Housing Developers’ Heterogeneous Decision-Making under Negative Shock after the High-Growth Era: Evidence from the Chinese Real Estate Economy

Abstract

This research uses difference-in-difference (DID) and other empirical methods to analyze firm-level real estate data to discover how heterogeneous firm characteristics affect managers’ decision-making about development expansion when a firm faces a temporary negative sales shock in the Chinese housing market. The manager’s decision is a utility maximization problem under uncertainty, determined by their risk aversion levels, which managers choose to optimize by considering other factors of interest, including career risk and personal wealth. Also, the advance payment rule encourages real estate developers to maintain high turnover, since new projects allow developers to collect cash first. The results show that state-owned enterprises (SOEs) are much more conservative than other types of developers. SOEs tend to focus on current developing projects. Firms with more concentrated management pursue expansion and seek to use new project sales to compensate for their slower growth. Larger developers with headquarters in large cities tend to slow their development speed when they observe negative signals, as they can quickly engage in new projects given these firms’ easy access to financial resources such as bank loans. This study makes a novel contribution to the literature since previous research has tended to focus on the macro market level rather than the firm level. The findings also have strong policy and regulation value. The results indicate that higher cashflow monitoring needs, especially to monitor family-owned developers, to prevent misuse and excessive project expansion.

1. Introduction

The Chinese real estate market has experienced aggressive growth, followed by steady growth, and is now entering a recession. Rapid urbanization and the population flow from rural areas to cities after 2005 led to a boom in the real estate economy and a significant increase in housing prices [1]. The real estate market shows a large development synergy efficiency effect [2], and it is associated with rapid infrastructure development [3], which significantly increases local social welfare. Even though the government implemented a purchase limit policy in 2010, this did not stop increasing prices. Growth slowed in approximately 2015 and has become steady, without too much volatility. Urbanization has also entered a new stage [4], and most homebuyers are now seeking better living and environmental conditions. Many economists started to question high housing prices and housing bubbles in 2017 [5], and real estate development firms have begun to form divergent opinions about the market’s future. The speed of development expansion started to change based on managers’ beliefs and support from financial institutions. A high urbanization rate has further slowed the real estate market. After COVID-19, the real estate market significantly decreased, and housing prices decreased more quickly than most economists expected.

The Chinese real estate market has unique prospects. The Chinese market has adopted the advance payment method, in which a developer firm purchases the land and uses the land as collateral to apply for development loans from commercial banks. Once a new project officially starts, development firms are allowed to charge the home buyer the full housing price [6,7]. Of course, home buyers may take out mortgages, and then the developer firm receives mortgage capital from the bank. A commercial bank would have monitoring responsibility if the developer’s project was funded by that commercial bank; then, the bank is required to monitor the project account to ensure that there is enough funding to complete the project and successfully transfer the housing title to the home buyer. This advance payment method increases overall efficiency, and nearly all developers seek to achieve high turnover during rapid growth [8]. However, once the real estate economy slows, rapid growth seeking requires high leverage and becomes disastrous.

Most larger developers are stock exchange-listed firms, and few of them are listed in Hong Kong. There are a few types of real estate development firms. The first type is the state-owned enterprise (SOE). Support is typically provided by the local government. The managers of SOEs closely follow policy, and they balance the firm’s reputation, social obligation and profit and performance. Although firms are usually large-scale and financially strong and may easily obtain support from banks, there are many soft constraints [9] which may cause managers to be conservative. The second most common type of stock exchange-listed real estate firm is the family-owned firm. Compared to SOEs, these firms are usually medium to small. The key characteristic of a family-owned firm is that its managers have greater power and authority, and, most often, the chair of the board and the general manager are the same person, a situation known as duality management. Their greater management power makes them more aggressive when seeking growth, and they often like to take risks; the results depend on their future market status [10]. The last type is the share-diversified stock exchange-listed firm. In this firm, there is no single shareholder who controls the firm, and managers do not have close political connections, so they focus more on shareholders’ interests.

This study contributes to the current literature in the two following ways. Most research on the real estate market has focused on macro-level observations, whereas this study uses micro firm-level data to show the different risk reactions that occur when different types of firms experience negative shocks. Second, we introduce a method for using advance payments rather than firm performance or profitability to measure shocks since accounting profitability is inappropriate and misleading if firms receive significant advance payments or unearned revenue in terms of accounting. Real estate firms realize a profit when they transfer property rights to the home buyer, but they have perfect control over project speed and the time of transfer, which makes earning smoothing and manipulation easy.

The rest of this paper is organized as follows. This section briefly introduces the background of the research. The second section focuses on past research and the differentiation of our work from previous research. Then, the next section shows the existing literature and formulates several hypotheses to investigate. The data and research methodology are explained in the following section. A report of the results, interpretations, and discussions follows. The final section concludes and provides policy suggestions and future research directions.

2. Previous Discussion and Research Gaps

Most of the past empirical data on real estate focus on aggregate market status and performance. Past discussions focus on market developments and urbanization, policy impact, and economic growth contributions. Past research mostly stays on the macro and market level rather than understanding how individual developers may have differentiated behaviors when they observe negative shocks and events. Past research also emphasizes the difficulties in evaluating the real estate projects’ fair value and market value, especially in structured projects involving intensified subjective accounting estimations and judgements.

From the perspective of real estate market development and urbanization, real estate market development helps and contributes to city infrastructure development and speeds up urbanization [11]. Real estate market development and housing prices are highly correlated with population mobility, especially rural-to-urban movement and population movement from smaller to larger cities [12]. The change in market expectations and the slowdown of the urbanization process led the real estate market to slower but more stable growth after 2015 [13]. Older societies lower housing demand [14]. The lower transparency of real estate market information and high government intervention become problematic once its growth slows down [15]. The supply side of the real estate market is questioned as being less efficient because the pending project development time is long [16]. The slowdown of the real estate market is also associated with non-performing loans in some regional banks and has become a financial stability issue [17].

From a policy perspective, increased housing prices and rapid growth cause the government to consider the trade-off and the policy needed to balance the stability of housing prices and urban development [1]. The government created some policies and guidelines to alleviate the rapid growth of housing prices during the high-growth period. Such policies include purchase restrictions, which show heterogeneous effects in different cities but are generally less effective in larger cities [18]. During the high-growth period, the boom of the real estate market was closely associated with the local government’s land finance strategy [19]. The growth of the real estate market is also closely associated with the expansionary Chinese monetary policy [20,21]. The expansionary monetary policy leads banks to issue more real estate development loans, which is the key factor that explains the significant growth of housing prices [22]. The evidence suggests past lower growth regimes experienced higher monetary effects [23].

