The administration of the recognition of high-tech enterprises is a critical policy for high-tech industry in China, as it reduces the income tax rate for high-tech firms to 15%—a fairly low rate compared to 25% for traditional firms. This policy stems from The Law of the People’s Republic of China on Enterprise Income Tax (Thirteenth Standing Committee of the National People’s Congress, 2018) and significantly affects the strategic arrangement, distribution, and development of high-tech enterprises. Although previous studies have examined the various impacts of this recognition on high-tech enterprises’ research and development (R&D) activities, such as R&D investment and patent output (e.g., Sun, 2018; Wan & Xu, 2019; Yang et al. 2017; Yang & Rui, 2020), empirical evidence on the policy’s impact on R&D disclosure remains limited (Han & Feng, 2021). In addition, this recognition incentive plays a critical role in shaping the core competitiveness of high-tech enterprises.

This study therefore focuses on key dimensions of high-tech recognition in China. This policy grants bonuses to firms whose technologies meet certain criteria and conditions for innovation. Our goal is to to understand its fundamental implications, specifically regarding R&D disclosure. We define high-tech enterprise and distinguish between authentic and pseudo-high-tech organizations, examining the differences in R&D disclosure between these two types of firms.

R&D disclosure plays a critical role in investors’ assessment of enterprise value and the design of corporate investment strategies. Firms typically adopt one of the two strategies in response to high-tech recognition. The first is to disclose more information, as the “high-tech designation” and the increasing number of R&D activities in these firms prompts a quick response from investors seeking more R&D information (e.g., Farrell & Klemperer, 2007; Huang et al. 2021; Xu & Zheng, 2016). The second strategy is to disclose less R&D information to protect proprietary knowledge and safeguard competitiveness and market share (e.g., Beyer et al. 2010; Merkley, 2014). In addition, since financial disclosure policies target specific industries and may encourage competitors to enter these fields (Liu et al. 2023), the more information firms release, the higher the risk of sharing it. Thus, empirical evidence is needed to determine how high-tech enterprise recognition affects R&D disclosure. Understanding this differential impact is crucial for policymakers and business leaders to effectively balance transparency with competitive advantage.

To test the impact of high-tech recognition on R&D disclosure, this study draws on the theoretical frameworks of stakeholder theory (Freeman et al. 2010) and institutional theory (Meyer & Rowan, 1977; Scott, 2005) and employs a difference-in-differences (DID) design, which estimates the changes in R&D disclosure before and after the recognition for high-tech enterprises (treatment group) and non-high-tech enterprises (control group). The sections of this paper are as follows: First, we discuss the theoretical and empirical contexts of this study and develop hypotheses. Next, we explain the research method and model design. Then, we explain the data analysis process and present the results. Finally, we discuss the theoretical and policy contributions and implications of our findings.

Literature review

Stakeholder expectation and R&D disclosure

Freeman et al. (2010) regard stakeholder theory as an integration of value creation, trade development, and business management. The theory suggests that while most stakeholder groups share a common interest in value creation, attention should be given to how value is generated for each specific group. Earlier research found that external stakeholders can impose information disclosure costs on the firm, including political costs, which means that engaging stakeholders is crucial to the management and operation of the company (Li et al. 1997). In the same vein, Freeman et al. (2010) also emphasized that addressing the concerns of serious stakeholders is crucial to company policy development. Drawing from stakeholder theory, various studies have demonstrated a strong connection between R&D disclosure and stakeholder engagement, highlighting how transparent communication can enhance trust and involvement among stakeholders. Gordon et al. (2020) have argued that effective disclosure of a company’s R&D activities can enhance its perceived competitive potential among external stakeholders, thereby reducing the information gap between corporations and stakeholders. Another study (Chen et al. 2022) found that voluntary R&D expenditure disclosure is positively related to innovation in emerging markets with transparent environments that critically emphasize stakeholder engagement. However, it has also been documented that market competition may reduce willingness to disclose R&D, thereby affecting stakeholders’ access to information. The study of Zhang (2023) demonstrated that firms adjust their R&D investments in response to competitors’ disclosures, with disclosures from strong competitors discouraging investment, while those from weaker competitors encouraging further investment. These studies contribute to the stakeholder theory by demonstrating how R&D disclosures can meet the information needs of stakeholders, including investors and the market, and influence a firm’s relationships with them, thereby improving performance. The rationale behind this argument is that disclosure can mitigate information asymmetry (limited or lack of information on various firms) for stakeholders, perhaps leading to multiple benefits for the company, including a providing greater value for the firm, reducing financing costs, and allowing better access to external financing (e.g., Healy & Palepu, 2001; Li &Yao, 2020). Furthermore, Chen and Yu (2022) found that vertical, horizontal, and competitor collaboration directly enhances the firm’s R&D intensity, R&D human capital, and growth capacity. Fedorova et al. (2022) also found that financial performance, R&D disclosure and innovation, patent portfolios, and patent citations all can have positive effects on the capitalization of companies. Recently, it has been argued that high-quality R&D disclosure can create competitive advantages for firms through product differentiation and cost-control leadership (Shrestha et al. 2023). Baruffaldi et al. (2024) demonstrated that firms increase scientific publications to mitigate information asymmetry, thereby enhancing transparency and meeting stakeholder expectations. Therefore, R&D disclosure becomes a crucial strategy for companies seeking a good relationship with their stakeholders.

Several factors can influence the decision to increase the level of R&D disclosure. Based on the explanation above, by analyzing data on publicly listed companies in China from 2008 to 2020, two main factors were identified: recognition from government and attractiveness to investors. Government recognition of high-tech enterprises enhances their commitment to R&D innovation and increases the likelihood of achieving R&D success. Figure 1 illustrates the trends in the number of enterprises, R&D investment, and R&D results, comparing high-tech firms with non-high-tech firms. High-tech enterprises display a more noticeable increase in the number of firms, R&D investment, and R&D achievements, especially after 2017, with the implementation of the Administration on Recognition of High-tech Enterprises. Therefore, this recognition policy has resulted in more enterprises applying for the “high-tech” designation and increased their willingness to disclose R&D innovations and prospects for future achievements. This increased transparency attracts investors and helps them better understand these firms and invest in them more confidently. Managers may choose to disclose more R&D information to address the traditional imbalance created by limited information on R&D investments, which leaves investors seeking greater transparency. Furthermore, increased investment often results in significant achievements, raising awareness of the company’s new products (Beyer et al. 2010) and attracting new investors (Huang et al. 2021). However, uncertainties remain about the appropriate approach for high-tech enterprises to disclose their R&D information while actively pursuing innovation.

