Research on the Influence of Enterprise Financialization on the Accuracy of Analysts’ Forecast -- Mediating Effect based on Fixed Asset Intensity

. This paper built a fixed effect model to study the influence of financialization of enterprises on the accuracy of analysts' forecasts and the mediating effect of fixed asset intensity based on the data of Chinese A-share listed companies from 2016 to 2020. The results showed that the rise in degree of financialization improved the accuracy of analysts' forecasts by reducing the intensity of fixed assets. Taking a further consideration into the influence of the property rights of enterprises, it found that in the state-owned enterprises with rich financial resources, the degree of financialization had little influence on the accuracy of analysts' forecasts. However, in non-state-owned enterprises with large financing constraints, the degree of financialization had a more significant effect on the accuracy of analysts' forecasts. In general, this paper expanded the research situation of enterprises’ financialization on the accuracy of analysts’ forecast and provided a certain empirical support basis for the transformation of enterprises to financial investment, which had a certain reference significance for promoting the long-term development of China's economy.


Introduction
In recent years, the trend of financialization of Chinese enterprises had become more and more intense and transformed them from substantial to fictitious. Existing studies showed that financialization of enterprises was a double-edged sword. On the one hand, enterprises could avoid liquidity risks by allocating more financial assets (Demir, 2009), which could also guarantee the production and operation of their main businesses and promote high-quality development of the real economy. On the other hand, if the return on financial assets was higher than that of industrial investment, enterprises' pursuit of financial profits might squeeze physical investment such as fixed assets (Hu Yiming et al., 2017; Zhang Chengsi and Zhang Butan 2016). It was not conducive to the improvement of production efficiency and long-term healthy development of enterprises. Chang et al. (2016) demonstrated that enterprises allocating financial assets provided economic complexity to their financial reports, leading to the decline of the accuracy of analysts' forecasts. But the introduction of relevant accounting standards helped analysts to improve the accuracy of their forecasts. However, there were few direct studies on the influence of financial degree of enterprises on the accuracy of analysts' forecasts, and few studies on the influence mechanism of financial degreed on the accuracy of analysts' forecasts, which provided an opportunity for the research of this paper.
Therefore, this paper took the data of Chinese A-share listed companies from 2016 to 2020 as samples and by using empirical analysis, it's found that financialization of enterprises helped to alleviate the information asymmetry between companies and analysts and improved the accuracy of analysts' forecasts. Also, it showed the intensity of fixed assets had a certain mediating effect on the relationship between them.
The contributions of this paper were as follows. First, this paper filled the gaps in existing research fields and explored the impact of financialization of enterprises on stakeholders from the perspective of analysts. Second, it provided some practical support for enterprises transforming to financial investment and the transformation was more conducive to ease the information asymmetry between enterprises and the outside world. Third, the financialization of enterprises might reduce the intensity of their fixed assets and had a tendency of being distracted from their intended purpose, which might inhibit the development of the real economy to some extent. Therefore, the relevant national departments still needed to carry out reasonable supervision on the financialization of enterprises.

Financialization of Enterprises and Accuracy of Forecast
Because financial assets had strong liquidity (Hu Yiming et al., 2017), it helped to diversify risks (Peng Yuchao et al., 2018) and enhanced financing ability (Yang Xingquan et al., 2017). Financialization of enterprises meant that enterprises conducted more investment activities and allocated more financial assets, which was beneficial for a better presentation of the company's finance. Thus, managers had higher enthusiasm to deliver real information of the enterprise (from and ruby leung, 2008), reduced the information asymmetry. Analysts' forecast accuracy was also increased. (Fang Junxiong, 2007;Bai Xiaoyu, 2009). Therefore, this paper proposed hypothesis H1: H1: Financialization of enterprises promoted the accuracy of analysts' forecasts.

The Mediating Effect of Fixed Asset Intensity
Different from the open market of financial assets, the management of fixed assets was basically controlled by enterprises themselves. In addition, existing enterprises had insufficient internal control of fixed assets, and there were many unreasonable situations in the acceptance and maintenance of fixed assets, leading to the discrepancy between the actual loss of fixed assets and accounting records (Cai Xiaofang, 2017). The financialization of enterprises reduced the intensity of fixed assets, which reduced the internal control errors of fixed assets, and increased the quality of enterprise information disclosure, thus helping to improve the predictability of analysts. Therefore, this paper proposed hypothesis H2: H2: Fixed asset intensity had a certain mediating effect on the relationship between enterprise financialization and forecast accuracy.

