Spatiotemporal Evolution and Cause Analysis of Urban Housing Investment Resilience: An Empirical Study of 35 Large and Medium-Sized Cities in China

Abstract

In this study, we explore the evolution and formation mechanism of urban housing investment resilience from three perspectives: theoretical analysis, model construction, and empirical testing. Based on the three-element theory of investment decision making and urban resilience system theory, a theoretical framework of urban housing investment resilience is constructed. Spatiotemporal analysis and qualitative comparison methods are used to divide 35 large and medium-sized cities into two categories, first-tier and non-first-tier cities, and their spatiotemporal characteristics and differences in terms of formation mechanism differences then explored. The results show that (1) the overall housing investment resilience of the 35 investigated cities is low, with the characteristics of periodic evolution, and there are obvious differences between the first-tier and non-first-tier cities as well as unbalanced development problems. (2) The three internal investment decision factors of returns, costs, and expectations and the five external support factors of economic growth, infrastructure development, labor market, policy regulation, and monetary policy do not, by themselves, constitute the necessary conditions for high levels of urban housing investment resilience, and there are three paths for the development of high levels of housing investment resilience in both first-tier and non-first-tier cities. (3) The twos types of cities have the same conformational path of "return- and cost-driven" but different dedicated conformational paths, including "cost-driven", "expectation- driven", and "return- and expectation-driven", and the core conditions in their driving paths are different, with real estate policy being the core condition in the path of first-tier cities and infrastructure development and labor markets being the core conditions in the path of non-first-tier cities. (4) There is a potential substitution relationship between the three configuration paths of first-tier and non-first-tier cities, and the substitution relationship between the two types of cities is also different. The findings of this study reveal that the complex interaction between the urban resilience system, represented by infrastructure construction, and multiple factors, including the three elements of investment, can have an impact on the resilience of real estate investment.

1. Introduction

As a pillar industry of the national economy, the real estate industry occupies an important position in terms of promoting regional economic growth and boosting domestic demand and urbanization in China. Housing investment is a core component of the healthy development of the real estate industry. Although the industry is currently facing uncertainties in terms of the external environment and policies, housing investment is characterized by a resilient evolution of alternating expansion and contraction. During the economic upturn, to promote the stable and healthy development of the real estate market, the government successively put forward the following policies: “houses are for living, not for speculation”; “stable land prices, stable house prices and stable expectations”; “real estate should not be used as a means of stimulating the economy in the short term”; and “solving the outstanding housing problems in big cities”. During the economic downturn, especially after the financial crisis in 2008, real estate housing investment remained at around 20% of fixed asset investment and 13% of GDP. Real estate housing investment grew at an average annual rate of 9.1% from 2011 to 2019, higher than the average annual GDP growth rate. It is evident that housing investment is an important driving force for the economic repair of cities, showing strong resilience. Accordingly, how can urban housing investment resilience be measured scientifically? What are the factors affecting housing investment resilience? Answering these questions will provide theoretical references to help strengthen urban housing investment resilience in China moving forward.

Resilience, as a concept in physics, refers to the process of recovery and growth of an object under the impact of external forces [1]. With the development of resilience theory has seen the concept of resilience introduced in various disciplines, such as urban planning, ecology, and management, with researchers conducting theoretical explorations [2]. In 2007, the Resilience Alliance proposed the concept of urban resilience, specifically including urban facilities and the urban environment, metabolic flows, governance networks, and social dimensions [3]. At present, research on urban resilience focuses mostly on issues such as socioeconomic resilience, spatial planning, urban clusters, and urban governance [4,5,6], and panel models, system coupling, comprehensive evaluation, network analysis, and other methods are used to perform conceptual analyses of urban resilience, spatiotemporal evolution, influencing factors, and formation mechanisms [7,8,9]. Most studies have focused on the developmental characteristics of regional urban resilience and the evolution of its industrial structure. Whereas accurately measuring industrial resilience and characterizing its evolutionary development pattern and influence mechanism are hot topics, there have been few studies on the measurement of resilience and the spatiotemporal evolution of urban housing investment; research on the formation mechanisms of housing investment resilience according to city type is also lacking. Scholars have mainly studied housing investment in terms of the investment environment and differences in investment types [10,11] and have argued that the level of housing investment is influenced by the economic growth rate, housing stock, total social investment, population size, and housing policies [12,13]. Research on housing investment in China can be broadly categorized into two types of influencing factors: housing supply (economic scale, land supply, ecological environment, etc.) and market demand (spatial game, ability to pay, market expectations, etc.) [14,15]. The choice of research objects occurs at the micro, meso, and macro levels, and the temporal characteristics and spatial aggregation evolution patterns of housing investment in major cities are explored with the help of spatiotemporal evolutionary analysis [16]. In addition, some scholars have analyzed cross-regional choices in real estate enterprise investment and the impact of population mobility on residents’ housing investment, starting with the housing investment behavior of enterprises and residents [17,18]. In general, there is a foundation of research on urban housing investment and its influencing factors but a lack of perspectives on resilience in the context of the cyclical pattern of housing investment. Furthermore, most studies have analyzed the net effect of various factors on housing investment as a whole from a regional perspective. Urban housing investment is closely linked to regional socioeconomic development, but few studies have considered identifying different city classes and further exploring the spatiotemporal differences in the development of housing investment in different types of cities, as well as their formation mechanisms.

In this study, the housing investment resilience of 35 large and medium-sized cities in China is measured using an economic resilience approach based on the three-factor theory of investment decisions and the requirements of resilient city development. The fuzzy set qualitative comparative analysis method is used to analyze the differences in housing investment resilience between first-tier and non-first-tier cities, and similarities and differences in the factors driving housing investment resilience in the two types of cities are further explored. On the basis of theoretical analysis, model construction, and empirical testing, this study comprehensively reveals the spatiotemporal evolution of urban housing investment resilience and the mechanism of its formation, providing a reference for the promotion high resilience and quality development in urban housing investment in China moving forward.

