Frontiers in Business, Economics and Management 
ISSN: 2766-824X | Vol. 17, No. 3, 2024 

 

225 

High Government Leverage and High Corporate 
Leverage: Crowding-in or Crowding-out 

Shiwen Wang * 

Nanjing University of Finance & Economics, Nanjing 210000, China 

 

Abstract: This paper uses panel data of non-financial listed companies between 2012 and 2018 to empirically analyze the 
impact of local government debt on corporate leverage. We find that as the scale of local government debt expands, it has a 

negative impact on corporate leverage. However, this impact varies a lot across industries. For the real estate industry, we 

surprisingly found that it has crounding-in effect which is different from others industry.Moreover, this impact on corporate 

leverage also varies by enterprise size and equity nature. The government's local debt affects the internal cash stock of enterprises, 

which in turn reduces the leverage of enterprises, verifying the channel through which the crowding-out effect arises. The analysis 

in this paper provides empirical guidance for improving the leverage structure of the Chinese economy. 

Keywords: Local debt, Corporate leverage, Crowding out effect. 

 

1. Introduction 

In recent years, the international financial crisis, the 

increase in government fiscal deficits, and continued rise in 

debt levels have resulted in local governments entering a 

phase of high leverage, causing widespread concern. Large-

scale fiscal and monetary stimulus programs have also led to 

a sustained increase in corporate leverage. However, the 

impact of the size of government debt on corporate leverage 

has not been thoroughly analyzed. 

The level of corporate indebtedness is measured using total 

liabilities to total assets to measure the level of leverage of a 

firm (Graham, 2014; Greenwood). While a firm reaps the 

benefits of leverage, it brings with it a certain amount of 

financial risk. Corporate leverage affects the development of 

enterprises in a proportional relationship; the higher the 

corporate leverage, the greater the financial risk of the 

enterprise and the more detrimental to the long-term 

development of the enterprise. If corporate leverage is used 

correctly, it is conducive to improving the efficiency of capital 

allocation and promoting stable and faster economic 

development. Conversely, if leverage is used excessively, it 

may lead to economic overheating, bubble accumulation debt 

risk increases and other problems.  

This paper uses the panel data of prefecture-level cities and 

enterprises, and finds that: in general, there is a negative 

correlation between local governments and corporate 

leverage, that is, local governments can squeeze out corporate 

leverage, the higher the local debt level, and the lower 

corporate leverage. However, from the perspective of 

different industries alone, local government debt has a 

completely opposite crowding effect on real estate enterprises, 

which is also an innovation point found in this paper. Under 

the crowding out effect of local government debt on the 

overall market, the real estate industry alone presents the 

crowding in effect. Moreover, this paper will also study their 

effects on different properties, different scales and whether 

there is foreign investment. 

The current leverage ratio of non-financial enterprises in 

China is extremely high, and the level of corporate debt will 

affect financial risk, and at the same time, as the external 

environment undergoes adverse changes, corporate operating 

risk and credit risk will increase, which may lead to financial 

crisis. China's corporate leverage alert line of 90%, according 

to the data, which is pointed out in China Financial, China's 

non-financial enterprises of China's non-financial corporate 

leverage ratio of 152.9%, has exceeded the alert line by 62.9 

percentage points. It is now a rapid economic recovery in the 

post-epidemic period, the overall economic recovery is slow, 

and in the context of China's economic restructuring, there is 

still some downward pressure on China's economy, and 

enterprises are facing a more unfavorable external 

environment for production and operation. In addition, the 

fiscal decentralization system and the Budget Law limit local 

government tax revenues. Local government financing 

platforms, which is referred local governments initiated the 

establishment, through allocating land, equity, fees, national 

bonds and other assets, quickly packaged a company whose 

assets and cash flow can meet the financing standards bear the 

burden of financing public projects; therefore, the security 

and liquidity of local government obligations are closely 

linked to local government fiscal policies and conditions (Ang 

et al., 2018). With a high government leverage ratio, 

maintaining stable economic growth is crucial to not incur 

systemic risks. Therefore, studying the impact of local debt 

level on corporate leverage is of great significance to solve 

the current problem. 

The remaining sections of the paper are structured as 

follows. Section 2 critically analyzes the extensive current 

literature and formulates the corresponding hypotheses. 

Section III describes the sample, data collection, and research 

methodology used in this paper. Section IV reports the 

analysis of the empirical results of the fixed effects model, 

including the underlying regression results, robustness tests, 

heterogeneity analysis, and mechanism effects. Section V 

reports the regression results of the PVAR model, including 

the unit root test, cointegration test, determination of the 

lagged optimal order, impulse analysis, and variance 

decomposition. Section 6 reports the conclude. 

2. Background 

To "improve the financial supervision system and guard the 

bottom line of no systemic financial risks", which is the 

economic and financial development goal proposed by China 



 

226 

in The National Development and Reform Commission, the 

key issue to be addressed in the current economic 

development is to actively deal with the risk of local 

government debt and solve the fiscal deficit problem. 

Currently, with the economic development coupled with the 

impact of the three-year COVID-19, China's economy is 

facing the problem of large local government debt, which is a 

long-term and structural problem faced in the country's 

economic operation. Local governments announced a RMB 4 

trillion stimulus package in 2008, which led to a rapid 

increase in bond issuance by local government financing 

platforms. Local governments absorbed three-quarters of the 

total funds, most of which were raised by local government 

financing platforms. Since 2015, the government has 

systematically opened up the local government's right to issue 

debt financing, and debt governance has begun to step into the 

stage of standardization, transparency and institutionalization, 

and the local government debt governance model has been 

transformed. Local government debt-related issues have 

received widespread attention (Zhang, 202; Graeve, 2023; Bo, 

2018), however, the impact of government debt on corporate 

leverage has not been studied in depth. This paper investigates 

whether local debt has a crowding-out effect on corporate 

leverage. 