Data Source and Data Processing
In this paper, Chinese A-share listed companies from 2016 to 2020 were selected as research samples. All data used in the study were from the CSMAR database. The samples were screened according to the following principles; (1) Samples from the financial industry, insurance industry and real estate industry were excluded; (2) Samples that were ST (* ST) or PT during the sample period were excluded; (3) Samples with missing and abnormal data were excluded. According to the above principles, 9125 observations were obtained. In order to eliminate the influence of extreme observations on regression results, all continuous variables were subjected to a bilateral Winsorize tail reduction at the 1.5% level.

Explained Variables (MFbias)
Referring to the article of Wu Yanran et al. (2016) and considering the factors of changes in the equity capital of enterprises between the forecast period and the actual period and the level of stock prices of different enterprises, the analyst prediction accuracy (MFbias) was calculated as follows:

BCP Business & Management
, +1 represented the accuracy of analysts' prediction of t+1 year, , +1 was the average forecast value of earnings per share of different analysts for t+1 year, , +1 represented the actual earnings per share of t+1 year, , +1 represented the net assets per share of enterprise i in year t+1, and , represented the total capital stock of enterprise i in year t.

Explanatory Variables (Fin)
With Song Jun and Lu Yang (2015) as well as Du Yong (2017) as reference, the calculation formula of enterprises' financialization degree (Fin) was Fin = (trading financial assets + derivative financial assets + net loans and advances + net financial assets available for sale + net held-to-maturity investment + net investment real estate)/total assets.

Intermediary Variables (Fix) and Control Variables
This paper adopted fixed asset intensity (Fix) as the intermediary variable and used the ratio of fixed assets to total assets as its representation.

Model Building
In order to verify H1, this paper established model (3) to investigate the influence of the degree of financialization of enterprises on the accuracy of analysts' forecasts: Industry represented the fixed effect of the industry, which was divided according to the Guidance on Industry Classification of Listed Companies published by CSRC (revised in 2012). Year was the time fixed effect, which was used to control the common impact of unobserved time factors on enterprises. ε was the random error term.
In order to investigate the mediating effect of fixed asset density and verify H2 with the "threestep method", this paper constructed the following mediating effect model (4), (5) and (6):

Descriptive Statistics
The descriptive statistics of the main variables studied in this paper were shown in Table 1. It could be found that the fluctuation range of analyst forecast bias (Mfbias) ranged from -88.82 to 193.2 with a standard BCP Business & Management

FEMS 2022
Volume 19 (2022) deviation of 2.450, and its mean value of 0.0959 was positive, indicating that the accuracy of analysts' forecasts varied greatly among different samples, but they tended to be optimistic bias on the whole. The mean value of financialization of enterprises (Fin) degree was 0.0426, the variance was 0.0851, and the maximum value was 0.809, indicating that the overall financialization degree of the samples did not differ much. The minimum value of fixed asset intensity (Fix) was 0.000211, the maximum value was 0.954, and the variance of the sample was 0.158, indicating that the intensity of fixed asset intensity varied greatly among different samples.   t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

Correlation Analysis
The correlation analysis of the main variables studied in this paper was shown in Table 2. The correlation coefficients between the main variables were all significant at the level of 1%, which BCP Business & Management

FEMS 2022
Volume 19 (2022) 105 preliminarily supported H1 and H2 proposed above. The correlation coefficients of other control variables were less than 0.5, which preliminarily indicated that there was no serious collinearity among variables. t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01  Table 3 reported the test results of the basic assumptions of this paper. It could be seen from Column (1) that the regression coefficient value of Fin was -0.0263 and significantly correlated at the 10% level, which preliminarily indicated that financialization of enterprises helped to improve the accuracy of analysts' forecasts. As can be seen from Table 4, the mean and median of Fin were both positive, which meant that analysts' forecasts were more optimistic bias. In order to verify the robustness of the results, the above regression was performed after deleting the samples with negative Mfbias from the original samples. According to the results in Column (2), the regression coefficient value of Fin was still significantly negative and decreased, proving once again that financialization of enterprises promoted the accuracy of analysts' predictions. Therefore, H1 was valid. Table 5 reported the test results of the mediating effect of fixed asset intensified on the relationship between enterprise financialization and forecasting accuracy. First of all, column (2) showed the influence of Fin on Fix: it was found that there was a significant negative correlation between them, which was, the increase of the degree of financialization will have a significant negative impact on the degree of fixed asset density. Secondly, as shown in Table (3), after the addition of Fix on the basis of baseline regression, Fin was still highly significant, and the regression coefficient of Fix was -0.0325 and significant at 1% level. Finally, the indirect effect (_bs_1) and direct effect (_bs_2) of Bootstrap in column (4) were positive and negative, respectively. Overall, we could draw: financialization degree rise will reduce the intensity of fixed assets, thus promoting effect on analysts forecast accuracy, the intensity of fixed assets to the enterprise financialization and predictive accuracy relations existed certain intermediary function (mediation effect part of the total effect rate was about 7.80%), then H2 was valid. t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

