by exposure to additional risk factors. Still, the reported improvement should foremost be viewed in the context of a market in which both the naive equal-weighting and the minimum volatility strategy failed and, therefore, did not yield results comparable to the developed markets. Nevertheless, to evaluate the improvement of this approach over the standard fundamental-weighting further out-of-sample testing has to be conducted.