Trade policy uncertainty and the patent bubble in China: evidence from machine learning

Abstract

ABSTRACT This paper draws upon resource dependence theory and investigates how trade policy uncertainty affects firm strategic innovation management in China. Adopting a novel machine learning approach called Word2Vec, we construct and validate a measure of firm-level managers’ perceived trade policy uncertainty (TPU). We find that TPU has a positive effect on the number of total patent applications, but this positive effect is totally driven by low-quality patents instead of high-quality patents. Moreover, we document that firms have stronger incentives for such strategic innovation behavior when the underlying firms are more financially constrained, and/or when the management is more myopic.