When considering that real estate is an important sector contributing to the overall economy, the discussion usually focuses on the real estate sector’s contribution to the development and growth of regional economies [24,25]. Real estate development has contributed differently to economic growth in different regions of China [26]. The eastern provinces have experienced much stronger economic growth effects from real estate development than midwestern regions [27]. The real estate market’s development is also associated with financial market development [28]. Development requires large capital, so developers have a large loan position in the banking industry. The real estate market is believed to greatly impact systematic economic risks [29].

As mentioned in the introduction section, the real estate industry’s construction procedures are complicated, frequently require accounting judgements, and may be subject to different accounting treatments [30]. For example, fair values are subject to accounting methods, and market values require estimations [31]. Accrual accounting and the nature of the business make earning smoothing easy. Income-decreasing accruals are common during high macroeconomic-control periods in the Chinese real estate industry [32]. Developers need proper risk control and access to diverted financial resources [33]. The growth of inventory caused by blind and less-considered expansion indicates the risk here [34,35]. The landing banking regulation, which requires the developer to start the project within a certain time after purchasing the land, indirectly increases the blind expansion risk [36]. The other reason for aggressive expansion could be revenue recognition, which is also related to the issue of earning manipulation. Developers are incentivized to have aggressive revenue recognition when their performance is unsatisfactory [37,38] and a slower recognition when they need to smooth their earnings. Based on the developers’ incentives, their accounting performance becomes a less reliable indicator of the developer’s firm-level behaviors.

Individual’s decision-making is a utility maximization problem, and only knowing their aggregated decision does not efficiently help in making specific targeted policy decisions. The recent period of slower development has had a larger impact on the Chinese real estate market, and developers need to get used to slower market growth and control their project expansion. As mentioned in the introduction, this research focuses on heterogeneous decisions and discovers the incentives behind them based on individual firms rather than capturing the aggregated market status. This empirical analysis will provide more value when adopting differentiated industrial regulations and policies.

3. Literature Review and Hypotheses

When making investment decisions, the risk-averse utility function of decision-makers can significantly affect their decision-making. The most common utility assumptions assume that people are risk averse, and, most commonly, the utility function assumes constant relative risk aversion (CRRA). The utility function is a logarithm-like function with diminishing returns, which means that a much higher payoff is required to increase the utility function as the utility level increases. The utility function and the calculation of the coefficient of risk aversion are shown in Equations (1) and (2), respectively.

$$U\left(x\right)=\frac{x^{1-\gamma }-1}{1-\gamma }$$

(1)

$$\gamma =-\frac{U^{″}\left(x\right)}{U^{\prime }\left(x\right)}x$$

(2)

Decision-making seeks to maximize utility when individuals are uncertain. If the returns of the investment projects follow a normal distribution with a mean μ and standard deviation σ for a time interval T, then the maximization problem with the decision-maker’s utility function assumed to be CRRA is given by Equation (3).

$$\mu T-\frac{1}{2}(1-\gamma ){\sigma }^{2}T$$

(3)

The decision-maker seeks a distribution that maximizes Equation (3). In Equation (3), the result is affected by the risk aversion coefficient, which is determined by many factors. To make decisions about real estate firm investment projects when managers experience a negative market shock, signaled by a reduction in advance payments from house buyers, the decision is based on project quality and the decision-maker’s risk aversion level. The factors that might affect a manager’s risk aversion level, which makes heterogeneity contingent on the firm’s characteristics and the manager’s career risk, for example, include whether the firm is private or an SOE. The risk aversion level could also depend on the manager’s power as the decision-maker and the level of the internal and external monitoring of project risks. Additionally, the firm’s capital structure and financial constraints may affect the decision.

3.1. Negative Shocks and Project Development

Current Chinese real estate market policy adopts advance payment rules. These rules are a double-edged sword, as mentioned in the introductory section. Managers who work for different firms have different concerns when they experience negative market shocks and decreasing sales. State-owned enterprises are usually required to closely follow policies and need to consider the balance between corporate income and social welfare. SOEs may have an information advantage when purchasing land from government-organized auctions [39] or expanding businesses through mergers and acquisitions [40], but SOE managers may risk their careers if SOEs inefficiently overinvest [41]. SOE managers are usually directly appointed by the government. They have fewer corporate performance concerns but more of their own career concerns. Family-owned corporate firms or corporate firms whose managers have excessive authority may behave more aggressively, making investments [42,43,44] and seeking quicker expansion to develop more projects in different locations to increase sales. Under current advance payment rules, if corporate real estate firms can pay initial land transfer fees to local governments through land auctions, they can use the purchased land as collateral and obtain initial development funds from banks. This initial investment is not too large considering the total amount of money required to finish the project, and the real estate firm could quickly collect this cash from home buyers through advance sales. It is also important for corporate real estate firms to have good relationships with banks to obtain project loan support from them [45,46,47]. Usually, firms with headquarters in larger cities have better connections with banks and are therefore more likely to obtain loan support. When a firm has better financial access, it can take more risks, but this flexibility can also allow managers to postpone the use of added leverage and wait to observe market behavior. Therefore, in firms that are less constrained and have greater financial accessibility, expansion occurs when they experience a negative shock, as we believe that managers are more likely to choose expansion. For corporate real estate firms, if decision-makers have more performance concerns, greater management power and authority, and greater financial resource accessibility, then they may choose to expand aggressively. Therefore, we propose the first set of hypotheses.

H1a. 

SOEs slow their development when they perceive a negative shock.

H1b. 

Firms with duality management do not slow their project development when they perceive a negative shock.

H1c. 

Firms with better financial access are not able to slow their development.