Fig. 1: Trends over time in observations, R&D investment, and patent applications for high-tech and non-high-tech enterprises.
figure 1

This figure illustrates the trends in the number of enterprises, R&D investment, and R&D results, comparing high-tech firms with non-high-tech firms. Data is obtained from the CNRDS database and manual extraction from annual reports.

Full size image

This study focuses on high-tech recognition in China. By examining the impact of this policy, which grants bonuses to firms meeting specific technological criteria and recognition conditions, we aim to explore its broader implications, particularly regarding R&D disclosure. As this R&D recognition policy grants bonuses to firms whose technologies meet certain criteria and conditions, it can draw more competitors into an industry (Liu et al. 2023). This increased competition can heighten the risk of rivals exploiting R&D information and “stealing” company achievements. To safeguard their R&D activities from competitors, companies often choose to limit public access to innovation after being certified as high-tech enterprises. Studies have found that innovation-driven firms tend to disclose less R&D information due to concerns about the cost of their innovations and giving away their advantages. For example, Jones (2010) suggested that innovation-driven companies do not need to disclose specific R&D information to meet investors’ expectations. The objective can be achieved by releasing information such as earnings forecasts to encourage investors’ hopes about the company’s future without exposing sensitive R&D details. Ramaboa and Chen (2017) found that firms with a strong commitment to R&D investment are systematically undervalued in equity markets. Similarly, Kalantonis et al. (2020) observed that R&D disclosures significantly influence investor decisions, enabling investors to identify and target companies effectively.

Enterprises tend to disclose R&D information only when the potential profits outweigh the associated development costs (Huang & Liang, 2024). As a result, significant uncertainty remains regarding the value of R&D disclosure in China’s capital market. Traditional economic theory posits that increased disclosure reduces information asymmetry by creating a more level informational playing field between insiders and external stakeholders (Marquardt & Wiedman, 1998). However, recent studies offer a more nuanced perspective. Huang and Liang (2024) found that extensive narrative R&D disclosures can unintentionally increase information asymmetry by allowing insiders to exploit the difficulties faced by investors in processing complex information. Their study reveals that insiders achieve higher profitability from stock sales when narrative disclosures are lengthy but unclear, as these disclosures impose significant processing costs on external investors, hindering their ability to fully comprehend the firm’s R&D activities. Another risk arises from the dominance of individual investors in this market, who often exhibit strong speculative intent and prioritize investing in concept stocks over considering R&D or non-financial information that reflects the firm’s intrinsic value (You, 2017). Even if more attention were paid to R&D information, investors could not interpret it well because technological jargon can obscure the details. These can include tester, lifting device, medical record, and other items that are difficult to interpret. Some scholars therefore, categorize critical information in annual reports as “technical references” (Jones & Shoemaker, 1994). Compared to investors in developed capital markets, individual Chinese investors generally lack investment experience, skills, and professionalism (Zhong & Lu, 2018), making it even more challenging for them to effectively interpret the vast amount of R&D information. As a result, it is hard for high-tech enterprises to determine whether the benefits of R&D disclosure outweigh the associated costs, leading to limited disclosure of R&D information. Drawing on the aforementioned literature and studies, we propose the following hypothesis:

H1: The recognition of high-tech enterprises is not associated with changes in R&D disclosure.

Industrial impact on R&D disclosure

Institutional theory (Meyer & Rowan, 1977; Scott, 2005) is a systematic framework to explain organizational behavior and change. It specifies that, in their formation and operation, organizations are subject to the limitations of industry systems and norms. This adaptation leads them to adopt specific behavioral strategies to meet social expectations, thereby gaining legitimacy and stability (Washington & Patterson, 2011). Kafouros et al. (2009) used firm-based datasets to show that economic payoffs for larger firms and for organizations that followed an R&D-intensive strategy are significantly higher, allowing such firms to improve their corporate performance. Li and Zou (2024) demonstrated that the 2012 regulation in China, mandating narrative innovation disclosures, helped reduce the cost of equity capital associated with R&D expenditure. The regulation had a more significant effect for firms facing high information asymmetry and risky R&D projects, highlighting how mandatory disclosure can enhance transparency and investor confidence in high-tech industries. More recently, Baruffaldi et al. (2024) showed that in the face of an information asymmetry (imbalance) in financial industries, firms tend to adapt their R&D disclosures by increasing scientific publications to enhance information transparency and gain legitimacy. Similarly, Chu et al. (2024) found that extensive innovation disclosures in product announcements not only increase transparency but also positively influence future sales, enhancing firm performance and investor confidence. This perspective supports the broader industrial belief that R&D disclosures serve as strategic tools for firms to showcase their innovative capabilities and secure competitive advantages. Grounded in institutional theory, high-tech companies strive to align their policies with institutional expectations and norms, both legally and ethically, to ensure their long-term development.

However, not all high-tech firms perform R&D activities. To obtain the title of high-tech enterprise and its preferential policies, pseudo-high-tech enterprises inflate their R&D investment by fabricating accounting items and manipulating innovative business activities (Wan & Xu, 2019). From Yang and Rui’s (2020) point of view, high-tech enterprise recognition inherently constitutes an incomplete contract between the government and the given firms, leading to significant information asymmetry. This imbalance creates opportunities for pseudo-high-tech enterprises to manipulate their R&D investment figures. Because some R&D activities conducted by pseudo-high-tech firms are fictitious and their innovation efforts are solely aimed at gaining investor recognition, their R&D disclosure strategies differ from those of authentic high-tech companies. Zhou et al. (2024) examined R&D information manipulation in the context of Chinese manufacturing firms, discovered that while R&D manipulation is common, the adoption of robots has a mitigating effect. By enhancing transparency, human capital, and governance, robot adoption reduces firms’ need or ability to manipulate R&D disclosures, promoting a more accurate representation of their innovation activities.