The overall leverage ratio (government debt/GDP) of the 

Chinese government has a fluctuating upward trend, as shown 

in Figure 1 below, which shows the level of urban debt as well 

as the level of urban GDP from 2012 to 2018, and it is obvious 

from the figure that the two are positively correlated in the 

direction of development, with the level of urban debt and 

GDP increasing year by year, and from the actual situation, 

the overall leverage ratio of the Chinese economy is moderate 

and manageable of the Chinese economy, but the leverage 

structure of the Chinese economy is extremely unreasonable, 

and the impact of government debt on corporate leverage is a 

crowding-in or crowding-out effect that requires in-depth 

analysis. Figure 2 is a line graph of the overall debt level and 

overall GDP of the country, from which it can be seen that the 

overall government debt and GDP of the country are growing 

year by year, and the growth rate of GDP has accelerated after 

2016, which is in line with the general trend of China's 

economy continuing to improve and the scale of government 

debt increasing year by year.  

 
Figure 1. Trend chart of local debt 

 

 

 
Figure 1. Trend chart of local debt 

3. Literate Review and Hypothesis 

First, our paper adds to the literature that Xiao et al. (2022) 

suggest that the trend of corporate leverage and macro 

leverage (which can be interpreted as the level of government 

debt) is the same before the current actual national situation 

in China, the level of both government leverage and corporate 

leverage is relatively high and shows an upward trend. 

Corporate financing and borrowing can cause changes in 

corporate leverage, which is affected by various factors. Liu 

(2023) argues that corporate financing is affected by its own 

financial characteristics, such as solvency and profitability, 

and also by macro-level effects, such as government financing 

behavior and economic growth conditions. Narayan et al. 

(2021), by studying the actual situation of corporate leverage 

in the U.S., point out that macroeconomic influences on 

corporate leverage are important. leverage in real terms, 

pointing out that macroeconomics affects important factors of 

corporate leverage, with positive correlation between short-

term leverage and interest rates and negative correlation 

between long-term leverage and inflation, while the impact 

varies across periods of economic development. Jaffe (1978) 

shows that inflation affects corporate leverage by influencing 

interest rates and thus corporate leverage. Demirci et al. (2019) 

point out that when firms are financed through debt financing, 

there is often a tendency to hold relatively stable bond 

investments, so an increase in government debt increases the 

return to financiers on alternative bonds, thereby increasing 

the cost of financing. This informs our discussion of control 

variables when examining the benchmark model. 

Second, many scholars have studied the relationship 

between government borrowing and corporate leverage in the 

past, and their conclusions have similarities and different 

aspects of the findings. We enrich the theory and propose our 

own hypothesis based on the theories of previous studies. 

Local governments have their own financing platforms, 

companies that are capitalized by local governments, with the 

purpose of financing not for profit but for building 

infrastructure and participating in public construction (Clarke 

and LU, 2017). Eberhardt and Predbitero (2015) examine the 

relationship between government debt problems and 

economic growth and find that there is a threshold for the 

positive effects of government debt. Zhang et al. (2014) study 

that local government debt was one of the most important 

forces driving the leverage boom and the surge in the shadow 

banking sector in China during the decade. On the one hand, 

expanding local government debt levels and increasing 

government leverage allow the government to have more 

funds to invest in infrastructure development, boost the 



 

227 

economy, and improve social development. Local 

governments tend to use increased local spending to increase 

GDP growth as a criterion for performance assessment (Mo, 

2018). The first reflection of Chinese enterprises in financing 

will choose bank loans, and the choice between the interests 

of banks, government and enterprises will lead to that local 

governments will make the enterprise debt to be squeezed out 

through the constraints of enterprise loans (Badoer et al,2016). 

In the short term, local issuance of debt is a response to the 

economic crisis, and in the long term, it can cause crowding 

out private investment. Local government borrowing has an 

impact on corporate financing in the case of increasing their 

own government leverage, which affects corporate leverage 

(Jong et al. 2008 and Fan et al. 2012). Krishnamurthy et.al 

(2012) point out that there is a substitution relationship 

between government debt and corporate leverage. Becker and 

Ivashina (2018) also concluded by studying the period of 

European debt crisis that government issuance of debt 

obligations led to a reduction in corporate debt, which 

resulted in a crowding out effect. There is a slight difference 

in the expected reporting of government debt in the short and 

long term, and firms respond in an elastic manner, suggesting 

that government debt influences firm debt choices 

(Greenwood et al., 2010). Firms with relatively unconstrained 

investment behavior bear higher interest rates and generate 

more pronounced crowding-out effects. Liang et al. (2017) 

conducted an empirical study on the impact of local 

government debt on corporate indebtedness using provincial 

and municipal-level data in China and found that local 

governments' preferential access to capital market financing 

worsens the allocative efficiency of capital markets and 

crowds out corporate debt. Hubbard (2012) demonstrates the 

impact of government debt on corporate finance by 

constructing a mathematical model in which the continued 

expansion of government debt raises the market interest rate, 

which in turn discourages corporate investment and thus 

reduces the level of corporate debt. Not only that, the increase 

in government bond holdings leads to a more pronounced 

crowding out effect on corporate lending, and studies have 

found that an increase in government bond holdings has a 

negative impact on private capital as a type of business 

(Becker and Ivashina, 2017). Friedman (1978) argues that in 

terms of the riskiness of investments, the government issues 

bonds, which are risk is less than that of corporate bonds, and 

the two become substitution, investors buying government 

bonds will reduce the purchase of corporate bonds, making 

corporate capital investment decrease and creating a 

crowding-out effect. 

