Introduction

Implementing innovation-driven strategies is crucial for sustainable economic development in China (Zhang and Yu, 2019). Innovation activities require substantial and continuous capital investments. However, owing to the high costs and significant risks involved, innovation does not yield immediate returns. This places firms in a dilemma of high capital demand and difficult financing (Hall, 1992; Wang and Sun, 2019). Driven by the profit-seeking nature of capital, non-financial firms increasingly favor financial investment activities, making financial assets an integral part of their asset portfolios (Demir, 2009). Capital flowing from the real economy into the financial sector has become an indisputable reality. This raises the question of whether innovation, which drives the continuous and stable development of the real economy, is influenced by firms’ financialization.

Scholars remain divided on how financialization impacts firm innovation. Some argue that financialization fosters innovation. Financialization enables firms to secure additional funding during external market fluctuations (Stulz, 1996), alleviate financing constraints (Yang et al. 2019), and reduce default risks (Deng et al. 2020). The high returns from financial activities help firms save on taxes by hedging transactions, lowering financial distress costs, and diversifying investment portfolio risks (Zhang and Zheng, 2019). This provides stable cash flow for innovation activities (Liu, 2017), increases the sustainability and conversion rate of innovation (Gehringer, 2013), and contributes to firms’ long-term development (Almeida et al. 2004; Kliman and Williams, 2015). However, others contend that the positive effects of financialization merely serve to mask the issue and cannot compensate for its encroachment on innovation resources (Dore, 2008; Orhangazi, 2008; Duan and Zhuang, 2021; Zhao et al. 2024). In pursuit of high returns from financialization, firms may preferentially allocate limited funds to the financial sector (Crotty, 2003; Demir, 2009; Dallery, 2009; Xiao et al. 2021), crowding out the capital needed for innovation and exacerbating financing constraints (Dore, 2008; Demir, 2009; Wang et al. 2017; Xu et al. 2023). Consequently, firms’ profits become increasingly dependent on financial asset returns, with the expansion of financial activities suppressing operating profit margins (Xie and Kuang, 2020). Moreover, managerial behavior tends to be short-term (Stockhammer, 2004; Peng et al. 2018). This leads to R&D departments being neglected and R&D projects with long cycles and high risks being rationally rejected (Palley, 2007; Orhangazi, 2008; Demir, 2009; Cupertino et al. 2019; Wu and Lu, 2023).

A possible reason for the above heterogeneous conclusions is that different types of financial assets vary in terms of income and risk characteristics, reflecting diverse motivations and goals behind firms’ financialization, and inevitably have varied impacts on firm innovation. Unified conclusions cannot be drawn without distinguishing between these forms and their motives. Following Demir (2009), Orhangazi (2008), and Yu et al. (2022), we separated transaction-oriented financial assets from firms’ overall financial assets. Transaction-oriented financial assets, which are significantly liquid assets on the balance sheets of non-financial listed firms, are cash assets with a certain investment income. They may provide reserve support for innovation activities, alleviate risk fluctuations caused by innovation investments, and promote firm innovation. However, they can also become firms’ main investment channels, thereby hindering innovation. Therefore, separating transaction-oriented financial assets from other financial assets and studying their transmission mechanisms concerning firm innovation is crucial.

The results of this study were similar to those of Yang et al. (2017) and Duan and Zhuang (2021). Duan and Zhuang (2021) examined the mechanism by which financial investment behavior affects firm innovation in China, arguing that the financial investment activities of non-financial firms had been found to negatively affect both their innovation efforts and outcomes; however, this negative effect persists. Conversely, we believe that the motives and returns for different types of financial assets vary, leading to varied effects on firm innovation. Therefore, identifying the impact mechanisms of various types of financial assets on firm innovation is more important. Additionally, the two studies differed in terms of model selection. Duan and Zhuang (2021) used ordinary least squares (OLS) estimation as the primary regression model. However, because the number of patents is a count variable, the Poisson and Negative Binomial (NB) models, which are more suitable for count data, were used in this study; they were used in addition to the general OLS estimation to analyze the impact of transaction-oriented financial assets on patents (Kortum and Lerner, 2000). This provided consistent results. Yang et al. (2017) identified the motives for firms to allocate transaction-oriented financial assets by examining the relationship between monetary policy, stock market cycles, and transaction-oriented financial assets as well as their impact on firm investment behavior. However, we analyzed transaction-oriented financial assets from the perspective of liquid reserves and risk mitigation. Yang et al. (2017) found that transaction-oriented financial assets can promote firm innovation input, which is consistent with the primary findings of this study.

Based on the identification of the motives for firms’ allocation of transaction-oriented financial assets, we empirically examined the impact mechanisms of transaction-oriented financial assets on firms’ innovation inputs and outputs from the perspectives of liquid reserves and risk mitigation. Consequently, we found that transaction-oriented financial assets alleviate firms’ external financing constraints by providing sustained financial support for innovation as a liquid reserve. Furthermore, combining transaction-oriented financial assets and innovation activities facilitates risk sharing, thereby allowing companies to engage in high-risk innovation projects while maintaining an acceptable level of risk.