3.2. Heterogeneous Corporate Characteristics, Sales Attitudes, and Sales Strategies

An important factor in the operating efficiency of Chinese real estate firms is the availability of advance sales. In contrast to traditional firms, the cash conversion cycle from the perspective of corporate real estate firm starts from land purchase and land banking [36], and the inflow finishes when the firm finishes collecting advance payments. Future operations after the advance payments entail expense and cash outflow from the firm to finish the project. Regulations require that at least a portion of the advance payments be managed by the supervision banks in which the project has its official bank account; the money is used to guarantee that the project will be finished on time. During periods of market growth, real estate firms seek high turnover and quick expansion, and sales expenses are an important indicator of the desire for turnover [48]. The development of projects and sales are usually separated. Sale is outsourced to real estate agent firms, and real estate corporate firms pay sales commissions to the agent. If corporate real estate firms need higher turnover, they tend to offer larger commissions to sales agents, similar to individual sellers who want to give the broker a strong incentive by providing additional rewards [49]. SOEs need to consider more about the speed at which they finish a project because their finish date is clearly contracted when they receive advance payments from home buyers. The higher corporate social responsibility (CSR) requirements of SOEs make their managers conservative when balancing turnover speed and the risk of project delay [50,51,52]. More performance-sensitive managers and managers with greater authority may be more eager and more likely to seek high growth and turnover for a higher profit, but managers may need rigid rules to successfully meet their growth goals [53]. Corporate real estate firms with fewer constraints should feel less pressure and a lower burden because of bank loan support, and they should also seek high growth rather than being conservative. This strategy would also support Hypothesis H1, as expansion- and high turnover-seeking corporate real estate firms tend to pay higher commissions. Following this logic, we propose the following hypotheses:

H2a. 

SOEs have less aggressive sales strategies.

H2b. 

Duality management firms adopt more aggressive sales strategies.

H2c. 

Firms with greater financial accessibility adopt more aggressive sales strategies.

3.3. Heterogeneous and External Monitoring in Corporate Real Estate Firms

To test our hypotheses of different decision-making practices and the level of aggressiveness concerning expansion in Section 3.1 and Section 3.2, we test the external monitoring of different types of corporate firms in the real estate market. The most common form of the external monitoring of operating status and the supervision of funds to meet the obligations of unearned revenue is auditing [54,55,56]. Audit checks address both the reliability and the timeliness qualities of financial reporting. The quality of auditing could be reflected by the reputation of the audit firm and the auditors who conduct the check [57,58]. Some research has shown that the Big Four auditing firms have a higher auditing quality, but there are arguments against this view [59,60]. Nevertheless, firm-size-adjusted auditing fees could reflect auditing services. Managers pay higher auditing fees when the quality of their financial reports is low, and a high auditing fee could lead to some unethical and misleading auditing reports [61,62]. In addition to external auditing services, internal auditing committees, independent board directors, and firm self-checks are important [63,64]. Managers with greater management power are more likely to conspire with auditing firms to report false financial reports and conceal their true performance. Following the above logic, we propose the following hypotheses:

H3a. 

SOEs experience stricter external monitoring.

H3b. 

Duality management firms experience less external monitoring.

H3c. 

Firms with fewer constraints and better financial accessibility experience less external monitoring.

3.4. Heterogeneous and Management Power in Corporate Real Estate Firms

Manager power is a traditional agency problem and is well connected to corporate real estate decision-making. One of the best indicators of management power is income [65]. If a manager has greater delegation power or is a key member of the team, the manager should receive a higher income than managers in other firms. The results of managers’ powers should reflect and confirm our hypotheses H1 and H2 well. Managers with greater power are more likely to take risks, and agency costs increase with management power [66]. SOEs experience more complicated agency problems [67], one of which involves shareholder–manager delegation. Another problem is the more complicated government and firm delegation. Managers in SOEs should have less power than those in family-owned firms and expect to receive lower compensation [68], especially in terms of the risk-taking aspect, but the opposite is also true. The efficiency of SOEs could be in question [69], especially when their investment is geographically removed from the reach of the local government, which could otherwise provide support as a shareholder [70]. Based on the above discussion, we propose the following hypotheses:

H4a. 

CEOs and managers in SOEs have less power than CEOs and managers of other types of firms.

H4b. 

CEOs and managers in duality management have greater power.

H4c. 

CEOs and managers in less constrained firms have greater power.

3.5. Heterogeneous Leverage in Corporate Real Estate Firms

As mentioned in earlier sections, financial constraints and financial resource accessibility can affect corporate managers’ decision-making. SOEs can easily obtain bank support and greater leverage. The constraints of SOEs are minimized by the support they receive from the local government [71,72,73]. A family-owned firm may not receive high levels of support from banks, and if its management power is concentrated, it could further reduce bank support. Banks treat concentrated management power as a signal of negative corporate governance and a higher firm risk indicator [74,75]. Firms with headquarters in larger cities should have lower leverage since their easy access to banks means that they can flexibly borrow and return [76,77]. Even SOEs have easy access, but government support could significantly lower their borrowing costs. Non-SOEs with better access to financial resources could receive a better rate if they have lower leverage, as preferred by banks that treat corporate leverage as the key measurement of firm credit and operating risk. Therefore, we propose the following hypotheses:

H5a. 

SOEs have higher debt and operating leverage.

H5b. 

Duality management firms cannot have high leverage.

H5c. 

Firms with easy access to financial resources and less constraints have lower leverage.

4. Data and Methodologies

4.1. Data

This study chose a sample period from 2016 to 2023 to capture real estate firm behaviors during periods of steady growth in the real estate market. We selected only mainland stock exchange-listed firms whose main business was classified as real estate by the China Securities Regulatory Commission. We then excluded any firm that was not established before 2016 or any firm classified as a financially distressed firm by the stock exchange. The final sample contained 97 qualified firms. The real estate project data were collected from the China Stock Market and Accounting Research Database (CSMAR), but many firms do not report their projection status, so there were only 200 firm-year observations. We estimated the firms developing decision incentives by calculating their net start developing project in thousand square meters, using the starting project minus the finished project. Further, by matching the listed firms’ stock tickers, we matched the firm-year information collected from the CSMAR with the firm-year operating and financial information collected from the Choice Database. Then, we merged the macroeconomic indicators from those years and obtain the final panel data. The child burden ratio for the year 2023 has not been published, so, for the variable “Child”, the information is missing for one year. The control variables were chosen based on past studies and their overall macro level impact on the real estate industry, which may affect development and project expansion rather than decisions due to firm-level heterogeneities. Chinese families have a higher preference for purchasing real estate than other commodities [1]. The economic growth and wealth effect make significant contributions to the Chinese real estate market’s growth [78]. Real estate is always population-associated, and childbirth and social well-being are highly associated with real estate development [79].