The increase in local government debt also makes the cost 

of financing and the cost of debt taxation increase, which 

leads to a decrease in corporate debt (McDonal, 1983). Taxes 

are the main melody of current life and investors must 

consider the tax situation, high tax investors represent a 

higher level of wealth and tend to hold equity, the increase in 

taxable income and equity causes an increase in tax costs, 

which leads to a decrease in corporate debt. Wu et al. (2015) 

state that local governments also indebtedness land supply, 

which results in land revenue. Under special circumstances, 

the government will provide land at low prices, offer tax 

incentives to enterprises, create a favorable financing 

environment, and enterprises receive incentives that reduce 

the financing costs they need to pay to purchase land. Man 

Zhang (2022) presents in more detail that central government 

debt has a security crowding out effect on enterprise debt and 

local governments have a financial base effect on enterprise 

debt, in the original Greenwood et al. (2010) conclude that 

there is a negative relationship between government debt and 

firm debt maturity by studying the relationship between 

government debt maturity and firm debt maturity. The capital 

structure of firms also makes the effect of local government 

debt on firm leverage different, with the crowding-out effect 

of local government debt on firms being more pronounced in 

countries with larger and more liquid stock markets, where 

firms rely more on markets and less on banks for their 

financing sources, and thus have more flexibility to adjust 

their capital structure and thus have a more pronounced 

crowding-out effect (Demirci et al., 2019). In summary, the 

above study forms the main research hypothesis of this paper: 

under high local government leverage and high corporate 

leverage, local government debt raising causes a crowding-

out effect on corporate leverage. 

In contrast, SOEs have easier access to development funds, 

i.e., credit resources are disproportionately allocated to 

operationally inefficient firms, a phenomenon This explains 

to some extent the resource distortion effect of land finance 

(Hu and Qian, 2017). This is because SOEs often act as 

instruments for local governments to implement policies and 

they operate not following market laws but policy guidelines 

(Tang et al., 2019). Ayturk (2017) argues that large reputable 

and financially sound firms are more sensitive to local 

government debt than smaller firms with poorer finances. The 

larger the size of the firm, the greater the ability of the firm to 

finance itself in general, producing a more pronounced 

crowding-out effect (Demiric et al., 2019). The degree of 

financial openness also affects the crowding out effect of local 

government debt on firm leverage, and foreigners financing 

firms will, to some extent, reduce firms' reliance on the local 

interest rate sector, thus mitigating the crowding out effect 

(Demirci et al., 2019) 

Firms are affected differently by local government debt 

depending on their ownership structure, their size and the 

industry they are in. Differences in GDP growth have a 

significant differential impact on local government debt 

across regions, with local governments tending to issue more 

local debt in regions with GDP growth compared to other 

regions (Qu et al., 2023). Thus, government leverage tends to 

be higher in lagging regions and the crowding out effect on 

firm leverage is more pronounced. 

Finally, referring to the above literature, the research 

contribution of this paper lies in We refer to this literature and 

put heterogeneity analysis in this paper's study, classifying 

firms according to ownership, firm size, and the presence of 

foreign investment input or not, and then derive the 

corresponding results as an extension to enrich the impact of 

local debt on different firms. And furthermore, we classify the 

firms according to the type of industry they are in, trying to 

see if the level of local government debt differs across 

industries for firms in different industries. 

4. Data and Empirical Strategy 

4.1. Data 

In this paper, 1288 A-share listed non-financial enterprises 

from 2012-2018 are selected because listed non-financial 

enterprises can better reflect the current leverage problem of 

Chinese non-financial enterprises as a whole, while we 

identify 128 prefecture-level cities with a total of 9016 

observations in this dataset. The main corporate data are 



 

228 

obtained from CSMAR, and government debt balances and 

municipal investment debt balances are obtained from Wind 

Information financial database and the National Bureau of 

Statistics. According to the availability and completeness of 

the data, the sample is processed as follows: firstly, the 

enterprises with financial category, abnormal financial data 

and serious missing data are excluded; secondly, the listed 

enterprises with special treatment of ST and *ST are excluded.  

4.2. Empirical Strategy 

In this paper, with reference to Demiric (2019), we want to 

study whether local government debt will have a crowding-in 

or crowding-out effect on corporate leverage output in the 

current economic context of China, with high local 

government leverage and high corporate leverage, and set up 

the model of equation (1) and use it for the benchmark 

regression. The Hausman test is conducted and the original 

hypothesis of a random effect is rejected at the 1% 

significance level. The following fixed effects model is 

developed in this paper: 

Leveragei,t = α0 + β0gov _ d ebtc,t−1 + β1Xi,t−1 +

β2Yi,t−1 + ηi + δt + εi,t         (1) 

In equation (1), leverage is firm leverage, gov_debt is the 

size of local government liabilities, which we follow Admati 

et al (2012) and Badoer and James (2016) and lag all 

independent to safeguard against endogeneity, X is the firm-

level control variable, which includes firm size (size), firm 

profitability (net_pro), and firm's guarantee 

capacity(tangibility), Y is the government-level control 

variable, which includes per capita gross domestic product 

(per_gdp), consumer price index (CPI), and unemployment 

rate (unemployment). i and t denote firm and year, 

respectively, and t-1 denotes a lagged period. We choose a 

lagged period to enter the regression equation is mainly 

because the business activities of firms have a certain lag in 

response to government policies, and the setting of lagged 

variables can address certain endogeneity (Irem Demirci, et, 

2019). and represent city and time fixed effects, respectively. 

is the error term. 

In addition, according to the previous theoretical analysis, 

local government debt will affect corporate government debt 

through the balance of cash equivalents at the end of the 

period cash thus affecting corporate leverage, corporate funds 

are represented by cash, the level of cash holdings is equal to 

money funds/total assets, the greater the amount of cash, 

indicating that the internal funds of the firm are sufficient and 

more secure debt service (Xiao et al., 2022). Therefore, we 

introduce intermediation effects to test the mechanism 

according to the way proposed by Baron and Kenny (1986). 