The contributions of this study are threefold. First, considering the different purposes and motives of financial assets with differences in risk, liquidity, and profitability, transaction-oriented financial assets are analyzed specifically. This can help curtail the differences in extant research conclusions on the relationship between financialization and firm innovation. Second, it analyzes the mechanism through which transaction-oriented financial assets affect firm innovation and demonstrates that transaction-oriented financial assets can promote firm innovation continuity as a liquid reserve. Finally, from a risk mitigation perspective, this study verifies that transaction-oriented financial assets help balance and hedge innovation investment risks and enhance firms’ risk-taking abilities. By refining the logical chain of transaction-oriented financial assets that influence firm innovation at the micro level, this study reveals the impact of transaction-oriented financial assets on firm innovation.

The remainder of this paper is organized as follows: Section 2 presents the literature review and research hypotheses; Section 3 details the model and data; Section 4 presents the empirical results; Section 5 addresses the robustness test; and Section 6 concludes the paper.

Literature review and research hypotheses

Transaction-oriented financial assets promote firm innovation as a liquid reserve

According to the precautionary savings theory, imperfect capital markets often exhibit credit discrimination. Additionally, the financial sector is stringent in considering the risks of firms and is cautious about the financing needs of firms’ innovation activities (Yang and Zeng, 2014). Firms are motivated to engage in smooth innovation activities (Opler et al. 1999).

Innovation activities are highly confidential. For reasons such as intellectual property protection, firms are reluctant to share detailed R&D information with their creditors. Consequently, creditors cannot accurately assess firms’ financial status and future profitability, leading to substantial information asymmetry. To ensure fund security, creditors tend to make conservative estimates of firms’ borrowing information (Lin and Li, 2001). This reverse selection results in higher external financing costs for firms, preventing them from obtaining the long-term stable funding required for high-risk innovative activities. Moreover, the long timelines and uncertain returns associated with innovation increase the risks involved, leading creditors to demand higher interest rates or implement credit rationing for firms’ innovation financing (Stiglitz and Weiss, 1981; Harhoff, 1998). The gap between internal and external financing costs makes it even more challenging for firms to secure external funding for innovation. Banks and other financial institutions often require the use of physical collateral assets. However, innovation primarily involves technological and human capital inputs, and innovation outcomes mainly consist of intangible assets such as proprietary technology or patents. These intangible assets generated via innovation investments lack physical collateral and do not meet the risk control requirements of financial institutions. Therefore, firms face more severe financing constraints for innovation than fixed asset investments (Liu et al. 2015).

To alleviate the difficulty of financing innovation, firms often choose to maintain a certain level of liquid reserves within their operations (Lu et al. 2013), thus ensuring relatively stable funding for R&D investments (Ju et al. 2013). Cash holdings are the primary form of liquid reserves. Although cash guarantees free control and convenience of use, it is susceptible to erosion by managers’ personal interests and incurs high opportunity costs. As financial markets evolve, the variety of financial products and tools has expanded significantly, providing firms with various types of liquid reserves. According to Tobin’s theory of cash asset demand, transaction-oriented financial assets have characteristics such as timing flexibility, frequent trading, immediate convertibility, and reasonable profits that ensure liquidity and safety. These low-risk, high-liquidity assets with relatively low returns on investment are more cost-effective than cash and have better liquidity and convertibility than fixed assets (Liu, 2017; Tori and Onaran, 2018). By allocating transaction-oriented financial assets, firms can preserve and potentially appreciate capital and reduce the adjustment costs associated with temporary shortages of innovation funds (Almeida et al. 2004; Liu, 2017; Yang et al. 2017). This reduces the volatility of cash flow for innovation funding, providing a more advantageous position for accumulating funds for innovation activities. Additionally, the excess returns from financial asset investments can improve performance and alleviate financial pressures on firms (Duchin et al. 2017). This can offer firms more financial support during external market fluctuations (Stulz, 1996) and prevent adverse effects on innovation owing to funding shortages (Yang et al. 2017), thus achieving a “reservoir” effect.

Krippner (2005) and Wen and Ren (2015) argued that financialization may cause changes in firms’ investment strategies, exacerbating investment myopia. In the pursuit of high returns from financialization, when firms’ resources are limited, managers and major shareholders prioritize investments in the financial sector that serve their own interests. This increases short-term profits through capital arbitrage but adversely affects R&D investment (Wang et al. 2017). However, compared with other financial asset investments, the returns on transaction-oriented financial assets are low. Allocating funds for highly liquid financial assets is more likely driven by a precautionary savings motive. This helps mitigate future cash flow uncertainties and provides a continuous funding source for investments, thereby improving firm performance (Bonfiglioli, 2008; Kliman and Williams, 2015). Therefore, holding transaction-oriented financial assets based on the “reservoir” motive can provide continuous funds for firm innovation. This can be crucial in coping with cash flow risks and alleviating external financing constraints (Liu, 2017). The development of the financial market provides a way for firms to alleviate financing constraints through transaction-oriented financialization and supports the sustainable development of firm innovation activities. Therefore, Hypothesis 1 is as follows.

