"AI and CSCM"

"The dataset includes independent variables (AI), dependent variables (CSCM), moderating variables (EME, SCME, CP, and MP), and control variables. Dependent variable: Circular Supply Chain Management (CSCM)From the ESG sub-database of CNRDS to measure the adoption of CSCM, this paper selects eight indicators related to corporate ecological advantages. The relevant indicators include circular economy, measures to reduce three types of waste, energy conservation and emission reduction, green office, environment-friendly products, ISO 14001 certification, environmental recognition, and other strengths. If a firm engages in a behavior that reflects ecological advantages, the corresponding indicator is assigned a value of 1; otherwise, it is 0. To assess the firm\u2019s adoption of CSCM, the values of these eight indicators are summed.Independent variable: Artificial Intelligence TechnologyConstructing AI Indicators Based on Corporate Patents. First, manually selecting AI patents and their CPC classification codes based on the \u201cPatents and the Fourth Industrial Revolution\u201d which was published by the EPO. Then, transferring CPC classification codes into IPC classification codes. Lastly, calculating the annual total number of AI patents enterprises apply to measure their AI technology (AI).Building AI Indicators Based on Corporate Annual Reports. Also, the machine learning method is used to construct AI technology indicators for subsequent robustness checks: First, crawling the annual reports of listed companies from the Cninfo website. Second, referring to the AI vocabulary provided by WIPO and Chen and Srinivasan, select 52 seed words. Third, using Word2vec and Skip-gram model to select the 10 most similar words for each seed word. Next, remove duplicate words, words with low frequency, or those not directly related to AI to get 73 words in a final AI dictionary. Last, incorporating the AI dictionary into the Jieba segmentation module and counting the occurrences of AI-related words in the annual reports. Use the logarithm of this count plus one to measure AI (AIcp).Moderating variablesEnvironmental Management Experience of TMTs (EME). If the executive\u2019s resume contains keywords like \u201cenvironmental protection\u201d or \u201csustainability,\u201d the value is 1; otherwise, it is 0. Then, calculate the proportion of executives with environmental management experience in TMTs, denoted as EME.Supply Chain Management Experience of TMTs (SCME). If the executive\u2019s resume includes keywords such as \u201csupply\u201d or \u201cprocurement,\u201d the value is assigned as 1; otherwise, it is 0. The proportion of executives with supply chain management experience in the TMTs is calculated to measure SCME.Coercive Pressure (CP). Measured by the proportion of environment-related word frequencies in the text of Chinese local government reports.Mimetic Pressure (MP). Excluding the focus firm, use the average CSCM level of other enterprises within the same industry as the measure.Control variablesFirm governance variables, firm characteristics, and external resource variables are included as control variables. These variables are measured as follows: (1) First shareholder ownership (fir), measured by proportion of shares held by the largest shareholder to total outstanding shares. (2) Board independence (ine), measured by the ratio of independent directors to total board members. (3) Return on equity (roe) is measured as net profit divided by net assets. (4) State ownership (soe), a dummy variable coded as 1 for state-owned enterprises and 0 otherwise. (5) Book-to-market ratio (bm) is measured as total assets divided by market value. (6) Financial leverage (lev) is measured as long-term debt divided by total assets. (7) Government subsidies (sub) are measured as the ratio of government grants to operating income."