We further use the real estate firm features collected from the Choice Database to explain the heterogeneities in their decision-making. There are three aspects we have special interest in: their different monitoring levels, management power, and accessibility to financial resources. External monitoring can efficiently lower risk and increase corporate governance, particularly in emerging markets [80]. When a manager has a concentrated high level of power, their firm makes inefficient investments and less transparent disclosures [81,82]. The firms located in larger cities with stronger bank competition have increased financial accessibility and could increase their risk and decision management efficiencies [83]. The variables’ definitions, variable symbols used in the Methodologies section, and variable treatments are provided in

Table 1. Variable definitions.

Variable	Symbol	Variable Treatment
The net project building area in ten thousand square meters	Diff	Start minus complete project, in ten thousand square meters
Sales cost, scaled by revenue	Saleexp	Sales cost/revenue
The audit fee, scaled by assets	Audit	Audit fee/total assets
Board chair income	CEO	Observed from the data
Top three managers’ total income	Manager	Observed from the firm’s financial report in the data
Total long-term debt amount	Debt	Observed from the firm’s financial report in the data
Capital structure, the liability ratio	Liabratio	Liability/total assets
State-owned enterprise	SOE	Dummy variable: if the firm is a state-owned enterprise, it equals 1
The manager and the board chair are the same person	Dual	Dummy variable that indicates whether the firm’s board chair and the general manager are the same person
If the headquarters are located in one of the big four cities	Big	Dummy variable that indicates whether the firm’s headquarters are located in Beijing, Shanghai, Shenzhen, or Guangzhou
The first advance payment decreases during the sample period	Shock	Dummy variable: if the firm in has experienced a decrease in the first advance payment in a given year, the shock in that year equals 1, otherwise 0
Return on asset	Asset	Net income/total assets
Interest rate	Int	Interest expense/interest bearing debt
Number of independent board members	Idp	Observed from the firm’s financial report in the data
GDP per capita in a given year	GDP	Observed from Choice data
Consumer price index	CPI	Observed from Choice data
Producer price index of durable goods	PPI	Observed from Choice data
Number of child burdens per 100 members of the working population	Child	Observed from Choice data

. Descriptive statistics are provided in

Table 2. Descriptive statistics.

Statistic	N	Mean	St. Dev.	Min	Pctl (25)	Pctl (75)	Max
Diff	200	123.814	546.799	−2060.900	−35.800	102.042	3977.526
Saleexp	776	4.205	6.556	0	1.858	4.143	64.579
Audit	776	159.068	710.14	0	33.32	121.701	12,371.18
Manager	776	3.959	3.135	0.552	2.018	4.29	17.006
CEO	776	0.957	1.27	0	0	1.2	6.834
Debt	776	12.166	29.877	0	0.345	8.271	242.482
Liab	776	63.659	25.202	7.477	50.191	77.947	454.289
SOE	776	0.446	0.497	0	0	1	1
Dual	776	0.175	0.38	0	0	0	1
Big	776	0.526	0.5	0	0	1	1
Shock	776	0.125	0.331	0	0	0	1
GDP	776	11.172	0.166	10.893	11.067	11.319	11.4
CPI	776	1.775	0.813	0.2	1.425	2.2	2.9
PPI	776	−0.800	0.656	−1.800	−1.275	−0.175	0.1
Child	679	24.34	1.131	22.9	23.4	25.6	26.2
Asset	776	8.776	23.609	0.006	0.784	4.939	193.864
Int	729	14.285	59.699	0	3.469	11.927	1414.467
Idp	776	3.189	0.567	1	3	3	7

.

4.2. Methodologies

4.2.1. Negative Shock and Project Development

The first set of tests involve using the difference-in-differences (DID) method to test the treatment effects on project development, speed expansion, and contraction due to the different characteristics of the firms. The decrease in the advance payment for a year is recorded by the dummy variable “Shock”, and its treatment effect is represented by the interaction term “Shock” and the firm characteristic variables. Equations (4)–(6) represent the characteristics of SOEs, firms with high management power, and firm headquarters located in tier-one cities, respectively.

$${Diff}_{i,t }={\beta }_0+{\beta }_1{SOE}_{i,t}+{\beta }_2{Shock}_{i,t}+{\beta }_3[{SOE}_{i,t}\ast {Shock}_{i,t}]+{\beta }_4{GDP}_{i,t}+{\beta }_5{CPI}_{i,t}+{\beta }_6{PPI}_{i,t}++{\beta }_7{Child}_{i,t}+{\epsilon }_{i,t}$$

(4)

$${Diff}_{i,t }={\beta }_0+{\beta }_1{Dual}_{i,t}+{\beta }_2{Shock}_{i,t}+{\beta }_3[{Dual}_{i,t}\ast {Shock}_{i,t}]+{\beta }_4{GDP}_{i,t}+{\beta }_5{CPI}_{i,t}+{\beta }_6{PPI}_{i,t}++{\beta }_7{Child}_{i,t}+{\epsilon }_{i,t}$$

(5)

$${Diff}_{i,t }={\beta }_0+{\beta }_1{Big}_{i,t}+{\beta }_2{Shock}_{i,t}+{\beta }_3[{Big}_{i,t}\ast {Shock}_{i,t}]+{\beta }_4{GDP}_{i,t}+{\beta }_5{CPI}_{i,t}+{\beta }_6{PPI}_{i,t}++{\beta }_7{Child}_{i,t}+{\epsilon }_{i,t}$$

(6)

4.2.2. Heterogeneous Characteristics, Sales Attitudes and Sales Strategies

The second set of tests replace the project expansion measure in Equations (1)–(3) with the brokerage or commission expense paid by the real estate developer. In addition, these tests change the method from DID to traditional ordinary least squares, with time controls to determine robustness and test whether similar results to those in Equations (4)–(6) can be obtained. Furthermore, the time control of a year is replaced by macroeconomic indicators, including GDP, CPI, PPI and working-age people’s child raising burden, to test for robustness again. Equations (7) and (8) show the OLS regression results. Note that firm characteristics are used to test this relationship using the variables “SOEs”, “Dual”, and “Big”.

$${Saleexp}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(7)

$$\begin{array}{ll}{Saleexp}_{i,t }={\beta }_0& +{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(8)

4.2.3. Heterogeneous and External Monitoring

Furthermore, the monitoring levels of developers with different characteristics are tested. Equations (9) and (10) show the relationship between these characteristics and the auditing fee, scaled by asset. The audit fee is the monitoring-level indicator, and a higher fee means lower auditing quality and lower monitoring. If a characteristic leads to a higher audit fee, then that characteristic leads to lower monitoring, and managers are more likely to take greater risks. As above, the firm characteristic term in the equation indicates the three different features of different firms, and the test is repeated by replacing the time control variable with macroeconomic indicators.