Hypothesis 1: Transaction-oriented financial assets, such as liquid reserves, promote innovation.

Risk mitigation effect of transaction-oriented financial assets on firm innovation

Portfolio theory suggests that rational investors can achieve optimal risk investment portfolios by meeting one of the following two conditions: choosing the portfolio with the highest expected investment return or selecting the portfolio with the lowest investment risk. Firms’ investment decisions essentially involve investors making reasonable choices between uncertain returns and risks by comprehensively assessing the risks and returns of various assets. This involves balancing risk and returns in business decisions, diversifying to spread operational risk, and finding the optimal balance point (Brown and Petersen, 2011; Bargeron et al. 2010), thereby ensuring that the firm takes on reasonable levels of risk.

Innovation is distinct from other strategic decisions. Although innovation can improve future market competitiveness, its outcomes are unpredictable. The initial phase of innovation requires significant financial investment from firms while being very likely to fail, which significantly increases the possibility of financial distress. Furthermore, firms must contend with intense market competition, significant financing constraints due to limited collateral, and face high liquidity and investment risks, which weaken their risk tolerance and make them risk-sensitive. Firms’ innovation activities are considered a form of risk investment, with innovation output being the successful result of undertaking risky investment projects (Singh, 1986; Griffin et al. 2009; Baumol, et al. 2007). Excessive risk-taking can also cause financial distress, leading to bankruptcy (John et al. 2008); this harms firms’ innovation capacity. To maintain stable income levels, firms attempt to diversify their operations to reduce business risk (Brown and Petersen, 2011).

The real options theory introduces financial market rules into firms’ investment decisions. This theory suggests that investment projects’ reversibility influences their sensitivity to uncertain risks (Li and Yang, 2015). Compared with transaction-oriented financial assets, firm innovation has relatively low reversibility. When uncertain risk increases, the option value of innovation fluctuates significantly, prompting firms to reduce innovation investments and avoid high losses from potential innovation failures. Transaction-oriented financial assets are not the preferred channel for firms seeking high investment returns (Yang et al. 2017); they have relatively high reversibility and are more beneficial for firms to address future uncertainties. These assets help achieve revenue growth while reducing the uncertainty of primary business income fluctuations, making them an important way for firms to reasonably allocate assets and manage risks (Tornell, 1990; Zhang and Zheng, 2019).

Firms purchase transaction-oriented financial assets with temporarily idle funds to increase their capital supply during periods of low funding, thereby allowing them to cope with sudden financial fluctuations. This reduces the pressure that firms face from the uncertain risk of innovation and balances the risk of failure in the innovation knowledge conversion process (Yang et al. 2019), ultimately reducing the cash flow risk associated with innovation investment. Enhanced risk tolerance for innovation also increases the flexibility of capital operations, positively impacting current business and innovation activities.

Incorporating low-risk transaction-oriented financial assets into firms’ investment portfolios with high-risk innovation activities is beneficial in the following ways: for balancing and hedging innovation risks; reducing the portfolio’s overall risk (Liu and Ju, 2019); enhancing firms’ tolerance for high-risk innovation behaviors; reducing the impact of innovation activities on the overall asset risk stability of the firm; and ensuring the safety of the entire balance sheet. Transaction-oriented financial assets are more likely to balance the risk of the main business, thereby mitigating firm risk. Therefore, Hypothesis 2 was proposed as follows.

Hypothesis 2: Transaction-oriented financial assets in an investment portfolio mitigate risk and promote innovation by enhancing firms’ risk-taking levels.

Research design

Data Source

The data were mainly obtained from the CSMAR and Wind databases. Firms’ financialization data were manually compiled from financial statements, with missing data supplemented by manually collected annual reports. The data for listed firms on the Shanghai and Shenzhen A-share stocks in China from 2008 to 2019 were processed by excluding the following: Financial, ST, and PT firms; firms with incomplete data or those that had undergone major asset transactions during the study period; and those with negative or zero equity values. This resulted in an unbalanced panel dataset comprising 18 industries, 2232 firms, and 12,086 valid observations. To eliminate the influence of outliers, continuous variables were winsorized at the 1st and 99th percentiles.

