Analysis of the Impact of Industrial Land Price Distortion on Overcapacity in the Textile Industry and Its Sustainability in China

Abstract

At present, China is faced with a situation of unbalanced regional economic development. The report of the 19th session of the national congress of the Communist Party of China pointed out that we should accelerate the implementation of the strategy of coordinated regional development. However, this study finds that, in the process of “leaving the cage and changing birds”, the number of enterprises in the textile industry in the eastern region has risen again, which, in turn, has led to overcapacity and has seriously affected the sustainable development of the textile industry. The present study finds that the increasingly distorted price of industrial land is the root cause of the “surge phenomenon” of enterprises in the textile industry and overcapacity. Therefore, China’s central government should be fully aware of the negative impact of the distorted price of industrial land on the sustainable development of textile industry. Efforts must be made to solve the problems of the excessive influx of textile enterprises, overcapacity, and unsustainability in the eastern region in three aspects. Specifically, the land transfer procedure should be standardized, the motivation of local governments’ land attracting investment should be curbed, and the introduction of high-quality enterprises should be paid attention to.

1. Introduction

As China’s economic restructuring and industrial transformation and upgrading processes accelerate, the manufacturing industry shows a trend of shifting from the eastern coast to the central and western regions. This trend first appeared in traditional labor-intensive industries, among which the textile industry is a typical example. The textile industry has indeed achieved rapid development over the years. However, the structural contradictions and problems accumulated over a long period of time have become increasingly prominent. The eastern coastal region, which is home to 85% of China’s textile industry, is facing serious factor constraints. Meanwhile, the comparative advantages of the central and western regions have not been fully exploited. Promoting the transfer of the textile industry would be conducive to easing factor constraints in eastern China. The comparative advantages of factor cost and labor force in central and western China would also come into play. Furthermore, the national textile industry would develop sustainably. However, the textile industry is characterized by a low equipment coefficient, low long-term investment, a short investment payback period, and fast capital turnover. These characteristics can help achieve rapid economic growth in the region in a short period of time, and this is undoubtedly attractive to competitive local governments [1]. Therefore, even if the textile industry transfer could form a pattern of resource integration, the local governments in the eastern region also have a competitive motivation to attract textile enterprises. This results in the influx of a large number of new textile enterprises into the eastern region (Figure 1) and overcapacity (Figure 2). Based on previous studies, the degree of overcapacity is measured in this study by capacity utilization. The lower the capacity utilization rate is, the more serious the overcapacity is. Figure 1a,b shows that, based on the number of enterprises entering the region, the number of newly established textile enterprises in eastern China showed a clear upward trend from 2001 to 2004. This stage was in line with the characteristics of the beginning of China’s textile industry. From the trend in Figure 1b,c, we can see that from 2004 to 2011, the textile industry began to gradually shift to the central and western regions. However, according to Figure 1d, it can be seen that in 2014, the number of newly established textile companies in the eastern region showed a rise again. Accordingly, from 2008 to 2011, the capacity utilization rate of the textile industry in eastern China gradually increased. However, the capacity utilization rate in 2014 was significantly lower than that in 2011, and overcapacity appeared again (the number of textile enterprises entering and the capacity utilization rate of the textile industry are all from the Database of Chinese Industrial Enterprises). The direct cause of excess capacity is usually the “tidal wave phenomenon” and enterprises’ over-investment [2,3]. The influx of textile enterprises and excessive investment is what leads to repeated construction projects and excessive competition, which in turn leads to overcapacity in the industry. What, then, is the root of the influx of textile enterprises in the east and overcapacity? Why is it that, when the central government guides the industrial transfer, there is still overcapacity? Answering these questions will help explain why labor-intensive enterprises choose to enter the eastern region, even in the face of high cost disadvantages. This understanding will help put forward relevant policy suggestions for further promoting the coordinated development of the eastern region and the central and western regions and the sustainable development of the textile industry.

