Exports of intermediate goods and innovation: Evidence from China’s firms

Abstract

AbstractBased on the experience of developed countries, the production and export of intermediate goods have been observed to foster innovation within firms. This study investigates whether China's export of intermediate products has significantly enhanced the independent innovation capability of China’s firms. Using microdata and patent applications as a measure of innovation, we adopt the staggered PSM-DID method to explore the influence of China’s intermediate exports on firm innovation. We find that China's export of intermediate goods positively influences firm innovation, particularly utility patents. Domestic firms benefit more from this effect than foreign firms, and medium-high technology industries experience the most significant innovation impact. The export of intermediate goods stimulates innovation through the following channels: scale economies and employment structure optimization. These findings shed light on China's evolving innovation landscape.KEYWORDS: Exports of intermediate goodsfirm innovationstaggered PSM-DIDJEL Classifications: F02F15O31 Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 Trade data are derived from the UN COMTRADE database.2 These are calculated by the authors from the UN database.3 This paper is mainly based on the results of the dataset from 2000 to 2007, and to ensure the robustness of this result, we also re-estimate using the period 2011–2013.4 However, one crucial issue with staggered DID is the presence of heterogeneous treatment effects, meaning that the same treatment's impact can vary among individuals. Traditional two-way fixed-effect models may introduce potential bias in such cases. To address this concern, in the robustness check conducted in this paper, we employ a method proposed by De Chaisemartin and D’Haultfœuille (Citation2020) to re-estimate the model and mitigate this problem.5 To ensure the reliability of the estimation results, the kernel matching, and Mahalanobis matching methods are also tried in the robustness analysis.6 The empirical results based on the dataset for the period 2011–2013 are shown in Appendix Table A6, which are consistent with those based on the period 2000–2007. It should be noted that we did not use TFP as a control variable when performing the robustness check, because this dataset lacks the relevant indicators for calculating TFP.Additional informationFundingThis work was supported by National Natural Science Foundation of China: [Grant Number 71673205]; the Fundamental Research Funds for the Central Universities: [Grant Number 2023YJ10501].