2. Research Methodology and Data Processing

2.1. Research Methodology

2.1.1. Housing Investment Resilience Measure

(1) Housing investment growth characteristics

In 2011, the “Eight New National Policy Measures” specified an increase in the down payment ratio for home purchases, and in 2013, the “Five New National Policy Measures” proposed strict restrictions on purchases. Furthermore, the growth rate of real estate investment in housing declined significantly between 2011 and 2013, indicating a shift in China’s housing investment from “overheated” to “slightly overheated” under the influence of national policy regulation. Subsequently, the PMI, an indicator of economic prosperity, was below the Ronggu line for six consecutive months in 2015, and China’s economy entered “three overlapping periods” (a period of shifting growth rates, a period of painful structural adjustment, and a period of digestion of economic stimulus). The growth rate of real estate housing investment remained in decline during 2014–2015, showing the characteristics of a cold phase. In 2015, the real estate industry entered a period of supply-side structural adjustment in order to dissolve real estate inventories and promote the healthy development of the industry, which helped the market heat up again. The 2016 Central Economic Conference proposed that “houses are for living, not for speculation”, and although real estate policy was still in a period of tightening, the growth rate of investment in real estate housing showed a steady rebound from 2016 to 2019 (Figure 1). In short, although the growth rate of real estate investment declined rapidly, it was able to rebound steadily under relevant policies and industrial structural adjustments, showing an overall trend of rebound after a decline, revealing the distinctive resilience of real estate.

(2) Housing investment resilience indicator

In this study, we draw on Martin’s measure of economic resilience [19] to construct a housing investment resilience indicator. Using the magnitude of fluctuations in the amount of housing investment at the time of the shock as a reference point, urban housing investment resilience is measured as the ratio of the change in actual and counterfactual investment fluctuations, representing the ratio of relative impediment to recovery. Counterfactual data refer to the difference in each city between the change in housing investment and the expected change, providing feedback on the tendency for housing investment to decline or rebound. This method not only enables an examination of the resilience of the urban housing investment system in the event of a shock but also provides a useful measure of housing investment resilience. It assumes that the volatility of housing investment resilience in a given city during a recession will follow the national rate of contraction after a shock and the national rate of expansion during a recovery period. In conjunction with the cumulative growth rate of housing investment in China discussed in the previous section, the ratio of actual and counterfactual changes in housing investment volatility will be scaled according to two cycles, 2011–2013 and 2014–2019, as calculated by the following formula:

$${Resilience}_i^t=\frac{\frac{\left(G_i^t-G_i^{t-k}\right)}{G_i^{t-k}}-\frac{\left(G_r^t-G_r^{t-k}\right)}{G_r^{t-k}}}{\left|\frac{\left(G_r^t-G_r^{t-k}\right)}{G_r^{t-k}}\right|}$$

(1)

where $Resilienc{e}_i^t$ is the housing investment resilience index for city i in year t; $G_i^t$ and $G_i^{t-k}$ are the amount of property development investment in city i in years t and t−k, respectively; and $G_r^t$ and $G_r^{t-k}$ are the sums of property development investment in cities in years t and t−k, respectively.

2.1.2. Qualitative Comparative Analysis Methods

The process of shaping the resilience of urban housing investments is complex and often difficult to explain with a single factor. Qualitative comparative analysis (QCA) can identify how a number of factors affect the final outcome in combination. Accordingly, in this study, we use the QCA approach to analyze the mechanisms that shape the resilience of urban housing investment. The QCA approach is based on the idea of set theory and Boolean algebra operations, combining the advantages of qualitative and quantitative analysis to investigate the ideal set of paths leading to the final outcome from a combination of multiple antecedent variables [20]. The method takes a holistic view of the relationship between factor configurations and outcome variables, emphasizing the complexity of causal relationships and the existence of multiple pathways that can produce the same outcome [21]. Consistency and coverage are calculated, and the antecedent configurations that have the greatest influence on the outcome variable are selected. The formula is as follows:

$$consistency\left(X⩽Y\right)={{\displaystyle \sum }}^{\sum \left(x_i,y_i\right)\sum x_i}min$$

(2)

$$coverage\left(X⩽Y\right)={{\displaystyle \sum }}^{\sum \left(x_i,y_i\right)\sum y_i}min$$

(3)

where X is the set of all antecedent variables, Y is the set of outcome variables, x_i is the individual antecedent condition, and y_i is the outcome variable corresponding to x_i. Consistency is a sufficient condition for determining whether X is a sufficient condition for Y. A sufficient condition is considered to hold if it is higher than 0.75; if the sufficiency of a single variable (x_i) is greater than 0.9, then x_i is a necessary condition for Y [22]. Coverage describes the strength of X’s explanation of Y. The greater the coverage, the stronger the explanation of the outcome variable (Y) by the histogram path (X).

2.2. Theoretical Framework Construction and Analysis of Influencing Factors

Housing investment resilience refers to the fact that against the backdrop of an economic downturn, real estate investment has maintained a high level of development, lending strong support to the smooth operation of the macro economy [23,24]. Specifically, housing investment resilience, as one of the manifestations of the regional economy, is an organic whole. After a shock, housing investment goes through a process of "slightly cold", "too cold", "slightly hot", and "too hot". To achieve the transition from decline to rebound, the housing industry needs to consolidate resources, restructure, and improve its ability to adapt to the external environment so as to maintain a stable level of housing investment. This process has both the common characteristics of urban resilience and the individual attributes of housing investment.

With respect to urban resilience, urban economic resilience is expressed in terms of economic diversity, employment levels, and economic stability in the event of risk. Urban social resilience is the reserve status and supply capacity of social resources, which determine a system’s ability to withstand the challenges of shocks. Urban institutional resilience is the ability of government institutions to govern; in particular, it is the ability of government to exercise organization, management, planning, and action following external risk shocks. Urban ecological resilience is the ability of urban ecosystems to recover from shocks, such as environmental pollution, ecosystem overload, and sharp reductions in public green space. Urban infrastructure resilience is a system’s ability to cope with and recover in the face of risky perturbations when population density increases, for example, the ability to secure facilities and lifelines, such as transport, water supply, electricity supply, and healthcare. To this end, the factors affecting urban resilience systems are integrated and combined with the reality of housing investment development. In this paper, we considers economic growth, infrastructure, policy support, and the labor market as important factors regulating the resilience of housing investment.

