Innovation policies have long been a focal point of academic research, as they reshape enterprises’ competitive landscapes, drive industrial advancement, and have a profound impact on national economic growth (34). The existing literature provides robust evidence that innovation policies positively impact enterprise performance across diverse sectors and institutional settings (35). Within China’s transition toward an innovation-driven economy, the government has implemented a suite of innovation policies. Including R&D subsidies, tax incentives, intellectual property protection measures, and targeted regulatory reforms—aimed at enhancing enterprises’ innovation capabilities (36). These policies are widely acknowledged as critical drivers of increased innovation outputs, such as patent applications, new product launches, and technological breakthroughs, which ultimately translate into improved overall enterprise performance (37). R&D subsidy policies, as a form of government intervention, provide financial support to enterprises engaged in innovative activities. By mitigating the financial constraints associated with high-risk, long-term R&D projects—endeavors often plagued by information asymmetry and positive externalities that deter private investment—these subsidies enable enterprises to allocate more resources to research and development (38). This includes investments in advanced equipment, recruitment of specialized R&D personnel, and execution of extensive clinical trials or experimental studies. Consequently, enterprises can develop high-value-added new products, optimize production processes to reduce costs, and strengthen market competitiveness through technological differentiation, ultimately leading to improved financial performance, including higher revenues, enhanced profit margins, and increased market valuation (39). Tax incentives play a pivotal role in alleviating the tax burden on enterprises involved in innovative activities. These measures encompass super-deductions for R&D expenses, preferential tax rates for high-tech enterprises, and tax holidays for technology-driven startups. By effectively boosting post-tax profits, these incentives equip enterprises with additional resources that can be reinvested in R&D and other innovation-related initiatives (40). This influx of financial resources enables enterprises to attract top talent through competitive salaries and research grants, invest in cutting-edge technologies and state-of-the-art laboratory facilities, and conduct in-depth research on emerging technologies and market demands. Over the long term, these efforts contribute to enhanced innovation capabilities and improved enterprise performance. Within the context of the TCM industry, all TCM products in China must undergo a registration process before entering the market, with sales profits directly influencing an enterprise’s performance (41). This makes the investigation of the interplay between TCMRIPs and enterprise performance particularly crucial. Building on existing evidence and considering the unique context of TCM enterprises, we propose Hypothesis 1(a): TCMRIPs positively influence TCM enterprise performance. Specifically, TCMRIPs significantly enhance the performance of listed TCM enterprises in China. Notably, the impact of innovation policies on enterprise performance may not be immediate (36). Policy implementation typically involves a complex process, requiring enterprises to adjust their strategies, reorganize internal resources, and develop new capabilities to align with evolving policy frameworks (42). For example, enterprises may need to establish dedicated R&D teams to leverage policy incentives, collaborate with research institutions to meet new regulatory standards, or revise product development pipelines to align with policy objectives. Additionally, the outcomes of innovation activities stimulated by these policies, such as the development of new TCM products. This may take time to materialize, due to inherent uncertainties and lengthy processes in R&D, clinical trials, and market penetration (43). This delay stems from the time required for R&D investments to yield tangible results, for new products to complete regulatory approval, and for market acceptance to develop. In the TCM sector, this lag may be more pronounced due to the extensive documentation required for traditional usage, complex quality control standards for natural ingredients, and the gradual accumulation of clinical evidence supporting holistic therapeutic effects. Thus, we propose Hypothesis 1(b): The impact of TCMRIPs demonstrates a time lag, with their beneficial effects on enterprise performance becoming more pronounced over time and peaking after a specific duration.

