Estimation of Odds Ratio as a Quality Indicator on Investment Recommendations - A Bayesian Approach

A Stock brokerage is a service-oriented agency whose primary objective is to buy or sell shares on behalf of their clients and they also deal with broking services, research, wealth management, retirement planning, depository services, mutual funds, etc., This study is about finding a pattern of profit and loss in two types of call recommendations (Buy/Sell) based on the data collected from a reputed stock brokerage firm. The inherent advantage in handling Bayesian modelling has been attempted with the necessary models using suitable transformation of underlying parameters. Quantifying the measure of associations between the variable of interest is achieved through odds ratio together with the measure of heterogeneity. Various models could be achieved through possible combination of variables and the results are presented both in numerical and graphical mode. This study has made an attempt in building a model based on the recommendations of a stock broker. Based on the data received from the stock broker, the response metric variable is treated as a categorical variable using appropriate rules. Identifying suitable associated variables to understand the variability quantification in a more better way and the summaries may be better in Random effect model approach compared to original treatments. This study has given a clear approach to Bayesian analysis which could be carried out on a fixed dataset relatively simple using MCMC to simulate posterior distributions. This study provides a direction to understand the recommendations given by the stock broker.