Existing research on enterprise site selection in the eastern and western regions is mostly conducted from the perspectives of trade cost and factor cost [4,5]. However, at present, the trade costs between the east and the west have been greatly reduced and the eastern regions are gradually losing the comparative advantage of labor cost [6,7]. Therefore, the existing theories cannot explain the re-influx and overcapacity of textile enterprises in the eastern region at this stage. What, then, is the reason for the number of rebounds and overcapacity in the textile enterprises in the eastern region? According to Figure 3, in 2007, the Ministry of Land and Resources of the People’s Republic of China began to implement the national minimum price standard for the transfer of industrial land. The distortion of industrial land prices in eastern China was curbed by this new policy, but then began to rebound after 2011 (the data on price distortion of industrial land were collected from China land market network (www.landchina.com, accessed on 14 August 2021)). This time coincides with the rebound of the textile industry’s import volume and overcapacity; so, is the rebound of industrial land price distortion the root cause of the textile industry’s issues? This problem needs to be tested both theoretically and empirically. We first explain how the distortion of industrial land prices leads to the excessive influx of textile enterprises by combining the geographical advantages of the eastern region and the characteristics of textile enterprises. The distortion of industrial land price is found to expand the location advantage of the eastern region, while weakening the relative cost advantage of the central and western regions. In addition, land price distortion increases the net land income of textile enterprises entering the eastern region; this is confirmed by the mechanism test. Secondly, this study makes a theoretical analysis of the relationship between the distortion of industrial land prices, excessive influx, and overcapacity in the textile industry. Overcapacity is found to be caused by the distortion of land prices and the resulting influx of textile firms. This finding is verified by a series of empirical analyses.

The contributions of this paper are as follows: (1) This is the first time the phenomenon and reasons for traditional enterprises’ re-influx and the revival of overcapacity in eastern China have been revealed. This information provides a new theoretical perspective for the study of industrial transfer policy and the low efficiency of production capacity utilization. (2) This study has enriched the research on the price of industrial land. We find that the national minimum price standard for assigning industrial land, which was first implemented in 2007, does reduce the distortion degree of industrial land price. However, this phenomenon still exists, and even shows a rebound trend. This finding shows that policy measures alone cannot fully achieve the expected effect of industrial land price control. Rather, it is also necessary to strengthen supervision and promulgate corresponding matching measures. As such, the research conclusion has certain reference significance for deepening the price control of industrial land.

2. Literature Review and Research Hypotheses

2.1. Industrial Land Price Distortion and Enterprise Entry

Under China’s unique land acquisition and transfer system, local governments play a decisive role in the mode, structure, and quantity of land resource allocation [8,9]. Therefore, land resources become an important source of fiscal revenue for local governments. Under the dual motivation of fiscal revenue maximization and political promotion, local governments’ motivation to attract land investment is increasingly strong. As this “bottom-to-bottom” competition becomes increasingly fierce, the sale of industrial land gradually evolves from the appropriately low price typically found at the beginning, to the “loss” sale price [10,11]. This directly leads to the distortion of China’s industrial land prices and has a far-reaching impact on the sustainable land use [12,13]. Low land prices reduce enterprises’ cost of entry to a large extent; low prices therefore become an effective behavior engaged in by local governments to attract enterprises [14,15]. Thus, distortions in the price of industrial land can attract companies. However, the distortion of industrial land prices is not limited to the reduction of plant construction cost, but also produces a substantial subsidy effect on enterprises. Thus, distortions in the price of industrial land can attract companies [16]. However, low-price land investment alone is not enough to explain the obvious increase in the number of textile enterprises in the eastern region. The central and western regions also compete with each other in land price for the purpose of development, and these regions even have a greater advantage in land cost, compared to the eastern region. To fully explain this phenomenon, we need to combine the geographical advantages of the eastern region.

Low industrial land prices are undoubtedly attractive to enterprises [17]. However, even given the low land price, enterprises will give priority to and choose areas with better economic development foundations as production centers [18]. If the degree of land marketization is high, the price of industrial land in cities with good economic foundations will increase in line with the increased demand. Enterprises will then move to the surrounding areas (for investment reasons) after considering the net income of land. However, local governments can sell industrial land at a lower price than can be obtained on the open market. As such, the development and location advantages of cities in the eastern region will be amplified, and the net land income will be raised. Meanwhile, the relative cost advantages of cities in the central and western regions will be weakened. If their decisions are based on the comprehensive consideration of land cost and urban economic development, enterprises will almost certainly move back to the eastern region. To be specific, the eastern coastal region has convenient transportation and other superior infrastructure, as well as a high degree of marketization and opening to the outside world. This business climate is more conducive for local governments seeking to both transfer commercial land in a market-oriented way and to horizontally subsidize the gap caused by the low-price transfer of industrial land [19]. At the same time, low-price industrial land can attract enterprises to invest in and set up factories, thus driving the development of the related local services. The development of the service industry will increase the demand for industrial land and bring about a new round of land transfer, forming what is termed a “virtuous” cycle. By contrast, restricted by geographical conditions and economic development, the central and western regions do not have advantages in land investment; it is also difficult for these areas to make up for the horizontal compensation between industrial land and commercial land. As such, it is impossible for these areas to attract enterprises by offering low-price industrial land [20]. At the same time, as a competitive market industry, the textile industry is more inclined to directly approach the market [21]. Only under the normal action of a market competition mechanism can enterprises improve their benefits and simultaneously improve the quality of textiles. Therefore, when the entry cost of the eastern region (with a higher degree of marketization) decreases, the net land income of textile enterprises will increase to an even greater degree. At that stage, some rational textile enterprises and manufacturers will choose to enter the eastern region. Therefore, this study argues that the distortion in industrial land prices in the eastern region magnifies the development and location advantages of that region and weakens the relative cost advantages of the central and western regions. These factors lead to a rebound in the number of textile enterprises entering the eastern region. Based on the above analysis, we propose:

Hypothesis 1 (H1).

