Explaining Narrow-Body aircraft depreciation and value dynamics with machine learning

Accurate forecasting of aircraft depreciation is critical for valuation, leas- ing, and risk management in aviation. Traditional appraisal and cost-based approaches often fail to capture the nonlinear effects of market cycles and macroeconomic conditions. This study applies machine learning to predict the current fair market value (CFMV) of Airbus and Boeing narrow-body air- craft using a rolling-origin evaluation framework. The feature set integrates appraisal-standard variables (age, delivery year, subtype) with macroeco- nomic indicators such as the consumer price index, jet fuel price, interest rates, and air traffic indices. We benchmark regularized linear models against ensemble methods, finding that gradient boosting (XGBoost) consistently de- livers the strongest performance, achieving mean absolute percentage error (MAPE) below 5% and R2 near 0.90. Residual analysis confirms stable accu- racy across aircraft types, while depreciation surface visualizations illustrate how lifecycle aging and market shifts interact to shape values. Results in- dicate that lifecycle and technical characteristics dominate predictive power. These findings demonstrate the potential of machine learning to enhance traditional appraisal practices.