Enterprise financing constraint influence factor analysis method and system
By building a dual-differential model, combining regulatory letters and reply text data, analyzing the influencing factors of corporate financing constraints, the problem of ignoring key information of regulatory letters in the existing technology is solved, and more accurate financing constraint analysis and strategic support is achieved.
Patent Information
- Application Number
- CN202411763044.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-03
- Publication Date
- 2025-01-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing technology ignores the key information in regulatory letters and reply letters when analyzing corporate financing constraints. There are limitations in data collection and processing, and it is impossible to accurately capture semantic information and contextual relationships, resulting in insufficient understanding of the deep reasons for financing constraints.
By building a dual-differential model, combining regulatory letters, reply text data, financial data and governance data, the KZ index and WW index are calculated, the reply deviation index and regulatory letter quality indicators are determined, and the supervision letter and reply deviation are analyzed. The impact of degree index and external institutional environment on financing constraints.
It realizes a more comprehensive and accurate analysis of corporate financing constraints, can deeply reveal deep reasons, provide support for the formulation of effective financing strategies, and improve analysis efficiency and reduce costs.
Smart Images
Figure CN119295224A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of enterprise management, and in particular relates to an analysis method and system for factors affecting enterprise financing constraints. Background Art
[0002] In the existing technology, the issue of corporate financing constraints has attracted much attention. With the strengthening of supervision, listed companies have received more and more regulatory letters involving many aspects. Current research methods mainly rely on financial data, governance data and market reactions to evaluate corporate financing constraints, but often ignore key information in regulatory letters and replies, which is crucial to understanding the deep-seated causes of financing constraints.
[0003] Specifically, existing methods usually collect basic information of enterprises and use statistical or econometric analysis to reveal financing constraints. Some studies construct complex models to predict the degree of financing constraints, but lack consideration of the actual operation and regulatory environment of enterprises. The existing technology has obvious defects: first, it ignores the key information in regulatory letters and replies; second, there are limitations in data collection and processing, such as the difficulty in obtaining data, missing data and outliers, and the simple text data processing method cannot accurately capture semantic information and contextual relationships, which limits its application in the analysis of corporate financing constraints. To this end, the present invention proposes a method and system for analyzing the influencing factors of corporate financing constraints. Summary of the invention
[0004] The purpose of the present invention is to provide an analysis method and system for factors affecting corporate financing constraints, so as to solve the problems raised in the background technology.
[0005] The present invention achieves the above-mentioned purpose through the following technical solutions: In a first aspect, the present invention proposes a method for analyzing factors affecting corporate financing constraints, the method comprising: S1. Set the listed companies that have received regulatory letters within the set period as target companies, obtain the text data of regulatory letters received by the target companies, the text data of the replies to the corresponding regulatory letters, the company's financial data and corporate governance data, and form a sample data set; S2. Calculate the KZ index and WW index of the target enterprise according to the sample data set, and use the KZ index and WW index as explained variables, use whether the target enterprise has received regulatory letters, the reply deviation index determined based on the reply text data, and the regulatory letter quality index as explanatory variables, use the external institutional environment as a moderating variable, and determine the first control variable and the second control variable according to the sample data set; S3. construct a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable, wherein the double difference model is used to analyze the relationship between whether the target enterprise receives regulatory letters and financing constraints, to analyze the relationship between the reply deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; S4. Select data corresponding to each of the variables in the sample data set and input them into the double difference model to obtain the interaction coefficient of each of the variables, and determine the influencing factors of the target enterprise's financing constraints based on the significance level of the interaction coefficient.
[0006] Furthermore, the step S1 also includes preprocessing the sample data set, specifically including: eliminating samples of financial companies; eliminating samples of companies with missing financial data and corporate governance data; and Winsorizing the upper and lower 1% quantiles of all continuous variables in the sample data set.
[0007] Furthermore, the reply deviation index determined based on the reply text data includes: (1) Text preprocessing: Perform text preprocessing on the regulatory letters and reply letters, including word segmentation, removal of stop words, and deletion of English letters, numbers, Chinese and English punctuation marks and separators; (2) Calculation of word frequency vectors: Calculate the word frequency vectors of each inquiry content in the regulatory letter and the word frequency vectors of the reply text data respectively; (3) Vector representation and similarity calculation: Use the pre-trained DistilRoBERTa model to represent the words in the text as high-dimensional vectors and calculate the soft cosine similarity between the query content and the reply text data; (4) Calculation of reply deviation index: The reply deviation index is calculated based on soft cosine similarity. The index is used to measure the degree to which the reply accurately focuses on the inquiry content. It is a continuous numerical variable between 0 and 1. The larger the index value, the greater the degree to which the reply deviates from the inquiry content and the worse the quality of the reply.