With respect to housing investment, houses, as commodities, are necessarily influenced by many factors. Revenue, cost, and expectations are important internal influences on housing investment. This is in line with China’s policy objective of "stable land prices, stable house prices, and stable expectations". The level of income is closely linked to the economic development of cities. When the economy experiences a downturn, investments, consumption, and savings among stakeholders, such as the government, enterprises, banks, and residents, are all affected, leading to impacts on investment resilience. Land costs are the primary cost involved in housing development, and the impact of shrinking land supply leads to higher housing development costs, which is not conducive to housing investment resilience. House price expectations provide feedback on the potential of housing market demand. When house price expectations suffer a negative shock, investors’ decisions and market expectations are seriously affected, ultimately affecting the development of housing investment resilience.

In conclusion, based on the synthesis of the above analysis, the three elements of investment and the urban resilience system work together in the development of housing investment resilience to form the conceptual connotation of housing investment resilience investigated in this paper. Housing investment resilience is characterized by complexity, openness, and comprehensiveness, including not only the supply and demand dimensions of urban development, such as the economy, policy, and labor but also sustainability investment dimensions, such as returns, land, and expectations. The three elements of investment are the internal factors affecting housing investment, whereas the urban resilience system comprises the external factors affecting housing investment, which constitute the internal generative logic and external driving mechanism of housing investment resilience development (Figure 2).

2.3. Data Processing

2.3.1. Data Sources and Scoping Study

The outcome variable in this paper is urban housing investment resilience. We use the commodity residential investment resilience index for 35 large and medium-sized cities as a proxy variable [25]. The three elements of investment—returns, costs, and expectations—are proxied by the average growth rate of house prices in the previous three years [26], the growth rate of land acquisition costs [27], and population density [28], respectively. As specified in the research of Hu [27], an increase in land acquisition costs increases house prices, and an increase in house prices promotes real estate investment. Thus, all three elements of investment decisions are positively related to housing investment resilience. Economic growth indicators are represented by the GDP growth rate [25], and infrastructure construction investment is represented by the growth rate of local general public budget expenditures [29]. Labor market indicators are expressed in terms of employment rates [30]. The policy factors are real estate policy and monetary policy [31]. The proxy variable for real estate policy is the weighted average lending rate for individual housing [32]. To capture the differences in real estate policies among the 35 large and medium-sized cities, the CPI of each city (with 2011 as the base period) was used to convert them into real interest rates. Monetary policy is defined as loose monetary policy according to the Monetary policy Implementation Report issued by the People’s Bank of China. If the report explicitly mentions “easing”, then the monetary policy is directly defined as loose. If there is no explicit statement, then it is classified as loose monetary policy if interest rates are reduced and as tight monetary policy if the opposite is true according to the actual operation of the year [33]. The years 2012, 2014, 2015, 2016, 2018, and 2019 were assigned a value of 1 for accommodative monetary policy, and the remaining years were assigned a value of 0. An increase in personal housing loans leads to an increase in house prices, which further leads to an increase in housing investment according to Hu [27]. Thus, there is a positive relationship between economic growth, the growth rate of local general public budget expenditures, employment rates, the weighted average lending rate for personal housing, accommodative monetary policy, and housing investment resilience.

In this study, we selected 35 large and medium-sized cities in China for urban housing investment resilience research. The research data are sourced mainly from the CEIC macroeconomic database and the National Bureau of Statistics, with supplementary data from the China City Statistical Yearbook and the China Real Estate Statistical Yearbook. According to the 2019 City Business Attractiveness Ranking, 35 large and medium-sized cities are divided into first-tier and non-first-tier cities based on the results for the 337 Chinese cities above the prefecture level. There are 17 first-tier cities, including 4 super-first-tier cities, namely Beijing, Shanghai, Guangzhou, and Shenzhen, and 13 new first-tier cities, namely Chengdu, Hangzhou, Chongqing, Wuhan, Xi’an, Tianjin, Nanjing, Changsha, Zhengzhou, Qingdao, Shenyang, Ningbo, and Kunming. The remaining 18 cities, namely Shijiazhuang, Taiyuan, Hohhot, Dalian, Changchun, Harbin, Hefei, Fuzhou, Xiamen, Nanchang, Jinan, Nanning, Haikou, Guiyang, Lanzhou, Xining, Yinchuan, and Urumqi, are all non-first-tier cities.

2.3.2. Measurement and Calibration

QCA methods based on set theory aim to identify sufficient or necessary subset relationships between the configurations of different antecedent variables and outcome variables. The QCA method is divided into a crisp set, a multi-value set, and a fuzzy set. The crisp set is mainly used to analyze binary variables, the multi-value set is mainly used to analyze multivariate discrete variables, and the fuzzy set is used to analyze continuous variables between 0 and 1. Because all variables investigated in this paper are continuous variables, except for monetary policy, which is a binary variable, it is appropriate to use the fsQCA method to study them. Therefore, performing fsQCA analysis, the individual antecedent variables are calibrated and transformed into an ensemble concept; that is, all raw data are converted into fuzzy affiliation scores within the range [0, 1]. In this paper, to calibrate the raw data into a fuzzy set, three anchor points need to be identified, namely fully affiliated, intersection, and fully unaffiliated. The intersection point is the intermediate point that distinguishes between fully affiliated and fully unaffiliated and indicates whether the case belongs to the maximum fuzziness point of a set. In this paper, the three-valued fuzzy set calibration method of Du et al. (2020) is adopted, with the variables using 75%, 50%, and 25% quantile values as the thresholds for completely affiliated, crossover point, and completely unaffiliated, respectively [34]. In this paper, to eliminate the effect of the time factor, the price type indicators are treated as constant prices, with 2011 as the base period; the calibration data of each variable are shown in

Table 1. Calibration of antecedent and outcome variables.