Cashi,t = α0 + β0gov _ d ebti,t-1 + β1Xi,t−1 + β2Yi,t−1 + ηi +

δt + εi,t          (2) 

Leveragei,t = α0 + β0gov _ d ebti,t-1 + β1casℎi,t−1 +

β2Xi,t−1 + β3Yi,t−1 + ηi + δt + εi,t (3) 

4.3. Variable Selection 

Regarding the dependent variable: leverage is corporate 

leverage, according to most scholars (Irem Demirci, et, 2019; 

Yusuf Ayturk, 2017), corporate leverage is usually measured 

using the corporate gearing ratio, i.e., gearing ratio = total 

liabilities/total assets, and we also use this measure in this 

paper for the benchmark regression, which is the traditional 

way of measuring corporate leverage. 

Regarding the independent variable: gov_debt is the size of 

local government debt and is measured using local 

government debt/GDP (Rangarajan & Srivastava, 2005). A 

possible concern with using the government debt/GDP ratio 

to measure the size of local government debt is that the 

relationship between corporate leverage and government debt 

may be driven by changes in GDP rather than changes in 

outstanding government debt. 

Regarding the control variables: the control variables in the 

paper include government level and firm level. For the 

selection of macro-level control variables, we mainly refer to 

relevant studies by Narayan (2021), including per_gdp (per 

capita gross regional product) which using is used logarithmic 

form in the text, inflation is measured by the consumer price 

index (CPI), and unemployment is derived from the 

unemployment registration rate of urban residents. The 

micro-level control variables refer to relevant studies by 

Baker and Wurgler, 2002; Frank and Goyal, 2003; Lemmon 

et al., 2008 and others are including firm's guarantee capacity 

(tangibility), which is used to guarantee the firm's ability to 

secure itself after borrowing, using the net value of firm's 

fixed assets over total assets; firm's size (size), which is 

expressed in terms of the firm's total assets; and the firm's 

profitability, which we use the gross operating margin to 

express. 

 

Table 1. Variable definition 

Variables Symbol Definition 

Corporate 

Leverage 
leverage total debt/total asset 

Government Debt gov_debt government debt/GDP 

Corporate Size size corporate asset 

Tangible Asset tangibility 
net fixed assets/total 

assets 

Return on Asset net_pro Profit/sales revenue 

Inflation CPI 
resident consumption 

index 

Unemployment unemployment Unemployment rate 

GDP per capita Lnper_gdp GDP/total population 

 

Note: Standard errors in parentheses. *, **, ***.  

4.4. Descriptive Statistics 
Table 2. Descriptive statistics. 

 (1) (2) (3) (4) (5) 

Variables obs. Mean S.D. Min Max 

leverage 9,016 0.448 0.301 -0.19000 
11.51

0 

gov_debt 9,016 0.137 0.109 0.00530 0.734 

size 9,016 2.470 9.562 0.00100 230.7 

tangibility 9,016 0.210 0.168 0.00001 0.948 

net_pro 9,016 0.278 0.180 -1.01100 1.154 

CPI 9,016 
102.3

00 
0.987 95.00000 113.3 

unemploym

ent 
9,016 2.782 0.770 0.90000 4.900 

Lnper_gdp 9,016 2.168 0.579 0.00896 3.845 

 

Table 2 below displays the descriptive statistics. The total 

sample size is 9016.The average ration leverage in our sample 

is 0.448, standard deviation of leverage is 0.301, and the 

maximum value is 11.51, which the minimum value is -0.19. 

The average gov_debt each city is 0.137, which standard 



 

229 

deviation is 0.109. The maximum value is 0.734, and the 

minimum value is 0.0053. On average, the size of each 

company is 2.47 billion. The standard deviation of size is 

9.562. The maximum value is 230.7 billion. The minimum 

value is 0.001 billion. The average tangibility is 0.21, the 

standard deviation of tangibility is 0.168. And the maximum 

tangibility value is 0.948, which the minimum value is 

0.00001. The average net_pro is 0.278. The standard 

deviation of net_pro is 0.18.The maximum value of net_pro 

is 1.154,and the minimum value is -1.011. The average level 

of CPI is 102.3,which the standard deviation is 0.987.The 

maximum value of CPI is 113.3,and the minimum value of it 

is 95.Moreover,the average of unemployment is 2.782.The 

standard deviation of the unemployment is 0.77,which the 

maximum value is 4.9 and the minimum value is 

0.9.Finally,the average Lnper_gdp is 2.168.The standard 

deviation of it is 0.579.The maximum value of Lnper_gdp is 

3.845,and the minimum value is 0.00896. 

5. Empirical Results  

5.1. Main Results  

In this paper, we want to study the impact effect of high 

government leverage and high corporate leverage, and use a 

double fixed effects template for regression estimation 

according to equation 1, and analyze the regression results by 

observing corporate leverage to determine whether there is a 

crowding-in or crowding-out effect between high government 

leverage and high corporate leverage.  

 

Table 3. Benchmark results 

 (1) (2) (3) 

 leverage leverage leverage 

gov_debt (t-1) -0.136** -0.149*** -0.150*** 

 (0.056) (0.053) (0.054) 

tangibility (t-1)  -0.113*** -0.113*** 

  (0.016) (0.016) 

size (t-1)  0.004*** 0.004*** 

  (0.000) (0.000) 

net_pro (t-1)  -0.418*** -0.418*** 

  (0.015) (0.015) 

Lnper_gdp (t-1)   -0.002 

   (0.012) 

unemployment (t-1)   -0.004 

   (0.009) 

cpi (t-1)   -0.002 

   (0.002) 

_cons 0.461*** 0.593*** 0.794*** 

 (0.007) (0.009) (0.251) 

Individual FE YES YES YES 

Time FE YES YES YES 

N 7728 7728 7728 

r2 0.101 0.245 0.245 

F 5.850 258.065 150.503 

 

Note: Standard errors in parentheses. *, **, ***.  

