The cost shock, margin gap and enterprise financialization: An exogenous shock based on minimum wage

ABSTRACT This study examines the exogenous institutional impact of minimum wage policy on enterprise financialization in China. Empirical results show that an increased minimum wage significantly promotes the financialization of real enterprises. Moreover, the impact is more pronounced in enterprises characterised by a more significant profit margin gap between tangible and financial assets and higher degrees of labour-intensiveness. Further tests reveal that the promotional effect of the minimum wage on the financialization of labour-intensive enterprises tends to be more substantial under the following circumstances: (i) a smaller gap exists between an enterprise’s average wage and the local minimum wage; (ii) an enterprise finds it more challenging to pass on costs to the market; (iii) following the implementation of the Labour Contract Law. Further investigation indicates that raising the minimum wage exacerbates the adverse effects of enterprise financialization on corporate value.


Introduction
Driven by disproportionate profit margins in the financial industry, many manufacturing enterprises have deviated from their primary business and begun to chase high returns on financial assets, creating a phenomenon known as the 'financialization of enterprises' (Hongjian Zhang & Zhang, 2016). Researchers have determined that financialization not only reduces enterprises' degree of physical investment but also damages their innovation and financial performance (Stockhammer, 2004;Orhangazi, 2008;Demir, 2009;Seo et al., 2012;Zhang & Zhang, 2016;Tori & Onaran, 2018). These 'improper works and duties' have become a significant constraint on the sustainable development of the real economy. In this regard, vigorously developing the real economy and avoiding 'de-realization' has become a key measure to alleviate current downward economic pressures and actively promote supply-side structural reform. An emphasis on enhancing the capacity of the financial sector to serve the real economy has been proposed to optimise the structure of industry and promote the transformation and upgrading of enterprises.
However, the financialization of enterprises remains a fundamental element in the current development of domestic enterprises. The statistics of sample companies listed in China from [2000][2001][2002][2003][2004][2005][2006][2007][2008][2009][2010][2011][2012][2013][2014][2015][2016][2017] show that the proportion of enterprises' allocation to financial assets exhibits a significant upward trend (Figure 1). The total average proportion of financial assets accounts for more than 18% of all assets, while the annual average proportion of financial assets is close to 20%. Moreover, empirical studies demonstrate the enterprise financialization substantially lowers the investment willingness of these enterprises. For example, Zhang and Zhang (2016) found that financialization, in conjunction with the mismatch effect between returns and risks of financial assets, can significantly suppress the industrial investment rate in China. Hence, the following question arises: what exactly is the motivation for the financialization of enterprises?
The existing literature has elaborated on the economic consequences of financialization for enterprises (e.g. Stockhammer, 2004;Demir, 2009;Liu, 2017;Tori & Onaran, 2018). The core conclusion drawn by these studies can be summarised as follows: given the existence of a significant profitability gap between financial assets and the real economy, managers have a greater tendency to act in a short-sighted manner and pursue short-term interests by allocating short-term financial assets in an attempt to secure profits, dividend distributions, share buybacks, and stock prices. Some literature explores the factors influencing the financialization of enterprises from an institutional perspective by examining the effects of credit constraint, risk constraint, and the nature of property rights (Farhi & Tirole, 2012;Wang et al., 2016Wang et al., , 2015. These authors conclude that the investment decisions made by enterprises regarding financial assets are restrained by the quality of the institution as seen from a macro perspective. In recent years, the literature has begun to focus on the motives behind corporate financialization. For example, Yiming Hu et al. suggest that financialization is primarily a 'reservoir' motive of enterprises based on precautionary savings, while the allocation of non-cash financial assets represents an 'alternative' motive to physical investment. From the perspective of economic uncertainty, Peng et al. (2018) argue that financialization in China's enterprises is more profit-driven rather than being motivated by precautionary savings.