Variables	Indicator	Calibration of First-Tier Cities			Calibration of Non-first-Tier Cities
		Completely Affiliated	Crossover Point	Completely Unaffiliated	Completely Affiliated	Crossover Point	Completely Unaffiliated
Outcome variables	Urban housing investment resilience	0.1605	0.0761	0.0150	0.1841	0.0709	-0.0262
Antecedent variables	Returns	0.1138	0.0658	0.0289	0.1090	0.0576	0.0287
	Costs	0.6596	0.0508	−0.3023	0.4361	−0.0205	−0.3758
	Expectations	1057.7664	838.5956	575.7955	670.8886	472.3437	248.8328
	Economic growth	0.0959	0.0792	0.0629	0.1050	0.0767	0.0544
	Infrastructure construction	0.1376	0.1000	0.0555	0.1548	0.1027	0.0527
	Labor market	97.8	97.2	96.7	97.4	96.9	96.4
	Real estate policies	6.6839	5.6985	5.3955	6.7011	5.6873	5.3739
	Monetary policy	1	/	0	1	/	0

.

3. Spatiotemporal Evolutionary Characteristics of Urban Housing Investment Resilience

3.1. Time Series Characterization

In terms of the time trend, the average value of housing investment resilience in the 35 large and medium-sized cities is above the 0 axis, but the overall average value of resilience is on the low side, showing an M-shaped fluctuation in its evolution (Figure 3). Specifically, the difference in housing investment resilience between first-tier and non-first-tier cities gradually increased after 2013, possibly because in 2013, the “Five New National Policy Measures“ called for the resolute suppression of speculative investment purchases and the strict implementation of purchase restrictions on commercial housing. Housing investment resilience in non-first-tier cities declined, and the difference between the two types of cities widened. China entered a period of supply-side reform and adjustment in 2015. To dissolve the real estate inventory and promote the healthy development of the real estate industry, the remaining cities, except for the four super-first-tier cities, i.e., Beijing, Shanghai, Guangzhou, and Shenzhen, lifted their purchase restrictions one after the other. The investment resilience in both types of cities improved significantly, and the difference decreased. In 2016, the Central Economic Work Conference first proposed that “houses are for living, not for speculation”, and cities that had lifted purchase restrictions gradually resumed them, leading to a downward trend in housing investment resilience in both types of cities. Real estate regulation and control policies continued to escalate in 2017, with a focus on city-specific policies, resulting in improved investment resilience in non-first-tier cities and a moderate difference in investment resilience between the two types of cities. In 2018, the policy of “stable land prices, stable house prices and stable expectations” was put forward, and the real estate regulation and control policy entered a long tightening period. Overall, the difference between the two types of cities did not show a trend of further expansion. Among non-first-tier cities, Lanzhou had the highest median resilience value of 0.995, whereas Xiamen, Nanchang, Jinan, Nanning, and Haikou showed strong investment resilience. The median resilience of Hohhot and Changchun was below 0, with peaks concentrated at low levels and relatively weak investment resilience. Shijiazhuang, Hefei, Fuzhou, Xining, Guiyang, and Yinchuan experienced a significant increase in the peak change of the curve. The curve fluctuated significantly in Taiyuan, Hohhot, Dalian, Changchun, Harbin, Nanning, and Yinchuan. Overall, cities with high housing investment resilience led the way in terms of real estate market development, economic growth, population, and purchasing power. Additionally, most of the cities that showed an upward trend in resilience were either new first-tier cities in the strong urban areas of Beijing–Tianjin–Hebei, the Yangtze River Delta, and the Pearl River Delta or rising core cities or provincial capitals in the central and western regions. They were supported by a number of policies under the national development strategy and were in a rapid growth stage in terms of real estate, economy, and population.

3.2. Spatial Characteristics

3.2.1. Spatial Distribution Characteristics

Figure 4 represents the kernel density distribution of housing investment resilience for 35 large and medium-sized cities from 2011 to 2019. The changes in the width of the curves in the figure represent the degree of agglomeration of regional housing investment resilience at different time series stages. The small white circles in the figure represent the median, indicating the overall development level of regional resilience values. A horizontally increasing kernel density curve indicates a crest phenomenon, representing an increased degree of agglomeration of housing investment resilience in the region. An elongated kernel density curve with a narrow vertical width indicates a trailing phenomenon, which means that the housing investment resilience values are only at that level for a few periods. There is significant variability in the shape and flatness of the curves across cities in Figure 4, and most cities have double-wave peaks, suggesting a bipolar distribution pattern in the evolution of housing investment resilience. The cities with sharp fluctuations in the resilience index are concentrated in the central and western regions and in the northeast. Specifically, among first-tier cities, Shenzhen has the highest median value of resilience, at 0.9374. Hangzhou, Chongqing, Wuhan, Ningbo, and Kunming all show strong investment resilience, with median and curve peaks converging around the medium–high range. Beijing, Chengdu, and Shenyang all have medians and crests below the value of zero, with a lower distribution of the curve, and their housing investment resilience is relatively poor. Shanghai, Guangzhou, Xi’an, Tianjin, Nanjing, Zhengzhou, and Qingdao show upward trends in the peak of the curve, with their resilience increasing. The resilience curves of Changsha, Zhengzhou, and Shenyang exhibit long upward and downward tails with large fluctuations.

3.2.2. Endogeneity of the Network Structure

Based on the above analysis, to explore the trajectory of the evolution of the spatial agglomeration of urban housing investment resilience, further kernel density calculations need to be carried out for housing investment in first-tier and non-first-tier cities. In this way, whether the development of housing investment resilience in the two types of city tends to improve or deteriorate and whether the gap in resilience values between cities tends to widen or contract can be clarified. In this paper, the kernel density evolution pattern of housing investment resilience is measured for the two types of cities from 2011 to 2019 (Figure 5).