Compared to authentic high-tech enterprises, pseudo-high-tech firms disclose less R&D information because they engage in fewer R&D activities. Yang et al. (2017) found that the ratio of R&D investment to sales revenue tends to diverge as firms approach the threshold for high-tech enterprise recognition, providing evidence of manipulative R&D behaviors. Heng et al. (2024) provided empirical evidence from China’s Science and Technology Innovation Board, showing that firms under financial pressure, like pseudo-high-tech firms, are more likely to inflate R&D spending figures. This manipulation enables firms to access policy benefits and address financing challenges, but it also poses risks of misleading investors and regulators.

In addition, pseudo-high-tech enterprises are unlikely to provide detailed descriptions of their often-fabricated R&D investments. Usually, this investment is allocated to employees’ salaries, machine supplies, or transportation, and in some cases, the R&D innovation is not put into operation (Wan & Xu, 2019). Therefore, firms are unlikely to disclose information related to such fabricated R&D activities to avoid being penalized. Compared to genuine high-tech enterprises, pseudo-high-tech enterprises are also more likely to cloud the readability of R&D information. R&D manipulation, similar to providing false information, conducting tax evasion, and committing other violations, it is illegal. Once caught, firms will be stripped of their high-tech enterprise status and may face substantial fines and criminal penalties. To avoid such punishment, they often conceal their behavior through complex or opaque transactional records or accounting practices. However, digital manipulation alone is not sufficient, so companies also resort to manipulating textual information. Lo et al. (2017) found that the readability of annual reports decreased when the company’s earnings for a given period barely exceeded those of the previous period. This suggests that companies obscure management practices related to earnings by making R&D information difficult to interpret. Similarly, Xu et al. (2021) found that the readability of annual reports decreased when a firm’s performance allegedly exceeded normal expectations or when the firm was under special treatment for financial difficulties or other crises. This increases the likelihood that such companies will manipulate earnings reports. In this context, pseudo-high-tech enterprises may deliberately complicate R&D information and reduce the readability of reports, making the information difficult to understand. This can distort authorities’ judgment regarding their behavior concerning such information. Based on the arguments above and institutional theory, we propose the following second hypothesis:

H2: Compared to authentic high-tech enterprises, pseudo-high-tech enterprises disclose less and vaguer R&D information.

Methods and model

Sample selection

The initial sample included all the firms listed on the Shanghai Stock Exchange (SSE) or Shenzhen Stock Exchange (SZSE) from 2008 to 2020. We excluded non-high-tech industries, such as finance, restaurants, hotels, motels, real estate, leasing, and business services. Following Li’s (2008) investigations and other research, we excluded the management discussion and analysis section of annual reports if it contains less than 100 words, as well as firms with total negative assets or missing financial data, as these cannot be used to objectively determine R&D disclosure. The method described above collected 24,883 enterprise-year observations from 3036 unique firms. R&D information was obtained from the CNRDS database, while other financial data required for the empirical analysis were sourced from the China Stock Market & Accounting Research Database (CSMAR) and the Wind Database. We used STATA 18 to process the data.

Measurement of R&D disclosure

R&D disclosure was measured using content analysis and the machine-learning technique of the MD&A section in annual reports. This method has several advantages. First, it allows sampling from a broad range of enterprises over a relatively long period (e.g., Merkley, 2014; Rawson, 2021). Second, it can more accurately capture the multiple meanings of Chinese words in financial documents. Third, prior research suggests that the MD&A section is a vital source of information and can better indicate managers’ strategies than other sections in the R&D disclosure (e.g., Li, 2010a, 2010b, 2012; Loughran & McDonald, 2011).

To measure R&D disclosure, we first selected terms from Chinese R&D keywords and phrases by translating Merkley’s (2014) English dictionaryFootnote 1 and then randomly read excerpts from 500 MD&A sections. Given the multiple meanings of Chinese words, we applied Mikolov et al. (2013) and the Word2Vec technique to identify synonyms and expand our dictionary with new entries. The Word2Vec technique is a word-embedding approach used within neural networks. It transforms words into multidimensional vectors and captures their semantic meanings in context. By evaluating the similarity between these vectors, Word2Vec adeptly identifies synonyms and contextual relationships, thereby offering a nuanced understanding of verbal semantics based on placement and usage in the text (Bengio et al. 2003). Specifically, we used the continuous bag-of-words (CBOW) model for the Word2Vec technique to train the words/phrases in all MD&A sections. The CBOW model is:

$${{Max}}\mathop{\sum}\limits_{{{w}}\in {{C}}}\log {{p}}\left({{w}}|{{Context}}({{w}})\right)$$

Here, C represents the words and phrases in the MD&A sections, w is the list of Chinese R&D keywords and phrases defined above, and Context (w) refers to the context for w. We first estimated the probability of C based to Context (w), maximized the model to obtain word vectors corresponding to w, and then derived synonyms for w by calculating the similarity of these vectorsFootnote 2. These synonyms are particularly suitable for financial documents; the CBOW model’s word training is based on a large corpus of words and phrases from such documents (Mikolov et al. 2013). To ensure the adequacy of our final dictionary, we invited three industry experts and three scholars to review the R&D keywords and phrases.

We carefully selected a dictionary comprised of R&D keywords and phrases to measure the enterprises’ R&D disclosure at the sentence level in the MD&A section. Specifically, we used three measures to capture the quantity of the enterprises’ R&D disclosure based on prior work by Li (2010a), Merkley (2014), and Huang et al. (2021): 1) the amount of R&D disclosure (R&DSent), 2) the amount of numerical R&D disclosure (R&DSentQ), and 3) the amount of forward-looking R&D disclosure (R&DSentFLS). R&DSent is the number of R&D-related sentences in the MD&A section. We defined a sentence as R&D if it contains Chinese R&D keywords and phrases. R&DSentQ is the number of numerical R&D-related sentences in the MD&A section. We classified an R&D sentence as numerical if it contains numerical information, excluding dates. R&DSentFLS is the total number of forward-looking R&D-related sentences in the MD&A section. We defined an R&D sentence as forward-looking if it contains words in the future tense (Li, 2010a).