The different conclusions drawn by Peng et al. (2018) regarding the motives behind enterprise financialization may be due to the omission of exogenous institutional variables from the analysis and the failure to effectively identify the causal relationships between them. Transnational evidence shows that determining the effect of financialization on economic development requires an examination of specific institutional characteristic variables (Arestis & Demetriades, 1997). Institutional characteristics can substantially impact both financialization and the development of the economy. Meanwhile, the institutional characteristics of micro-enterprises are ultimately reflected by the institutional costs they incur, causing exogenous shocks to the investment decisions of enterprises to some extent. In light of the above, it is of theoretical and practical significance to explore the strategic motives behind the financialization of enterprises in conjunction with the exogenous system in order to comprehend the current excessive financialization of enterprises and ascertain the institutional incentives for economic de-realisation. The minimum wage system is a specific example of costs to the business that are 'designed from the top down by the political power structure of the ruling community, imposed on society and implemented' and are somewhat exogenous (Ke & Shi, 2002;Sun et al., 2013). As an essential policy instrument of the labour market, the minimum wage is a national regulation set by the government to guarantee the primary income of employees, as well as a vital element of the transitional state system. Studies show that the minimum wage system not only affects employment, resident income, factor market price (Autor et al., 2016;Neumark et al., 2005) but also increases the employment costs of enterprises, thus squeezing the profits of corporations and increasing the uncertainty of their decision-making behaviours (Cuong, 2017;Magruder, 2013). In particular, the implementation of the Labour Contract Law in 2008 further strengthened the protection of workers' rights and interests; in turn, this significantly strengthened the implementation of the minimum wage system, leading to a significant increase in the cost of illegal and non-compliant employment (Pan & Chen, 2017). The rise in labour costs caused by strengthened minimum wage standards will thus directly impact the short-term earnings of enterprises and increase the short-term operational risks of tangible assets. In order to obtain the established returns and diversify investment risks, enterprises will tend to not only adopt cost reduction measures (such as reducing human capital input), but also take advantage of the difference in return rate between financial assets and tangible assets to conduct portfolio management, with a consequent impact on corporate financial investment decisions.
Based on this, using a sample comprising the listed manufacturing companies in China's A-share stock market from 2000-2017, this study examines the exogenous institutional cost impacts of the minimum wage system on the financialization of enterprises. Specifically, this study answers the following two questions: first, whether and how does the minimum wage affect the financialization of enterprises? Second, does the minimum wage system promote the financialization of labour-intensive enterprises differently from non-labour-intensive enterprises? The findings of this paper conclusively support the inference presented above, namely that a more stringently enforced minimum wage system significantly promotes enterprise financialization, and further concludes that this promotion effect is increased in enterprises with higher labour intensity and a more significant gap between the profitability of tangible and financial assets.
The potential academic contributions of this paper are as follows. First, this paper identifies the causal relationship between cost shocks and enterprise financialization using the exogenous event of staggered adjustment of China's minimum wage at the regional level. Compared to previous literature that has examined the economic consequences of enterprise financialization -e.g. enterprise innovation (Seo et al., 2012;Hongjian Wang et al., 2017), enterprise performance (song & Lu, 2015), and capital accumulation (Gehringer, 2013;Orhangazi, 2008;Stockhammer, 2004) -this paper expands and deepens the research related to the motivations behind enterprise financialization in China. The conclusions of this study differ from those of Yang et al. (2019). Specifically, we provide a complementary and interpretative approach to the controversies within theory on the motives of enterprise financialization raised by Peng et al. (2018). We further add new empirical evidence to explain the myth of China's declining industrial investment rate (Zhang & Zhang, 2016).
Second, this study enriches and expands the research related to the economic consequences of the minimum wage system on microeconomic agents. Previous studies have revealed the economic consequences of the minimum wage system from different perspectives, including structural adjustments to human capital and its costs, firm profits or firm exit rates, and the behaviour of exporting enterprises (Sun et al., 2013;Xu & Wang, 2016;Zhao et al., 2018). This paper provides further information on the economic impacts of the minimum wage system from an enterprise financialization perspective. We find that stringently enforced increases to the minimum wage significantly promote the financialization of enterprises, and moreover that this promotion effect is more substantial for labour-intensive enterprises. These findings enrich the literature pertaining to the economic consequences of an increased minimum wage on companies' investment behaviour and expand the boundaries of theory discussing the minimum wage system's impact on the investment decisions of microeconomic agents.
Third, this study reveals the specific mechanisms driving the motivation to engage in enterprise financialization in China. We find that the exogenous cost impact caused by minimum wage increases intensifies the profit margin gap between real capital and financial assets. The conclusions of this study may have significant implications for the identification of the mechanisms driving the enterprise financialization in China.