Table 3 reports the regression results of the data on the 

effect of local government debt on corporate leverage: 

column 1 is the regression result without adding control 

variables, and we can see that the coefficient -0.136 is 

significant at the 5% level, concluding that the relationship 

between local government debt and corporate leverage is 

negative and generates a crowding-out effect, with each 

percentage point increase in local government debt 

decreasing corporate leverage by 13.6%. Column 2 adds 

micro firm-level control variables, and the final regression 

comes out to be significant at 1% with a coefficient of -0.149, 

where each percentage point increase in local government 

debt decreases firm leverage by 14.9%, and the crowding out 

effect is even more pronounced. Column 3 adds the macro-

level control variables and the estimated coefficient is 15%, 

again significant at the 1% level, and for every one percentage 

point increase, the corporate market leverage decreases by 

15%. This indicates that the larger the size of government debt, 

the lower the market leverage of firms, i.e., government debt 

has crowding-out effect on the market leverage of firms, 

validating the previous hypothesis. 

Moreever, different industries are affected by government 

debt raising differently depending on the nature of the 

environment they are in, policies, and industry characteristics. 

We divide the industries in which the enterprises are located 

according to 2012 Edition SEC Industry Classification in 

CSMAR, excluding financial enterprises in the sample 

screening, so the industries involved in the sample are the 

public utilities sector, the real estate sector, the industrial 

sector, the commercial sector, and the Conglomerates sector, 

and we divide the sample into five groups to conduct 

regressions separately to obtain Table 4. Table 4 Column 1 in 

order is Public Utilities, Column 2 is Real Estate, Column 3 

is Comprehensive Industry, Column 4 is Industrial sector, and 

Column 5 is Business2 regression results. The results reported 

that: the local government debt has the most significant 

crowding out effect on the enterprises in the industrial sector, 

the reason is that industrial enterprises will undertake more 

research and development, the most important material 

production sector, providing material products for the life of 

the population and economic activities of various industries, 

this important role is irreplaceable by any other industrial 

sector. Industry is the industrial sector with the most R&D 

investment, the most active technological innovation, and the 

strongest radiation drive, so the relationship between 

enterprises in the industrial sector and government debt is the 

closest, and the issuance of government debt will often 

become the main influence of the industrial sector, so the 

crowding out effect is the most obvious. The regression 

results for the real estate sector appear in column 2 of Table 4 

are opposite to those for other industry types. We believe the 

reason for this is that: this in turn may trigger potential risks 

such as market volatility and credit rating downgrades, 

curbing financing channels and raising funding costs for local 

governments and real estate enterprises, thereby exacerbating 

the latter's leverage levels. Additionally, the larger debt 

issuance may also shift the investment focus of banks and 

other financial institutions towards real estate and property 

developers as their preferred lending targets, potentially 

lifting the leverage ratio of the real estate industry. 

 

 

 

 

 



 

230 

Table 4. Benchmark resultⅡ 

 Public Utilities sector  Real Estate sector Comprehensive sector Industrial sector Business 

gov_debt (t-1) -0.075 2.878*** 0.676 -0.423** -0.196 

 (0.174) (0.975) (0.477) (0.193) (0.409) 

tangibility (t-1) -0.355* 0.166 -0.333*** -0.254*** -0.124 

 (0.191) (0.251) (0.104) (0.069) (0.172) 

size (t-1) 0.002 0.239*** 0.108* 0.002*** 0.000 

 (0.002) (0.062) (0.056) (0.001) (0.005) 

net_pro (t-1) -0.409*** -2.593*** -0.313*** -0.551*** -0.478*** 

 (0.100) (0.413) (0.065) (0.078) (0.125) 

Lnper_gdp (t-1) -0.008 -0.016 0.010 0.023 0.036 

 (0.065) (0.131) (0.082) (0.036) (0.056) 

unemployment (t-1) -0.079 -0.869*** -0.239* -0.091*** -0.086 

 (0.049) (0.285) (0.135) (0.035) (0.099) 

cpi (t-1) 0.012 0.007 0.005 -0.008 -0.040 

 (0.011) (0.010) (0.013) (0.008) (0.024) 

_cons -0.303 2.253* 0.530 1.759** 4.804* 

 (1.147) (1.088) (1.253) (0.876) (2.477) 

Individual FE YES YES YES YES YES 

Time FE YES YES YES YES YES 

N 144 36 78 612 66 

r2 0.674 0.719 0.650 0.333 0.911 

F 4.223 12.347 5.657 20.770 3.674 

Note: Standard errors in parentheses. *, **, ***.  
5.2. Heterogeneity Analysis 

5.2.1. Corporate ownership structure 
 

Table 5. Heterogeneity: corporate ownership structure 

 NON—SOE SOE NON—FC FC 

 leverage leverage leverage leverage 

gov_debt (t-1) -0.189** -0.056 -0.153*** 0.020 

 (0.075) (0.076) (0.055) (0.262) 

tangibility (t-1) -0.094*** -0.146*** -0.116*** -0.072 

 (0.023) (0.022) (0.016) (0.066) 

size (t-1) 0.007*** 0.003*** 0.004*** 0.003*** 

 (0.001) (0.000) (0.000) (0.001) 

net_pro (t-1) -0.382*** -0.396*** -0.420*** -0.384*** 

 (0.019) (0.028) (0.016) (0.056) 

Lnper_gdp (t-1) 0.001 -0.001 -0.004 0.043 

 (0.014) (0.023) (0.013) (0.045) 

unemployment (t-1) -0.008 0.004 -0.004 -0.010 

 (0.012) (0.012) (0.009) (0.039) 

cpi (t-1) -0.002 -0.002 -0.002 0.004 

 (0.003) (0.003) (0.002) (0.009) 

_cons 0.749** 0.862** 0.831*** 0.023 

 (0.343) (0.346) (0.256) (0.958) 

Individual FE YES YES YES YES 

Time FE YES YES YES YES 

N 4920 2808 7212 516 

r2 0.257 0.263 0.246 0.385 

F 75.860 46.533 136.792 16.250 

Note: Standard errors in parentheses. *, **, ***.  