[0008] Furthermore, the quality index of the regulatory letters is determined as follows: The quality index of regulatory letters is determined based on the proxy variable determined by the number of reply words in the reply text data corresponding to the regulatory letters, the frequency measurement text of the numbers appearing in the reply text data, and the submission date of the reply text.
[0009] Furthermore, determining the first control variable and the second control variable based on the sample data set includes: taking enterprise size, return on assets, financial leverage ratio, accounts receivable turnover rate and asset turnover rate as the first control variable, and taking the number of directors, proportion of independent directors, management shareholding ratio, years of listing and institutional investors' shareholding ratio as the second control variable.
[0010] Furthermore, constructing a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable includes: Fc i,t =β0+β1Take i,t ×Clquality i,t +Controls i,t +Time t +Indu i +ε i,t Among them, Take i,t ×Clquality i,t Indicates the cross-product of whether the inquiry letter was received and the quality of the reply letter. i,t Indicates the first control variable or the second control variable, Time t Indicates the control year, Indu i represents the control industry, β0 represents the constant term, β1 represents the regression coefficient, ε i,t represents the disturbance term.
[0011] Furthermore, before selecting data corresponding to each of the variables in the sample data set and inputting them into the double difference model, the method further includes: subjecting the double difference model to a reliability test, wherein the reliability test includes an endogeneity test and a robustness test.
[0012] Furthermore, the determining of the target enterprise's financing constraint influencing factors based on the significance level of the interaction term coefficient includes: If the coefficient of the interaction between the target enterprise receiving regulatory letters and the enterprise's financing constraints is significantly positive, it is judged that the enterprise receiving regulatory letters has higher financing constraints; If the interaction coefficient between the target enterprise's reply deviation index and the enterprise's financing constraints is significantly positive, it is judged that the enterprise that receives the regulatory letter has lower financing constraints; If the coefficient of the interaction term between the external institutional environment and corporate financing constraints is significantly positive, it means that the level of institutional strictness of the external environment has a positive regulatory effect on the target enterprise's receipt of regulatory letters and the target enterprise's reply deviation index has a positive regulatory effect on corporate financing constraints.
[0013] In a second aspect, the present invention proposes an analysis system for factors affecting corporate financing constraints, which is applied to execute any of the above analysis methods, and the system comprises: A data acquisition module is configured to set listed companies that have received regulatory letters within a set period as target companies, acquire text data of regulatory letters received by the target companies, text data of replies to corresponding regulatory letters, company financial data, and corporate governance data, to form a sample data set; a variable determination module, configured to calculate the KZ index and the WW index of the target enterprise according to the sample data set, and use the KZ index and the WW index as explained variables, use whether the target enterprise has received a regulatory letter, a reply deviation index determined based on the reply text data, and a regulatory letter quality index as explanatory variables, use the external institutional environment as a moderating variable, and determine a first control variable and a second control variable according to the sample data set; A model building module is configured to build a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable, wherein the double difference model is used to analyze the relationship between whether the target enterprise receives regulatory letters and financing constraints, to analyze the relationship between the reply letter deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; The result analysis module is configured to select data corresponding to each of the variables in the sample data set and input the data into the double difference model to obtain the interaction coefficient of each of the variables, and determine the influencing factor of the target enterprise's financing constraints based on the significance level of the interaction coefficient.
[0014] The beneficial effects of the present invention are: The present invention combines the financial data and governance data of the enterprise to construct a double difference model to deeply analyze the influencing factors of the enterprise's financing constraints. Compared with the existing methods, the model is more comprehensive and accurate, and can comprehensively consider multiple aspects such as the company's financial status, governance structure, business strategy, and regulatory environment. Through the analysis of the double difference model, the present invention can more accurately reveal the deep-seated reasons for the enterprise's financing constraints and provide strong support for the formulation of effective financing strategies. In addition, the method and system are also intelligent and automated, which can improve analysis efficiency, reduce analysis costs, and make positive contributions to the stability and development of the financial market. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 A schematic diagram of a flow chart of a method for analyzing factors affecting corporate financing constraints provided in an embodiment of the present application; Figure 2 Another system architecture diagram of the system for analyzing factors affecting enterprise financing constraints provided in an embodiment of the present application. DETAILED DESCRIPTION
[0017] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.