Empirical model and main variable setting

Transaction-oriented financial assets promoting firm innovation as a liquid reserve from the dimensions of financing constraints and innovation sustainability

Fazzari et al. (1988) proposed a significant positive correlation between financing constraints and firms’ investment-cash flow sensitivities. Investment-cash flow sensitivity is effective in measuring financing constraints (Whited, 1992). Therefore, firms’ financing constraints were assessed using the investment-cash flow sensitivity model. The following regression model was proposed to empirically verify Hypothesis 1:

$$\begin{array}{l}{{INV}}_{i,t}={\alpha }_{0}+{\alpha }_{1}{{AM}}_{i,t-1}+{\alpha }_{2}{FC}_{i,t}+{\alpha }_{3}{{AM}}_{i,t-1}* {FC}_{i,t}+{\alpha }_{4}{{SIZE}}_{i,t-1}+{\alpha }_{5}{{DBR}}_{i,t}\\\qquad\quad +\,{\alpha }_{6}{{CUR}}_{i,t}+{\alpha }_{7}{{TOBI}{N}^{{\prime} }Q}_{i,t-1}+{\alpha }_{8}{{GROW}}_{i,t}+\sum _{j}{\alpha }_{j}\,{Industrydummies}\\\qquad\quad +\,\sum _{k}{\alpha }_{k}{Yeardummies}+{\varepsilon }_{i,t}\end{array}$$
(1)

Based on Demir (2009), transaction-oriented financial assets (SF) were categorized into two groups: transactional financial assets (SF1) and wealth management and trust products (SF2). SF1 includes financial assets for trading, derivative financial assets, net short-term investments, net financial assets available for sale, net held-to-maturity investments, and net long-term debt investments. The data of SF2 was obtained by manually sorting the details of “other liquid assets” in notes to firms’ financial statements. In the four models used in our study, FC represents SF, SF1, and SF2.

Model (1) focused on the coefficient \({\alpha }_{3}\), that is, the cross-term of cash flow and transaction-oriented financial assets. If \({\alpha }_{3}\) is significantly negative, this means that transaction-oriented financial assets can significantly reduce the financing constraints of firms, partially supporting Hypothesis 1.

$$\begin{array}{l}{{RD}T}_{i,t}={\beta }_{0}+{\beta }_{1}{{RD}T}_{i,t-1}+{\beta }_{2}{FC}_{i,t}+{\beta }_{3}{{RD}T}_{i,t-1}* {FC}_{i,t}+{\beta }_{4}{{SIZE}}_{i,t}+{\beta }_{5}{{DBR}}_{i,t}\\\qquad\quad\; +\,{\beta }_{6}{{EBIT}}_{i,t}+{\beta }_{7}{{AM}}_{i,t}+{\beta }_{8}{{TOBI}{N}^{{\prime} }Q}_{i,t}+{\beta }_{9}{{CONC}}_{i,t}+{\beta }_{10}{{AGE}}_{i,t}\\\qquad\quad\; +\,\sum _{j}{\beta }_{j}{Industrydummies}+\sum _{k}{\beta }_{k}{Yeardummies}+{\varepsilon }_{i,t}\end{array}$$
(2)

Model (2) tested the impact of transaction-oriented financial assets on innovation sustainability. Innovation input (R&D expenditures) and output (patents) are commonly used to measure firms’ innovation behavior. Furthermore, Cao and Chen (2019) highlighted that design and utility model patents belong to “incremental innovation,” which is only a small improvement over existing products without strict scrutiny. Compared with “incremental innovation,” invention patents are “radical innovation,” which have higher certainty and can better represent firms’ innovation ability. Therefore, both R&D expenditure (RD) and the number of invention patents (LnPAT) were used to measure firm innovation (RDT) in our study.

As patents are count variables bounded by zero, we opted for count models to assess the effect of transaction-oriented financial assets on patents, in addition to the standard OLS estimates (Maddala, 1983; Cameron and Trivedi, 2013). Commonly used count models include the following two: the Poisson and NB (Hausman et al.1984; Cincera,1997; Hu and Jefferson, 2009). Therefore, we estimated the results of both the Poisson and NB models.

A one-period lag in firm innovation was included in the model. The greater the coefficient \({\beta }_{1}\) is, the higher firm innovation sustainability is. In model (2), we mainly focused on the coefficient \({\beta }_{3}\), that is, the cross-term of firm innovation and transaction-oriented financial assets with one-period lag. If \({\beta }_{3}\) is significantly positive, this means that transaction-oriented financial assets can increase firm innovation sustainability, further supporting Hypothesis 1.

Verifying the risk mitigation effect of transaction-oriented financial assets on firm innovation with firms’ risk-taking levels as the mediating variable

Freedman and Schatzkin (1992) posited that when the relationship between independent and dependent variables can be explained by the same variable to a certain extent, it can be concluded that this variable plays a mediating role. To test the risk-mitigation effect of transaction-oriented financial assets on firms’ innovation behavior, this study introduced firms’ risk-taking levels into the mediation model and examined how transaction-oriented financial assets affect firm innovation under portfolio motivation. According to the mediating effect model, Model (3) focuses on whether transaction-oriented financial assets improve firms’ risk-taking levels.