Ownership attributes are actually a standard, exogenous 

division based on financing constraints. We classify firms into 

state-owned and non-state-owned enterprises based on their 

ownership structure, making state-owned enterprises 1 and 

non-state-owned enterprises 0. We conduct a heterogeneity 

analysis of firm ownership, as shown in Table 5, column 1, 

column 2. According to the results reported in Table 5, 

column 1, column 2, the estimated coefficient of non-state-

owned enterprises is -0.189, which is significant at the 5% 

level, while the estimated coefficient of state-owned 

enterprises is -0.056, and the crowding out of non-state-

owned enterprises the crowding out effect of non-SOEs is 

more significant than that of SOEs. The reason is that, on the 

one hand, the main shareholder of SOEs is the government, 

and government intervention makes SOEs have policy 

advantages in terms of credit, and high-quality resource 

factors tend to flow to SOEs, forming a soft budget constraint, 



 

231 

resulting in a significantly weaker financing constraint for 

SOEs than for non-SOEs, and the crowding-out effect is less 

pronounced. On the other hand, non-state-owned enterprises 

are mainly subject to market allocation of resources, and their 

operating conditions are influenced by the market. They tend 

to choose their financing structure according to the local 

government's debt raising, and when the government's debt 

raising scale is larger, the interest rate rises, and the 

enterprise's comprehensive financing cost and return risk, etc. 

will reduce the enterprise's debt and lower the enterprise's 

leverage, so the crowding out effect is more obvious. 

In the enterprise ownership structure, we also divided the 

enterprises with foreign investment and without foreign 

investment, as shown in Table 5 column 3 and column 4. 

when foreign participation in enterprise investment, no 

foreign participation in investment enterprises with financial 

openness than with foreign participation will be lower, more 

dependent on local government debt, that is, more 

significantly affected by the negative impact of local 

government debt, that is, column 3 is more significant, the 

crowding out effect is more obvious. 

5.2.2. Firm size 

Table 6. Heterogeneity: firm size 

 S M L 

 leverage leverage leverage 

gov_debt (t-1) -0.153* -0.186** -0.117 

 (0.079) (0.088) (0.078) 

tangibility (t-1) 0.054** -0.087** -0.231*** 

 (0.025) (0.038) (0.022) 

size (t-1) 0.043*** 0.014*** 0.002*** 

 (0.008) (0.004) (0.000) 

net_pro (t-1) -0.399*** -0.355*** -0.362*** 

 (0.019) (0.031) (0.032) 

Lnper_gdp (t-1) -0.000 -0.001 -0.005 

 (0.019) (0.020) (0.018) 

unemployment (t-1) 0.006 -0.010 -0.015 

 (0.012) (0.015) (0.014) 

cpi (t-1) -0.001 -0.005 -0.001 

 (0.004) (0.004) (0.003) 

_cons 0.580 1.107** 0.822*** 

 (0.366) (0.454) (0.303) 

N 3348 1650 2730 

r2 0.259 0.264 0.293 

F 79.189 27.133 49.957 

 

Note: Standard errors in parentheses. *, **, ***.  

The total sample was divided into three subsamples and 

regressed separately. The classification criteria were as 

follows: according to the total assets of enterprises in 2018, 

enterprises with total assets less than 500 million yuan were 

classified as small-scale enterprises, enterprises with total 

assets greater than 1 billion yuan were classified as large-scale 

enterprises, and the rest were classified as medium-sized 

enterprises. According to the size of enterprises, the 

regressions were grouped for the sample of enterprises, and 

the results are shown in Table 6 columns 1 to 3 are small, 

medium and large enterprises, respectively, and it can be 

concluded that the crowding-out effect is most obvious for 

small and medium enterprises, while large enterprises are not 

significant instead. The reasons are: first, large enterprises 

have more financing options compared to small-scale 

enterprises, and these enterprises rely on their own strength 

tend to issue their own bonds for financing, so the default cost 

is lower and they are less likely to be replaced by government 

bonds. Secondly, the financing structure of large enterprises 

will be more complex than that of small and medium-sized 

enterprises, with their own financial strength, they will 

consider using internal funds first before using external funds. 

Considering the high cost of external financing, large 

enterprises tend to consider internal financing first, so large 

enterprises are less sensitive to the expansion of government 

debt scale. In summary, the crowding-out effect is more 

significant for small and medium-sized enterprises than for 

large-scale enterprises. 

6. Robustness Checks 

Although we include explanatory variables, explained 

variables, and six micro- and macro-level control variables in 

our baseline regression model, the estimation results of the 

model may still be affected by model setup, omitted variables, 

and differences in sample background, making the results 

unconvincing and robust. Therefore, we include robustness 

tests to remove the interference of other factors to the model 

as much as possible through various ways. 

 

Table 7. Robustness checks 

 (1) (2) (3) 

    

province_debt (t-1) -0.154***   

 (0.055)   

gov_debt (t-1)  -0.153*** -6.123** 

  (0.057) (2.745) 

tangibility (t-1) -0.113*** -0.091*** -8.139*** 

 (0.016) (0.018) (0.753) 

size (t-1) 0.004*** 0.008*** 0.365*** 

 (0.000) (0.001) (0.043) 

net_pro (t-1) -0.418*** -0.425*** -6.597*** 

 (0.015) (0.015) (0.558) 

Lnper_gdp (t-1) -0.002 0.001 -0.161 

 (0.012) (0.013) (0.416) 

unemployment (t-1) -0.004 -0.005 0.076 

 (0.009) (0.009) (0.265) 

cpi -0.002 -0.003 -0.083 

 (0.002) (0.002) (0.099) 

_cons 0.794*** 0.889*** 17.638* 

 (0.251) (0.257) (10.085) 

Individual FE YES YES YES 

Time FE YES YES YES 

N 7728 6174 7728 

r2 0.245 0.273 0.238 

F 150.547 129.694 43.032 

 

Note: Standard errors in parentheses. *, **, ***. 