$${Audit}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(9)

$$\begin{array}{ll}{Audit}_{i,t }=& {\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(10)

4.2.4. Heterogeneous and Management Authority

The next test reconfirms management power and authority. Usually, SOEs expect managers to have less authority and power, and firms with dual or more family-owned features should have greater management authority. Management authority is reflected by the managers’ income. Here, we use two indicators to represent management authority, the CEO’s wage and the income sum of the top three managers. A higher manager income should indicate greater management authority. Equations (11)–(14) show these relationships:

$${CEO}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(11)

$$\begin{array}{ll}{CEO}_{i,t }=& {\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(12)

$${Manager}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(13)

$$\begin{array}{ll}{Manager}_{i,t }& ={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(14)

4.2.5. Heterogeneity, Liability, and Capital Structure

The last set of tests evaluates the relationship between firms’ characteristics and their debt and capital structures. Equations (15) and (16) show the firms’ characteristics and their long-term debt relationship, and Equations (17) and (18) show the characteristics and capital structure relationship. For firms with government support, SOEs are expected to have higher debt and a higher liability ratio. With more family-owned firms, when the manager has greater power, banks are less likely to issue loans due to concentrated management power, and the firms are expected to have less debt. Firms with financial flexibility are also expected to have less debt and a lower liability ratio since they can easily add leverage by borrowing from banks.

$${Debt}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(15)

$$\begin{array}{ll}{Debt}_{i,t }=& {\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(16)

$${Liab}_{i,t }={\beta }_0+{\beta }_1{[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+\sum Year+{\epsilon }_{i,t}$$

(17)

$$\begin{array}{ll}{Liab}_{i,t }={\beta }_0+{\beta }_1& {[Firm Charateristic]}_{i,t}+{\beta }_2{Asset}_{i,t}+{\beta }_3{Int}_{i,t}+{\beta }_4{Idp}_{i,t}+{\beta }_4{GDP}_{i,t}+{\beta }_4{CPI}_{i,t}+{\beta }_4{PPI}_{i,t}\\ & +{\beta }_4{Child}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(18)

5. Results

5.1. Negative Shock and Project Expansion Speed

Table 3. Real estate firm characteristics and net expansion of their projects.

	Dependent Variable
	Diff
	(1)	(2)	(3)
SOE	−136.971 *
	(81.78)
Dual		80.953
		(108.274)
Big			298.039 ***
			(105.647)
Shock	375.964 **	66.435	283.738 **
	(155.833)	(132.447)	(127.215)
SOE*Shock	−418.382 *
	(217.287)
Dual*Shock		526.777 **
		(266.528)
Big*Shock			−352.051 ***
			(130.146)
GDP	−747.115	−728.260	−246.318
	(630.947)	(636.196)	(659.042)
CPI	111.139	122.075	109.661
	(86.044)	(86.752)	(86.457)
PPI	90.101	80.805	56.203
	(98.507)	(99.101)	(98.985)
Child	7.925	14.403	15.277
	(85.731)	(86.503)	(86.013)
Constant	8120.801	7654.336	2221.142
	(5179.260)	(5218.944)	(5570.276)
Observations	200	200	200
R2	0.108	0.091	0.101
Adjusted R2	0.075	0.057	0.068
Residual Std. Error (df = 192)	525.781	530.866	527.961
F Statistic (df = 7; 192)	3.318 ***	2.732 **	3.065 ***

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

shows the results of the firms’ characteristics and expansion after a negative advance payment decrease using the difference-in-differences methods. The interesting variable is the interaction term between the three firm characteristics and the “Shock” term. The coefficient for SOEs shows a negative net project expansion. The managers of SOEs are more conservative when they experience a negative shock. As seen in the second column, firms with more concentrated management power or that are more family-owned, when there is a negative shock to advance payments, expand their projects and seek higher income performance. When a firm has headquarters in a tier-one large city, its development speed decreases when a negative shock occurs. Firms with high financial accessibility may be very confident about their speed of expansion because receiving bank support more easily allows them to rapidly switch from being conservative to expansionary. Regarding our hypotheses, H1 and H2 are supported by the results, but H3 is not.

5.2. Firm Characteristics and Sales Expenses

Firm characteristics and sales expenses also measure firm expansion and contraction. In

Table 4. Firm characteristics and sales expenses.

	Dependent Variable
	Saleexp
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	−2.499 ***			−2.478 ***
	(0.463)			(0.494)
Dual		4.191 ***			4.171 ***
		(0.584)			(0.624)
Big			−0.542			−0.593
			(0.473)			(0.503)
Asset	−0.019 *	−0.015	−0.009	−0.018 *	−0.014	−0.008
	(0.010)	(0.010)	(0.010)	(0.011)	(0.010)	(0.011)
Int	−0.004	−0.002	−0.002	−0.003	−0.002	−0.002
	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)
IDP	−1.743 ***	−1.346 ***	−1.767 ***	−1.832 ***	−1.406 ***	−1.826 ***
	(0.416)	(0.416)	(0.429)	(0.452)	(0.451)	(0.466)
GDP				1.491	1.066	1.018
				(3.657)	(3.603)	(3.723)
CPI				0.061	0.054	0.075
				(0.519)	(0.512)	(0.529)
PPI				0.008	0.03	0.051
				(0.614)	(0.605)	(0.625)
Child				0.038	0.027	0.037
				(0.491)	(0.484)	(0.501)
Constant	11.188 ***	7.898 ***	10.261 ***	−6.417	−4.661	−2.01
	(1.468)	(1.473)	(1.486)	(30.629)	(30.176)	(31.187)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.071	0.098	0.035	0.07	0.097	0.035
Adjusted R2	0.057	0.084	0.02	0.059	0.086	0.023
Residual Std. Error	6.103 (df = 717)	6.013 (df = 717)	6.220 (df = 717)	6.099 (df = 632)	6.011 (df = 632)	6.213 (df = 632)
F Statistic	4.987 *** (df = 11; 717)	7.096 *** (df = 11; 717)	2.376 *** (df = 11; 717)	5.983 *** (df = 8; 632)	8.498 *** (df = 8; 632)	2.904 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

, SOEs have a negative significant coefficient in the first column, and firms with high management concentrations generate significant positive results. The location of the headquarters does not have a significant effect. With the exception of fewer constraints and easily financially accessible firms, the remaining results confirm the findings reported in the previous section using the DID method. SOE managers behave conservatively, and more family-owned firms with concentrated management power have various negative coefficients, which indicates that SOEs are not in a hurry to expand their projects when they have future concerns, but that firms with a high concentration of management power increase their sales commission to seek greater growth. Furthermore, the time control was replaced with macroeconomic indicators, and the regressions were retested. The results in Columns (4) to (6) are similar to those in Columns (1) to (3), which indicates the robustness of the test.