On the whole, both types of cities show an evolution from low to high values of resilience development, and both develop from a “single-peak” agglomeration to a multipolar differentiation. With respect to the contour map, the denser the kernel density contour around the central circle, the greater the amplitude of the kernel density value. Specifically, the number of contour circles of resilience values in first-tier cities declines during the observation period, with the main peak moving right from the region of 0.0–0.2 to the region of 0.2–0.4, indicating that the gap in the volatility of housing investment toughness values between first-tier cities is shrinking. Furthermore, the decrease in the number of contour circles causes the main peak height values to fall, indicating that the initial intercity gradient effect is gradually weakened by the divergent development of individual city governments through city-specific policies over the years. Although driving the overall improvement in urban housing investment resilience, as of 2019, the overall resilience values are still low, with significant potential for future upturns. The number of contour circles in non-first-tier cities also decreases during the observation period, although their single-peak slope is significantly flatter than that of the contour circles in first-tier cities, indicating that the phenomenon of differentiated internal development is more prominent. As of 2019, only a few cities jump to within the 0.2 to 0.4 range, showing the characteristics of the Matthew effect of strong growth in high values and low growth in weak values. For example, Shijiazhuang, Hefei, and Fuzhou show strong upward trends and significant overall improvement, whereas the overall improvement in other cities is not significant, indicating that this divergent development does not contribute to a significant increase in housing investment resilience in non-first-tier cities. This may be because, on the one hand, housing investment between cities stems from a multipronged competition between regional human resources, public resources, and the ecological environment. The cities that developed first have a clear first-mover advantage, showing a “long-tail effect” of housing investment, with most of the population and resources converging on a few cities but not yet spreading out in multiple directions. On the other hand, in the context of the “six guarantees” and “six stabilizations”, various cities introduced policies to restrict purchases and sales and curb speculation in response to increasing house prices, with the vigorous regulation of housing market demand ultimately affecting the overall development of the housing investment market. In conclusion, the spatial clustering of housing investment resilience in the two types of cities shows polarized development, with the difference in the overall housing investment resilience of first-tier cities gradually narrowing but the divergence between cities in non-first-tier cities becoming increasingly visible under the influence of the Matthew effect.

4. An Analysis of the Mechanisms That Shape Urban Housing Investment Resilience

4.1. Analysis of the Need for Individual Conditions

Before a fuzzy set truth table analysis can be conducted, a necessity analysis of all antecedent variables, that is, an analysis of whether a particular individual antecedent variable constitutes a necessary condition for high levels of urban housing investment resilience, is required. The criterion is a consistency level greater than 0.9 [21]. As is shown in

Table 2. Analysis of necessary conditions.

	Housing Investment Resilience in First-Tier Cities		Housing Investment Resilience in Non-First-Tier Cities
	Consistency	Coverage	Consistency	Coverage
Returns	0.5538	0.5419	0.6683	0.6864
~Returns	0.5355	0.5315	0.6246	0.6061
Costs	0.6027	0.6110	0.6559	0.7224
~Costs	0.4920	0.4717	0.6618	0.6037
Expectations	0.5227	0.5138	0.5807	0.6707
~Expectations	0.5568	0.5503	0.7127	0.6261
Economic growth	0.6208	0.5977	0.7507	0.7097
~Economic growth	0.4656	0.4699	0.6046	0.6388
Infrastructure construction	0.5355	0.5412	0.7398	0.7348
~Infrastructure construction	0.5575	0.5360	0.6144	0.6160
Labor market	0.6213	0.6107	0.6327	0.6972
~Labor market	0.4831	0.4773	0.6446	0.5877
Real estate policies	0.6478	0.6245	0.7317	0.6449
~Real estate policies	0.4465	0.4501	0.5166	0.5941
Monetary policy	0.5811	0.4295	0.6284	0.4703
~Monetary policy	0.4189	0.6192	0.3716	0.5563

Note: "~" stands for "not" in logical operations. In the case of returns, for example, the absence of "~" indicates a high return, and the presence of a "~" indicates a low return. Qualitative comparative analysis is a form of asymmetric analysis. For example, high returns are the cause of high housing investment resilience, but it does not follow that low returns are the cause of low housing investment resilience. That is, the reasons for the emergence or otherwise of the desired outcome are asymmetric and need to be analyzed separately.

, the consistency level of all the antecedent variables for first-tier and non-first-tier cities is less than 0.9; thus, the necessary condition for high levels of urban housing investment resilience does not exist. This implies that further analysis of the effect of conditional configurations on housing investment resilience in high-tier cities is needed.

4.2. Sufficiency Analysis of Conditional Configurations

The analysis of the sufficiency of a conditional configuration and the analysis of the necessity of individual conditions can be distinguished by the level of consistency, but the difference is that a level of consistency greater than 0.75 is usually accepted as sufficient for a conditional configuration. In previous studies, the threshold of consistency varied from 0.76 [35] to 0.8 [36] to 0.89 [37], depending on the studied sample. Frequency thresholds should also be determined according to the size of the sample. When the sample size is small, 1 or 2 is generally chosen as the threshold value [38]. In this study, the threshold value was selected based on the four principles that truth table results 0 and 1 should be covered and roughly balanced, the frequency threshold should cover at least 75% of the observed samples, the minimum value of PRI consistency should be greater than or equal to 0.75, and that a contradictory configuration should be avoided [38]. The frequency threshold used in this study was ultimately determined to be 2; the original consistency threshold was 0.85, and the PRI consistency threshold was 0.75.