Theoretical analysis and hypothesis development
Considered broadly, the cost effect of an increased minimum wage drives up labour prices and leads directly to a relative decline in short-term returns on investment within business entities. These findings are supported by both the domestic and international literature. In a study of the implementation of the minimum wage system in the UK in 1999, Draca et al. (2011) find that raising the minimum wage directly squeezes the profit margins of enterprises, especially those with stronger market power. Gan et al. (2016), in a study of large and medium-sized domestic manufacturing enterprises from 1999-2007, also show that a 10% increase in minimum wage results in a 0.9% decline in the profitability and size of exported goods. In terms of micro-mechanisms of action, minimum wage increases change the demand for and prices of different levels of skilled labour; this not only directly increases the relative price of low-skilled labour (Rebitzer & Taylor, 1995;Shuang Ma et al., 2012), but also increases the demand for higher-skilled labour, consequently raising the price of hiring higher-skilled workers (Slonimczyk & Skott, 2012). Coupled with the fact that the minimum wage system balances the bargaining power of employers and employees in the labour market (Flinn, 2006) and the possible 'climbing effect'  on the distribution of income among employees, this ultimately leads to an overall upward trend in the enterprises' employment costs. At present, China's demographic dividend is gradually disappearing (Cai, 2010), while the minimum wage standard in all regions is expected to steadily increase, suggesting that an incremental increase in labour costs will be an established trend for companies in the future.
Faced with the above, rational entrepreneurs will adopt various measures to cope with the short-term revenue impact of minimum wage increases from a multi-dimensional perspective, including costs and benefits. In response, companies will adopt strategic measures to hedge or mitigate the incremental labour costs resulting from the minimum wage increase. Studies have shown that, on the one hand, an increase in the minimum wage induces enterprises to adjust and optimise their existing human resources allocation and its extended costs; typically, this involves downsizing and cutting the relative levels of payroll expenses, such as by reducing the size of the labour force and tending to dismiss male employees or employees without social insurance (Nguyen, 2012), freezing salary increases for employees, extending workers' overtime hours and reducing nonwage benefit expenditures (Jing and Gunderson, 2015), laying off low-skilled and temporary workers, and eliminating employee welfare expenditures such as pensions and insurance (Long & Yang, 2016). On the other hand, enterprises may also reduce their capital investment in human resources (Xiahai Wei et al., 2018), increase the marginal output rate of employees (Riley & Bondibene, 2017), increase firm productivity (Liu, Jun, et al., 2017), reduce the relative unit marginal cost of products and suppress their capital investment rate in human resources (Haepp & Lin, 2016), and reduce the number of new apprentices in training and the funding allocated thereto (Hara and Hiromi, 2017).
In addition to the 'cost-bearing' perspective, enterprises also respond on the 'revenue' dimension; i.e. the extrusion effect of gradually increasing labour costs on enterprises' short-term revenue, which prompts rational entrepreneurs to seek short-term arbitrage opportunities by making portfolio investment decisions under the existing resource constraints. In fact, the theories of enterprise financialization motivation discussed by Peng et al. (2018) address resource allocation competition between investment projects, i.e. how the resources available to enterprises are allocated between financial assets and real assets. If financial assets and real assets are considered as two separate types of 'investment projects', then the reasonable allocation of available resources between these two types of 'investment projects' represents an enterprise's 'portfolio' decision. Classical portfolio theory (Markowitz, 1952) suggests that a 'rational person' will always seek to maximise the expected return for a given level of risk or minimise the risk for a given level of expected return. The cost effect of minimum wage adjustment directly reduces the profitability of real assets and changes the relative profitability relationship between investment in real assets and investment in financial assets. Under the premise of constant relative investment risk, enterprises are incentivised to optimise the portfolio of both types of assets and engage in short-term arbitrage to cope with the direct short-term return shock brought about by a minimum wage increase.
From the perspective of return on assets, the current level of profitability for real assets is both generally low and much lower than the return on financial assets in the same period. The statistical results of listed companies' data from 2000-2017 ( Figure 2) show that the total average operating profit margin of real assets is only 4.77%, while the average annual profit margin largely hovers between 4% and 5%; by contrast, the total average operating profit margin of financial assets in the same period is as high as 11.38%, or 2.39 times higher than that of real assets (especially after 2006), while its average annual profit rate is over 13%.
In terms of the level of asset risk, the relative risk of financial assets in China is currently low. On the one hand, financial assets offer superior liquidity and ease of investment; this can not only effectively support the maintenance (or expansion) of the physical boundaries necessary when companies need funds, but also facilitates adaptation to economic uncertainty to a certain extent, thereby effectively reducing the business risks faced by companies. On the other hand, studies have shown that the volatility of domestic financial asset yields is currently lower compared to real asset yields (Zhang & Zhang, 2016). As risk hedging instruments, financial assets also have the function of risk diversification. Based on this, given an existing risk, entity enterprises tend to construct investment decisions that will maintain (or expand) a combination of necessary physical boundaries and highermargin asset items driven by the short-term arbitrage motive, with the goal of improving corporate performance in general, such that increased allocation of funds to financial assets with higher returns naturally becomes an appealing choice for enterprises.
In order to present the above theoretical logic more visually, the following Figures 3 and 4 have been developed.
Based on these considerations, we present the following hypothesis: Hypothesis 1: A higher minimum wage will significantly increase the financialization of enterprises.
Following research hypothesis 1, the decline in the rate of return on corporate real assets due to minimum wage adjustments will lead to an increase in the profitability gap between real and financial asset investments. Given the objective fact that there is currently a large profitability gap between the rate of return on real assets and financial assets, rational entrepreneurs working in the portfolio theory framework are strongly motivated to use financial assets with higher yields for arbitrage activities. This suggests that the size of the gap between the return on real assets and the return on financial assets directly affects the magnitude of the firm's incentive to allocate resources to financial assets. For example, Song and Lu (2015) find that the level of the firm's own performance directly affects the strength of the enterprise's incentive to allocate resources to financial assets, since the performance of the enterprise itself directly determines the profitability gap between the enterprise and its financial assets . Specifically, when the margin gap between real and financial assets is small, the exogenous shock brought about by a minimum wage increase will have a weaker impact on the margin gap between the two, such that enterprises will have a weaker incentive to allocate resources to financial assets. Conversely, when the   margin gap between real and financial assets is large, the exogenous shock caused by an upward minimum wage adjustment will have a stronger impact on the margin gap between the two, meaning that enterprises will have a stronger incentive to allocate resources to financial assets. Based on the above analysis, we propose the following research hypothesis: Hypothesis 2: The greater the gap between the margin of real and financial assets, the more significant the contribution of a minimum wage increase to the degree of enterprise financialization.
Generally speaking, compared with non-labour-intensive enterprises, labour-intensive enterprises both employ more labour and rely more on labour, such that labour costs make up a higher proportion of these enterprises' total costs. Therefore, a higher minimum wage will generally increase the labour costs of labour-intensive enterprises to a greater extent. In an empirical study using data from above-scale manufacturing enterprises from 1998-2007, Ma et al. (2012) find that a 10% increase in the minimum wage occurred, with the average wage of manufacturing enterprises raised by 0.4%-0.5%, and that the minimum wage pushes up the average wage level to a greater extent in labour-intensive or low-capital-per-capita enterprises. Jiang Wei et al., among other studies, show that the implementation of China's Minimum Wage Regulation has a significantly larger weakening effect of cost stickiness in labour-intensive industries and enterprises with lower wages per employee. This implies that labour-intensive enterprises are more exposed to earnings shocks due to minimum wage increases than non-labour-intensive enterprises; this further widens the margin gap between financial and non-financial assets of enterprises in labour-intensive industries, resulting in stronger incentives for labour-intensive industries to allocate financial assets for short-term gains. Based on the above analysis, we propose the following research hypothesis: Hypothesis 3: An increased minimum wage promotes the financialization of labourintensive enterprises significantly more than is the case for non-labour-intensive enterprises.