5.3. Heterogeneous and External Monitoring

Using the audit fee as an indicator of the level of monitoring, a higher fee indicates a lower monitoring level, and a lower fee indicates a higher monitoring level. The results for firms’ characteristics and their monitoring are shown in

Table 5. Firm characteristics and monitoring.

	Dependent Variable
	Audit
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	−41.718 ***			−40.800 ***
	(8.491)			(8.708)
Dual		0.613			−0.723
		(11.041)			(11.365)
Big			−14.829 *			−10.858
			(8.629)			(8.858)
Asset	−1.272 ***	−1.117 ***	−1.085 ***	−1.244 ***	−1.092 ***	−1.071 ***
	(0.180)	(0.181)	(0.181)	(0.186)	(0.187)	(0.187)
Int	0.012	0.032	0.033	0.01	0.025	0.027
	(0.070)	(0.071)	(0.071)	(0.069)	(0.070)	(0.070)
IDP	−19.380 **	−21.035 ***	−18.924 **	−18.890 **	−20.294 **	−18.623 **
	(7.628)	(7.859)	(7.835)	(7.968)	(8.216)	(8.193)
GDP				23.713	15.827	15.937
				(64.501)	(65.589)	(65.511)
CPI				−5.216	−5.027	−4.97
				(9.155)	(9.313)	(9.302)
PPI				0.696	1.354	1.411
				(10.825)	(11.011)	(10.998)
Child				0.799	0.828	0.771
				(8.668)	(8.818)	(8.807)
Constant	202.138 ***	185.938 ***	186.905 ***	−86.89	−14.466	−14.271
	(26.905)	(27.848)	(27.104)	(540.281)	(549.406)	(548.709)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.109	0.079	0.083	0.107	0.076	0.078
Adjusted R2	0.096	0.065	0.069	0.096	0.064	0.067
Residual Std. Error	111.833 (df = 717)	113.699 (df = 717)	113.466 (df = 717)	107.584 (df = 632)	109.436 (df = 632)	109.306 (df = 632)
F Statistic	8.004 *** (df = 11; 717)	5.621 *** (df = 11; 717)	5.912 *** (df = 11; 717)	9.473 *** (df = 8; 632)	6.504 *** (df = 8; 632)	6.707 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

. SOEs have a negative coefficient, which indicates their lower audit fees. For firms with concentrated management power, these coefficients are negative but not significant. Firms with fewer constraints also have negative but nonsignificant coefficients. Therefore, the only conclusion we can draw from the audit fees is that SOEs are especially monitored. Thus, H3a is supported, but H3b and H3c are rejected.

5.4. Firm Characteristics and Wages

Furthermore, the wages of CEOs and managers were used to measure management power. In

Table 6. Firm characteristics and CEO wages.

	Dependent Variable
	CEO
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	−0.261 ***			−0.264 ***
	(0.092)			(0.098)
Dual		0.756 ***			0.744 ***
		(0.114)			(0.123)
Big			−0.144			−0.13
			(0.092)			(0.099)
Asset	0.017 ***	0.017 ***	0.018 ***	0.017 ***	0.017 ***	0.018 ***
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Int	0.002 ***	0.002 ***	0.002 ***	0.002 ***	0.003 ***	0.002 ***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
IDP	0.073	0.152 *	0.083	0.039	0.121	0.05
	(0.082)	(0.081)	(0.084)	(0.090)	(0.089)	(0.091)
GDP				−0.371	−0.412	−0.421
				(0.728)	(0.712)	(0.731)
CPI				−0.002	−0.004	0.0004
				(0.103)	(0.101)	(0.104)
PPI				0.016	0.017	0.02
				(0.122)	(0.120)	(0.123)
Child				0.012	0.01	0.011
				(0.098)	(0.096)	(0.098)
Constant	0.609 **	0.088	0.516 *	4.631	4.63	5.104
	(0.290)	(0.289)	(0.289)	(6.102)	(5.966)	(6.126)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.138	0.178	0.131	0.131	0.169	0.123
Adjusted R2	0.124	0.165	0.118	0.12	0.158	0.112
Residual Std. Error	1.207 (df = 717)	1.179 (df = 717)	1.212 (df = 717)	1.215 (df = 632)	1.188 (df = 632)	1.220 (df = 632)
F Statistic	10.402 *** (df = 11; 717)	14.107 *** (df = 11; 717)	9.816 *** (df = 11; 717)	11.891 *** (df = 8; 632)	16.032 *** (df = 8; 632)	11.114 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

and

Table 7. Firm characteristics and manager wages.

	Dependent Variable
	Manager
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	−1.566 ***			−1.534 ***
	(0.206)			(0.221)
Dual		1.057 ***			1.058 ***
		(0.270)			(0.291)
Big			0.610 ***			0.621 ***
			(0.213)			(0.228)
Asset	0.043 ***	0.047 ***	0.048 ***	0.042 ***	0.047 ***	0.047 ***
	(0.004)	(0.004)	(0.004)	(0.005)	(0.005)	(0.005)
Int	0.001	0.002	0.002	0.002	0.003	0.002
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
IDP	1.355 ***	1.417 ***	1.201 ***	1.337 ***	1.416 ***	1.197 ***
	(0.185)	(0.192)	(0.193)	(0.202)	(0.210)	(0.211)
GDP				−0.14	−0.423	−0.442
				(1.637)	(1.680)	(1.688)
CPI				0.027	0.029	0.031
				(0.232)	(0.239)	(0.240)
PPI				−0.007	0.013	0.014
				(0.275)	(0.282)	(0.283)
Child				−0.04	−0.042	−0.036
				(0.220)	(0.226)	(0.227)
Constant	−0.065	−1.250 *	−0.686	2.488	4.531	5.157
	(0.651)	(0.682)	(0.668)	(13.711)	(14.074)	(14.136)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.288	0.247	0.239	0.275	0.236	0.229
Adjusted R2	0.277	0.235	0.228	0.266	0.226	0.219
Residual Std. Error	2.707 (df = 717)	2.785 (df = 717)	2.798 (df = 717)	2.730 (df = 632)	2.803 (df = 632)	2.816 (df = 632)
F Statistic	26.405 *** (df = 11; 717)	21.353 *** (df = 11; 717)	20.517 *** (df = 11; 717)	30.004 *** (df = 8; 632)	24.395 *** (df = 8; 632)	23.467 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

, the measurements of different firm characteristics show their impact on the CEOs’ and top three managers’ total wages. SOEs have negative coefficients in both tables. This result indicates that the managers of SOEs do not have excessive authority. The wage of managers reveals two things, it indicates a greater concentration of management power and confirms that the managers received higher incomes, which also indicates greater management power. Even for firms in large cities, the coefficient of the CEO’s wages is not significant, but the combined wage of their top three managers is significant and positive. Therefore, firm managers can make decisions more easily if the firm has fewer constraints and better financial access.

