Abstract:

Based on the pricing behavior of micro-enterprises, this paper theoretically analyzes the nonlinear relationship between different markup enterprises in terms of price transmission under cost shocks. The theoretical research finds that high markup enterprises show a stronger price-raising behavior under positive cost shocks, i.e., price transmission shows spillover effects. However, under negative cost shocks, product pricing has not been significantly adjusted with cost reductions, but the impact of cost shocks has been absorbed by markup, which shows the absorption effect. In contrast, for low markup enterprises, price transmission shows the absorption effect under positive cost shocks and the spillover effect under negative cost shocks. This paper further empirically tests the above research conclusions at the industry level by using the double/debiased machine learning (DML) model based on nonparametric inference with continuous treatments. The research in this paper provides a new perspective for understanding price transmission behavior, which in turn has important reference significance for the prediction and prevention of inflation in China.

Key words:

markup, price transmission, spillover effect, absorption effect, double/debiased machine learning (DML) model

摘要：

原材料等成本冲击是否必然导致我国物价的普遍上涨？基于微观企业的定价行为，从理论上分析了成本冲击下不同成本加成企业在价格传递方面所呈现出的非线性关系。理论研究发现：高成本加成企业在正向成本冲击下表现出更强的涨价行为，即价格传递表现为溢出效应；在负向成本冲击下，企业产品定价并未随成本下降做出显著调整，而是由成本加成吸收了成本冲击的影响，表现出吸收效应。与之相对照的是，低成本加成企业在正向成本冲击下表现出价格传递的吸收效应，而在负向成本冲击下表现出溢出效应。进一步使用基于连续处理变量的非参数推断双重/无偏机器学习模型，从行业层面进行了实证检验。对洞悉价格传递行为提供了新的视角，进而对我国通货膨胀的预测和防控具有重要参考意义。

关键词:

成本加成, 价格传递, 溢出效应, 吸收效应, 双重/无偏机器学习模型