Sample selection and data sources
Our sample includes the CSMAR database of Chinese listed companies from 2000-2017. We apply the following restrictions to the sample: (1) we exclude non-manufacturing listed companies from the sample; (2) we exclude companies with a ratio of liabilities higher than 100% from the sample; (3) we exclude listed firms in receipt of Special Treatment or Particular Transfer; (4) we exclude observations missing data values necessary to calculate the main regression variables. Our final sample consists of 17,533 observations. The minimum wages in each region were compiled by hand. To control for possible outliers, all continuous variables are winsorised at the 1% level at both tails of their distributions.

Research design and variable definition
In line with Peng et al. (2018), to test the research hypotheses, we estimate the following regression model: where Fratio is the explanatory variable that indicates the degree of financialization of enterprises. The existing literature mainly measures financial assets allocated and their returns. Drawing on the research results, enterprise financialization can be defined as the proportion of financial assets to total assets. It should be noted that, given the difference between cash-based financial assets and other types, we do not include the former in the present analysis. Due to adjustments made to the accounting system of the Ministry of Finance in 2007, the data from 2000-2017 are divided into two segments for calculation: specifically, for 2000-2006, the degree of enterprise financialization = (short-term investments + interest receivables + long-term bond investments)/total assets; moreover, for data from 2007-2017, the degree of enterprise financialization = (trading financial assets + short-term investments + financial derivatives + interest receivables + bought return financial assets + available-for-sale financial assets + held-to-maturity investments + long-term receivables)/total assets. Robustness tests are also performed for enterprise financialization, including cash-based financial assets and financial income measures.
The explanatory variable, minimum wage, is the upper limit of the monthly minimum wage set by the government at the national prefecture level. According to Ma et al. (2017), its natural logarithm is taken as the proxy variable (Lnwage) for the minimum wage system; other defined methods are also used for robustness testing. Their data sources were obtained by manually collecting the minimum wage standards from 2000-2017 in prefecture-level administrative regions across the country.
In model (2), Gap_dum denotes the dummy variable of the difference in return between real asset and financial asset. The observations with the difference above the sample median are recorded in the 'high margin gap' subgroup; all observations below the sample median are recorded in the 'low margin gap' subgroup, where the margin gap is defined as the difference in return between non-financial assets and financial assets. The return of financial assets is the ratio of profit from the financial channel to financial assets, and the return of non-financial assets is the ratio of profit on non-financial assets to non-financial assets. According to Zhang and Zhang (2016), profit from the financial channel is obtained via the sum of net investment income, gain or loss on fair value changes, and net exchange gain of non-financial enterprises after deducting their investment income in associates and joint ventures; moreover, operating profit minus profit on financial assets is deemed profit on non-financial assets. Financial assets are defined in the same way as above, while non-financial assets are determined by deducting financial assets from total assets. The income data above is obtained from information related to the income statement, while the asset data is obtained from information related to the balance sheet.
In model (3), Lab_dum denotes the dummy variable of whether the enterprise belongs to a labour-intensive industry. The median of the labour-intensive index is used as the critical value to divide the sample into two groups. If an enterprise's labour force indicator is higher than the median, it is recorded as a labour-intensive enterprise and assigned a value of 1; otherwise, it is classified as a non-labour-intensive enterprise and a value of 0 is recorded. As for the construction of labour-intensive indicators, this is performed by drawing on the work of Pan and Chen (2017). In more detail, the median number of employees per unit (number of employees/total assets) of enterprises in an industry is used to determine the median of that industry (Lab1), from which the median of all industries is calculated (Lab2). If the median of an industry (Lab1) is higher than the median of all industries (Lab2), the industry is deemed labour-intensive, the enterprises belonging to that industry are classified as labour-intensive industries, and Lab_dum takes the value of 1; otherwise, it takes the value of 0.
CV represents the control variables. Drawing primarily from the work of Peng et al. (2018), we introduce the variables of capital structure (Lev), firm size (Size), profitability (ROA), and sales revenue growth rate (Growth). Among these, Lev represents the gearing ratio, defined as the ratio of total liabilities to total assets at the end of the period; Size is defined as the natural logarithm of total assets at the end of the period; ROA is characterised by the return on total assets, defined as the ratio of net profit to total assets; finally, Growth is calculated as the ratio of the difference between current and prior period sales revenue to the current period sales revenue. Moreover, some prior works (Liu, 2017;Song & Lu, 2015) have also investigated the impact of enterprise size or corporate governance on investment behaviour, including enterprise financialization; thus, further variables are introduced, such as equity concentration (Stk). Here, enterprise age (Fage), which refers to the period during which the enterprise is listed, is expressed as the natural logarithm of the difference between the year of examination minus the year in which the enterprise is listed plus one. Stk represents the level of corporate governance and is defined as the percentage of shares held by the top five shareholders. Variables such as the nature of enterprise property rights, regional economic development level and financial status are also controlled: here, the nature of enterprise property rights (Property) is a dummy variable that takes a value of 1 if the sample enterprise is a stateowned enterprise and 0 otherwise; the regional economic development level (GDP_per) is the natural logarithm of the GDP of the regional city in which the sample enterprise is located; the fiscal status (Fis_rev) is the natural logarithm of the fiscal revenue of the region and city where the sample enterprises are located. In addition, the year, industry (taking three digits according to the 2012 SEC industry code), prefecture-level urban areas and company fixed effects are also controlled. Table 1 presents the summary statistics for the main variables. The mean value of enterprise financialization (Fratio) is 0.021; this indicates that the average ratio of financial assets to total assets for sample companies reaches 2.10%, which exceeds the median value of 0.10%. This in turn indicates that more companies in listed companies have made financial asset allocations. Moreover, its maximum and minimum values are 0.301 and 0.000, respectively, indicating that the degree of financialization varies substantially between samples. This also demonstrates that enterprise financialization is a widespread phenomenon that is of great concern. For the mean value of the minimum wage (Lnwage), taking the natural logarithm yields a value of 6.840, which means that the average monthly minimum wage in the cities where the sample enterprises are located is 940 yuan. The median of this value is 7.003, closer to the mean, indicating that the sample is basically normally distributed; moreover, its maximum value (7.692) and minimum value (5.521) show that there is a significant gap in the minimum wage in each region. Table 2 presents the results for the benchmark regression based on Equation (1), i.e. the strength of minimum wage standards and the financialization of entity enterprises. Model (1) is the original regression result of the strength of minimum wage standards (Lnwage) and enterprise financialization (Fratio), while model (2) is the regression analysis of the strength of minimum wage standards and enterprise financialization based on model (1) after controlling for year, industry, region and firm fixed effects. We further introduce the control variables on the basis of model (1) in model (3). The regression results in Table 2 all indicate that the strength of minimum wage standards is positively correlated with enterprise financialization within the 5% significance level. This finding provides preliminary evidence that the strength of minimum wage standards significantly promotes enterprise financialization, which supports research hypothesis 1. In economic terms, on average, each increase of standard deviation (0.553) in the logarithm of the minimum wage (Lnwage) results in an increase in the level of enterprise financialization (Fratio) equal to 14.378% of the standard deviation (= 0.013*0.553/0.050).   Table 3 show that the regression coefficient of the high profitability gap group is 0.017 (t-value of 2.38), indicating significant positive correlation at the 5% statistical level, while the regression coefficient of the high margin gap group is −0.011, which is not significant at the 10% statistical level. The difference between the coefficients of the two groups is significant at the 5% statistical level, indicating that the greater the margin gap between the return on real assets and the return on financial assets, the more significant the promotional effect of the strength of minimum wage standards on the financialization of enterprises. Thus, research hypothesis 2 is supported.