The interplay between policies, R&D investment, and enterprise performance has been extensively explored in academic research, with a growing body of evidence supporting the view that R&D investment acts as a critical mechanism through which innovation policies exert an impact on firm-level outcomes (44). Existing literature consistently demonstrates that well-designed policies can incentivize enterprises to augment their R&D expenditures, which in turn drives performance improvements, thereby forming a sequential chain of influence: policy → R&D investment → performance (45). Innovation policies, including R&D tax incentives and government-funded R&D subsidies, have been empirically validated as practical tools for boosting R&D investment across diverse industrial sectors. R&D tax incentives, such as tax credits, function by reducing the marginal cost of R&D activities for enterprises (46). When companies can benefit from lower tax rates on R&D expenditures or can deduct a specified percentage of their R&D costs from taxable income, the net cost of each additional R&D project diminishes. This creates a strong economic incentive for firms to allocate a greater share of resources to R&D initiatives, as the post-tax return on such investments becomes more appealing (47). On the other hand, government R&D subsidies provide direct financial support for research and development projects, spanning the entire innovation lifecycle from early-stage basic research to late-stage commercialization efforts. These subsidies play a pivotal role in alleviating financial constraints, particularly for small and medium-sized enterprises that often face challenges in accessing external capital markets (48). For TCM enterprises, investments in R&D domains, such as standardizing quality control protocols or conducting evidence-based clinical research, can also strengthen brand reputation and foster consumer trust, which in turn drives further performance improvements. Based on this conceptualization of the relationship between policy, R&D investment, and performance, we propose Hypothesis 2: TCMRIPs affect enterprise performance through the proportion of R&D investment, with R&D investment serving as a mediating variable in this relationship.

Policy effectiveness is not static; it evolves through revisions that reflect shifting industry demands, regulatory lessons, and societal expectations (49). In 2020, China issued the updated Drug Registration Administration Measures, representing a significant milestone in the nation’s drug regulatory framework. A comparison of the 2020 version with earlier iterations, such as the 2005 and 2007 versions, reveals that the new edition incorporates more targeted, refined, and industry-adapted policy measures, particularly regarding TCM (50). These enhancements are expected to enhance its effectiveness in improving enterprise performance. Firstly, the 2020 version integrates the achievements of reforms in the drug review and approval system over the past decade. It prioritizes innovative drugs and establishes a more science-driven evaluation system. The updated version clarifies the basic requirements for drug registration management, outlining the fundamental systems, principles, procedures, and key responsibilities of all involved parties in the process. This heightened transparency and predictability in the regulatory framework reduce uncertainties for enterprises, enabling them to plan their R&D and registration strategies more effectively and minimizing the risk of costly delays or rejections. Secondly, the 2020 version explicitly defines China’s TCM development path as focusing on “inheritance and innovation.” This dual emphasis addresses the unique needs of the TCM industry. By highlighting inheritance, the new version recognizes the value of traditional knowledge and classic formulas, simplifying registration requirements for products with well-documented traditional use. Simultaneously, it encourages innovation by promoting the application of modern scientific methods to validate the efficacy of TCM, develop new formulations, and expand clinical applications. This balanced approach provides enterprises with clear guidance on preserving traditional strengths while pursuing technological progress, thereby fostering a more supportive environment for R&D investment. Lastly, the classification of TCM registration has been significantly streamlined from nine categories to four, alleviating administrative burdens and optimizing the registration process. The previous classification system was criticized for its excessive complexity, featuring overlapping categories and ambiguous criteria that increased the time and costs associated with registration for TCM enterprises. The new classification—comprising innovative TCM, improved TCM, classic TCM prescriptions, and imported TCM—better aligns with the actual development needs of the industry, making it easier for enterprises to navigate the registration process and bring new products to market more rapidly. By reducing administrative barriers, aligning with industry characteristics, and providing clearer incentives for R&D and registration, the 2020 version is expected to create a more favorable policy environment that accelerates the translation of R&D efforts into market success. Therefore, we propose Hypothesis 3: The newly revised Drug Registration Administration Measures are more effective than previous versions in enhancing the performance of TCM listed enterprises.