[0018] Example 1
[0019] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows: Traditional solutions mostly focus on the impact of inquiry letters received by enterprises on financing constraints. As an important means of capital market supervision, exchange inquiry letters not only require enterprises to disclose information on specific matters, but also evaluate the adequacy, accuracy and timeliness of their information disclosure through the quality of the enterprise's reply. The quality of the enterprise's reply not only reflects the degree of attention the enterprise pays to regulatory requirements, but also reflects its internal governance level and the soundness of its information disclosure system. Therefore, measuring the impact of inquiry letters on corporate financing constraints from the perspective of the quality of corporate replies has important theoretical significance and practical value. High-quality replies mean that enterprises can respond to regulatory requirements fully, accurately and promptly, reduce information asymmetry, and enhance investor confidence, which is conducive to reducing the financing costs and difficulties of enterprises. On the contrary, low-quality replies may aggravate information asymmetry, damage corporate reputation, and further aggravate corporate financing constraints. Therefore, through the impact of empirical analysis, it can reveal the important role of information disclosure quality in alleviating corporate financing constraints.
[0020] Based on this, this application takes the exchange's annual report inquiry letter (regulatory letter) as the perspective, empirically examines the impact of the annual report inquiry letter on corporate financing constraints, and further examines whether the quality of the reply to the annual report inquiry letter has an inhibitory effect on financing constraints.
[0021] like Figure 1 As shown, this embodiment proposes a method for analyzing the factors affecting corporate financing constraints, and the method includes: S1. Set the listed companies that receive regulatory letters within the set period as target companies, obtain the text data of regulatory letters received by the target companies, the text data of the replies to the corresponding regulatory letters, the company's financial data and corporate governance data, and form a sample data set; regulatory letters specifically refer to various letters issued by stock exchanges, securities regulatory commissions, etc., such as regulatory letters, inquiry letters, and concern letters.
[0022] S2. Calculate the KZ index and WW index of the target enterprise based on the sample data set, and use the KZ index and WW index as explained variables, whether the target enterprise has received regulatory letters, the reply deviation index determined based on the reply text data, and the regulatory letter quality index as explanatory variables, and use the external institutional environment as a moderating variable. Determine the first control variable and the second control variable based on the sample data set.
[0023] The KZ index is used as a measure of corporate financing constraints; the larger the KZ index, the higher the degree of financing constraints faced by listed companies and the lower the financing efficiency; the WW index is an index to measure corporate financing constraints. The nonlinear GMM method is used to estimate the parameters of the Euler equation to obtain the financing constraint value at the corporate level. More specifically, the variables are set as follows: Explained variables: financing constraints. In this example, the measurement indicators include KZ index and WW index. Explanatory variables: whether the target enterprise received regulatory letters (1 if received, 0 otherwise), the reply deviation index determined based on the reply text data, and the regulatory letter quality index; Moderating variable: external institutional environment. The level of external institutional environment is measured by the market intermediary development and legal system environment index in the enterprise's "China Provincial Marketization Index Report".
[0024] It should be noted that the setting of the first control variable and the second control variable in this application can help us to have a deeper understanding of the influencing mechanism of corporate financing constraints. These control variables cover a number of important factors at the corporate financial level, corporate characteristics and management level, such as corporate scale, debt-to-asset ratio, profitability, governance structure, etc. By incorporating these factors into the regression model (double difference model), the causes of corporate financing constraints can be analyzed more comprehensively, and strong support can be provided for the formulation of effective financing strategies.
[0025] S3. A double difference model is constructed based on the explained variable, the explanatory variable, the first control variable and the second control variable. The double difference model is used to analyze the relationship between whether the target enterprise has received regulatory letters and financing constraints, to analyze the relationship between the reply deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; S4. Select the data corresponding to each variable in the sample data set and input it into the double difference model to obtain the interaction coefficient of each variable. Based on the significance level of the interaction coefficient, determine the influencing factors of the target enterprise's financing constraints.
[0026] The analysis method in this embodiment can make full use of the key information in the regulatory letters and their reply texts, which are often ignored by existing methods but are crucial for understanding the deep-seated reasons for corporate financing constraints. By combining the financial data and governance data of the enterprise, the present invention constructs a double difference model to deeply analyze the influencing factors of corporate financing constraints.
