Robust Forecasting Models for Highway Construction Cost Indices during Pandemic-Era Inflation

Abstract

The highway construction cost index (HCCI) is a critical benchmark for managing highway construction project budgets. However, unpredictable economic and market changes, such as inflation due to the pandemic and supply chain disruptions, challenge accurate HCCI forecasts. In this context, an innovative index prediction approach is proposed, using advanced forecasting models such as vector error correction model (VECM), long short-term memory (LSTM) networks, and seasonal autoregressive integrated moving average (ARIMA). Key economic indicators, such as crude oil prices, gross domestic product (GDP), unemployment rates, and so forth, are identified as the explanatory factors. Their data are collected for 2010–2023, including prepandemic stability, pandemic volatility, and postpandemic recovery. The data exhibit distinct differences during this period and are used to train and calibrate forecasting models. The prediction results reveal that VECM is the most effective model for short-term forecasting under volatile conditions and achieves the lowest average mean squared error (MSE) of 0.0205 in the walk-forward validation, compared with 0.05200 for LSTM and 0.1410 for seasonal ARIMA. By leveraging these comparative insights, the study advances the development of adaptive forecasting methods for HCCI and provides insight into the responses of VECM to economic disruptions, especially during the pandemic era.