$$\begin{array}{l}{{RISK}}_{i,t}={\gamma }_{0}+{\gamma }_{1}{FC}_{i,t}+{\gamma }_{2}{{SIZE}}_{i,t}+{\gamma }_{3}{{DBR}}_{i,t}+{\gamma }_{4}{{EBIT}}_{i,t}+{\gamma }_{5}{{AM}}_{i,t}\\\qquad\qquad +\,{\gamma }_{6}{{TOBIN}^{\prime}\, Q}_{i,t}+{\gamma }_{7}{{CONC}}_{i,t}+{\gamma }_{8}{{AGE}}_{i,t}+\sum _{j}{\gamma }_{j}{Industrydummies}\\\qquad\qquad +\,\sum _{k}{\gamma }_{k}{Yeardummies}+{\varepsilon }_{i,t}\end{array}$$
(3)
$$\begin{array}{l}{{RD}T}_{i,t}={\delta }_{0}+{\delta }_{1}{FC}_{i,t}+{{\delta }_{2}{RISK}}_{i,t}+{\delta }_{3}{{SIZE}}_{i,t}+{\delta }_{4}{{DBR}}_{i,t}+{\delta }_{5}{{EBIT}}_{i,t}+{\delta }_{6}{{AM}}_{i,t}\\\qquad\quad\; +\,{\delta }_{7}{{TONBIN}^{\prime}\, Q}_{i,t}+{\delta }_{8}{{AGE}}_{i,t}+{\delta }_{9}{{CONC}}_{i,t}+\sum _{j}{\delta }_{j}{Industrydummies}\\\qquad\quad\; +\,\sum _{k}{\delta }_{k}{Yeardummies}+\,{\varepsilon }_{i,t}\end{array}\,$$
(4)

If \({\gamma }_{1}\) is significantly positive, this indicates that transaction-oriented financial assets can significantly improve firms’ risk-taking levels. We further examined the significance of \({\delta }_{2}\) in Model (4). If \({\delta }_{2}\) is significantly positive, this indicates that transaction-oriented financial assets affect firm innovation by improving firms’ risk-taking capacities, which supports hypothesis 2.

The measurement indicators of firms’ risk-taking levels include performance volatility, stock return volatility, debt ratio, and possibility of firm survival. Adams et al. (2005) specified that performance volatility not only reflects firms’ profit stability but also describes the size of the risks faced by firms by their performance deviating from normal levels. Therefore, we used performance volatility to measure firms’ risk-taking levels.

Performance volatility is usually calculated using the financial performance indicator ROA and the market performance indicator TOBIN’Q. However, the existence of non-tradable shares in China’s stock market will lead to certain biases in the calculation of TOBIN’Q. The market performance index TOBIN’Q is not accurate in China (Zhang and Li, 2012). Therefore, referring to Adams et al. (2005), we considered every three years as an observation period and used the degree of deviation between annual real performance and the mean value adjusted by industry and year to reflect firms’ performance volatility. The greater the performance volatility is, the higher the firm’s risk-taking level is. The calculation process is as follows.

$$\begin{array}{c}{{ADJ}{{\_}}{ROA}}_{i,t}=\frac{{{PBIT}}_{i,t}}{{{ASSETS}}_{i,t}}-\frac{1}{{x}_{n}}\mathop{\sum }\limits_{k=1}^{x}\frac{{{PBIT}}_{k,i}}{{{ASSETS}}_{k,t}}\\ \begin{array}{cc}\sigma \left({ROA}\right)=\sqrt{\frac{1}{N-1}{\sum }_{t=1}^{N}({{ADJ}{{\_}}{ROA}}_{i,t}-\frac{1}{N}\,{{\sum }_{t=1}^{N}{{ADJ}{{\_}}{ROA}}_{i,t}})^{2}\,} & N=3\end{array}\end{array}$$

i and t refer to the ith firm in the tth year of the nth industry. \({x}_{n}\) refers to the number of firms in the nth industry. k and t refer to the data for year t for each firm in the nth industry. After calculating \({{ADJ}{\rm{\_}}{ROA}}_{i,t}\) for each year, the standard deviation of periods t-1, t, and t + 1 was calculated as the risk-taking level (RISK) of period t.

Control variables

In addition to transaction-oriented financial assets, firm innovation is influenced by a variety of factors. Following of Liu (2017), we included indicators related to firm characteristics and governance as control variables. Detailed descriptions of these variables and their calculation methods are presented in Table 1 as per format of Ameer et al. (2025).

Table 1 Variable definitions and calculation methods.
Full size table

Firm size (SIZE)

Larger firms with economies of scale, relatively abundant operating cash flows, fewer financing constraints, better access to external funding at lower costs, and better positioned to sustain their competitive edge are inclined to invest in high-risk projects. As firm size increases, the likelihood of investing in innovation also grows. The regression results were expected to be positive.

Asset–liability ratio (DBR)

Firms’ asset–liability ratio may also affect their innovation input. A higher debt ratio leads to increased financing costs for a firm. Moreover, firms with high debt ratios are often subject to greater constraints imposed by creditors, who may limit risky investments to safeguard their own interests. As a result, a higher asset-to-liability ratio tends to discourage firms from investing in innovation. The regression results were expected to be negative.

Firm profitability (EBIT)

Firm with high profitability has more funds, which affects future innovation activities. The higher a firm’s profitability level is, the more funds it will use for innovation. However, a high return on capital may also reduce the demand for innovation (Liu, 2017). Therefore, the direction of the regression results was uncertain.