The fsQCA software calculates a total of three solutions based on counterfactual assumptions: a complex solution, parsimonious solution, and intermediate solution. The complex solution contains only the configuration of the actual observed cases; the parsimonious solution contains both the configuration of the actual cases and all the simple and complex logical residues; the intermediate solution contains only the configuration of the actual observed cases and the simple logical residues. The intermediate solution achieves a balance between the complex and parsimonious solutions in terms of complexity; therefore, the analysis of path configurations presented in this paper is the result of the intermediate solution. If the antecedent condition in the path appears in both the intermediate and the parsimonious solution, it is the core condition and is denoted by "●". Those that only appear in the intermediate solution are auxiliary conditions, denoted by "🞄" [36]. The results for high levels of housing investment resilience in first-tier and non-first-tier cities are shown in

Table 3. Configuration path analysis of high-level housing investment resilience in first-tier and non-first-tier cities.

	Path	First-Tier Cities			Non-first-Tier Cities
		Return- and Cost-Driven	Expectation-Driven	Cost-Driven	Return- and Cost-Driven	Return- and Expectation-Driven	Expectation-Driven
		A1	A2	A3	B1	B2	B3
Internal decision-making factors	Returns	●			●	🞄
	Costs	●		●	●
	Expectations		●			●	●
External support factors	Economic growth	●	●		●	●
	Infrastructure construction	●	●		●	●	●
	Labor market	●		●	●	●	●
	Real estate policies	🞄	●	●	●	🞄
	Monetary Policy			🞄			🞄
	Raw coverage	0.0440	0.0332	0.0411	0.1350	0.1070	0.1565
	Unique coverage	0.0389	0.0139	0.0233	0.0516	0.0463	0.0104
	Consistency	0.8198	0.8170	0.9172	0.9963	0.9331	0.9583
	Solution coverage	0.1976			0.3471
	Solution consistency	0.8806			0.9567
	Cases	Shenzhen, Zhengzhou, Nanjing, Hangzhou, Chongqing	Kunming, Beijing, Chengdu	Ningbo, Changsha	Changchun, Yinchuan, Xining, Nanning	Lanzhou, Nanchang	Haikou, Jinan, Fuzhou, Hefei

Note: “●”means the core condition; “🞄”means the auxiliary condition.

. Each of the two types of cities forms three paths, and the consistency between the overall and individual solutions is greater than 0.75, suggesting that each of the three paths can be considered a sufficient combination of conditions for high levels of urban housing investment resilience to drive high levels of housing investment resilience.

4.2.1. Path Analysis of High Levels of Housing Investment Resilience in First-Tier Cities

Path A1: Return- and cost-driven. Path A1 implies that higher returns, higher land costs, faster economic growth, improved infrastructure development, a fully employed labor market, and tightened real estate policies are important conditions to promote resilience in housing investment. Representative cities include Shenzhen, Zhengzhou, Nanjing, Hangzhou, and Chongqing. For example, Hangzhou is the core city of the Yangtze River Delta city cluster and was approved by the State Council as a National Independent Innovation Demonstration Zone in 2015, which accelerated the improvement of its overall infrastructure and the attractiveness of its resource factors. The net population inflow is consistently positive and increasing year over year, with an active labor market. In the first half of 2016 alone, Hangzhou ranked first in the country in terms of land turnover, prompting more capital to flow into the housing investment market. Furthermore, along with the increase in housing market demand, house prices in HCM city jumped to the fifth-highest in the country in 2019, creating a positive boost for housing investment resilience.

Path A2: Expectation-driven. Path A2 implies that higher expectations, faster economic growth, improved infrastructure development, and tightened real estate policies are important conditions that promote resilience in housing investment. Representative cities include Kunming, Beijing, and Chengdu. In the case of Chengdu, for example, demand was expected to grow the most between 2011 and 2020, and the city proposed a strategy to establish a “western capital of numbers”, further increasing its population attraction capacity and ability to become the fourth city in China with a population of “20 million+”. Furthermore, population growth was inseparable from the city’s sustainable economic development, with GDP increasing by 219% in just ten years, considerably contributing to its resilient investment development. In addition, the city’s real estate policy effectively prevented real estate market risks through the implementation of a residential land disposal method that limited housing prices, set quality standards, and competed for land prices, providing a strong policy guarantee for the resilient development of housing investment.

Path A3: Cost-driven. Path A3 implies that higher land costs, a fully employed labor market, tighter real estate policies, and accommodative monetary policies are important conditions to promote resilience in housing investment. Representative cities include Ningbo and Changsha. In Changsha, for example, the cost of land was an important component of urban housing investment, with 13.99 million square meters of land sold in 2020 and land premiums of CNY 59.5 billion. The increase in this indicator is bound to drive up housing investment resilience in the city. After the “318” and “520” control policies were issued, Changsha City introduced a series of control policies, such as “limited house price, competitive land price” and “directionally priced commercial housing”. These policies favorably contributed to the development of a virtuous cycle of housing investment resilience in Changsha.

4.2.2. Path Analysis of High Levels of Housing Investment Resilience in Non-first-Tier Cities

Path B1: Return- and cost-driven. Path B1 implies that higher returns, higher land costs, faster economic growth, well-developed infrastructure, a fully employed labor market, and tightened real estate policies are important conditions to promote resilience in housing investment. Representative cities include Changchun, Yinchuan, Xining, and Nanning. For example, Nanning, as the center of the Beibu Gulf Economic Zone and the hub city of the southwest’s access to the sea, has absolute advantages in terms of economy and infrastructure and has been effective in stabilizing employment. In recent years, the Nanning government played a positive role in stabilizing land prices and promoting housing investment resilience through the implementation of the “housing price, land price restriction and transfer of ownership” policy.

Path B2: Return- and expectation-driven. Path B2 implies that higher returns, higher expectations, faster economic growth, well-developed infrastructure, a fully employed labor market, and tightening real estate policies are important conditions that promote resilience in housing investment. Representative cities include Lanzhou and Nanchang. For example, Nanchang, as the core city of the Poyang Lake city cluster and an important central city in the middle reaches of the Yangtze River, has a much higher GDP than other cities in the province. As the capital city of the province, Nanchang constantly improved its infrastructure construction, and land transactions showed a steady increasing trend, making it an “absorber” of regional housing investment and ultimately promoting the stable development of the city’s housing investment resilience.

