International Journal of Scientific Research and Engineering Development

Forecasting Total Quality Management (TQM) performance in hospitality remains underdeveloped despite its strategic role in pricing, staffing, and service quality. Existing approaches, dominated by univariate and deterministic models such as ARIMA, struggle to capture the volatility and interdependencies that define hotel operations. This study addresses this gap by employing a simulation-based framework to evaluate the Bayesian Autoregressive Moving Average (BARMA) and Vector Autoregression (VAR) models using three core TQM indicators—defect rates, quality costs, and customer satisfaction. The objectives were to (i) demonstrate the methodological integration of BARMA and VAR, (ii) compare their predictive and interpretive capacities, and (iii) generate actionable managerial insights for quality management under uncertainty.Using simulated monthly time-series data calibrated to industry benchmarks, BARMA (1,1) consistently outperformed VAR in predictive accuracy. For example, BARMA reduced RMSE for defect rates (0.37 vs. 0.39) and delivered lower MAPE for customer satisfaction (1.08% vs. 1.11%). In contrast, VAR revealed statistically significant causal pathways, such as defect rates predicting quality costs (β = 0.38, p < 0.05), and customer satisfaction feeding back into reduced defects (β = –0.22, p < 0.05). These findings highlight methodological complementarity: BARMA as a robust predictor for short-term operational planning, and VAR as a diagnostic tool for tracing structural feedback loops. The study contributes to forecasting research by showing how integrating Bayesian and multivariate approaches, even within a simulated environment, can overcome the limitations of linear, univariate models. For practitioners, the results emphasise the need to balance predictive accuracy with structural diagnosis when managing quality costs and customer outcomes in luxury hotel operations.