Firm growth ability (Tobin Q)

Firms with strong growth potential experience higher market demand for their products, which in turn promote their innovation efforts. The regression results were expected to be positive.

Net cash flow (AM)

Firms with abundant cash flow have a stronger capacity to manage risks, making them more inclined to invest in high-risk projects. The regression results were expected to be positive.

Firm age (AGE)

The longer a firm has been established, the higher its risk resistance and the more willing it is to undertake high-risk investment projects. However, some research indicates that mature firms often adopt a more cautious approach to investments with uncertain returns, as they may lack the competitive drive of their younger counterparts. The direction of the regression results was uncertain.

Firm governance (CONC)

Ownership concentration is used as the firm governance variable. Shleifer and Vishny (1986) proposed that high ownership concentration can reduce free-rider behavior caused by ownership dispersion, which is conducive to investment activities that can increase firm value. However, Zhou and Song (2016) argued that, in firms with concentrated ownership, controlling shareholders have strong control over the firm, which can easily induce them to seek personal benefits, mismanage the resources of the firm, and have a negative impact on firm innovation. Therefore, the direction of the regression results was uncertain.

Dummy variables. Additionally, to account for the effects of industry-specific differences and macroeconomic conditions, this study includes two dummy variables: Industry and Year. Detailed definitions and calculations for all variables are provided in Table 1 as per the variable description format of Ameer et al. (2025).

Empirical analysis

Descriptive statistics and correlation analysis

Descriptive statistics of variables

Table 2 reports descriptive statistics for the main variables. Although the mean of RD in sample firms is 2.32%, the median is only 1.91%, which indicates that most firms had less R&D expenditure. Furthermore, innovation input among firms varied substantially, and the situation of LnPAT is similar. The mean of SF in sample firms is 12%, in which the mean of SF1 is 9.67% and that of SF2 is 2.31%. The medians of SF, SF1, and SF2 are all lower than their means, indicating that most sample firms had low transaction-oriented financial assets.

Table 2 Descriptive statistics of main variables.
Full size table

Correlation analysis

This study used Pearson’s correlation coefficient tests to identify the relationships between variables, thereby avoiding the influence of multicollinearity on the results. The results are shown in Tables 3 and 4. Most correlation coefficients are small, except for the correlation coefficient between SF and SF1, which is 0.9728. However, SF and SF1, the two different explanatory variables in our study, did not appear in the same model. Therefore, the interference of multicollinearity in the results was eliminated.

Table 3 Pearson correlation coefficient matrix for major variables.
Full size table
Table 4 The coefficients of VIF.
Full size table

The significance level of the main variables is below 10%; therefore, the analysis shows that our hypothesis of no correlation between the variables cannot be rejected. We further tested for multicollinearity by calculating the Variance Inflation Coefficient. The Variance Inflation Coefficient of each variable was less than 3, with a maximum of 2.7. Additionally, there was no significant multicollinearity among the key variables, ensuring that the regression results were reliable and accurate.

Analysis of empirical results

As a liquid reserve, transaction-oriented financial assets alleviating firms’ financing constraints

To verify that transaction-oriented financial assets have a positive impact on firm innovation through the mechanism of liquid reserves, a panel fixed-effect model was used to estimate investment cash flow sensitivity (AM). Table 5 presents the regression results of the impact of transaction-oriented financial assets on firms’ external financing constraints. FC in Columns (1)–(3) represents SF, SF1, and SF2, respectively. All AMt-1 *FC in the three models were significantly negatively correlated with firm innovation, indicating that transaction-oriented financial assets can significantly ease firms’ financing constraints, consistent with the findings of Yang et al. (2019) and Yang and Chen (2023) and partially supporting Hypothesis 1.

Table 5 The regression results of transaction-oriented financial assets easing firms’ financing constraints.
Full size table

Tables 6 and 7 present the regression results for the long-term effects of transaction-oriented financial assets and their subcategories on firm innovation (including R&D expenditure and patents).

Table 6 The regression results of transaction-oriented financial assets promoting the sustainability of firms’ R&D expenditures.
Full size table
Table 7 The regression results of transaction-oriented financial assets promoting the sustainability of firms’ patents.
Full size table

Innovation in the previous year significantly positively influences the current innovation behavior of firms. Table 6 presents the coefficient of RDTt-1*FC, which is significantly positive. The coefficients of the interaction terms between the subcategories (SF1 and SF2) and the previous year’s R&D expenditure are also significantly positive, indicating that transaction-oriented financial assets support firms’ ongoing innovation investments, consistent with the findings of Yang et al. (2017). Table 7 reports the results of the sustained effect of transaction-oriented financial assets on patents using OLS, Poisson, and NB estimation models. Other than the non-significant results of LnPATt-1*SF2 in the Poisson and NB models, the results are consistent with expectations, indicating that transaction-oriented financial assets promote the sustainability of firm innovation, supporting Hypothesis 1.