H2 predicts that pseudo-high-tech enterprises are more likely to provide opaque R&D information. We considered the readability of R&D disclosures (R&DSentFOG) as a measure of full disclosure. According to Merkley (2014) and Xu et al. (2021), we measured R&DSentFOG by using a modified fog index in Chinese, where FOG = 0.40× (average words per R&D-related clauseFootnote 3 + average complex words per R&D-related sentence). Following Wang et al. (2018), we defined complex words as sub-common, adverse, and accounting. Sub-common words were those that increase the difficulty of reading posed by uncommon characters in the text, and a higher sub-common words value signifies a more complex annual report text. Adverse words are based on the complexity of logical relationships in the text. More adverse components (such as however, but, even if, etc.) indicate more significant contradictions or differences among sentences and paragraphs, those that which complicate common thinking and make comprehension more challenging. A higher value of adverse words indicated greater complexity. As for accounting words, which reflect the frequency of financial and accounting terminology, the value of higher accounting words indicates higher complexity. It is noteworthy that many accounting words, such as accounts receivable, inventories, raw materials, long-term investments, idle assets, and R&D expenses, are generally easy to understand in financial documents. Li et al. (2020) also explained that most investors can understand accounting words in annual reports. To address these concerns, we compiled a list of accounting words that are easy to understand. Using the Jieba package in Python, we segmented the text, collected all accounting terms, and generated a list of 3000 high-frequency words. These words were excluded from the list of complex words.

Research design

To test H1, we estimated the changes in R&D disclosure before and after identification as high-tech enterprises. Since the recognition of each high-tech enterprise typically occurs in different years, we followed Moscviciov et al. (2010), and Beck et al. (2010), using a time-varying DID design to estimate the relationship. The advantage of the DID-estimating method is that it allows for control of omitted variables, such as unobservable trends and natural changes (Abadie, 2005), changes that can be considered part of the normal evolution of the groups being studied, independent of the intervention. Specifically, our first regression model isFootnote 4:

$$\begin{array}{c}{{{R\& D\_\,Disclosure}}}_{{{it}}}{={{{\beta }}}_{1}{{HTE}}}_{{{it}}}+\sum {{\beta }}{{{Controls}}}_{{{it}}}+{{Enterprise\; fixed\; effects}}+{{Year\; fixed\; effects}}\\ +{Province\,fixed\, effects}\times {Year\,fixed\,effects}+{{{\varepsilon }}}_{{it}}\end{array}$$
(1)

where i and t indicate enterprise and year, respectively. R&D_Disclosure is one of the three R&D disclosure quantity variables defined above (R&DSent, R&DSentQ, R&DSentFLS). We used the logarithm of R&D_Disclosure as our dependent variable to address concerns about skewed distribution. The recognition for high-tech enterprise (HTE) is an indicator that signifies that a firm has been identified as a high-tech enterprise with a validity period of three years from the initiation of the Administration on Recognition of High-tech Enterprises, otherwise indicated as zero. We focus on the coefficient of HTE (β1), which captures the changes in R&D disclosure before and after the firms have been identified as high-tech enterprises.

The control variables included enterprise-level factors associated with R&D disclosure. Following Merkley (2014) and Huang et al. (2021), we included Model (1) variables to measure the firms’ information environments (Size, Analyst, Inst, MFCount, ARInfo, DA), their maturity (Age, Age2), investment mix (R&DIntensity, R&DIntensity2, INTANG, B/M), information uncertainty (ROAVot, RetVot), profitability (ROA, Loss), and financing incentives SEO, Lev). Research suggests that the strategies of R&D information disclosure or innovation activities differ between state-owned and non-state-owned firms (e.g., Wen & Feng, 2012; Wei, 2020). Therefore, we also included state-owned firms (SOE) in Model (1), with an indicator variable of one if the firm is state-owned, of zero otherwise. (See Appendix A for detailed definitions of these variables).

Finally, we included enterprise and fiscal-year fixed effects to further control time-invariant heterogeneity at the enterprise level. Research suggests that firms’ innovation activities are affected by macro-economics and institutional policies, such as those facilitating economic development, intellectual property protection, and patent policies of different provinces (Qiu & Tao, 2021). We also included the interaction terms of province and year fixed effects in Model (1). Throughout the paper, standard errors were clustered at the enterprise level to mitigate heteroscedasticity, serial, and cross-sectional correlation.

To test H2, we estimate the difference of R&D disclosure between authentic and pseudo-high-tech enterprises before and after being identified as high-tech. Our second regression model is:

$$\begin{array}{c}{{{R\& D\_\,Disclosure}}}_{{{it}}}={{{{\beta }}}_{1}{{HTE}}}_{{{it}}}+{{{\beta }}}_{2}{{{HTEnterprise}}}_{{{it}}}\times {{{PseudoHT}}}_{{{it}}}+\sum {{\beta }}{{{Controls}}}_{{{it}}}\\ +{Enterprise\,fixed\,effects}+{Year\,fixed\,effects}\\ +{Province\,fixed\,effects}\times {Year\,fixed\,effects}+{{{\varepsilon }}}_{{it}}\end{array}$$
(2)

Here, R&D_Disclosure is one of the three R&D disclosure quantity variables in Model (1), and R&DSentFOG is the readability of R&D disclosure. Yang et al. (2017) argued that pseudo-high-tech enterprises might dilute R&D investment and exceed the threshold of the designated high-tech enterprises. Therefore, we followed Yang et al. (2017), Wang et al. (2019), and Yang and Rui (2020) to define pseudo-high-tech enterprises (PseudoHT) as an indicator variable equal to one if the ratio of R&D investment to sales revenue exceeds the high-tech firm threshold by 1%, and zero otherwise. Specifically, we defined PseudoHT as equal to one d of three conditions: (1) The ratio of R&D investment to sales revenue is 5–6%, when sales revenue is less than 50 million yuan; (2) The ratio of R&D investment to sales revenue is 4 to 5%, when sales revenue is more than 50 million but less than 200 million yuan; and (3) The ratio of R&D investment to sales revenue is 3–4%, when sales revenue is 200 million yuan or more. Other dependent, independent, and control variables were consistent with Model (1). We focused on the coefficient on PseudoHT (β2), which captures the difference between the quantity and readability of the R&D disclosure of authentic high-tech enterprises and pseudo-high-tech enterprises before and after they have been recognized as high-tech. We expected β2 to be negative for the R&D disclosure quantity model and positive for the R&D disclosure readability model.

