Trade Policy Uncertainty, Financial Frictions, and the Sources of Export Hysteresis

Abstract

Firms that could profitably export often do not, and firms that stop exporting often fail to return when conditions improve. Two literatures explain this hysteresis through different frictions. The real options literature holds that entry is irreversible and that uncertainty about future trade policy creates an option value of waiting, so firms delay entry even when current expected profits are positive. The financial frictions literature holds that entry requires sunk expenditure that constrained firms cannot fund, so they do not enter regardless of what they expect. The two produce observationally similar reduced forms, and empirical work has often been unable to separate them. This article argues that the failure to separate them has consequences, because the implied remedies are not merely different but are frequently treated as substitutes when they are complements. Binding trade agreements address uncertainty and do nothing for a firm that cannot raise the entry capital. Trade finance facilities address constraints and do nothing for a firm that is rationally waiting. The article's central claim is that the two frictions compound rather than add: uncertainty raises the return a lender requires on sunk-cost lending, tightening the constraint, while the constraint removes the firm's ability to exercise its option when uncertainty resolves favourably, so the option becomes worthless precisely when it should pay off. Under this reading, the standard practice of estimating one channel while treating the other as background does not merely omit a term. It biases the estimated magnitude of whichever channel is measured, in a direction that depends on the correlation between policy volatility and financial development across the sample. The article specifies the evidence that would separate the channels, argues that most of it already exists in unlinked datasets, and draws out the policy consequences of continued conflation.