In the baseline regression model of equation 1, we use local 

debt data of prefecture-level municipalities for our sample 

data, and to test the robustness of the regression results, we 

use provincial-level debt data for the regression. The relevant 

estimates are shown in column 1 of Table 7. the regression 

results using the provincial sample can be found: the 

coefficients are very similar to the baseline regression results; 

the estimated coefficients differ by 0.003, the results are very 

small, and the robust standard errors are also very close, 

indicating the robustness of the estimation results in this paper. 

Due to the special political and economic factors of Direct-

administered municipalities of China, local government debt 



 

232 

tends to be different from that of ordinary prefecture-level 

cities, which will have an impact on the results when 

considered. To test the robustness of the regression results, we 

exclude the data of municipalities directly under the central 

government and use the data of the remaining prefecture-level 

cities for the regression. The results in column 2 of Table 7 

show that the estimated coefficients and robust standard 

errors are very similar. Therefore, the results of the 

benchmark regression model are still significant and robust in 

terms of robustness tests. In the benchmark regression model, 

we use total liabilities/total assets (gearing ratio) to measure 

firm leverage, and based on the related literature by 

Greenwood (2015), Irem Demirci (2019), we use the market 

leverage ratio of liabilities/market value added instead of the 

original indicator of firm leverage, and the results are 

presented in column 3 of Table 7.Although the estimated 

coefficients and robustness criteria are in error with the 

coefficients in Table 3, the results are still significant, 

implying that to some extent, the results of the benchmark 

regression model are still robust and local government debt 

still has a crowding-out effect on corporate leverage. 

 

Table 8. Robustness checks: endogeneity 

 Basic IV 

 leverage leverage 

gov_debt (t-1) -0.150*** -0.685*** 

 (0.054) (0.235) 

tangibility (t-1) -0.113*** 0.046** 

 (0.016) (0.021) 

size (t-1) 0.004*** 0.002*** 

 (0.000) (0.000) 

net_pro (t-1) -0.418*** -0.212*** 

 (0.015) (0.018) 

Lnper_gdp (t-1) -0.002 0.020 

 (0.012) (0.013) 

unemployment (t-1) -0.004 -0.007 

 (0.009) (0.005) 

cpi (t-1) -0.002 -0.003** 

 (0.002) (0.002) 

_cons 0.794*** 0.819*** 

 (0.251) (0.159) 

Individual FE YES YES 

Time FE YES YES 

N 7728 7728 

r2 0.245  

F 150.503  

Underidentification test  490.713 

Weak identification test  530.611 

Sargan statistic  0.000 

 

Note: Standard errors in parentheses. *, **, ***. 

We next use the instrumental variable approach to deal with 

endogeneity, referring to Irem Demiric et al. (2019) who use 

defense spending as an instrumental variable for local 

government debt, and in this paper, we use the sum of defense 

spending and health spending. In terms of the choice of 

instrumental variables, defense spending and health care 

spending are closely related to local fiscal spending, and 

defense spending and health care spending are basic 

livelihood spending, which are less affected by 

macroeconomic influences and are relatively fixed 

expenditures in government finance, and can be considered as 

exogenous variables with weak correlation with the economic 

influences on corporate leverage. We use IV estimation 

method for regression, as shown in column 2 of Table 8, and 

the results show that the results are significant at 1%, 

indicating that the instrumental variables are valid, and we 

conduct Underidentification test, Weak identification test, and 

Sargan statistic, and the results are 490.713, 530.611 and 

0.000, from which we can conclude that there is and our 

baseline regression results are robust between corporate 

leverage and instrumental government debt. 

The trend of mass data in power system provides a basis 

for load characteristic analysis and prediction model 

establishment, but the classical load forecasting method can 

not afford such a huge time and computing resource 

consumption. The problem of over fitting in large sample set 

will affect the prediction accuracy. In this paper, a power load 

forecasting model is built by using the BP neural network 

model, making full use of the powerful data processing 

function of Clementine and preventing the over fitting 

function. The experimental results show that the BP neural 

network model has good predictability and robustness, and 

has a certain practical application value. 

Acknowledgment 

Project Fund: “Postgraduate Research & Practice 

Innovation Program of Jiangsu Province”  

Project coding <KYCX23_1877> 

References 

[1] Admati, A.R., DeMarzo, P.M. Hellwig, M.F. & Pfleiderer, P.C., 
2012. Debt Overhang and Capital Regulation., Rock Center for 
Corporate Governance at Stanford University, Working Paper, 
No. 114.  

[2] Ang, A., Bai, J., & Zhou, H., 2018. The Great Wall of Debt: 
Real Estate, Political Risk, and Chinese Local Government 
Financing Cost, Georgetown McDonough School of Business 
Research Paper, PBCSF-NIFR Research Paper No. 15− 02.  

[3] Badoer, D.C., James, C.M., 2016. The determinants of long-
term corporate debt issuances, Journal of Finance, 71 (1), 457–
492. 

[4] Baker, M., Wurgler, J., 2002. Market Timing and Capital 
Structure. Journal of Finance, 57, 1–32. 

[5] Baron R M, Kenny D A., 1986. The Moderator-mediator 
Variable Distinction in Social Psychological Research: 
Conceptual, Strategic, and Statistical Considerations. Journal 
of Personality and Social Psychology, 51(6). 