TCM registration, review, and approval system operates under a supervisory framework composed of relevant laws and registration measures. The Drug Registration Administration Measures, first issued in 2002, have been updated three times, specifically in 2005, 2007, and 2020. Against this background, this study selects all A-share listed TCM enterprises in Shanghai and Shenzhen stock exchanges from 2002 to 2024 as the original sample. To ensure the accuracy and completeness of the conclusions, the original samples in this study were processed as follows: (1) Excluding ST/PT enterprises, as these entities typically adopt bankruptcy accounting standards, rendering them incomparable with other listed companies. Specific elimination steps: If the enterprise data in the dataset belongs to ST/PT in the current year, then in the downloaded data, the enterprise name will have the prefix “ST/PT” before it. Such data with the “ST/PT” prefix needs to be excluded. (2) Screening to include only enterprises specializing in proprietary Chinese medicines and TCM decoction pieces, while excluding those primarily engaged in TCM materials. This is because the TCM materials cannot be sold directly as products. They need to be further processed into Chinese herbal decoctions or Chinese patent medicines before they can be sold. (3) Excluding enterprises that received special treatment in the corresponding year (Contains missing values). (4) Perform winsorization at the 1% level to mitigate the impact of outliers. Enterprise-level data were sourced from the CSMAR database and the annual reports of individual companies (released at the end of each fiscal year). In this study, the relevant policies on the review and evaluation of all TCMRAPs issued by the Chinese government/NMPA are reviewed. The time period selected for this article is from 1 January 2002, to 31 December 2024. Policy texts were mainly retrieved and obtained from the official portal website of China’s central government,¹ the Law and Regulation Database of Peking University,² the China Food and Drug Administration),³ China National Knowledge Infrastructure,⁴ and other common retrieval platforms. In addition, relevant supplementary documents were searched from Baidu, Google, Bing, and other websites. The selected policy is related to the registration, review, and approval policy of TCM. In addition to “TCM registration,” “TCM review,” “TCM approval,” “New medicine review,” “Drug registration,” the search scope was also expanded by entering keywords such as “TCM innovation” to ensure a comprehensive and careful search. The types of policies retrieved included laws, ordinances, methods, rules and requirements, opinions, announcements, guiding principles, and notifications. Data analysis was conducted using Stata 18.0.

Table 2 presents the descriptive statistics of the main variables in this study. The results indicate that the sample consists of 1,406 observations with no missing values for any variable. For the dependent Variable ROA, the Mean (0.067) and Median (0.066) are close, suggesting a relatively symmetric distribution of ROA. This indicates that the overall asset return level of enterprises is stable, with most firms exhibiting comprehensive asset utilization effects close to the average. Standard deviation (SD = 0.069) reflects a moderate degree of dispersion, indicating certain differences in asset returns among enterprises without extreme polarization. Extremes (Min = −0.549; Max = 0.355) means the negative minimum value indicates that some enterprises have negative asset returns (potentially facing operational difficulties or asset impairments), while the maximum value of 0.355 reflects high asset operational efficiency and prominent profitability in some enterprises. For Independent Variable Pol: The Median is close to the Mean, indicating an approximately symmetric distribution. Most observations show innovation-oriented word frequencies in TCM policies clustered around the Mean, with no significant skewness. A small standard deviation suggests low dispersion in Pol across the sample, reflecting either consistent emphasis on innovation in TCM registration policies from 2002 to 2024 (with minor annual/sample differences) or relatively balanced “exposure” to innovation-oriented policy terms among the sampled enterprises. For the First Control variable, Size, the Mean (21.356) and Median (21.404) are close, indicating a symmetric distribution of firm size. Most enterprises cluster around the 21.3–21.4 range, with no significant skewness in size differences within the industry. For the second Control variable, Age: Mean (17.312) and Median (17.000) are close, reflecting a symmetric distribution of firm age. The average establishment duration of sample enterprises is approximately 17 years, with most operating for around 17 years, indicating relative maturity in the industry. For the last Control Variables Lev: Mean (0.362) is slightly higher than the Median (0.329), suggesting a mild right-skewed distribution. This indicates the presence of some enterprises with high asset-liability ratios, which inflate the Mean; overall, the liability levels of sample enterprises remain within a reasonable range. All variables are deemed suitable for further empirical analysis.

As shown in Table 3, the correlation coefficient between ROA and Pol is significant at the 1% level with a positive sign, preliminarily validating Hypothesis 1(a). Specifically, a higher frequency of innovation-oriented terms in TCM policies (indicating stronger policy emphasis on innovation) is associated with a higher ROA. This may be explained by innovation-oriented policies that encourage the R&D of new TCM products and optimize registration processes, thereby enhancing enterprises’ product competitiveness and accelerating commercialization. Alternatively, policy resources may be allocated toward innovative enterprises, reducing operational costs and improving profitability. All control variables except Lev exhibit significant correlations with ROA at the 1% level. The non-significant correlation for Lev may stem from the complex relationship between capital structure and profitability in TCM enterprises. High-leverage firms face financial risks but may expand production, R&D, and other activities through debt financing, while low-leverage firms maintain stability but may be constrained by insufficient funds. These offsetting effects could weaken the linear relationship. Numerous domestic and foreign studies have confirmed that Lev has a certain impact on ROA. Therefore, in this study, this variable is still included in the control variables. Multiple covariance analysis (variance inflation factor, VIF) was conducted for all variables. As shown in Table 4, the VIF values of all independent variables are less than 10, indicating that there are no severe multiple covariance issues among the variables.

