Investigating commercial banks profitability in Nigeria using Bayesian statistical approach

Abstract

This paper investigates the profitability of Nigeria commercial banks using internal and external variables considered to be determinants of bank profitability. The well celebrated Bayesian model averaging is employed systematically. And it reveled that from the thirteen variables considered EQTY, AST, INV, CAR, CRR, BSZ, EXM, LTA, DTA, GDP, IRR, EXR and IR which gave 8192 plausible models  generated by MCMC algorithm a sample of 1393 models were sampled. Only EQTY, AST, INV, CAD, CRR, BSZ and EXM showed evidence of strong explanatory power in explaining commercial
Banks profitability, with EQTY as the most probable predictor. It is found that the mean recorded number of covariates is seven with a shrinkage statistics 0.9988 meaning that for any commercial bank to improve in their profitability they must give more attention and improve greatly on their returns on EQTY, AST, INV, CAD, CRR, BSE and EXM.

Keyword: Shrinkage, Posterior Probability. MCMC, Profitability.