[0027] Further preferably, step S1 also includes preprocessing the sample data set, specifically including: eliminating samples of financial companies; eliminating samples of companies with missing financial data and corporate governance data; and Winsorizing (truncation) the upper and lower 1% quantiles of all continuous variables in the sample data set.
[0028] Further preferably, the reply deviation index determined based on the reply text data includes: (1) Text preprocessing: Perform text preprocessing on regulatory letters and reply letters, including word segmentation, removal of stop words, and deletion of English letters, numbers, Chinese and English punctuation marks and separators; (2) Calculation of word frequency vectors: Calculate the word frequency vectors of each inquiry content in the regulatory letter and the word frequency vectors of the reply text data respectively; (3) Vector representation and similarity calculation: Use the pre-trained DistilRoBERTa model to represent the words in the text as high-dimensional vectors and calculate the soft cosine similarity between the query content and the reply text data; (4) Calculation of reply deviation index: The reply deviation index is calculated based on soft cosine similarity. The index is used to measure the degree to which the reply accurately focuses on the inquiry content. It is a continuous numerical variable between 0 and 1. The larger the index value, the greater the degree to which the reply deviates from the inquiry content and the worse the quality of the reply.
[0029] As an example, first, after text preprocessing, the word frequency vector of each inquiry content (Question) and the word frequency vector of the company's reply content (reply text) (Answer) are calculated respectively: Question=(q1,q2,…,qn) Answer=(a1,a2…,an) Among them, n is the number of non-repeated words appearing in the annual report inquiry letter and reply letter, and the elements qi and ai in the vector are the frequency of word i appearing in the document.
[0030] Again, the DistilRoBERTa model is used to represent the words in the corpus as 100-dimensional vectors, and the soft cosine similarity function is used to calculate the soft cosine similarity metric Softcosinesimilarity between the "inquiry content" and the "company response content".
[0031] Finally, by subtracting the soft cosine similarity ratio from 1, we get the “irrelevant answer” index (Irrelevant) of this article, that is, Irrelevant=1-Softcosinesimilarity.
[0032] Further preferably, the quality index of regulatory letters is determined as follows: The quality index of regulatory letters is determined based on the proxy variable determined by the number of words in the reply text data corresponding to the regulatory letters, the frequency measurement text of the numbers in the reply text data, and the submission date of the reply text; more specifically: Response word count: (1) Calculation of the number of questions: Number of questions = len(re.findall(r'\d、', question)) If the number of questions is 0, set it to 1: Number of questions = max(number of questions, 1) (2) Calculation of the number of words in the reply: Reply word count = len(answer) (3) Calculation of average number of question responses: Average number of words in response to a question = number of words in response / number of questions (4) Specific naturalization: Specificity = {average number of words in the response to a question - min(average number of words in the response to a question)} / {max(average number of words in the response to a question) - min(average number of words in the response to a question)} Calculate the frequency of numbers appearing in the reply text: .Answer the number of digits: (1). Number of answer digits = len(re.findall(r'\d + d +|\d+',answer)) (2) Accuracy normalization: Accuracy = {number of answers - min(number of answers)} / {max(number of answers) - min(number of answers)} Calculation of reply text date: (1) Countdown calculation of timeliness response: If the timeliness response is 0, set it to 1: Timely response = max(timely response, 1) Timely response situation inverse = 1 / timely response situation (2) Timely naturalization: Timeliness = the inverse of the timeliness response - min (the inverse of the timeliness response) / max (the inverse of the timeliness response) - min (the inverse of the timeliness response) (3).Indicator calculation: Indicator = (Specificity 2 + Accuracy 2 +Timeliness 2 ) / (Specificity + Accuracy + Timeliness) Further preferably, the first control variable and the second control variable are determined based on the sample data set, including: taking the enterprise financial level (enterprise size, return on assets, financial leverage ratio, accounts receivable turnover rate and asset turnover rate) as the first control variable, and taking the enterprise characteristics and management level (number of directors, proportion of independent directors, shareholding ratio of management, years of listing and shareholding ratio of institutional investors) as the second control variable.
[0033] Further preferably, a double difference model is constructed based on the explained variable, the explanatory variable, the first control variable and the second control variable, including: Fc i,t =β0+β1Take i,t ×Clquality i,t +Controls i,t +Time t +Indu i +ε i,t Among them, Take i,t ×Clquality i,t Indicates the cross-product of whether the inquiry letter was received and the quality of the reply letter. i,t Indicates the first control variable or the second control variable, Time t Indicates the control year, Indu i represents the control industry, β0 represents the constant term, β1 represents the regression coefficient, ε i,t represents the disturbance term.