Baseline results
In order to test research hypothesis 3, we further run grouped regressions based on whether or not the sample is labour-intensive. The results of these tests are reported in Table 4. From these results, we can determine that the coefficient of minimum wage (Lnwage) is significantly positive at the 10% statistical level in the labour-intensive sample group, but not significantly positive at the 10% statistical level in the non-labour-intensive sample group; moreover, the difference between the coefficients of the two groups is significant at the 1% statistical level, indicating that relative to non-labour-intensive Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
enterprises, a higher minimum wage more strongly promotes financialization among labour-intensive enterprises than among their less labour-intensive counterparts. Thus, hypothesis 3 is supported.

Mechanism test
To reveal the mechanism by which the implementation of a stronger minimum wage system impacts the extent of enterprise financialization, this paper will further examine the relationship between the strength of minimum wage standards and enterprise financialization from the perspective of average regional wage, cost-shifting ability, and the implementation of the Labour Contract Law, due to their direct influences on the extent to which the minimum wage affects profit shocks. The details are as outlined below.

Average wage level
The extent to which the strength of minimum wage standards impacts enterprise behaviour is influenced by the current average wage levels of enterprises. When the average wage level of enterprises is low, an increase in minimum wage will produce a more substantial increase in the current average wage level of enterprises, thus squeezing the corporate revenue space more significantly. Consequently, enterprises with relatively low average wage levels have more incentives to allocate resources to more profitable financial assets in order to hedge against potential profit loss. The extent of this squeeze effect is also higher for labour-intensive enterprises, where labour costs account for a higher share of total costs. Conversely, when the existing average wage level of enterprises is higher, the cost effect of a higher minimum wage produces a relatively moderate squeeze on enterprises' profits; as a result, the effect of a higher minimum wage on the financialization of enterprises is diminished. Based on the above, in order to further assess the influence of enterprises' existing wage levels on the relationship between the strength of minimum wage standards and financial assets, the ratio of the enterprise's per capita wage level to the average minimum wage standard of the region (Avg_MW index) is used as the distinguishing criterion of the average wage level. Subsamples differentiated by the size of the median of the complete sample's Avg_MW index were subjected to regression analysis, and the empirical results are presented in Table 5. The results of the test show that the regression results of the group with lower average wages are significantly positively correlated at the 5% statistical level relative to the standard ratio of average local minimum wage; moreover, the regression coefficient of the higher Avg_MW index group is negative and not significant at the 10% statistical level, while the test of the difference between the regression coefficients of the two sample groups is significant at the 1% statistical level. This suggests that the strength of minimum wage standards has a significantly stronger facilitating effect on the financialization of lower-average-wage enterprises relative to those with higher average wages, further supporting the rationale articulated in this research.