To avoid potential bidirectional causality between ROA and Pol, this study conducts an endogeneity test using lagged periods of the explained variable with model (4). Specifically, while Pol in a given year may affect the current year’s ROA (as indicated by the strong positive correlation in the regression equation), it is also possible that the current year’s ROA influences the current year’s Pol. Therefore, a forward-looking regression analysis was conducted on the ROA, given that ROA in a given year cannot affect Pol in the previous year. As shown in Table 10, a 0–3 period forward-looking regression analysis was conducted on the ROA. The results indicate that, except for the significant correlation between the current year’s ROA and Pol in Column (1), ROA fronted by 1–3 periods in Column (2–4) shows no correlation with Pol. This confirms that there is no strong endogeneity among the variables selected in this study. However, Companies with higher levels of profitability can invest more in R&D, regardless of policy. Therefore, in this paper, a two-stage system GMM is also used to recheck and verify the correlation between the dependent variable and the independent variables. The GMM uses lagged 1-period and 2-period ROA as proxies to construct the dynamic panel, and uses the lagged period polynomial as an instrument variable. As shown in Table 11, AR(1) (0.000) and AR(2) (0.127) indicate that the first-order difference exhibits autocorrelation, while the second-order difference does not. Therefore, the GMM estimation is a practical approach. The Hansen J test statistic (0.238) is not significant, indicating that the model does not suffer from over-identification, the instrumental variables are valid, and the model set-up is reasonable. The results show that, after considering the possible endogenous problems, the promotion effect of TCMRIPs on the performance increase of TCM listed enterprises remains established, and the regression results are positively correlated at the 5% significance level.

A fixed-effect model regression was performed using ROE (Return on Equity) instead of ROA, with Pol and control variables included. The results remained significant, confirming robustness. As shown in Table 12, the empirical analyses of the previous study are robust.

This study conducts an empirical analysis on the data of A-share TCM enterprises listed on the Shanghai and Shenzhen Stock Exchanges from 2002 to 2024 by constructing an unbalanced panel fixed-effect model, aiming to explore the impact of TCMRIPs on enterprise performance. The results show that there is a significant positive correlation between TCMRIPs and the performance of TCM listed enterprises in China, which strongly proves that innovation-oriented policies have a positive regulatory effect on enterprise performance. This means that TCMRIPs can effectively promote enterprises to improve their performance through a series of mechanisms, such as encouraging the research and development of new drugs and optimizing the registration process. The study finds that the impact of the policy on enterprise performance has a lag, and it takes a certain period of time for the policy effect to be fully manifested. According to the results, the positive impact of the policy on the enterprise’s performance reaches its peak in the third year. From the perspective of the relationship among policies, R&D investment, and enterprise performance, the study confirms that TCMRIPs affect enterprise performance by influencing the proportion of enterprises’ R&D investment. That is, policies can encourage enterprises to increase R&D investment, and the increase in R&D investment is positively correlated with the improvement of enterprise performance. This indicates that R&D investment plays a key mediating role between policies and enterprise performance.

In addition, beyond the statistical finding that TCMRIPs exert a significant positive lagged effect—with this impact peaking in the third year—and that R&D investment serves as a critical mediating factor, TCMRIPs ultimately influence enterprise performance through three closely intertwined mechanisms. First, as a formal institutional arrangement under the umbrella of industrial policy, TCMRIPs effectively mitigate market uncertainty and information asymmetry. For instance, the new version of the Drug Registration Administration Measures clarifies the development path for TCM and simplifies registration classification, which sends unambiguous policy signals to enterprises. These signals help enterprises avoid blind investment in R&D projects that do not align with policy priorities, while also enhancing their access to more stable external financing (e.g., equity investment or credit support), as investors gain greater confidence in the industry’s policy-backed development prospects. Second, TCMRIPs drive adjustments to enterprises’ internal resource allocation. Backed by implicit incentives such as accelerated registration for innovative TCM drugs and potential policy preferences (e.g., tax breaks or subsidies), enterprises reallocate resources away from low-value-added production of mature products toward high-value R&D activities. This realignment of resources directly aligns with a core goal of industrial policy: promoting industrial upgrading by optimizing resource distribution in strategically important sectors. Third, TCMRIPs facilitate the conversion of R&D achievements into tangible enterprise performance. Shortened registration cycles enable innovative TCM drugs to enter the market more quickly, allowing enterprises to recoup R&D costs promptly and secure first-mover advantages in competitive segments. This mechanism-centered analysis not only clarifies why the statistical correlation between TCMRIPs and enterprise performance exists but also embeds the study’s empirical findings within the broader theoretical framework of industrial policy. In doing so, it strengthens the research’s theoretical depth and enhances its explanatory power for both academic and practical audiences.