Summary statistics

Panel A of Table 1 presents the sample distribution for authentic high-tech enterprises, pseudo-high-tech enterprises, and non-high-tech enterprises by year. The number of high-tech enterprises has consistently increased and has exceeded that of non-high-tech enterprises since 2010. This indicates that more and more firms are drawn to the recognition policy of high-tech enterprises and more likely to become high-tech. The observations of pseudo-high-tech enterprises show a trend from rising to declining over time. Panel B of Table 1 demonstrates the sample distribution by industry sector. Manufacturing industries, including food and beverage, apparel, paper, machine, and other manufacturing (CSRC code C1–C4), contain the most high-tech enterprises and overall observations. This is not surprising because most manufacturing industries, such as pharmaceutical, motor, and electronic equipment, are often engaged in a substantial number of R&D activities (Huang et al. 2020). The ratio of pseudo-high-tech enterprises to authentic high-tech enterprises in the construction industry (CSRC code E) is 52%, beating all other industries.

Table 1 Sample Distribution.
Full size table

Baseline tests

Results from estimating Model (1) are presented in Table 2. Column 1 presents the results using the R&D disclosure quantity (LnR&DSent) as the dependent variable. The coefficient on HTE (β1) was positive and insignificant at less than 1%. The magnitude of the coefficient (0.035) suggests that high-tech enterprises on average disclosed 3.7% more information after they have been identified as such, relative to non-high-tech enterprises.Footnote 5 Columns 2 and 3 of Table 2 show where LnR&DSentQ and LnR&DSentFLS are specified as dependent variables in regressions. The coefficients of HTE (β1) were positive and significant at less than 1%, indicating that, after these firms were identified as high-tech, both the numerical and forward-looking information of R&D disclosure of high-tech enterprises, on average, rose by 3.6%. These statistics suggest that this recognition positively impacted the overall quantity of R&D-related information for such enterprises relative to non-high-tech enterprises.

Table 2 The Impact of the High-tech Enterprise Recognition on R&D Disclosure.
Full size table

Results from estimating Model (2) are presented in Table 3, where the quantity (LnR&DSent, LnR&DSentQ, LnR&DSentFLS) and readability (R&DSentFOG) of R&D disclosure are specified as dependent variables in the regressions. The coefficients of HTE×PseudoHT (β1) were negative (positive) and significant at less than 1% in the first three columns (column 4), consistent with our prediction that pseudo-high-tech enterprises, on average, disclose 3.6% less overall, 3.6% less numerical, 3.2% less forward-looking information, and 17.4% less in readable R&D disclosure, compared to authentic high-tech enterprises. These results strongly support the prediction of H2 that pseudo-high-tech enterprises are likely to cover up their R&D manipulation with less and vaguer information.

Table 3 The difference of R&D disclosure between real high-tech enterprises and pseudo-high-tech enterprises.
Full size table

These findings align with prior research on disclosure behaviors in high-tech firms, further contextualizing the differences between authentic and pseudo-high-tech enterprises. Consistent with Huang et al. (2021), our findings suggest that government recognition motivates authentic high-tech firms to increase transparency. However, pseudo-high-tech firms exhibit opportunistic behaviors, aligning with Yang and Rui (2020), who document inflated R&D figures to gain tax benefits. Pseudo-high-tech firms’ lower scores on readability metrics align with findings by Xu et al. (2021), who observed decreased readability in annual reports under financial stress. These behaviors underscore the dual impact of high-tech enterprise recognition, promoting transparency among genuine innovators while enabling strategic manipulation among less innovative firms.

Unlike prior research, we differentiate between authentic and pseudo-high-tech firms, demonstrating their distinct R&D disclosure strategies. Authentic high-tech firms enhance transparency to meet stakeholder demands, aligning with Freeman et al. (2010), who emphasize trust-building through disclosure. This approach not only reinforces stakeholder confidence but also aligns with institutional expectations of high-tech recognition (DiMaggio & Powell, 1983). In contrast, pseudo-high-tech firms obscure information to avoid scrutiny, a behavior that highlights the strategic nature of stakeholder management as they seek to minimize regulatory and public examination.

The role of information demand in high-tech enterprise

Although the previous results were consistent with the information demand motivation of H1, they also might relate to proprietary cost concerns. In fact, H1 suggested that both the information demand and information disclosure costs of R&D disclosure would increase after the firms have been identified as high-tech enterprises. Therefore, information demand and information disclosure costs affect the impact of high-tech enterprise recognition on R&D disclosure. Because measuring investor information demand and information disclosure costs directly in the setting was challenging, we used empirical proxies from prior research to create interaction variables to test these hypotheses.

In this section, we explore the conditions under which the impact of high-tech enterprise recognition of the firms’ R&D disclosure is expected to vary. We began by investigating the role of information demand in affecting the relation between high-tech enterprise recognition and R&D disclosure. H1 suggests that high-tech enterprise recognition is related to the quantity of R&D disclosures, primarily due to investors’ demand for R&D information. If this is true, then we should expect the strength of this relationship to vary based on investor information demand. For example, market opinion leaders, such as financial analysts and institutional investors, could amplify the impact of high-tech enterprise recognition on R&D disclosure. Therefore, we predict the relationship between the high-tech enterprise recognition and the quantity of R&D disclosure would be more positive when information demand is higher. Using Merkley (2014), we incorporated the logarithm of the number of analysts (Analyst) and the percentage of institutional ownership (Inst) as proxies for information demand in the firms’ R&D disclosure. Table 4 reports the results based on adding terms for the interaction of high-tech enterprise recognition (HTE) and information demand (Analyst, Inst) in Model (1). The coefficients on HTE × Information Demand were positive and significant for the analyst and institutional investors in all columns, consistent with our prediction for information-demand motivation.

Table 4 Information demand, high-tech enterprise recognition, and R&D disclosure.
Full size table