[6] Bo Becker, Victoria Ivashina, 2018. Financial Repression in 
the European Sovereign Debt Crisis. Review of Finance, 83–
115 

[7] Clarke, D., Lu, F., 2017. The Law of China’s Local 
Government Debt: Local Government Financing Vehicles and 
their Bonds. Am. J. Comp. Law 65 (4), 751–798. 

[8] De Jong, A., Kabir, R. & Nguyen, T.T., 2008. Capital Structure 
Around the World: the Roles of Firm-and Country-specific 
Determinants. J. Bank. Finance 32 (9), 1954–1969. 

[9] Demirci, I., Huang, J. & Sialm, C., 2019. Government Debt and 
Corporate Leverage: International Evidence, Journal of 
Financial Economic, 133(2), 337-356.  

[10] Eberhardt, M., A.F. 2015. Presbitero, Public Debt and Growth: 
Heterogeneity and Non-Linearity.Journal of International 
Economics, 97(1):45—58. 

[11] Fan, J.P., Titman, S., & Twite, G., 2012. An International 
Comparison of Capital Structure and Debt Maturity Choices. 
Journal of Financial and Quantitative Analysis 47 (01), 23–56. 



 

233 

[12] Ferre De Graeve, Giulio Mazzolini, 2023. The Maturity 
Composition of Government Debt: A Comprehensive Database, 
European Economic Review, Volume 154. 

[13] Frank, M.Z., Goyal, V.K., 2003. Testing the Pecking Order 
Theory of Capital Structure. Journal of Financial Economics. 
67, 217–248 

[14] Friedman, B. M.,1978. Crowding Out or Crowding In: 
Economic Consequences of Financing  

[15] Government Deficit, Brookings Papers on Economic Activity, 
9(3), 593-641.  

[16] Graham, M.T. Leary, M.R. Roberts., 2014. How does 
Government Borrowing Affect Corporate Financing and 
Investment? J. R.  

[17] Greenwood, S. Hanson, J.C. Stein., 2010. A Gap-Filling 
Theory of Corporate Debt Maturity Choice. The Journal of 
Finance, 65 (3), 993-1028. 

[18] Hailian Xiao, Ying Zhao, & Meihua Zhou, Can Financial 
Factors Affect Corporate Debt Leverage Convergence? 
Pacific-Basin Finance Journal, 72 (2022) 101728. 

[19] Jaffe, J.F, 1978. A Note on Taxation and Investment. Journal 
of Finance, 33, 1439-1445. 

[20] Keefe, M.O.C., Yaghoubi, M., 2016. The Influence of Cash 
Flow Volatility on Capital Structure and the Use of Debt of 
Different Maturities. Journal of Corporate Finance, 38, 18–36.  

[21] Krishnamurthy, Vissing-Jorgensen., 2012. The Aggregate 
Demand for Treasury Debt. Journal of Political Economy, 20 
(2): 233-267. 

[22] Lemmon, M.L., Roberts, M.R. & Zender, J.F., 2008. Back To 
The Beginning: Persistence and the Cross-section of Corporate 
Capital Structure. Journal of Finance, 63, 1575–1608. 

[23] Liang,Y, K. shi & L. Wang, 2017. Local Government Debt and 
Firm Leverage: Evident from China. Asian Economic Policy 
Review, 12(2):210—232 

[24] Man Zhang, Oscar T. Brookins & Xiaowei Huang, 2022. The 
Crowding Out Effect of Central Versus Local Government 
Debt: Evidence from China. Pacific-Basin Finance Journal, 
72.101707. 

[25] McDonald, R. L. 1983. Government Debt and Private Leverage: 
An Extension of the Miller  

[26] Theorem. Journal of Public Economics, 22(3), 303-325.  

[27] Mengtao Zhang, Wenchuan Chen, Aidi Kou & Yanjun Wu, 
2023. Promotion Incentives, Tenure Uncertainty, and Local 
Government Debt Risk, Finance Research Letters,Volume 
56.104136. 

[28] Mo, J., 2018. Land Financing and Economic Growth: Evidence 
from Chinese Counties, China Economic Review, 50, 218-239. 

[29] Qiongzhi Liu, Yun Bai, Hexin Song., 2023. The Crowding Out 
Effect of Government Debt on Corporate Financing: Firm-level 
Evidence from China, Structural Change and Economic 
Dynamics. 

[30] Rangarajan C, Srivastava D K., 2005. Fiscal Deficits and 
Government Debt: Implications for Growth and Stabilisation. 
Economic & Political Weekly, 40(27):2919-2934 

[31] Robin Greenwood, Samuel Hanson, & Jeremy C. Stein, 2010. 
A Gap-Filling Theory of Corporate Debt Maturity Choice. The 
journal of Finance. Volume 60, No. 3. 

[32] Seema Narayan, Minh Ngoc Thi Bui, Yishuai Ren & Chaoqun 
Ma., 2021. Macroeconomic Determinants of US Corporate 
Leverage, Economic Modelling, 104 105646. 

[33] Tang, P., Shi, X., Gao, J. Feng, S. & Qu, F., 2019. 
Demystifying the Key for Intoxicating Land Finance in China: 
An Empirical Study through the Lens of Government 
Expenditure, Land Use Policy, 85, 302-209.  

[34] Wu, Q., Li, Y., & Yan, S. 2015. The Incentives of China's 
Urban Land Finance, Land Use Policy, 42, 432-442. 

[35] Xi Qu, Zhiwei Xu, Jinxiang Yu & Jun Zhu, 2023. 
Understanding Local Government Debt in China: A Regional 
Competition Perspective, Regional Science and Urban 
Economics, Volume98, 103859. 

[36] Yusuf Ayturk. 2017. The Effects of Government Borrowing on 
Corporate Financing: Evidence from Europe, Finance Research 
Letters 20 (2017) 96–103 

[37] Zhang, Yuanyan Sophia & Barnett, Steven, 2014. Fiscal 
Vulnerabilities and Risks from Local Government Finance in 
China, IMF Working Paper.