Predictive Analytics for Inflation Forecasting and Data-Informed Monetary Policy Optimization in the United States

Abstract

Effective monetary policy formulation requires accurate forecasting of inflation since the dynamics of inflation directly affect interest rates, financial stability and the overall economic performance. Nevertheless, most of the traditional inflation forecasting models have commonly been linearized and they have been based on stable economic relationships that might not reflect nonlinearities, structural breaks and regime shifts that are observed in long term macroeconomic data. This paper builds a predictive analytics based model to study inflationary behavior and help rationalize the monetary policy in the US using data. The study is based on historical annual data that has been used since the year 1929 to the year 2024 and can therefore be said to have been used in the analysis of inflation dynamics across different economic regimes such as those of sustained stability, significant economic shocks, and increased inflation volatility. The research combines both descriptive statistical analysis and machine learning in the evaluation of inflation momentum, shock deviations, and volatility trends which cannot be evaluated by the traditional methods of econometrics. The model to be used is the Support Vector machine (SVM) to classify low and high inflation regime by the rate of inflation and the most important indicators of monetary policy especially the federal funds rate. In order to assure methodological rigor, model performance has been assessed with several measures, such as confusion matrices, receiver operating characteristic analysis, F1-score, and recall, which are ideal in the conditions of macroeconomic data such as limited observations and class imbalance. The empirical evidence shows that predictive analytics has the capability to detect the inflation regimes and can better detect the inflationary risk, especially when longer historical windows are taken into account. The findings also indicate the importance of regime-driven models and nonlinear analysis methods in improvement of inflation monitoring and forecasting models. Policy-wise, the research shows that the informed use of data provided by predictive analytics can decrease policy response lags, enhance situational awareness, and make more adaptive and evidence-based monetary policy choices in the United States, thereby enhancing the existing literature on the use of machine learning in macroeconomic policy analysis.