In a perfect market, the investment decision of a company depends on the net present value of a project, and its expenditure is determined only by investment opportunity sets, and independent of other factors (Modigliani and Mille, 1958) [2] . Hayashi (1982) theoretically proved that under the assumption of perfect market, the optimal capital expenditure of a company is the function of marginal Q [3] . It has also been the judgment of project investing in corporate financial theory, namely when making investment decisions, expected future cash flows should be discounted at the capital cost of the project, and the discount rate depends on size of the project, future cash flows and its risk. Only when the project in the future can provide sustained, stable and realistic cash flows, after actual execution the investment produce enough and effective supply of high quality to meet social demand. As the first major decision of corporate finance, the high quality capital investment decision must be conducive to create enterprise value and increase shareholders’ wealth. Based on these, capital allocation of listed companies is positively correlated with operating performance. The current accounting standards classify capital investment as equity investment, fixed assets, intangible assets and development expenditure by function and character, thus the following three hypotheses are proposed:

H1-1: Equity investment in listed companies is positively correlated with operating performance.

H1-2: Fixed assets investment in listed companies is positively correlated with operating performance.

H1-3: Intangible investment in listed companies is positively correlated with operating performance.

In the real world, however, due to asymmetric information, agency conflict and so on, company operating environment may make important influence on enterprise investment decisions for managers, which affects efficiency of capital allocation and enterprise value (Jensen and Meckling, 1976) [4] . National development strategy, external governance environment factors (Li Yanxi, et al., 2015) [5] and other macroeconomic environment factors, industry competition (Chen Xinyuan, et al., 2013) [6] , industry characteristics (Pan Yuxiang, et al., 2015) [7] and other medium industry factors, control power gain grabbing drive (Zhang Shuo and Zhao Xi, 2016) [8] , equity concentration (Song Xiaobao) [9] , manager features, management power (Wang Maolin, 2014) [10] , bank-enterprise relations (Zhai Shengbao, et al., 2014) [11] and other more corporate internal characteristics all will affect enterprise investment strategy.

When studying Chinese enterprises, we can’t ignore the characteristic of property rights. In China, the government has the strong control ability to state-owned enterprises, where state-owned enterprises have huge advantages in getting resources (such as financing, project bidding, etc.) relative to non-state-owned enterprises. As the main participants of economic activities, the main business goals should also be to promote enterprise value, therefore, in the capital allocation, they should be based on investment efficiency as the judgment standard, so as to promote the improvement of operating performance. Based on these, the following three hypotheses are proposed for the capital allocation in state-owned listed companies:

H2-1: Equity investment in state-owned listed companies is positively correlated with operating performance.

H2-2: Fixed assets investment in state-owned listed companies is positively correlated with operating performance.

H2-3: Intangible investment in state-owned listed companies is positively correlated with operating performance.

The non-state-owned listed companies are built according to the modern enterprise system, and flourishing in the process of marketalization, where the resource allocation is naturally optimal. At the same time, in non-state-owned listed companies underinvestment is mainly manifested with serious financing constraints. In order to obtain financing, enterprises also choose projects with better expected investment returns. Based on these, the following three hypotheses are proposed for the capital allocation in non-state-owned listed companies:

H3-1: Equity investment in non-state-owned listed companies is positively correlated with operating performance.

H3-2: Fixed asset investment in non-state-owned listed companies is positively correlated with operating performance.

H3-3: Intangible investment in non-state-owned listed companies is positively correlated with operating performance.

1) Sample and data

Our sample includes all listed companies in Chinese A stock market. We also exclude firms with missing or non-positive book value of relevant accounts, and where index calculation value is invalid (for instance, the denominator is zero), as is standard in the literature. Considering the financial crisis of 2008-2013, we use data from 2014 to 2016 in CSMAR database. Finally the full sample is 422. We winsorize regression variables at the 1% level to remove extreme outliers.

2) Model assumption and variable description

Perf = α + β 1 ( Δ M A / M A ) + β 2 ( Δ F i x / F i x ) + β 3 ( Δ I n t / I n t ) + β 4 Size + β 5 Lev + ε

The above multiple linear regression model is established. The dependent variable is operating performance (Perf), which uses the weighted average return on equity after deducting non-recurring gains and losses. There are three independent variables, namely, the growth rate of equity investment ( Δ M A / M A ) , the growth rate of fixed asset investment ( Δ F i x / F i x ) and the growth rate of intangible investment ( Δ I n t / I n t ) . ΔMA is the change value of equity investment in the current period, MA is the beginning value of equity investment in the current period, Δ F i x is the change value of fixed asset investment in the current period, Fix is the beginning value of fixed asset investment in the current period, fixed assets investment includes net value of fixed assets, engineering materials and net value of construction in process; Δ I n t is the change value of intangible investment, Int is the beginning value of intangible investment, including intangible assets investment and the average R&D expenditure. The control variable adopts the enterprise Size (Size) and financial leverage (Lev) at the end of the term (Table 1).

The correlation analysis results of full sample in the model are shown in Table 2. There is a significant positive correlation between equity investment and operating performance, fixed asset investment is so. The correlation coefficient of intangible investment on operating performance is positive, but not significant. In terms of control variables, there are significant linear correlations between enterprise scale, financial leverage and operating performance. In addition, the values of variance expansion factor of each variable are less than 5, which are far

Note: **1% significant level; *5% significant level.

less than 10, indicating that there are no multiple mutual problems among the variables.

R² is 18.4% in the model, and the fitting degree of the model is better. Regression results are shown in Table 3, equity investment, fixed assets investment and intangible investment on operating performance have significant positive correlations respectively, showing that the higher the capital growth is, the better operating performance, hypotheses H1-1, H1-2, H1-3 are verified.

In Table 4 for the subsample of state-owned listed companies, the relation between equity investment and operating performance is not significantly correlated, fixed assets investment and intangible investment are so, and the Pearson correlation coefficient of equity investment and operating performance is negative, H2-1, H2-2, H2-3 are not verified. In state-owned listed companies, there is no significant linear relationship between capital allocation and operating performance, which shows that the increase of capital does not bring the increase of operating performance, and capital allocation do not agree with operating performance. State-owned enterprises are the mixtures of administration management and market. Besides the pursuit of performance, they take into account tax, employment and social stability, and many other targets. Corresponding to these, senior executives are appointed and appraised by economic factors, political factors, social responsibility and other factors. Under multiple targets of state-owned enterprises, operating performance is not the sole decision criteria for selecting an investment project. Therefore, investment surplus or underinvestment occurs, and the investment efficiency is not high.

In Table 5 of the sample of non-state-owned listed companies, there is a significant positive correlation between equity investment and operating performance, fixed assets investment is so, and the correlation coefficient between intangible investment and operating performance is positive, but not significant. There are also significant linear relationships between enterprise scale, financial leverage and operating performance in control variables.

In the model design, R² is 20%, and the fitting degree of the model is better. As shown in Table 6, equity investment, fixed assets investment, intangible investment and operating performance are significantly positively correlated respectively, in non-state-owned listed companies the higher growth of capital allocation, the better operating performance, and every capital allocation is consistent with enterprise performance, hypotheses of H3-1, H3-2, H3-3 are verified.

Note: **1% significant level, *5% significant level.

Note: **1% significant level, *5% significant level.

There are many indicators to express operating performance. The above results have been demonstrated by Perf. In order to improve our study, the robustness test is carried out here. We change Perf of the dependent variable to ROA (net profit/total assets) to describe operating performance. The results of the robustness test are consistent with the previous conclusions.

Note: Dependent: Perf.