Detection of Effective Factors on Stock Screening Using Analytic Network Process

Investment managers of financial institutions are favor to identify superior companies to select those in their stock portfolio and hedge probable risks and achieve predetermined return. The main question is how we can sieve companies and basically which companies are superior. Therefore in this thesis we designed a novel sieving model by the use of multi criteria decision making and analytic network process. The model contains of 5 clusters including operating efficiency, profitability, risk, growth and market, which are consist of 21 nodes and subcriteria that indicate companies performance. After consulting with experts and applying analytic network process technique, subcriteria reduced to 18 effective criteria and used to score each company. The model ran for 10 active car manufacturing companies in Tehran Stock Exchange (TSE). The consequences of proposed stock sieving model compared with results of Iran Tahlil stock ranking model which is popular in (TSE). It showed on the 95% confidence level, results of two models have weak correlation. To assess validity and performance of each models, we defined profitability index because of stock yield importance for investors, and assessed two models respect to index. Results indicated performance of proposed stock sieving model is greater than Iran Tahlil ranking model.