The role of opportunistic motivation in pseudo-high-tech enterprises

In this section, we estimate the role of opportunistic motivations in pseudo-high-tech enterprises. H2 suggests that pseudo-high-tech enterprises have opportunistic incentives to adjust the quantity and readability of R&D disclosure. Specifically, pseudo-high-tech enterprises seek opportunistic incentives to construct pseudo-R&D investments to exceed the designated threshold of the high-tech enterprise. To divert regulators’ attention from this R&D manipulation, managers might provide less and vaguer R&D information. Under the opportunistic explanation, we would expect the quantity and readability differences in R&D disclosures between authentic and pseudo-high-tech enterprises to increase when opportunistic motivation is higher. Previous research suggests that a critical opportunistic motivation for managers to adjust disclosure information is to extract rents from the firm, including compensation over their equilibrium or base salary (Core et al. 1999; Wang et al. 2018). Therefore, we followed Core et al. (1999) and the excess managers’ compensation (ABMANAComp) as a proxy for their opportunistic motivation. Specifically, we used the following regression model to estimate ABMANACompit:

$${{{MANAComp}}}_{{{i}},{{t}}}={{{\beta }}}_{0}+{{{\beta }}}_{1}{{{Size}}}_{{{i}},{{t}}}+{{{\beta }}}_{2}{{{ROA}}}_{{{i}},{{t}}}+{{{\beta }}}_{3}{{{ROA}}}_{{{i}},{{t}}-1}+{{{\beta }}}_{4}{{{Lev}}}_{{{i}},{{t}}}+{{{\beta }}}_{5}{{{Dual}}}_{{{i}},{{t}}}+{{{\beta }}}_{6}{{{Boardsize}}}_{{{i}},{{t}}}+{{{\beta }}}_{7}{{{Manahold}}}_{{{i}},{{t}}}+{{{\beta }}}_{8}{{{SOE}}}_{{{i}},{{t}}}+{{\varepsilon }}$$

Here, managers’ compensation (MANAComp) is the logarithm of the average salary and the top three annual bonuses for managers. The independent variables include enterprise-level factors associated with managers’ compensation, such as economic determinants (Size, ROA, Lev), board structure (Dual, Boardsize), and ownership structure (Manahold, SOE). (See Appendix A for a detailed definition of these variables.) We then estimated Model (4) and obtained the estimated coefficients for all independent variables, then calculated the predicted and excess compensation.

Table 5 presents the results based on adding terms for the excess managers’ compensation (ABMANAComp); the interaction of high-tech enterprise recognition (HTE); and excess managers’ compensation (ABMANAComp)—along with the interaction of high-tech enterprise recognition (HTE), pseudo-high-tech enterprise (PseudoHT), and excess managers’ compensation (ABMANAComp) in Model (2). The coefficients on HTE×PseudoHT×ABManacomp were positive (negative) and significant at less than 1% in the first three columns (Column 4), aligning with our prediction that pseudo-high-tech enterprises are more likely to provide less and vaguer R&D disclosure when managers’ opportunistic motivation is higher.

Table 5 The Role of Opportunistic Motivations in the Difference of R&D Disclosure between Real High-tech Enterprise and Pseudo-high-tech Enterprise.
Full size table

Discussion

This study is one of the first attempts to empirically explore how high-tech recognition from the government affects high-tech enterprise R&D disclosure. Our findings provide empirical evidence that companies disclosed more R&D information after being recognized as high-tech enterprises and reveal that, compared to authentic high-tech enterprises, pseudo-high-tech enterprises disclosed less detailed and vaguer R&D information.

Theoretical implications

This study makes several significant theoretical contributions to the fields of stakeholder theory, institutional theory, and corporate disclosure behavior.

The findings enrich the literature on institutional theory, particularly in its application to industrial policies and firm innovations. Institutional theory, specifically through coercive, normative, and mimetic pressures (DiMaggio & Powell, 1983), offers a comprehensive lens to understand corporate R&D disclosure. Coercive pressures, such as regulatory mandates tied to high-tech certification, compel firms to meet disclosure standards to qualify for tax incentives and gain legitimacy. Our findings show that authentic high-tech firms respond to these pressures by providing detailed R&D disclosures, signaling their commitment to innovation and strengthening their credibility as industry leaders. In contrast, pseudo-high-tech firms—those lacking genuine R&D capabilities—meet only minimum disclosure requirements, using vague information to secure benefits while masking their limited innovation, thus reflecting a superficial compliance with regulatory demands.

Drawing on stakeholder theory, our findings indicate that authentic high-tech firms enhance transparency to meet stakeholder expectations, reinforcing trust and engagement (Freeman et al. 2010). Pseudo-high-tech firms, however, obscure information, strategically managing disclosures to avoid scrutiny. These results refine institutional theory by demonstrating how coercive pressures from government recognition drive divergent disclosure behaviors. Authentic firms align with institutional norms to gain legitimacy (DiMaggio & Powell, 1983), while pseudo-firms mimic compliance, highlighting the dual impact of recognition policies. Furthermore, consistent with information asymmetry theory, authentic firms reduce asymmetry to attract investment, whereas pseudo-firms exploit asymmetry to conceal limited innovation capabilities (Stiglitz & Weiss, 1981).

Normative pressures, driven by industry expectations (DiMaggio & Powell, 1983), further influence these firms’ disclosure behaviors. Authentic firms align with these norms by providing comprehensive R&D information to build trust with stakeholders, particularly when information demand is high. This commitment to transparency helps reinforce their standing within the high-tech ecosystem. Pseudo-high-tech firms, however, only meet stakeholder expectations superficially, presenting selective disclosures that project compliance without fully addressing information needs, highlighting their limited R&D resources.

At last, mimetic pressures encourage high-tech firms to emulate the disclosure practices of their peers. Authentic firms, influenced by these pressures, often exceed industry standards, positioning themselves as leaders in transparency. Pseudo-high-tech firms attempt to mimic authentic firms’ disclosures to appear legitimate, but they typically fall short in quality and depth, selectively withholding information that could reveal their lack of innovation.

The interplay of coercive, normative, and mimetic pressures creates a complex environment where high-tech recognition drives divergent disclosure behaviors. Authentic high-tech firms fully embrace transparency to bolster their legitimacy, while pseudo-high-tech firms engage in strategic, surface-level disclosures to meet minimum standards without revealing their resource limitations. This nuanced understanding of institutional pressures enriches institutional theory by demonstrating how certification-driven disclosure practices vary according to firms’ capabilities and motivations.

By incorporating the R&D information disclosure behavior of Chinese high-tech companies, our study adds a new dimension to the existing literature. It provides a nuanced understanding of how institutional factors, such as various policies and regulations, shape corporate behavior, especially in innovation and technology. Li and Zheng’s (2016) work underscores the importance of evaluating industrial policy through the lens of industry performance, suggesting that such an approach can foster innovation, optimize the mix of industries, and drive economic growth. In addition, Cocis et al. (2021) showed that economic growth, proxied by the GDP growth rate, was substantially influenced by economic indicators such as imports, exports, and gross capital formation and was mainly triggered by predictors such as interest rates, business angels, bank support, and public support.