Cost-shifting capacity
In this subsection, we investigate how cost-shifting capacity affects the relationship between the minimum wage system and enterprise financialization. Profit is the difference between an enterprise's revenue and its costs. The cost effect of the strength of minimum wage standards directly increases the production costs of enterprises. If enterprises can afford to transfer part or all of these additional cost components to consumers under the premise of established demand, this also becomes a crucial way in which enterprises can hedge against increased labour costs and cope with the shrinkage of profits. Of course, this is largely dependent on enterprises' bargaining power in the market. A lower degree of constraint in the product market indicates that enterprises have more bargaining power in the market, meaning that they can better afford to pass on the additional labour costs to consumers. A study by Lu et al. (2017), using a sample of Chinese listed companies, indicates that the cost effect of the strength of minimum wage standards leads to a reduction in enterprises' book profits; thus, enterprises with a poor ability to pass on their costs to the market have more incentives to engage in surplus manipulation and raise the level of their book earnings. As a result, we can expect that enterprises that find it more difficult to pass on labour costs to consumers will have a greater incentive to financialize in the face of rising labour costs due to an increase in the minimum wage.
In line with the above, the degree of product market constraint is introduced to further test the extent to which enterprises' market bargaining power influences the relationship between the strength of minimum wage standards and financial assets. The indicator design draws on the approach adopted by Lu et al. (2017) to construct the Lerner index (L) to measure the magnitude of an enterprise's market power, then subtracts the industry average Lerner index (weighted by sales from the individual enterprise's Lerner index) to obtain an indicator of the degree of product market constraint (MC): TQ=MainP ¼α 0 þα 1 Lnwageþα 2 Fratioþα 3 Lnwage�Fratio In the above, i denotes a firm, j denotes an industry, and t denotes a year. Moreover, Sale i,t represents the sales revenue of enterprise i in year t; Cost i,t represents the cost of sales of enterprise i in year t; MC i,j,t measures the degree of product market constraints on enterprise i in industry j in year t; L i,j,t measures the enterprise's Lerner index as calculated by model (3); finally, Weight i,j,t measures the share of sales revenue of enterprise i in the total sales revenue of industry j in year t. Table 6 reports the results of the grouped regressions for the magnitude of product market constraints (grouped by median). It can be readily determined that the regression results of the group with high product market constraints are positive and not significantly significant at the 10% statistical level, while those of the group with high product market constraints are significantly positive at the 5% significance level; moreover, the test of variance of the regression coefficients is significant at the 5% statistical level. This indicates that the strength of minimum wage standards more significantly promotes the financialization of enterprises when their cost-shifting ability is weaker.

Labour laws
The efficiency of the execution of the strength of minimum wage standards affects the implementation of the minimum wage within enterprises and its extent, which in turn yields different degrees of impact on enterprise earnings. Currently, some regional variability can be observed in the regulation of minimum wages nationwide (Ding, 2010;Ye et al., 2015), which may lead to different minimum wage cost effects across regions, enterprises, and enterprises. In this regard, the enforcement of the minimum wage directly affects the labour costs of enterprises, resulting in changes in their investment decision behaviour. It can be expected that the stricter the degree of minimum wage enforcement, the more likely enterprises will be to increase their allocation to financial assets in order to compensate for the loss of corporate earnings resulting from a possible increase in labour costs. Among existing labour regulations, China's Labour Contract Law, which came into force in 2008, has greatly strengthened protections for ordinary workers while objectively enhancing the enforcement of the minimum wage system in enterprises, especially labour-intensive enterprises with a wide impact. We predict that labour-intensive companies will tend towards increased financialization following the implementation of the Labour Contract Law. Accordingly, in order to further test the impact of more stringent minimum wage system implementation on enterprises' labour costs and earnings, and subsequently on enterprise financialization, we divide the regression test into two groups: namely, before and after the implementation of the Labour Contract Law (using 2008 as the boundary), where Law = 0 before the implementation of the Labour Contract Law and Law = 1 after the implementation. The regression results are reported in Table 7. From this, we can determine that the regression coefficient of the sample before the implementation of the Labour Contract Law is not statistically significant at the 10% level; by contrast, the regression coefficient after the Labour Contract Law was implemented is statistically significantly positive at the 1% level, indicating that the implementation of this legal framework notably strengthens the exacerbating effect of a more stringently enforced minimum wage system on enterprise financialization.

Endogeneity
There may be unobservable variables or omitted variables affecting the relationship between the strength of minimum wage standards and enterprise financialization, meaning that the regression model may produce biased results. Alternatively, the financialization of enterprises may exacerbate the uneven distribution of income among enterprise employees, causing governments at all levels to raise the local minimum wage to eliminate this income inequality; i.e. there may be an endogeneity problem between the strength of minimum wage standards and enterprise financialization. To address this potential endogeneity problem, we implement the processing steps outlined below. First, the identification strategy of the 2SLS instrumental variable method is used to more accurately assess the impact of a stronger minimum wage system on the financialization of enterprises. Drawing on existing research results Zhao et al., 2018), this paper sets 'the average minimum wage in the same province other than the region' (denoted as IV_wage) as an instrumental variable. Generally speaking, resource endowments, cultural background, geographic location and economic development levels within a certain province tend to be relatively similar. Moreover, within the framework of a unified provincial-level political system, the minimum wage rates in each region of the province tend to be highly correlated. Furthermore, governments in other regions of the same province typically do not consider the extent of financialization of individual enterprises when setting minimum wage rates, meaning that the exogeneity hypothesis is again satisfied. Table 8 reports the regression results of the instrumental variables method. From these results, we can conclude that the instrumental variables in the first stage are significantly positive at the 1% statistical level, while the strength of minimum wage standards (Lnwage) in the second stage is also significantly positive at the 1% statistical level, passing the weak instrumental variables test. This indicates that the results of the regression using the 2SLS instrumental variables method continue to support the basic research hypothesis. Second, drawing on the rationale of the research design developed by Geng et al. (2021), a difference-in-differences model is developed for endogeneity testing, which uses county administrative reorganisation in China as an exogenous event: If the company's region has experienced boundary adjustment and the minimum wage standard has been increased, this is recorded in the control group and Treat is assigned a value of 1, otherwise it is 0; moreover, the year in which the company's region experienced boundary adjustment (Post) is taken as the standard value 0. Take the value of the first two years (i.e. −2, −1) and the next four years (i. e. 1, 2, 3, 4); the meanings of the remaining variables are the same as above. Moreover, the control group sample is assigned to a new county with the same minimum wage as the original county for a placebo test. The regression results are reported in Table 9. As is evident, the coefficient of the Treat and Post interaction term in column (1) is significantly positive, while the placebo test in column (2) is not significant at the 10% statistical level, thus supporting the basic findings above. Third, the implementation of the Minimum Wage Regulation in 2004 was used as an exogenous event and a difference-in-differences (DID) test was conducted using data  Table 10 reports the results of these tests. Our findings show that the coefficient of the Lnwage*Post regression in column (1) is positive and significant Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
at the 5% statistical level, and moreover, that the coefficients of the triple interaction terms Lnwage*Post*Gap_dum and Lnwage*Post*Lab_dum in columns (2) and (3) are significantly positive; these findings are consistent with the results of the main test in the previous section, indicating that the findings of this paper still hold true after the DID model is applied. Fourth, we further mitigate the potential endogeneity problem by treating the control variables with a one-period lag. The results of these tests are reported in Table 11. From these results, we can determine that the findings on the role of minimum wage in exacerbating enterprise financialization remain valid, as do those suggesting that this phenomenon is more significant in cases of high profitability gaps and in labour-intensive industries.

