Energy Transportation Uncertainty and Global Economic Activity: State-Dependent Evidence from Multivariate Quantile-on-Quantile Regressions

This study examines whether energy transportation uncertainty (ETU) contains state-dependent information about global economic activity. Using monthly data, world industrial production (WIP) is treated as the primary activity measure and the Global Economic Conditions (GECON) indicator as an alternative outcome. The empirical design employs multivariate quantile-on-quantile regression (m-QQR), which maps the response of each conditional quantile of economic activity to each quantile of ETU while accounting for persistence and, where conceptually appropriate, broader risk and financial conditions. WIP growth is modeled with lagged geopolitical risk (GPR), whereas the preferred GECON specification excludes GPR because geopolitical-risk information already enters the information set used to construct GECON. The 0.05–0.95 quantile grid reveals substantial heterogeneity in the relationship between ETU and global economic activity, with negative estimates concentrated in several middle and upper WIP states. These local patterns are visible under pointwise inference but do not remain statistically significant after multiple-testing correction or moving-block-bootstrap inference. Financial controls do not alter this conclusion, and ETU provides little improvement in average out-of-sample WIP forecast accuracy. Scenario estimates nevertheless associate high ETU with weaker conditional median WIP growth. Overall, ETU contains state-dependent information about global activity, although the evidence does not support stable cell-specific effects once multiplicity and serial dependence are taken into account.