Path B3: Expectation-driven. Path B3 implies that higher expectations, well-developed infrastructure, a fully employed labor market, and accommodative monetary policy are important conditions for promoting resilience in housing investment. Representative cities include Haikou, Jinan, Fuzhou, and Hefei. For example, Hefei, one of the "Four Little Dragons" of the property market, directly drove up housing investment through rising land prices in the early stages of development. Furthermore, the abolition and relaxation of restrictions on the settlement of talented people has catapulted Hefei to become a city with a population of more than 10 million, boosting employment rates sufficiently to facilitate the anticipated expansion of demand for housing investment and the healthy development of a resilient system.

4.3. Robustness Tests

Changes in both the level of consistency and the calibration criteria lead to changes in the configuration in the truth table. Therefore, in this study, a robustness test was conducted by adjusting the level of consistency and changing the calibration threshold of the variables.

4.3.1. Adjusting Consistency Levels

By increasing the PRI consistency from 0.75 to 0.8, the original consistency level threshold is increased to 0.95. The frequency threshold remains at 2, and the results of the configuration analysis are shown in

Table 4. Robustness tests for adjusting the level of consistency.

	Path	First-Tier Cities			Non-First-Tier Cities
		Path 1	Path 2	Path 3	Path 1	Path 2	Path 3
Internal decision-making factors	Returns	🞄				●	●
	Costs			🞄			●
	Expectations		🞄		●	●
External support factors	Economic growth	●	●	●		●	●
	Infrastructure construction	●	●	●	●	●	●
	Labor market			🞄	●	●	●
	Real estate policies	●	●	●		●	●
	Monetary Policy				🞄
	Raw coverage	0.1393	0.1009	0.1040	0.0779	0.0349	0.0355
	Unique coverage	0.0431	0.0139	0.0168	0.0685	0.0349	0.0261
	Consistency	0.9697	0.9720	0.9403	0.9145	0.9613	0.9619
	Solution coverage	0.1738			0.1389
	Solution consistency	0.9427			0.9355

Note: “●”means the core condition; “🞄”means the auxiliary condition.

. Both first-tier cities and non-first-tier cities form three paths. The first-tier city path is a subset of the three paths presented in Table 3. The non-first-tier city paths are generally consistent with the three paths presented in Table 3. Therefore, after the level of consistency is adjusted, the findings of this paper do not change substantially, and the results are robust.

4.3.2. Change in Calibration Standard

The calibration thresholds of 75% and 25% were replaced with calibration thresholds of 95% and 5%, respectively, with the original consistency threshold of 0.95 and the frequency threshold of 2 being maintained for the conformational analysis. The results are shown in

Table 5. Robustness tests for varying calibration thresholds.

	Path	First-Tier Cities			Non-First-Tier Cities
		Path 1	Path 2	Path 3	Path 1	Path 2	Path 3
Internal decision-making factors	Returns	🞄				●	●
	Costs			🞄			●
	Expectations		🞄		●	🞄
External support factors	Economic growth	●	●	●		●	●
	Infrastructure construction	●	●	●	🞄	●	●
	Labor market			🞄	●	●	●
	Real estate policies	●	●	●		●	●
	Monetary Policy				🞄
	Raw coverage	0.0755	0.0450	0.0411	0.1565	0.08079	0.1349
	Unique coverage	0.0588	0.0450	0.0244	0.0753	0.0808	0.0538
	Consistency	0.9499	0.9339	0.9172	0.9583	0.9732	0.9963
	Solution coverage	0.1448			0.2911
	Solution consistency	0.9301			0.9695

Note: “●”means the core condition; “🞄”means the auxiliary condition.

, with three paths forming in both first-tier and non-first-tier cities. Thus, the test results remain robust.

4.4. Analysis of Potential Substitution Relationships between the Two Types of Urban Configuration Path Conditions

A comparison of the differences in high-level housing investment resilience profiles between first-tier and non-first-tier cities reveals that first-tier and non-first-tier cities share the same path, in addition to their own dedicated paths (

Table 6. Comparative analysis of grouping paths in first-tier and non-first-tier cities.

	Path	High-Level Development Configuration Path for Housing Investment Resilience in Two Types of Cities
		Same Path	Own Dedicated Paths
			First-Tier Cities		Non-First-Tier Cities
		Return- and Cost-Driven	Expectation-Driven	Cost-Driven	Return- and Expectation-Driven	Expectation-Driven
		A1/B1	A2	A3	B2	B3
Internal decision-making factors	Returns	●			🞄
	Costs	●		●
	Expectations		●		●	●
External support factors	Economic growth	●	●		●
	Infrastructure construction	●	●		●	●
	Labor market	●		●	●	●
	Real estate policies	🞄(First-tier)/●(Non-first-tier)	●	●	🞄
	Monetary policy			🞄		🞄

Note: “●”means the core condition; “🞄”means the auxiliary condition.

). The same conformational path for first-tier and non-first-tier cities is the “return- and cost-driven” path (A1, B1), indicating that high levels of housing investment resilience in both types of cities are driven by a combination of the internal factors of investment, expectations, and costs and five external factors. There are two dedicated paths for first-tier cities, namely the "expectation-driven” and “cost-driven” paths, both of which have the core variable of real estate policy, indicating that it is an extremely important conditioning variable for housing investment resilience in first-tier cities. There are also two dedicated paths in non-first-tier cities, “return- and cost-driven” and “expectation-driven”, both with two variables of infrastructure development and labor market, indicating that these two variables are extremely important condition variables for the resilience of residential investment in non-first-tier cities. In general, although the “expectation-driven” path is present in both types of cities, the external support conditions differ.

