Analysis of Risk in the Logistics Sector Using Non-linear Time Series in Türkiye

Yıldırım S. K., Can M.

Journal of Transportation and Logistics, cilt.11, sa.1, ss.51-66, 2026 (TRDizin)

Özet

This study examines the risk level of the logistics sector, which is a key sub-component of the service sector in Türkiye, by comparing it with other service sub-sectors that have a high share within the sector, namely retail trade, electricity, gas, and steam, and telecommunications indices. Daily data from the period between March 2, 2020, and October 8, 2025, when volatility in financial markets increased significantly after COVID-19, was used for the analysis. First, returns for the indices were generated, and various non-linearity tests were used to determine whether these returns were non-linear. Next, the stationarity properties of the series were tested using non-linear unit root tests, and the appropriate ARMA(p,q) model was determined for each series based on information criteria. To examine the volatility structure of the return series, symmetric conditional heteroskedasticity models (GARCH, GARCH-M) and asymmetric effect models (T-GARCH, E-GARCH, and GJR-GARCH) were applied. The most suitable model was selected by considering error criteria and log-likelihood values. The findings reveal that the risk levels among the service sector indices are close to each other but show sectoral differentiation. According to the analysis, the telecommunications sector had the highest risk during the period under review, followed closely by the logistics sector. In contrast, the retail trade sector was identified as the sub-sector with the lowest volatility.