[0034] Further preferably, before selecting data corresponding to each variable in the sample data set and inputting them into the double difference model, it also includes: subjecting the double difference model to a reliability test, the reliability test including an endogeneity test and a robustness test.
[0035] Further preferably, the influencing factors of the target enterprise's financing constraints are determined based on the significance level of the interaction coefficient, including: If the coefficient of the interaction between the target enterprise receiving regulatory letters and the enterprise's financing constraints is significantly positive, it is judged that the enterprise receiving regulatory letters has higher financing constraints; If the interaction coefficient between the target enterprise's reply deviation index and the enterprise's financing constraints is significantly positive, it is judged that the enterprise that receives the regulatory letter has lower financing constraints; If the coefficient of the interaction term between the external institutional environment and corporate financing constraints is significantly positive, it means that the level of institutional strictness of the external environment has a positive regulatory effect on the target enterprise's receipt of regulatory letters and the target enterprise's reply deviation index has a positive regulatory effect on corporate financing constraints.
[0036] According to the above embodiment of the present invention, a double difference model or other statistical model is constructed by combining the key information in the regulatory letters and their reply texts with other corporate financial data, governance data, etc. In these models, the interaction term coefficient reflects the interaction relationship between different variables. Secondly, to determine the significance level of the interaction term coefficient, calculate the t value or z value of the interaction term coefficient and compare it with the critical value. If the calculated value is greater than the critical value, it indicates that the interaction term coefficient is statistically significant, that is, its impact on the corporate financing constraints is significant. Finally, based on the significance level of the interaction term coefficient, it can be determined which factors have a significant impact on corporate financing constraints and which factors have no significant impact. This helps investors and regulators to more accurately understand the causes of corporate financing constraints.
[0037] like Figure 2 As shown, this embodiment also proposes an analysis system for the factors affecting corporate financing constraints, which is applied to execute the above analysis method. The system includes: The data acquisition module is configured to set the listed companies that have received regulatory letters within a set period as target companies, obtain the text data of the regulatory letters received by the target companies, the text data of the replies to the corresponding regulatory letters, the company's financial data and corporate governance data, and form a sample data set; a variable determination module, configured to calculate the KZ index and the WW index of the target enterprise according to the sample data set, and use the KZ index and the WW index as explained variables, use whether the target enterprise has received regulatory letters, the reply deviation index determined based on the reply text data, and the regulatory letter quality index as explanatory variables, use the external institutional environment as a moderating variable, and determine the first control variable and the second control variable according to the sample data set; A model building module is configured to build a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable, wherein the double difference model is used to analyze the relationship between whether the target enterprise has received regulatory letters and financing constraints, to analyze the relationship between the reply deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; The result analysis module is configured to select data corresponding to each variable in the sample data set and input it into the double difference model to obtain the interaction coefficient of each variable, and determine the influencing factors of the target enterprise's financing constraints based on the significance level of the interaction coefficient.
[0038] It should be noted here that each module in the above analysis system corresponds to steps S1 to S4 in implementing the above analysis method, and the instances and application scenarios implemented by multiple modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment 1.
[0039] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0040] In addition, each functional module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0041] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in the embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.
[0042] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for analyzing factors affecting corporate financing constraints, characterized in that: The method comprises: S1. Set the listed companies that have received regulatory letters within the set period as target companies, obtain the text data of regulatory letters received by the target companies, the text data of the replies to the corresponding regulatory letters, the company's financial data and corporate governance data, and form a sample data set; S2. Calculate the KZ index and WW index of the target enterprise according to the sample data set, and use the KZ index and WW index as explained variables, use whether the target enterprise has received regulatory letters, the reply deviation index determined based on the reply text data, and the regulatory letter quality index as explanatory variables, use the external institutional environment as a moderating variable, and determine the first control variable and the second control variable according to the sample data set; S3. construct a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable, wherein the double difference model is used to analyze the relationship between whether the target enterprise receives regulatory letters and financing constraints, to analyze the relationship between the reply deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; S4. Select data corresponding to each of the variables in the sample data set and input them into the double difference model to obtain the interaction coefficient of each of the variables, and determine the influencing factors of the target enterprise's financing constraints based on the significance level of the interaction coefficient.
2. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The step S1 also includes preprocessing the sample data set, specifically including: removing samples of financial companies; removing samples of companies with missing financial data and corporate governance data; and Winsorizing the upper and lower 1% quantiles of all continuous variables in the sample data set.
3. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The reply deviation index determined based on the reply text data includes: (1) Text preprocessing: Perform text preprocessing on the regulatory letters and reply letters, including word segmentation, removal of stop words, and deletion of English letters, numbers, Chinese and English punctuation marks and separators; (2) Calculation of word frequency vectors: Calculate the word frequency vectors of each inquiry content in the regulatory letter and the word frequency vectors of the reply text data respectively; (3) Vector representation and similarity calculation: Use the pre-trained DistilRoBERTa model to represent the words in the text as high-dimensional vectors and calculate the soft cosine similarity between the query content and the reply text data; (4) Calculation of reply deviation index: The reply deviation index is calculated based on soft cosine similarity. The index is used to measure the degree to which the reply accurately focuses on the inquiry content. It is a continuous numerical variable between 0 and 1. The larger the index value, the greater the degree to which the reply deviates from the inquiry content and the worse the quality of the reply.
4. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The quality indicators of regulatory correspondence are determined as follows: The quality index of regulatory letters is determined based on the proxy variable determined by the number of reply words in the reply text data corresponding to the regulatory letters, the frequency measurement text of the numbers appearing in the reply text data, and the submission date of the reply text.
5. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The determining of the first control variable and the second control variable based on the sample data set includes: taking enterprise size, return on assets, financial leverage ratio, accounts receivable turnover rate and asset turnover rate as the first control variable, and taking the number of directors, proportion of independent directors, shareholding ratio of management, years of listing and shareholding ratio of institutional investors as the second control variable.
6. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The constructing of a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable includes: Fc i,t =β0+β1Take i,t ×Clquality i,t +Controls i,t +Time t +Indu i +e i,t Among them, Take i,t ×Clquality i,t Indicates the cross-product of whether the inquiry letter was received and the quality of the reply letter. i,t Indicates the first control variable or the second control variable, Time t Indicates the control year, Indu i represents the control industry, β0 represents the constant term, β1 represents the regression coefficient, ε i,t represents the disturbance term.
7. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: Before selecting data corresponding to each of the variables in the sample data set and inputting them into the double difference model, the method further includes: subjecting the double difference model to a reliability test, wherein the reliability test includes an endogeneity test and a robustness test.
8. The method for analyzing the factors affecting corporate financing constraints according to claim 1, characterized in that: The determining of the target enterprise's financing constraint influencing factor based on the significance level of the interaction term coefficient includes: If the coefficient of the interaction between the target enterprise receiving regulatory letters and the enterprise's financing constraints is significantly positive, it is judged that the enterprise receiving regulatory letters has higher financing constraints; If the interaction coefficient between the target enterprise's reply deviation index and the enterprise's financing constraints is significantly positive, it is judged that the enterprise that receives the regulatory letter has lower financing constraints; If the coefficient of the interaction term between the external institutional environment and corporate financing constraints is significantly positive, it means that the level of institutional strictness of the external environment has a positive regulatory effect on the target enterprise's receipt of regulatory letters and the target enterprise's reply deviation index has a positive regulatory effect on corporate financing constraints.
9. An analysis system for factors affecting corporate financing constraints, characterized in that: Applied to perform the analysis method according to any one of claims 1 to 8, the system comprises: A data acquisition module is configured to set listed companies that have received regulatory letters within a set period as target companies, acquire text data of regulatory letters received by the target companies, text data of replies to corresponding regulatory letters, company financial data, and corporate governance data, to form a sample data set; a variable determination module, configured to calculate the KZ index and the WW index of the target enterprise according to the sample data set, and use the KZ index and the WW index as explained variables, use whether the target enterprise has received a regulatory letter, a reply deviation index determined based on the reply text data, and a regulatory letter quality index as explanatory variables, use the external institutional environment as a moderating variable, and determine a first control variable and a second control variable according to the sample data set; A model building module is configured to build a double difference model based on the explained variable, the explanatory variable, the first control variable and the second control variable, wherein the double difference model is used to analyze the relationship between whether the target enterprise receives regulatory letters and financing constraints, to analyze the relationship between the reply letter deviation index and financing constraints, and to analyze the relationship between the external institutional environment and financing constraints; The result analysis module is configured to select data corresponding to each of the variables in the sample data set and input the data into the double difference model to obtain the interaction coefficient of each of the variables, and determine the influencing factor of the target enterprise's financing constraints based on the significance level of the interaction coefficient.