Although this study verifies the positive impact of the TCMRIPs on corporate performance and the mediating role of R&D investment through basic regression, endogeneity tests, and robustness tests using ROE instead of ROA, with clear logic in statistics and correlation analysis, there remains room for deepening the interpretation of the correlation between control variables (Size, Age, and Lev) and corporate performance as well as policy effects. The existing research only incorporates these three variables into the model as independent control variables to eliminate interference, without fully analyzing their own mechanism of action on corporate performance and the potential correlation with policy effects. Therefore, we need to separately interpret and discuss the results of these three variables.

For Size, as a core reflection of the resource endowment of TCM enterprises, firm size does not have a simple linear correlation with corporate performance. Instead, it exerts an indirect influence through the chain of “resource integration capability—policy response efficiency—performance conversion” and has implicit adaptability to the effects of TCMRIPs.

From the perspective of the basic support for corporate performance, the sample data shows that the mean value of Size is 21.356 and the median is 21.404, with a symmetric and concentrated distribution. This indicates that most listed TCM enterprises are at a medium scale in the industry, and there is no extreme differentiation in size. This scale characteristic determines the enterprises’ ability to acquire and utilize resources: large enterprises usually have more complete R&D infrastructure, wider market channels, and stronger compliance management capabilities. Under the guidance of TCMRIPs, large enterprises can quickly convert policy support into practical actions. TCM R&D is characterized by high investment, long cycles, and high uncertainty. Although TCMRIPs can reduce enterprises’ R&D costs through subsidies, tax reductions, and other means, they cannot completely eliminate risks such as technical failures and low market acceptance. With sufficient cash flow reserves, large enterprises can withstand the losses from a single R&D failure and continue to invest in innovative projects in line with policy orientation. In contrast, if small enterprises experience R&D failures, they may face the risk of capital chain rupture and be forced to terminate R&D activities related to policy response, thereby weakening the role of policies in promoting performance.

For Age, which reflects the industry experience and operational maturity of TCM enterprises. Its impact on corporate performance has a “double-edged sword” characteristic and also indirectly affects enterprises’ response efficiency to TCMRIPs. The sample data shows that the mean value of Age is 17.312 and the median is 17.000, indicating that most sample enterprises have been established for more than 15 years and are in the mature stage of operation. Long-term industry involvement has enabled these enterprises to accumulate advantages in three aspects: first, experience in traditional technologies and formulas. The core competitiveness of TCM enterprises is often related to classic prescriptions and traditional processing techniques. Older enterprises have deeper technical accumulation in these fields and can develop improved new drugs (such as the secondary development of classic famous prescriptions) based on traditional advantages, reducing R&D risks. Second, accumulation of market trust. Through long-term product quality control, older enterprises have formed high brand awareness among consumers. After newly registered products are launched, they are more likely to gain market acceptance, accelerating performance conversion. Third, experience in policy adaptation. From the first version of the Measures for the Administration of Drug Registration in 2002 to the new version in 2020, older enterprises have gone through multiple rounds of policy adjustments, making them more sensitive to changes in registration processes and approval standards. They can quickly adjust R&D and declaration strategies to avoid performance losses caused by insufficient policy adaptation.

For Lev, the result that the asset-liability ratio has no significant correlation with corporate performance is the most important part requiring in-depth interpretation in the analysis of control variables. The sample data shows that the mean value of Lev is 0.362 and the median is 0.329, presenting a slightly right-skewed distribution. This indicates that there are some enterprises with high liabilities, but the overall liability level is within a reasonable range (the reasonable range of asset-liability ratio in the pharmaceutical industry is generally considered to be 30–50%). This result actually stems from the “offset of two-way effects” and “scenario dependence” of liabilities on performance, and there is an implicit interaction with the role of TCMRIPs.