Robust Test
In addition, this paper used the natural logarithm of total fixed assets to measure the intensity of fixed assets (lnPPE), and made regression estimation according to the original model to further verify the mediating effect of the intensity of fixed assets. According to the results in Table 6, not only was the Fin regression coefficient of column (2) -0.516 and significant at the 1% level, but the regression coefficients of Fin and lnPPE were still significantly negative after the addition of lnPPE in column (3) on the basis of the baseline regression. According to the test results of Bootstrap method in Column (4), the mediation effect was still highly significant. In conclusion, this paper once again proved that fixed asset intensity had a mediating effect on the relationship between enterprise financialization and forecast accuracy, and the mediating effect accounted for 23.54% of the total effect. t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

Heterogeneity Test
In order to study the influence of differences in property rights of enterprises on the relationship between financialization of enterprises and the accuracy of analysts' predictions, samples were divided into state-owned enterprises and non-state-owned enterprises according to property rights. The test results were shown in Table 7: (1) Fin regression coefficient of state-owned enterprises was 0.0161 and was not significant; The Fin regression coefficient of the non-state-owned enterprise group in Column (2) was still negative (-0.0393) and the significant degree increased. Column (2), (3), (4) and (5) were the test of the mediation effect of fixed asset intensity, and the results showed that the mediation effect was significantly established. To sum up, it could be found that in the highly centralized financial system dominated by state-owned commercial banks, credit allocation existed in the order of institutional principal and subordinate, and most domestic financial resources flew to state-owned enterprises (Song et al.,2011), resulting in greater financing constraints of non-stateowned enterprises than state-owned enterprises. In order to alleviate financing constraints, non-stateowned enterprises allocated more financial assets to enhance financing capacity, and the intensity of fixed assets decreased accordingly. At the same time, financial assets had a more transparent regulatory mechanism than fixed assets, which increased the quality of information disclosure of nonstate-owned enterprises. Therefore, the grouping results had significant heterogeneity, and the intermediary effect of fixed asset intensity still existed. t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

Conclusion
Based on the data of A-share listed companies from 2016 to 2020, this paper constructed a fixed effect model to test the internal relationship between financialization of enterprises and analyst forecast accuracy and the mediating role of fixed asset intensity. The paper also analyzed the heterogeneity of corporate property rights on the above relationship. The results showed that: (1) financialization of enterprises helped to alleviate information asymmetry and improve the accuracy of analysts' forecasts; (2) the increasing degree of financialization promoted the accuracy of analysts' forecasts by reducing the intensity of fixed assets; (3) due to differences in financing constraints, the degree of financialization of state-owned enterprises had no significant effect on the accuracy of analysts' forecasts, but the financialization of non-state-owned enterprises had a significantly greater effect on the accuracy of analysts' forecasts.
Based on the above conclusions, this paper put forward the following suggestions: first, enterprises could give full play to the function of financial assets to diverse risks through financialization, enhance financing capacity and the quality of external information disclosure, so as to weaken information asymmetry; second, the country should improve the supervision system for the financialization of enterprises and gave timely warning when enterprises triggered risks. In addition, considering the heterogeneity of the financialization of existing enterprises, it was suggested that regulators had a definite target and understood the advantages and disadvantages of financialization to enterprises themselves and the macro economy in various aspects.
This paper also had some insufficiencies: it only tested the mediating effect of fixed asset intensity on the relationship between financialization of enterprises and the accuracy of analysts' forecasts, and other factors such as non-financial information were not considered. Therefore, the economic consequences of the financialization of enterprises needed to be further studied in the future.