Risk mitigation effect of transaction-oriented financial assets on firm innovation

According to the theoretical analysis above, transaction-oriented financial assets may promote firm innovation by increasing risk-taking levels. According to the theory of mediation effects, if firms allocate transaction-oriented financial assets based on investment portfolio motives, they can maintain the risk level that they can bear by combining low-risk transaction-oriented financial assets with higher-risk innovation investments. Tables 8 and 9 indicate the regression results of the mediating effects of risk taking as an intermediary and the influence path of transaction-oriented financial assets on firms’ R&D expenditures and patents.

Table 8 The regression results of the risk mitigation effect of transaction-oriented financial assets on firms’ R&D expenditures.
Full size table
Table 9 The regression results of the risk mitigation effect of transaction-oriented financial assets on firms’ patents.
Full size table

The results in Column (1)–(3) Table 8 show that SF, SF1, and SF2, respectfully, are significantly positively correlated with RISK. This implies that transaction-oriented financial assets have better liquidity and greater security. The more transaction-oriented financial assets firms hold, the greater the volatility of earnings they can withstand, which is consistent with Zhang and Zheng’s (2019) findings.

Columns (4)–(6) in Tables 8 and 9 present the combined impact of transaction-oriented financial assets and risk-taking levels on firms’ R&D expenditures and patents, according to model (4). Both the coefficients of transaction-oriented financial assets (SF, SF1, and SF2) and the risk-taking level (RISK) are significant, indicating that transaction-oriented financial assets improve firms’ risk-taking levels and enable them to bear the high risks brought about by innovation activities and that risk-taking is the mediating variable of transaction-oriented financial assets affecting firm innovation. Excluding risk mitigation, transaction-oriented financial assets still significantly affect firm innovation, indicating that in addition to risk mitigation, other factors also impact the relationship between transaction-oriented financial assets and firm innovation behavior. Furthermore, risk mitigation only partly contributes to the impact of transaction-oriented financial assets on firm innovation. These results confirm that highly liquid financial assets help firms mitigate future uncertainty and increase their income stability. Firms engage in financial investment not only because of the high returns on financial assets but also as an investment decision to reduce the risks associated with physical investments (Tornell, 1990).

Robustness test and endogeneity processing

Owing to the high uncertainty in the outcomes of innovation activities, firms often strive to balance innovation and financial investment activities, aiming to manage risks while pursuing short-term profits. This dynamic can create a trade-off between financialization and R&D expenditure in non-financial firms. To ensure the reliability of the regression results, this study conducted several robustness checks and addressed potential endogeneity concerns.

Firms’ financialization strategies can affect their R&D expenditures, and innovation-driven strategies can influence their financialization decisions. Additionally, the financialization strategies of other firms within the same industry may have affected the sample firms’ financialization strategies. Therefore, potential endogeneity concerns, including omitted variables, bidirectional causality, and self-selection, may exist. To mitigate the risk of biased conclusions regarding the impact of financialization on firm innovation caused by selection issues, following the method of Yang et al. (2019), we introduced two instrumental variables for the robustness test: M2, which measures the macro money supply, and AFC, representing the average transaction-oriented financial assets of other firms in the same industry.

Heckman two-stage model

The Heckman two-stage model was used to address endogenous problems and sample selection bias caused by omitted variables. First, a probit model was used to select the factors influencing transaction-oriented financial assets. The Inverse Mills Ratio (IMR) was then calculated and entered into the second-stage regression as the control variable to investigate the impact of transaction-oriented financial assets on firm innovation. As presented in Table 10, the regression results align with those of the main analysis, supporting the robustness of the findings.

Table 10 The results of the Heckman two-stage regression.
Full size table

Testing endogenous problems with instrumental variables

The main regression results show that transaction-oriented financial assets can alleviate firms’ financing constraints. However, it is also possible that firms facing lower financing constraints are more inclined to prioritize transaction-oriented financial assets. To eliminate this effect, further control for endogenous problems, and ensure the reliability of the empirical results, this study selected AFC as the instrumental variable and employed the instrumental variable generalized moment estimation method (IV-GMM). The second-stage regression results, as shown in Table 11, are consistent with the main regression findings.

Table 11 The endogenous test of transaction-oriented financial assets easing firms’ financing constraints.
Full size table

Additionally, the main regression results show that transaction-oriented financial assets can improve firms’ risk-taking levels and promote innovation behavior. However, it is also possible that firms with higher levels of risk-taking tend to favor transaction-oriented financial assets. To examine this, the same method described earlier was applied. The second-stage regression results, presented in Table 12, align with the main regression findings.

Table 12 The endogenous test of transaction-oriented financial assets increasing firms’ risk-taking levels.
Full size table

Conclusions and suggestions

Conclusions

Presently, the relationship between financialization and firm innovation is a subject of ongoing debate with differing perspectives. This may be because different types of financial assets held in the financialization process have different income and risk attributes, representing different motivations and purposes that inevitably have different impacts on firm innovation. A unified conclusion cannot be reached without differentiating between the forms and motives of financialization.

