Analysis of Passenger Flight Distance as an Indicator of Economic Activity

Abstract

Understanding macroeconomic dynamics in the United States requires advanced forecasting techniques capable of capturing both seasonal structures and external shocks. This study investigates the relationship between passenger flight distance and the unemployment rate through the implementation of the Seasonal Autoregressive Integrated Moving Average with Exogenous Variables (SARIMAX) model—an enhancement of the SARIMA framework. While SARIMA accounts for autoregressive, differencing, and moving average components with seasonal integration, SARIMAX further augments this structure by incorporating exogenous predictors, enhancing explanatory and predictive power.

Monthly time series data from 2015 to 2024 were utilized, with flight distance as the endogenous variable and the unemployment rate as the exogenous regressor. The modeling procedure involved rigorous stationarity testing via the Augmented Dickey-Fuller (ADF) test, model selection using the Akaike Information Criterion (AIC), and residual diagnostics employing the Box–Ljung and Shapiro–Wilk tests. SARIMAX(0,1,0)(0,1,1)[12] + X emerged as the optimal specification, with all parameters statistically significant and a MAPE of 3.68%, denoting excellent forecast accuracy.

Empirical findings reveal a significant and negative association between unemployment and air travel activity, emphasizing the role of labor market dynamics in shaping mobility trends. These results reinforce the utility of SARIMAX as a robust tool in macroeconomic forecasting and evidence-based policy formulation.