From the perspective of the “offset of two-way effects,” the impact of liabilities on the performance of TCM enterprises has both positive and negative effects: on the one hand, moderate liabilities can supplement R&D and operating funds to help enterprises respond to TCMRIPs. For example, enterprises can expand R&D teams and conduct clinical trials through bank loans, or invest funds to optimize registration materials to meet the requirements of the new policies, thereby promoting performance improvement. On the other hand, excessive liabilities will increase financial risks and squeeze the space for R&D investment. TCM R&D has a long cycle. If an enterprise has a too high liability level, it needs to give priority to repaying interest, which may lead to the misappropriation of R&D funds, or even the termination of innovative projects in line with policy orientation due to short-term debt repayment pressure, dragging down performance. The slightly right-skewed distribution of Lev in the sample means that the negative effects of some high-liability enterprises offset the positive effects of most enterprises with reasonable liabilities, ultimately resulting in no significant correlation between Lev and ROA in the linear model.

From the perspective of “scenario dependence,” the impact of Lev on performance is also related to the “purpose of liabilities” and “policy adaptability” of enterprises. If enterprises use debt funds for R&D activities encouraged by TCMRIPs, liabilities may be converted into performance through policy dividends. If liabilities are only used for the expansion of traditional production, it will be difficult to enjoy policy support, and performance may decline due to increased market competition. The existing model does not distinguish between “liability purposes,” making it impossible to capture this scenario difference. For example, after the implementation of the 2020 new version of the Measures for the Administration of Drug Registration, some enterprises raised funds to carry out R&D of “innovative TCM drugs,” and such liabilities were positively correlated with performance. In contrast, some enterprises still raised funds to expand the production capacity of TCM decoction pieces, and such liabilities were negatively correlated with performance. The combination of these two situations leads to the non-significance of the Lev coefficient. In addition, the impact of Lev may also vary by enterprise size: large enterprises are more likely to obtain R&D funds through liabilities and have strong debt repayment capabilities, so liabilities may be positively correlated with performance; small enterprises mostly use liabilities for operating capital turnover, which may be negatively correlated with performance. This size difference further exacerbates the “apparent no correlation” between Lev and performance.

Based on the above conclusions, we can put forward the following recommendations:

(1) For the government

Continuous efforts should be directed toward promoting innovation in the registration of TCM while further streamlining procedural simplifications. Considering that the newly version of the Drug Registration Administration Measures has proven to be highly effective in enhancing enterprise performance, it is essential for the government to ensure continuity and stability in these policy refinements. This approach can significantly boost enterprise performance, which can then translate into improved products through various avenues such as upgrading product quality, expanding product lines, and enhancing clinical efficacy, thus fostering the advancement of TCM. Moreover, given the lag effect of policies, with their positive impacts peaking in the third year, the government should establish a robust long-term policy implementation and evaluation mechanism. It is essential to avoid short-term policy fluctuations, ensuring that enterprises have stable policy expectations to facilitate long-term R&D planning. Concurrently, policies should be dynamically adjusted based on the actual development of the TCM industry. This could involve regularly collecting feedback from enterprises regarding the effects of policy implementation, tracking changes in R&D investments, and monitoring the performance of listed TCM enterprises. Such information can help modify and enhance relevant policies in a timely manner to better meet the needs of industrial development. Additionally, since TCMRIPs influence enterprise performance through R&D investment, the government could introduce supportive measures to encourage enterprises to increase their R&D expenditures. For example, this could include providing tax incentives for R&D costs. Such as increasing the percentage of R&D expense deductions. Or establishing special funds to subsidize R&D projects for TCM enterprises, particularly small and medium-sized enterprises with relatively limited R&D capabilities. These initiatives could further amplify the role of R&D investment as a mediating factor, thereby enhancing the positive impact of policies on enterprise performance.

(2) For enterprises:

Policies have the potential to significantly enhance enterprise performance, with their effects often reaching a peak after 3 years. Consequently, businesses should develop long-term R&D investment strategies that align with the policy cycle. Specifically, enterprises should consider increasing their R&D investment intensity in line with their own development cycles. This approach can improve enterprise performance, boost output, and further advance the growth of the TCM industry. Enterprises must recognize the crucial mediating role that R&D investment plays and incorporate it into their core development strategies. Establishing a dedicated R&D management department can help conduct comprehensive research on trends in TCM new drug development and policy guidance. It is essential to formulate scientific R&D plans to ensure the effective allocation of R&D funds. Additionally, given the inherent lag in policy effects, enterprises should exhibit patience and persistence in their R&D investments. They should avoid reducing or halting R&D funding due to short-term performance fluctuations and instead commit to long-term investments, allowing sufficient time for the maximum impact of policies to materialize.