5.5. Heterogeneity and Leverage

The last set of tests are concerned with how firm leverage is affected by different firm characteristics. The results in

Table 8. Firm characteristics and long-term debt.

	Dependent Variable
	Debt
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	2.396 **			2.282 **
	(1.019)			(1.065)
Dual		−4.337 ***			−3.315 **
		(1.299)			(1.365)
Big			1.528			1.569
			(1.023)			(1.069)
Asset	1.143 ***	1.139 ***	1.131 ***	1.113 ***	1.109 ***	1.101 ***
	(0.022)	(0.021)	(0.021)	(0.023)	(0.022)	(0.023)
Int	−0.009	−0.011	−0.01	−0.008	−0.01	−0.009
	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
IDP	0.192	−0.227	0.066	0.403	0.075	0.248
	(0.916)	(0.924)	(0.929)	(0.975)	(0.987)	(0.988)
GDP				−5.875	−5.477	−5.449
				(7.891)	(7.880)	(7.904)
CPI				0.966	0.971	0.947
				(1.120)	(1.119)	(1.122)
PPI				1.139	1.116	1.093
				(1.324)	(1.323)	(1.327)
Child				0.807	0.816	0.813
				(1.060)	(1.059)	(1.063)
Constant	0.918	4.251	1.764	44.931	42.984	40.811
	(3.230)	(3.276)	(3.214)	(66.099)	(66.010)	(66.198)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.811	0.813	0.81	0.806	0.806	0.805
Adjusted R2	0.808	0.81	0.807	0.804	0.804	0.803
Residual Std. Error	13.425 (df = 717)	13.374 (df = 717)	13.456 (df = 717)	13.162 (df = 632)	13.148 (df = 632)	13.187 (df = 632)
F Statistic	279.999 *** (df = 11; 717)	282.687 *** (df = 11; 717)	278.424 *** (df = 11; 717)	328.192 *** (df = 8; 632)	329.026 *** (df = 8; 632)	326.636 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

and

Table 9. Firm characteristics and their liability ratio.

	Dependent Variable
	Liabratio
	(1)	(2)	(3)	(4)	(5)	(6)
SOE	4.464 ***			4.235 ***
	(1.272)			(1.334)
Dual		−2.878 *			−2.768
		(1.637)			(1.721)
Big			−6.177 ***			−6.571 ***
			(1.264)			(1.320)
Asset	0.219 ***	0.206 ***	0.215 ***	0.220 ***	0.208 ***	0.218 ***
	(0.027)	(0.027)	(0.026)	(0.029)	(0.028)	(0.028)
Int	−0.019 *	−0.022 **	−0.021 **	−0.021 **	−0.023 **	−0.022 **
	(0.010)	(0.011)	(0.010)	(0.011)	(0.011)	(0.010)
IDP	3.517 ***	3.358 ***	4.612 ***	3.962 ***	3.763 ***	5.057 ***
	(1.142)	(1.165)	(1.147)	(1.220)	(1.244)	(1.221)
GDP				2.705	3.487	3.583
				(9.880)	(9.934)	(9.765)
CPI				−0.369	−0.376	−0.352
				(1.402)	(1.411)	(1.387)
PPI				−0.687	−0.743	−0.718
				(1.658)	(1.668)	(1.639)
Child				−0.722	−0.716	−0.758
				(1.328)	(1.336)	(1.313)
Constant	49.322 ***	52.626 ***	51.279 ***	35.859	30.125	28.782
	(4.029)	(4.128)	(3.969)	(82.755)	(83.215)	(81.791)
Year Control	Y	Y	Y	N	N	N
Observations	729	729	729	641	641	641
R2	0.128	0.117	0.141	0.134	0.123	0.153
Adjusted R2	0.114	0.103	0.128	0.123	0.112	0.142
Residual Std. Error	16.748 (df = 717)	16.855 (df = 717)	16.616 (df = 717)	16.479 (df = 632)	16.576 (df = 632)	16.293 (df = 632)
F Statistic	9.547 *** (df = 11; 717)	8.601 *** (df = 11; 717)	10.733 *** (df = 11; 717)	12.190 *** (df = 8; 632)	11.126 *** (df = 8; 632)	14.276 *** (df = 8; 632)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses.

show that in both cases, SOEs have positive and significant coefficients, which indicates that SOEs have higher levels of debt and leverage. For firms with concentrated management, which are most likely more family-owned firms, the debt and leverage are significantly smaller. Banks do not prefer lending to firms with concentrated management or to family-owned firms, since bankers believe that high levels of management power indicate higher risk-taking probabilities. For firms with headquarters in large cities, even their debt level is insignificant, but they do have significant negative leverage ratios. Most firms with headquarters in large cities are large, and larger firms tend to have more long-term debt. They also have greater equity, which leads to a smaller leverage ratio. Moreover, this finding is in line with banks having better financial accessibility, so these firms can easily obtain bank loans to support them when they have short-term liquidity problems.

5.6. Summary of Findings

Table 10. Summary of findings.