Building on this foundation, this study expands the perspective by demonstrating how R&D disclosure practices, shaped by institutional frameworks, can serve as both a catalyst and a barometer for innovation in the high-tech sector. The exploration of R&D disclosure dynamics underscores the critical role of transparency and information flow in fostering an environment conducive to innovation. Information asymmetry theory (Akerlof, 1970; Stiglitz & Weiss, 1981) further enriches our understanding of these motivations by highlighting the imbalance in information between a company and its external stakeholders, such as investors or customers. While these stakeholders possess less knowledge about the company’s inner workings, risks, or prospects, high-tech firms face a dual challenge: reducing information asymmetry to lower investor uncertainty while safeguarding proprietary knowledge to maintain a competitive edge.

High-tech recognition encourages firms to be more transparent, aligning with institutional theory (Scott, 2005), which posits that organizational practices are influenced by regulatory, cognitive, and normative pressures. However, our findings reveal that pseudo-high-tech firms strategically diverge from normative expectations. These firms disclose less readable R&D information, aligning with information asymmetry theory as they attempt to conceal their lack of substantive R&D activities while superficially complying with coercive pressures. Stakeholder theory explains this behavior as a misalignment with stakeholder demands, where pseudo-high-tech firms prioritize short-term gains over long-term trust and transparency.

In contrast, authentic high-tech firms leverage recognition to disclose more comprehensive R&D information, enhancing their innovative capacity and contributing to the broader innovation ecosystem within the industry. This duality illustrates how institutional pressures interact with information asymmetry, shaping diverse organizational responses. Ultimately, our study enriches the understanding of how transparency in R&D disclosures can promote innovation while also exposing the potential for strategic obfuscation.

This study’s exploration of government as a critical stakeholder also extends stakeholder theory (Freeman et al. 2010), particularly regarding corporate disclosure behavior. Stakeholder theory emphasizes that different stakeholder groups—such as investors, customers, regulatory bodies, competitors, and the government—have distinct and often conflicting demands for R&D information disclosure, creating a complex environment for high-tech firms. Investors demand detailed and transparent R&D disclosures to assess a firm’s innovation capabilities and growth potential, which helps them make informed financial decisions. Transparent disclosures can build investor trust, reduce information asymmetry, and enhance the firm’s valuation (Shahid, 2024). Customers are also interested in R&D disclosures, viewing them as indicators of the firm’s commitment to quality, technological advancement, and innovation, which influences their purchasing decisions and brand loyalty.

The government plays a dual role as both a regulatory authority and a critical stakeholder. On one hand, the government requires transparency to ensure that high-tech firms comply with the terms of certification, especially when such certification provides tax incentives and other financial benefits. High-tech recognition mandates a certain level of R&D disclosure to justify these incentives, pushing firms to report on their innovation activities as a measure of accountability. On the other hand, as a stakeholder, the government seeks to promote genuine innovation and technological progress within the industry. This pressure influences recognized firms to disclose authentic R&D information to demonstrate their alignment with national policy objectives, supporting broader economic growth and industrial development.

However, competitors present a conflicting set of interests; they may perceive disclosed R&D information as an opportunity to gain strategic insights, potentially weakening the disclosing firm’s competitive edge. High-quality disclosures can inadvertently reveal proprietary knowledge or strategic priorities, which competitors could exploit. This creates a balancing act for firms, who must satisfy the government’s regulatory demands and investor expectations for transparency while protecting sensitive information from competitors.

To navigate these diverse and sometimes conflicting interests, high-tech firms often adopt selective disclosure practices. Authentic high-tech firms are likely to disclose detailed R&D information to satisfy investor, customer, and government demands, using transparency to signal innovation capacity and legitimacy. By demonstrating compliance with government standards, these firms benefit from the credibility associated with certification and government support. However, they carefully withhold proprietary details to protect their competitive position. Pseudo-high-tech firms—those that lack substantial R&D resources but seek certification benefits—respond differently, often meeting only the minimum disclosure requirements to maintain certification. These firms provide vague or superficial information to appear compliant while protecting themselves from deeper scrutiny by both the government and investors.

This strategic approach enables authentic high-tech firms to satisfy various stakeholder expectations, addressing demands for transparency from the government, investors, customers, and regulatory bodies, without revealing information that could be exploited by competitors. Pseudo-high-tech firms, meanwhile, navigate these demands superficially, offering minimal disclosures that project compliance without fully addressing the information needs of the government and other stakeholders. By analyzing these diverse stakeholder demands and firms’ varied responses, this study extends stakeholder theory by illustrating how high-tech firms manage competing interests among stakeholder groups. The government’s role as a critical stakeholder further highlights how R&D transparency can serve both public policy objectives and firm-specific strategic goals, particularly within a high-certification context.

Prior research (e.g., Yang et al. 2017; Sun, 2018; Wan & Xu, 2019; Yang & Rui, 2020) primarily focused on the influence of policy on tangible aspects of enterprises’ R&D, such as investment levels and patent output. We shift the focus by examining the impact of high-tech enterprise recognition on R&D disclosure, a critical but often overlooked aspect in discussions of traditional stakeholder theory (e.g., Sun, 2018; Wan & Xu, 2019; Yang & Rui, 2020; Yang et al. 2017). The current study also provides a more holistic view of the role of stakeholders, particularly the government, in shaping corporate disclosure practices, and underscores that stakeholders did not just influence the actual outcomes of R&D activities but also had a profound impact on how these activities are represented before stakeholders and the public. This nuanced approach to stakeholder theory acknowledges that the government, as a stakeholder, plays a dual role: first, as a regulator that sets the parameters for corporate achievement and disclosure, and second, as an influence on strategic decisions of enterprises regarding their R&D activities. This dual role is particularly evident in high-tech sectors, where technological advancements and innovations are closely tied to regulatory frameworks and government policies.