This study focused on transaction-oriented financial assets. This is because, as a very important liquid asset in the balance sheets of non-financial listed firms, transaction-oriented financial assets offer low adjustment costs, high liquidity, ease of cash out, and relatively stable returns. They not only help preserve and enhance firm value, but also provide additional capital when firms face a shortage of funds for innovation. As an effective alternative to holding cash, they can alleviate financial constraints and lower the costs of external financing.

Additionally, to reduce the negative impact of innovation investment funds on the flexibility of business adjustment, firms also allocate other types of assets, including financial assets, to balance, reduce, or hedge the risks of innovation investment, as Aiming to improve the security of the asset portfolio. Innovation investment and financialization are not choices, but a matter of portfolio optimization. Therefore, it is important to separate transaction-oriented financial assets from aggregate financial assets and study their impact on firm innovation based on different motivations.

This study selected non-financial listed firms in the Shanghai and Shenzhen A-share markets from 2008 to 2019 as the research sample. From the perspectives of liquid reserves and risk mitigation, we examined the impact mechanism of easing financing constraints and improving the risk-taking level through which transaction-oriented financial assets affect firms’ innovation behavior. The empirical regression results show that firms that invest idle funds in transaction-oriented financial assets can significantly alleviate financing constraints, reduce the adjustment costs associated with temporary shortages of innovation funds, and provide ongoing financial support for innovation activities through a liquid reserve.

Transaction-oriented financial assets significantly improve firms’ risk-taking levels, helping them engage in more innovation activities. Firms can bear high-risk innovation activities while maintaining an acceptable level of risk, through a portfolio combining transaction-oriented financial assets and innovation investment to achieve risk allocation. The risk-taking level is the mediating variable in the relationship between transaction-oriented financial assets and firm innovation activities.

This study identified the different forms and complex motives behind firms’ financialization, subdivided financial assets, and revealed the impact mechanism of transaction-oriented financial assets on firms’ innovation behaviors. This study helps understand the reasons for the continuous growth of firm innovation investment against the backdrop of an increasing level of financialization. It also contributes to a deeper understanding of the micro-consequences of financialization, which is crucial for advancing supply-side reforms in China and accelerating the achievement of the goal of becoming a technology powerhouse.

Suggestions

Financialization is a necessary stage in the global financial development process. Amid the relatively sluggish development of China’s real economy, transaction-oriented financial assets can broaden firms’ profit channels, ease financing constraints, diversify investment risks, and support them in transforming and upgrading their conduct. Based on our findings, this study proposes the following suggestions:

From the perspective of liquid reserve, non-financial firms use transaction-oriented financial assets as a channel and means to meet the long-term funding needs of innovation activities. Non-financial firms face a challenging financing environment for innovation investment and hope to maintain liquidity through transaction-oriented financial assets. First, the Chinese government should promote financial reforms, establish sound and diversified financing channels, enhance the efficiency and capacity of financial institutions to serve the real economy, and effectively address the financing difficulties of innovation activities. Second, the Chinese government should focus on improving basic financial services and markets, actively guiding and supporting the growth of emerging financial sectors, improving financial service efficiency, reducing loan costs, enhancing financial inclusiveness, accurately and effectively aligning financial resources with the needs of the real economy, and providing greater support for the innovation activities of non-financial firms. Finally, to encourage the development of innovative and entrepreneurial investment funds in the capital market, the Chinese government should actively guide the securities market toward innovative firms and reduce the financing difficulties faced by non-financial firms in the capital market.

From a risk mitigation perspective, non-financial firms’ innovation is closely linked to risk. Various unpredictable outcomes can arise during innovation. Firms aim to mitigate operational risks and manage the uncertainty caused by innovation through transaction-oriented financial assets. This study argues that the role of venture capital in shaping firms’ innovation capabilities should be fully leveraged. Establishing and improving mechanisms to compensate for innovation risks in firms will enhance inclusiveness in physical innovation and alleviate concerns about engaging in innovation activities. Second, firms should enhance risk-warning systems to better predict, assess, and control technological innovation risks. Firms are more willing to innovate by deepening the integration of next-generation information technology applications, focusing on market information mining and demand understanding, and identifying the optimal pathways for technological innovation succession.

Further discussions

By identifying firms’ motivation to allocate transaction-oriented financial assets, this study explored both theoretical and empirical analyses to examine the relationship between transaction-oriented financial assets and firm innovation. However, there are two limitations of the present study that need further discussion. First, this study did not consider the influences of internal governance mechanisms and external market environments. Internal governance mechanisms, including agency problems and the quality of firms’ financial reports, influence financial asset selection. External market environments such as market competition and government intervention levels also influence firms’ innovation strategies. Future studies should focus on enhancing the efficiency of capital markets in serving the real economy. Second, detailed differentiation in the level of firm risk-taking is lacking. This study used the overall risk-taking levels of firms when studying the risk-mitigation effects of transaction-oriented financial assets on innovation without a more nuanced characterization of risk-taking levels. Future research should consider separating non-systemic and systemic risks from total risks to provide a more accurate depiction of how risk-taking affects the relationship between transaction-oriented financial assets and firm innovation.