Alternative explanations
In addition to the theoretical paths discussed above, there may be other alternative paths by which the strength of minimum wage standards affects the financialization of enterprises. For example, the strength of minimum wage standards raises the business risk of the enterprises but may also increase their risk-taking capacity, which would in turn Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
promote enterprise financialization. Moreover, it may be that the strength of minimum wage standards reduces the number of low-skilled employees and increases that of highskilled employees, thus improving the level of enterprise financialization. These findings may also be due to the improvement of business risk, resulting in a complementary relationship between financial assets and physical asset investment. To exclude these alternative explanations, the following tests are carried out in this paper. First, we examine the relationship between the strength of minimum wage standards and enterprises' level of risk-taking. Referring to the related research, the variable Risk_taking of an enterprise is measured by the volatility of enterprise earnings (ROA); here, ROA is the ratio of enterprise EBIT to ending assets, adjusted for the mean of the industry in which the enterprise operates, and then calculated by multiplying its standard deviation by 100 as a proxy for Risk_taking. The results of these tests are reported in Table 12. It can be seen that the strength of minimum wage standards (Lnwage) coefficients are not significant, indicating that the strength of minimum wage standards does not significantly promote enterprise risk-taking and thus eliminating the alternative 'risktaking' hypothesis.
Second, we control 'employee structure' (Emply_stru) to test the robustness of the promotion impact of the strength of minimum wage standards on enterprise financialization. Theoretically, an enterprise that has more highly skilled employees would enjoy Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. higher labour productivity. However, due to the insufficient disclosure of employee structure data of listed companies in China, it is here necessary to use labour productivity as a proxy for employee structure. Referring to the work of Li et al. (2018), labour productivity is defined as output per employee, which is equal to the natural logarithm of the ratio of Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. the sum of sales revenue and inventory changes to the total number of employees. The median is used as the boundary for classifying high and low groups for subgroup testing purposes. The results of these tests are reported in Table 13. We can determine that there is no significant difference in the contribution of the strength of minimum wage standards to the financialization of entity enterprises in different employee structure configurations. Third, we test whether the strength of minimum wage standards has a heterogeneous facilitating effect on enterprise financialization under different levels of business risk. Drawing on the findings of , we use the volatility of corporate earnings -i.e. the standard deviation of rolling values of return on net assets -to measure the magnitude of business risk (Oper_risking). Enterprises are then categorised into high and low business risk groups according to the volatility of their earnings. The results of group tests are shown in Table 14. It is reported that no significant differences exist in the coefficients of the strength of minimum wage standards with respect to enterprise financialization under different levels of operating risk. That is, the regression coefficients of the group with higher business risk and the group with lower business risk are similar, and both are significant at the statistical level of 5%. These findings indicate that business risk does not have a differential promoting effect on the relationship between the strength of minimum wage standards and corporate financialization, thereby eliminating the alternative explanation of 'business risk'. Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
Fourth, to investigate how the strength of minimum wage standards affects the degree of enterprise financialization before and after the Labour Contract Law, we need to eliminate possible impacts brought about by the financial crisis. Referring to Pan and Chen (2017), to control for the impact of the financial crisis, we classify our sample into export enterprises (Export = 1) and non-export enterprises (Export = 0), indicating the presence or absence of foreign market sales before the financial crisis respectively. The results of these tests are reported in Table 15. Results of the subgroup regressions reveal that the relationship between the strength of minimum wage standards and the financialization of enterprises is not affected by the Labour Contract Law, which eliminates the prospect of interference from the financial crisis to a certain extent.