Lastly, in this study, we built a scenario to estimate the multiple motives of voluntary disclosure. Relevant literature (e.g., Amore, 2020; Huang et al. 2020; Kultti et al. 2007; Wang, 2007) argues that enterprises voluntarily disclose more information if the information demand increases, while they disclose less when information disclosure costs increase. Strategic disclosure may also exist because of management self-interest (e.g., Glaeser et al. 2020; Huang et al. 2014; Lo et al. 2017; Wang & Wang, 2018; Wang et al. 2018; Xu et al. 2021). By studying the relationship between an enterprise’s innovation and management performance reports, Huang et al. (2021) provided a scenario where both information demand and information disclosure costs are high. In our research, while the context of high-tech enterprise recognition not only increases enterprises’ need for R&D information and the cost of its disclosure, it also induces pseudo-high-tech enterprises to disclose their R&D information for incentives in the preferential policy. From the perspective of resource-based theory (Barney, 2001), which explains how firms can achieve and sustain competitive advantage through valuable, rare, inimitable, and non-substitutable resources, we argue that R&D information can be framed as a strategic asset. In the high-tech sector, recognized firms see their R&D achievements as valuable resources that contribute to competitive advantage. Consequently, these firms may selectively disclose information to maximize strategic benefits while safeguarding proprietary knowledge. In the context of high-tech enterprise recognition discussed in this paper, the market demand and information disclosure costs of enterprise R&D information may change. Besides, pseudo-high-tech enterprises are likely to strategically disclose their R&D information out of managerial self-interest. Our study provides a scenario that allows the co-existence of three motivations: information demand, information disclosure costs, and opportunistic or legitimate managerial self-interest. This highlights the complex interplay of factors influencing R&D disclosure practices and. underscores the need for nuanced regulatory frameworks.

In sum, this study integrates multiple theories to advance a holistic understanding of R&D disclosure dynamics in high-tech firms. The interaction between institutional pressures, stakeholder expectations, resource signaling, and information asymmetry reveals that R&D disclosure is not merely a response to regulatory requirements but a complex, strategically managed process influenced by certification status, resource constraints, and stakeholder relationships. These insights contribute to the literature on corporate disclosure by illustrating how high-tech certification can simultaneously drive transparency and enable selective disclosure, offering valuable implications for policymakers aiming to refine certification programs to foster genuine innovation.

Practical implication

Our study also provides empirical evidence to inform discussions on industrial policy, yielding practical implications. At the end of 2016, there was an intense debate about whether the government should enact industrial policies, focusing on whether the market’s role as the “invisible hand” should predominate or whether there should be an active role by the government on behalf of the people (Zhang & Lin, 2016). Yang et al. (2017) believe that the government may fail to achieve the goals set in an industrial policy because independent enterprises are often more motivated to meet the requirements of industrial policy to obtain benefits from it. Our findings indicate that industrial policy can both encourage enterprises to disclose more R&D information or, conversely, lead some firms to disclose selectively and strategically to obscure manipulation in R&D disclosure.

For policymakers, our findings suggest two main implications. First, the observation that genuine high-tech enterprises disclose more valuable R&D information compared to pseudo-high-tech firms can be leveraged to identify and address companies that may falsely claim high-tech status merely to benefit from government incentives. By more precisely analyzing the R&D disclosures of enterprises, policymakers can enhance the efficiency of government fund allocation and foster a more competitive market environment. Second, there is a need to further regulate and refine R&D information disclosure standards within annual reports, social responsibility reports, and ESG reports. Authorities should encourage industry organizations to develop disclosure standards and templates that balance competitive transparency with commercial confidentiality, ensuring that disclosures are both accessible and valuable to investors. This approach would enhance the availability and quality of R&D information, offering investors clearer references for making investment evaluations.

For corporate governance, our findings offer insights for shareholders in monitoring and mitigating potential R&D manipulation by high-tech enterprises. For instance, pseudo-high-tech firms may disclose vague or ambiguous R&D information to mislead investors. Thus, shareholders have a responsibility to encourage enterprises to elevate the strategic role of R&D within long-term business plans, ensuring that R&D spending is used responsibly and ethically, including in tax planning.

This duality—the market’s “invisible hand” (Smith, 1776, p.456) and a “promising government role” (Spiegel, 1999, p.222)—suggests that they can effectively coexist to foster genuine R&D innovation or, without sufficient oversight, fail to achieve these goals (Dong et al. 2020). Our findings also address the call by Sun (2018) and Hou and Yang (2019) to shift the focus of the debate from whether the government should enact industrial policy to how to enact industrial policy.

Batrancea et al. (2022) also highlights the importance of government support for small and medium-sized businesses due to their market significance and their contributions to national budgets. In recent years, Chinese regulators have gradually increased R&D disclosure requirements in annual reports, aiming to offer investors greater transparency. This aligns with a broader recognition of economic sustainability as a critical aspect of corporate governance. Previous studies have emphasized that robust R&D information disclosure supports sustainable economic development, informs governance strategies, and aids organizational decision-making. Our study reinforces these findings by demonstrating that the recognition of high-tech enterprises has positive implications.

Conclusion and limitations

Using a difference-in-differences (DID) method, this study revealed the impact of high-tech enterprise recognition on companies’ R&D information disclosure. Specifically, we examined the differences in information disclosure between two types of companies: authentic high-tech companies and pseudo-high-tech companies. Results revealed that authentic companies tended to disclose more R&D information, while pseudo companies disclosed less and vaguer R&D information. Our study generates insights into the intricate relations among government policy, company types, and companies’ R&D disclosure behaviors, offering both theoretical and policy implications.

This study has several limitations that should be acknowledged to provide a balanced perspective on the findings and identify directions for future research. First, despite our efforts to control for observable factors associated with R&D disclosure, we acknowledge that endogeneity concerns may still be present due to unknown or omitted variables (Lennox et al. 2012). These unobserved factors could potentially influence both R&D disclosure practices and stock-crash risk, leading to biased estimates in our analysis. Future studies could address this limitation by employing instrumental variable approaches, conducting robustness checks, and using more comprehensive datasets that capture a broader range of influencing factors.

Second, this study did not analyze the specific placement of R&D information within the management discussion and analysis (MD&A) section. R&D disclosures located closest to R&D expense entries may contain more detailed and specific information than disclosures elsewhere in the MD&A, potentially influencing stakeholder perception and information value. Future research could further analyze the MD&A section by examining distinct subsections to evaluate variations in the level of detail and transparency in R&D disclosures.

Third, the study focused primarily on R&D disclosure within official financial reports and did not consider other communication channels through which companies may release R&D information after being recognized as high-tech enterprises. To gain a more comprehensive understanding, future research could examine R&D information disseminated through alternative outlets, such as press releases, conference calls, or official social media accounts. By addressing these limitations, future studies could strengthen the robustness of findings and offer deeper insights into the complex dynamics of R&D disclosure, as well as its broader implications for stakeholder relationships and organizational outcomes.