Using genetic algorithms in panel Data modeling: the relationship between stock prices and financial performance of Saudi arabia’s listed companies

Abstract

This study aims to use Genetic Algorithms in Panel Data modeling optimization, so we applied it to measure the relationship between the stock price and financial performance indicators of ten of leading listed companies in Saudi Arabia depending on annual time series of the last twelve years. Thus, we formulated a Mixed Effects Panel Data model containing Stock price related to: Return on Equity, Return on Assets, Earning per Share, Financial Leverage and Current Ratio; then we estimated this model by maximizing an adapted Quasi-Likelihood function using Genetic Algorithms depending on Evolver software, In addition we estimated Fixed Effects and Random effects models using usual econometric methods, all the models indicate significant impact of: ROE, ROA and EPS. The optimum Mixed Effects model estimated by Genetic Algorithms shows more accurate results with relatively better statistical characteristics and allows more usage flexibility. Special issue: The use of decision-making tools and methods in economics