Robustness tests
To improve the reliability and robustness of the conclusions drawn in this paper, we also conduct the following robustness tests. Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
First, we re-test the models with a reduced sample. Specifically, we selected samples from 2008 to re-test the benchmark model, as most financialization processes in China's enterprises began after 2008. The results of the above test are reported in Table 16. It can be seen that the coefficient of the strength of minimum wage standards (Lnwage) remains significantly positive in the complete sample in column (1), in the high-profitability-gap enterprises in column (2), and in the labour-intensive industry group in column (4). These results indicate that the conclusions of the study remain valid after compressing the sample interval.
Second, we re-test the models using different criteria for classifying labourintensive enterprises. The content above uses industry standards to distinguish between labour-intensive and non-labour-intensive enterprises. In this section, referring to the work of Ni and Zhu (2016), we further adopt sales revenue per capita (Persales) as the criterion for differentiating labour-intensive enterprises and conduct a robustness test, in which sales revenue per capita = natural logarithm of number of employees/natural logarithm of sales revenue. The results of the tests are reported in Table 17. As can be seen from the table, the minimum wage and the degree of enterprise financialization are positively correlated in both labour-intensive and non- Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.  labour-intensive enterprises. However, the regression coefficients of the labourintensive group are significantly higher than those of the non-labour-intensive group, indicating that the findings of this paper for which labour-intensiveness was classified according to enterprise standards still hold. Third, we use substitute variables to measure enterprise financialization. Referring to the work of Zhang and Zhang (2016), we adopt return on financial assets (Fratio_profit) as a substitute measurement of enterprise financialization. Fratio_profit = (financial channel Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. profit -operating profit)/| operating profit|. Specifically, financial channel profit is the sum of net investment income, fair value change profit and loss, and the net exchange income of a non-financial enterprise after deducting its investment income and joint ventures. The results of the above tests, reported in Table 18, show that the results of the complete sample and the group tests with margin gap and labour intensity are consistent with the expectations of the research hypotheses. Fourth, we replace the measurement of the strength of minimum wage standards. In our benchmark model, the strength of minimum wage standards is measured by the value of the natural logarithm of the minimum wage, which potentially introduces the risk that the influence of price factors will be ignored. Referring to , in this section, we take the natural logarithm of the monthly minimum wage adjusted by the consumer price index as a proxy variable for the strength of minimum wage standards (Lnwage_CPI) and conduct robustness testing. At the same time, inconsistencies between an enterprise's place of business and its place of registration might introduce measurement error for minimum wage-related factors. Accordingly, to further test the research hypothesis, we take the natural logarithm (Lnwage_office) of the minimum wage standard of prefectural cities where the offices of listed companies are located as another proxy Note: Standard errors are clustered at the firm level. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.
variable of the strength of minimum wage standards. The results of these tests are reported in Tables 19 and 20. We find that the coefficient of the strength of minimum wage standards is significantly positive in the complete sample, high margin gap group, and labour-intensive industry group, indicating that the findings remain valid following the substitution of the independent variables. Fifth, given that differences in the place of operation and incorporation of parent and subsidiary companies may cause further errors when measuring minimum wage-related factors, we conduct the tests using only the data of parent companies to validate our results. The results of this regression are reported in Table 21. It can be found that the coefficient of the strength of minimum wage standards remained significantly positive in the complete sample, high margin gap group, and labour-intensive industry group, indicating that the findings of the study still hold when the data of parent companies is used.

Additional tests
Previous studies have demonstrated that the financialization of entity enterprises has a range of negative economic consequences (e.g. Seo et al., 2012;Tori & Onaran, 2018;Zhang & Zhang, 2016). Based on the institutional settings of minimum wage adjustment, this paper further discusses the impact of enterprise financialization brought about by minimum wage increases on firm value and primary business performance using the following model: In the above equation, TQ represents Tobin's Q, calculated as the ratio of market capitalisation to total assets, while MainP is the primary business performance of the enterprise. In line with Conghui Hu et al. (2015), MainP = (operating profit -investment income -change in fair value of trading financial assets + gain obtained or recognised from associates or joint ventures)/total assets. The meanings of the remaining variables are consistent with the content above.
The regression results of the impact of enterprise financialization caused by minimum wage increases on enterprise performance are reported in Table 22. As can be seen from the table, in columns (1) and (3), enterprise financialization (Fratio) is negatively correlated with enterprise value (TQ) and main business performance (MainP), which are significant at the statistical levels of 1% and 5%, respectively. Consistent with previous literature, these results indicate that financialization of entity enterprises damages enterprise performance. Minimum wage adjustment variables are further added to models in columns (2) and (4). It is reported that the regression coefficient and P value of the interaction term between Fratio and Lnwage are −0.520 (p <0.10) and −0.026 (p <0.01), respectively. These results indicate that an increase in the minimum wage standard intensifies the value-damaging effect of financialization in the short term.

Research conclusions and implications
Using the data of listed manufacturing companies in China's A-share stock market from 2000-2017, this paper empirically examines the impact of the minimum wage system on the financialization of entity enterprises. From the perspective of institutional costs, this paper reveals the strategic motivation behind enterprise financialization: that is, the financialization of enterprises is an investment behaviour targeted at pursuing profits and buffering the increasing cost of exogenous institutions while also balancing the return of financial assets and maintaining (or expanding) the necessary entity boundary. The empirical results show that increases in minimum wage standards significantly promote financialization among entity enterprises. Further tests reveal that this promotional effect of the minimum wage system on enterprise financialization is stronger for labour-intensive enterprises. Moreover, this effect is more substantial under the following circumstances: (i) when a smaller gap exists between an enterprise's average wage and the local minimum wage; (ii) when an enterprise finds it more challenging to pass on costs to the market; (iii) following the implementation of the Labour Contract Law. Further investigation reveals that a minimum wage increase will worsen the restraining effect of financialization on firm value. These findings remain robust after endogeneity and robustness tests.
The findings of this paper have several potential policy implications. First, it is necessary to establish and implement policies at a national level to curb excessive speculation in the capital market: specifically, by increasing the institutional costs of enterprises deviating from the real economy, as well as by guiding enterprises to actively pursue transformation and upgrading under the circumstances of a 'costbenefit' trade-off. Second, the government should reduce enterprises' overall institutional costs by strengthening the mechanism by which the market allocates resources and establishing a benign institutional operations platform. In addition, the government could lift the institutional incentive to engage in industrial structure upgrading, lay a solid foundation in support of the rational return of the pragmatic entity and avoid the 'disengagement from reality' of entity enterprises. Finally, when formulating relevant policies or systems, governments at all levels should comprehensively consider the characteristics of regional institutions, the local economic structure, and the specific development level. Ensuring a match between policy and regional development is vital for both the success of policy implementation and the reduction of policy costs driven by blind comparison between regions at different levels.

Disclosure statement
No potential conflict of interest was reported by the author(s).