Comparative analysis of the differences between the two types of city configuration paths shows that there is a potential substitution relationship between the variables (Figure 6). The left side of Figure 6 (A1, A2, and A3) represents first-tier city configuration paths. B1, B2, and B3 (on the right side) represent non-first-tier city configuration paths. The two-way arrows in the figure represent potential substitution relationships between multiple variables. Taking the contrasting relationship between paths A1 and A2 as an example (Figure 6a), paths A1 and A2 have the same three antecedent variables of economic growth, infrastructure development, and real estate policy, so the three antecedent variables of residual returns, costs, and labor market in path A1 form a potential substitution relationship with the expectation variable in path A2, indicating that expectations are an important variable in terms of enhancing the resilience of urban housing investment.

A comparison of the three paths in the two types of cities reveals that real estate policy is an important condition for resilience in urban residential investment in first-tier cities, whereas it is not important for non-first-tier cities, likely because compared to non-first-tier cities, first-tier cities have a superior economic environment, well-developed infrastructure, and higher population density. However, real estate regulation policies are commonly associated with restrictions on purchases and loans, as well as government restrictions on the approval of land for real estate development. At this point, housing no longer follows the law of supply and demand of ordinary commodities but is considered a financial asset, causing prices to rise rather than fall in the short term. Investment returns in the real estate market increase, thus creating a situation in which the stricter the regulation of the real estate market, the more resilient housing investment becomes. Infrastructure construction and the labor market are important conditions for housing investment resilience in non-first-tier cities but not in first-tier cities, possibly because compared to first-tier cities, non-first-tier cities require better infrastructure construction, more industrial support, and the release of sufficient jobs to attract the inflow of capital and population, which further enhance the competitiveness of the cities while the real estate market enters a new phase of destocking and increasing housing demand.

5. Conclusions and Discussions

Based on the theory of the three elements of investment decisions and urban resilience theory, a research framework of urban housing investment resilience and its influencing factors was constructed. With 35 large and medium-sized cities in China as the research objects, in this study, we investigated the spatiotemporal evolution characteristics, multiple influencing factors, and driving paths of housing investment resilience in first-tier and non-first-tier cities using spatiotemporal analysis and qualitative comparison methods. The findings are as follows.

First, the overall resilience value of urban housing investment during the observation period was low and can be divided into three development stages, showing an M-shaped fluctuating evolution with stage and cyclical developmental characteristics and a clear gap between non-first-tier and first-tier cities. Spatially, there are distinct polarization differences and imbalances. The gradient effect between first-tier cities is gradually weakening, whereas non-first-tier cities are characterized by the Matthew effect, with the strongest cities becoming stronger.

Second, the three internal decision conditions of returns, costs, and expectations cannot, by themselves, constitute the necessary conditions for high levels of resilience in urban housing investment decisions. There are three paths to high levels of housing investment resilience in first-tier cities, namely "return- and cost- driven", "cost-driven", and "expectation-driven". There are also three paths to high levels of housing investment resilience in non-first-tier cities, namely "return- and cost-driven", "return- and expectation-driven", and "expectation-driven".

Third, the resilience paths to high levels of housing investment in first-tier cities and non- first-tier cities are the same, in addition to different dedicated paths for each city type. The same path is "return- and cost-driven". Whereas both types of cities are "expectation-driven”, the external support conditions differ considerably.

Finally, there are potential substitution relationships between all three paths in first-tier cities and non-first-tier cities. Overall, real estate policy is an important condition for housing investment resilience in first-tier cities, and infrastructure development and labor markets are important conditions for housing investment resilience in non-first-tier cities.

Most studies have focused on regional housing investment growth, expansion trends, spatiotemporal characteristics, influencing factors, and their impact, but there has been a lack of in-depth integration with the cyclical patterns of housing investment ups and downs. Few studies have investigated housing investment resilience, and few have examined the impact mechanisms for different types of cities, making it difficult to illuminate the “black box” of urban housing investment. On the basis of previous research, in the present study, we combined the three elements of investment decisions with the theory of urban resilience systems, expanding the seven conditions affecting urban housing investment and constructing a theoretical research framework for urban housing investment resilience. On the other hand, this study extends the comparative study of groupings in explaining complex causal relationships through the “configuration perspective”, analyzing the reasons for differences in the driving paths of housing investment resilience between first-tier and non-first-tier cities to compensate for the shortcomings in the explanation of the influencing factors of housing investment under the weighting perspective. Analysis of the linkage effects of multiple factors reveals the complex causal relationships behind urban housing investment.

The findings reported in this paper have the following policy implications. First, the evolution of urban housing investment resilience in the face of external shocks is cyclical in nature, going through a process of "slightly cold", "too cold", "slightly hot", and "too hot". Following the evolution of urban housing investment resilience requires advanced planning in the face of the unknown risks of external shocks. As demonstrated, good urban housing investment resilience is not just about being resilient in the face of shocks. It should be reflected in the continuous optimization of the internal structure and order of the housing investment resilience system before external shocks and the continuous improvement of the stability and resilience of the resilience system. Ultimately, this effectively promotes the sustainable and healthy development of urban housing investment markets.

Second, costs and returns are important factors influencing the development of urban housing investment resilience. Combined with the cyclical evolution pattern of housing investment resilience, costs and returns become the direction of early intervention and focused regulation within the investment resilience system. The relationship between the costs and returns of urban housing investment needs to be integrated and balanced with a focus on the linkage effect between the two. That is, there is not always a positive correlation between costs and returns, and a competing relationship between the two is always present. It is important to avoid an imbalance in the development of the costs of housing openness and the return on investment. To this end, a focus on adjusting the cost–benefit structure of the housing investment market is necessary to effectively enhance the resilience of urban housing investment.

Third, there are significant differences in the evolution of housing investment resilience in different types of cities. There are also differences in terms of the focus of attention on strengthening the resilience of urban housing investment in the future. Specifically, for high-grade cities with better development infrastructure, real estate policy is the core condition affecting the development of their resilience systems. This category of cities needs to focus on housing policy regulation to promote the stable and healthy development of land prices, house prices, and expectations through localized and multipronged approaches. However, for low-grade cities with poor development infrastructure, infrastructure development and labor markets are the core conditions for the development of resilient systems. Therefore, infrastructure development and an adequate labor job market are also important conditions for attracting capital investment.