制造业向智能制造转型升级是中国实现制造强国战略的必由之路。本文采用双重差分法，以《中国制造2025》政策出台为准自然实验，利用A股上市公司2010-2019年数据实证检验了智能制造政策对企业全要素生产率的作用效果及机制。研究发现：（1）智能制造政策能有效促进智能制造领域企业全要素生产率的提高；（2）智能制造政策一方面诱导企业增加无意义或低效率的研发投入，降低企业全要素生产率，另一方面引导企业增加有效发明专利数，进而提高企业全要素生产率；（3）智能制造政策作用效果在不同所有制性质、不同市场化程度地区的企业中存在显著差异。上述结果证实了智能制造政策的作用效果，并部分打开其作用机制黑箱，对下一阶段智能制造产业政策的制定与落实具有重要借鉴和启示意义。

The transformation and upgrading of manufacturing industry to intelligent manufacturing is the inevitable choice for China to realize the strategy of manufacturing power. According to the double difference method, this paper takes the introduction of "made in China 2025" policy as the quasi-natural experiment, and empirically tests the effect and mechanism of intelligent manufacturing policy on enterprise total factor productivity by using the data of A-share listed companies from 2010 to 2019. First, "Made in China 2025" has effectively promoted the improvement of total factor productivity of smart manufacturing enterprises. As an incentive policy, this policy will give target companies preferential benefits in terms of capital, taxation, land use, and talents, which can effectively guide and help companies reduce costs and enhance competitiveness, thereby promoting the improvement of corporate total factor productivity. Second, "Made in China 2025" has two opposite mechanisms for the total factor productivity of enterprises. On the one hand, incentive policies induce enterprises to increase meaningless or inefficient R&D investment and reduce the total factor productivity of enterprises. This partly stems from Adverse selection behaviors caused by information asymmetry between enterprises and governments; on the other hand, they guide enterprises to increase the number of effective invention patents, thereby increasing their total factor productivity. Third, compared with state-owned enterprises, the effect of policy implementation on the total factor productivity of non-state-owned enterprises is more obvious. Due to the inherent lack of political resources and derivative advantages of non-state-owned enterprises, they are more eager to get support from policies, and their response speed and degree to policies will be higher. Therefore, policies have a stronger marginal promotion effect on their corporate factor productivity. Fourth, compared with enterprises in low-market areas, the impact of policy implementation on the total factor productivity of enterprises in high-market areas is more obvious. In regions with a high degree of marketization, government policies will be more inclined to consider supporting companies with more competitive market advantages, so as to give full play to the role of policy resources in promoting total factor productivity. The results confirm the effect of intelligent manufacturing policy, and partially open the black box of its mechanism, which has important reference and enlightening significance for the formulation and implementation of intelligent manufacturing industry policy in the next stage.

强调自主研发创新的技术减排固然重要，但在资源环境约束加剧背景下，通过绿色技术转移（GTT）纠偏R&D要素配置扭曲以实现效率减排，可能更易达到事半功倍的效果。基于资源错配理论，剖析并检验R&D要素配置扭曲对碳排放的影响及GTT的纠偏作用机制。研究表明：（1）长三角城市的R&D人员与资金均存在配置扭曲现象，且人员错配程度高于资金；（2）R&D要素配置扭曲显著加剧了碳排放，资金错配的碳排放影响更大；（3）GTT对R&D要素配置扭曲所致的碳排放，具有显著调节效应与纠偏作用；（4）GTT “数量”和“质量”的发展水平，都可影响R&D要素配置扭曲与碳排放间的非线性门槛效应，目前长三角已可通过GTT数量增加的纠偏路径实现碳减排，而在质量提升方面的纠偏效能仅能对碳减排起到一定缓解作用。因此盘活存量技术、促进城际GTT的“增量提质”，充分发挥其对R&D要素配置的纠偏效能，是突破资源约束，提高创新能力，有效推进“双碳”目标实现的务实路径。

It is important to emphasize self-developed and innovative technologies to reduce carbon emissions, but in the context of intensified resource and environmental constraints, it would be more effective using Green Technology Transfer (GTT) to correct distorted allocation of R&D factors to achieve efficient carbon emissions reductions. Based on the resource misallocation theory, this paper analyzes and empirical tests the mechanism of GTT on R&D factors configuration distortion and CO2 emission. This study shows that: (1) R&D personnel and funds in the Yangtze River Delta cities are distorted, and the mismatch of personnel is higher than that of funds; (2) The distortion of R&D personnel and funds has significantly aggravated CO2 emissions, and distorted R&D funds has a greater impact on CO2 emissions; (3) The significant adjustment of GTT on distorted allocation of R&D factors is conducive to reduce CO2 emission; (4) Development in both GTT “quantity” and “quality” could affect the nonlinear threshold effect between distorted allocation of R&D factors and CO2 emissions. Cities in the Yangtze River Delta could achieve the suppression of CO2 emission through the correction path of “quantity increase”, while for “quality improvement”, the correction effect could only mitigate CO2 emission reduction to some extent. Thus, revitalizing the stock technology, promoting the “incremental quality improvement” of intercity GTT, and giving full play to its effectiveness in correcting allocation of R&D factors are the pragmatic path to break through resource constraints, improve innovation capabilities, and effectively promote the realization of the “dual carbon”.

在心理学和其他社科研究领域, 大量实证文章建立中介效应模型, 以分析自变量对因变量的影响过程和作用机制。检验中介效应最流行的方法是Baron和Kenny的逐步法, 但近年来不断受到批评和质疑, 有人甚至呼吁停止使用其中的依次检验, 改用目前普遍认为比较好的Bootstrap法直接检验系数乘积。本文对相关的议题做了辨析, 并讨论了中介分析中建立因果关系的方法。综合新近的研究成果, 总结出一个中介效应分析流程, 并分别给出显变量和潜变量Mplus程序。最后介绍了中介效应模型的发展。

<p>Mediation models are frequently used in the research of psychology and other social science disciplines. Mediation indicates that the effect of an independent variable on a dependent variable is transmitted through a third variable, which is called mediator. In most applied research, Baron and Kenny&rsquo;s (1986) causal steps approach has been used to test mediating effect. In recent years, however, many methodological researchers questioned the rationality of the causal steps approach, and some of them even attempted to stop its use. Firstly, we clarify the queries on the causal steps approach one by one. Secondly, we propose a new procedure to analyze mediating effects. The new procedure is better than any single method that constitutes the procedure in terms of Type I error rate and power. The proposed procedure can be conducted by using observed variables and/or latent variables. Mplus programs are supplied for the procedure with observed variables and/or latent variables. Finally, this article introduces the development of mediation models, such as mediation model of ordinal variables, multilevel mediation, multiple mediation, moderated mediation, and mediated moderation.</p>