Measuring the Impact of Managerial Overconfidence on Firm Value Using a Hybrid Model Based on Longitudinal Data Models and the Random Forest Algorithm

Abstract

The research attempts to find the degree to which extreme administrative confidence has impact on the value of industrial firms that are listed on the Iraq stock market, with the use of a hybrid model that integrates longitudinal data obtained through linear associations between variables, and machine learning frameworks (random forest) that obtain non-linear and tricky connection between variables. To attain the study objectives, a purposeful sample was selected, which consist of ten industrial firms for the period of 2016-2025. The findings of the study indicated that disproportionate administrative has a positive effect on the values of firms, as well as the supremacy of the hybrid framework over the fixed effects model and the random model to represent correlation between variables. This needs the boards of directors of these firms to avoid the appointment of managers who lack the confidence to present genuine decisions. The study recommend that investors should direct their investments towards the firms that have managers with reliable leadership features and traits who are capable of facing all the potential circumstances and make necessary decisions without delay.