Hypotheses	Validation	Discussion
H1a. SOEs slow their development when facing negative shocks.	Supported	The DID tests show that SOEs reduce their speed of development when facing negative shocks, but duality managers choose to expand to collect larger advance payments.
H1b. Firms with duality management do not slow their development when facing negative shocks.	Supported
H1c. Firms with better financial accessibilities do not slow their development speed.	Rejected
H2a. SOEs have less aggressive sales strategies.	Supported	SOEs do not pursue aggressive sales strategies, whereas duality management firms adopt more aggressive sales strategies.
H2b. Duality management firms adopt more aggressive sales strategies.	Supported
H2c. Firms with more financial accessibility adopt more aggressive sales strategies.	Rejected
H3a. SOEs experience stricter external monitoring.	Supported	Only SOEs were found to have much lower auditing fees, which indicates a lower level of monitoring.
H3b. Duality management firms experience less external monitoring.	Rejected
H3c. Firms with fewer constraints and better financial accessibility experience less external monitoring.	Rejected
H4a. CEOs and managers in SOEs have less power than in other types of firms.	Supported	Managers in SOEs have less management power, as reflected by their lower wages. Duality management firms and firms with fewer constraints have more management power and pay higher wages to management.
H4b. CEOs and managers in duality management firms have more power.	Supported
H4c. CEOs and managers in firms with fewer constraints have more power.	Supported
H5a. SOEs have higher debt and operating leverage.	Supported	SOEs receive support from local government, so they have higher debt and leverage. Duality management is not preferred by banks, so duality management firms have low leverage. Better access to bank loans leads to low leverage in firms with fewer constraints and better financial access.
H5b. Duality management firms cannot have high leverage.	Supported
H5c. Firms with easy access to financial resources and fewer constraints have less leverage.	Supported

summarizes the findings in terms of our hypotheses.

5.7. Discussion

SOE real estate developers receive greater support from the government. Their managers are more conservative and have limited management power compared to managers in other firms. Under negative shock and uncertainty, they face significant career risk if they overinvest and if the expansion causes a performance crisis, while the reward of taking the risk and experiencing success is limited compared with that for other types of developers, which leads them to be more conservative. This finding confirms the past study that found that SOEs consider their social obligation and tend to make more socially responsible decisions [84]. Also, the recent emphasis on corporate social responsibility (CSR) has efficiently shaped corporate decision-making, and SOEs have shown that it is their mission to bridge the gap between shareholder value maximization and social obligation fulfillment [85].

Duality management real estate firms are usually more family-owned. They do not have much external support, and it is more difficult for them to obtain bank loans. Managers in these firms have high levels of concentrated power. They pursue more aggressive expansion to start new projects when they experience a decrease in advance housing payments, which are a negative shock in the steady-growth real estate market. Our finding confirms that managers’ concentrated power could increase their firm’s costs. In family-owned firms, significant family members in the top management team can adversely affect firm performance [86]. Their larger controlling interest in emerging markets increases their likelihood of paying management highly and lowers their disclosure quality [87]. However, the advance payment method requires transparency disclosure for home buyers to trust a family-owned developer, and then there is a significant conflict of interest.

Firms with headquarters in larger cities are usually large developers. They have easier access to financial resources. When they experience a negative shock in a steady-growth market, their reaction is to slow down and observe. Managers in these firms have greater power. These firms have lower leverage than SOEs, but they can quickly obtain bank support and expand to start new projects when they determine it is appropriate to do so. This finding confirms traditional firm–bank relationship studies. Firms located in larger cities have more diversified bank connections, and it is easier to access loans at better lending rates and receive other financial services [88,89].

These empirical results show that family-owned firms are more likely to expand project development and seek higher leverage. Recent policies have emphasized that their total leverage should be controlled, and these regulations have set a threshold. Only relying on leverage control is not enough for them to manage their incentive well. The focus should shift more to advance payment monitoring. The developers’ main bank or local branch should execute monitoring and ensure the cash outflow is well controlled. Hence, the leftover amount in the account should be enough to cover the current project costs. Also, project cost estimation should be conservative to avoid cost management errors.

6. Conclusions

This research revealed that different real estate firm characteristics led to different project development attitudes under negative shock after the real estate market slowed following its transition from high growth to moderate growth in China. We differ from past researchers, who focused more on market- and macro-level analyses. Our research explores the micro-level: firms’ heterogenous decision-making due to different firm-level characteristics. Also, we follow on from the discussion that the accounting performance of real estate developers could be incentive-driven and biased. We adopt the reduction in advance payment as the negative shock signal and use construction projects rather than profitability and other accounting information to test expansion incentives. This research minimizes the involvement of accounting estimation. It is more focused on directly observable evidence, which is used to run DID tests to demonstrate the heterogeneous development expansion incentives of managers from different types of firms. The results show that SOEs, as developers, have the most conservative attitudes toward risk expansion. Firms with duality management, which are more likely family-owned firms, will increase their number of projects to boost their sales when they have slower growth. Firms headquartered in large cities tend to wait because they have better financial access, and they can always quickly engage in new project development. Our findings contribute to valuable policymaking and policy differentiation decision-making. The firms’ different characteristics should be well addressed when designing industrial policies.

Under advance payment regulations, policy and regulation should prevent developers from engaging in quick expansion because developers have a higher incentive to use funds for the future development of existing projects to start new projects. From this perspective, it appears that SOEs are better candidates for housing development in China. However, SOEs suffer from efficiency problems, and when there is no or low market competition, their efficiency could further decrease, while their costs could significantly increase. Banks may be able to play a more essential role when monitoring project development and releasing funds from the sale of houses, which developers can withdraw from their current development projects. This policy should encourage banks to support more non-SOE developers but with slightly differentiated treatment when issuing loans. Banks can set strict withdrawal threshold levels, and these threshold levels and lending costs could be related to firms’ leverage and debt levels. When both leverage and debt are large, not only do borrowing costs increase, but the monitoring of funds withdrawn from developing projects also becomes difficult.

The recent downturn of the real estate market in mainland China has led to the default of few large Hong Kong-based family-owned developers, many of which are in crisis. The current research data are limited as they involve only the developers listed on the Chinese stock exchange. They do not cover real estate firms with headquarters in Hong Kong due to limitations in gathering accounting information and accounting recognition differences. Future research could examine geographical differences and development strategies, as most developers from southern China seem to have higher financial and operating leverage. Is this the case because developers from southern China have better local policy support and bank support? Also, even the current empirical results show that managers of SOEs make conservative decisions; however, the age difference of these managers could indicate that they have different future political careers, which should be considered along with their political connections when exploring their decision-making processes. Additionally, developer–bank relationships could be further examined. A better bank relationship increases the likelihood of a firm having higher leverage, but the relationship could also increase their engagement speed, similar to developers in larger cities. If it is possible for rapid engagement to lessen blind expansion, then policy and regulations should consider the power of these developer–bank relationships.