The distortion of industrial land prices leads to a rebound in the number of textile enterprises entering the eastern region.

2.2. Industrial Land Price Distortion, Enterprise Entry, and Overcapacity

Chinese and Western scholars’ studies on overcapacity emerge one after another. Western scholars mostly discuss the causes of overcapacity from the perspective of enterprise strategic behavior. First, economic fluctuation is an important factor that affects whether companies choose to store production factors. When the economy fluctuates greatly, rational enterprises will store certain idle production factors in order to deal with external economic uncertainty, resulting in overcapacity [22,23]. Secondly, when there is enough production capacity in an industry, potential competitors will hesitate to enter the industry. Therefore, strategic overcapacity can effectively inhibit the entry of potential competitors [24,25,26]. These studies may be more applicable to Western countries with a higher degree of marketization. For China, which has a high degree of government intervention, these may not be the dominant factors affecting overcapacity. In view of the Chinese-style overcapacity problem, most existing explanations are centered around market failure theory and government intervention theory. The market failure theory is represented by the “tidal wave phenomenon” [27]. As China is an investment-driven growth country and in a period of industrial transformation, it has certain “late-comer advantages” [28]. This means that Chinese enterprises can easily reach a consensus that current industrial development prospects are good, and for that reason, they will increase investment in these industries. This forms the investment surge phenomenon, which in turn results in overcapacity. That is, overcapacity is a collective irrational phenomenon caused by individual rationality. The government intervention theory holds that, local governments will use all policies and economic means available to them to intervene in micro-enterprises, thus promoting economic growth. However, improper intervention will distort the market price of production factors, resulting in blind investment and an increase in the number of related enterprises. This behavior will cause serious overcapacity [29,30,31]. Why is there a “tidal wave phenomenon” in China? What kind of improper measures by the government would cause overcapacity? Existing studies do not give consistent conclusions, this paper will uncover the reasons behind it from the perspective of industrial land price distortion. In addition, the impact of the international market will also lead to overcapacity. Colleges and universities related to textile majors in southeast Asian countries are stepping up cooperation and accelerating the pace of production, study, and research. The textile industry of these countries is also relatively strong. These factors have gradually shifted a large number of production orders from China to these emerging countries, and the international demand for China’s textiles has decreased. This is also one of the reasons for overcapacity in China’s textile industry [32].

The land investment behavior adopted by the government, if based on the target of developing the local economy, will lead to over-investment and overcapacity. As the only possible land transferor, local governments use low land prices to attract enterprise investment [33]. This strategy not only helps local governments to realize indirect income (as emphasized by the land finance hypothesis), but also helps to promote regional economic growth and achieve the goal of political promotion [19,34]. This approach naturally becomes the rational choice of local officials [35,36]. At the beginning of the current project, when the local government sold industrial land to enterprises at low prices, the strategy actually had a direct subsidy effect on enterprises [37]. As land is an excellent form of loan collateral, in the process of project construction, enterprises can mortgage to the bank at a market price higher than the land purchase cost to obtain low interest loans [38]. This is actually how the government provides low-cost land financing subsidies to enterprises. After the completion of the project, the enterprise can also transfer the rights to use the land at a market price higher than the cost of the land purchase. The huge intermediate price difference accrues huge profits for enterprises, forming substantial subsidies for enterprise investment [39]. As the distorted price of industrial land has brought a series of huge subsidies to enterprises, the enterprises can realistically expect to make profits, even if the industrial project loses money. Therefore, they continue to invest in the project, which leads to over-investment and overcapacity. At the same time, because the textile industry is at the lower end of the industrial structure, banks will expect lower profit margins and poorer growth in the future. When banks limit the allocation of credit funds, they are likely to reduce the amount of credit rationed to the textile industry. Therefore, the financing effect of textile enterprises caused by low-priced land is relatively stronger. Textile enterprises are also more likely to be attracted by low-priced land, and therefore, the phenomenon of overcapacity is more likely to occur. Based on the above analysis, we propose:

Hypothesis 2 (H2).

Distorted industrial land prices cause the revival of the excess capacity of the textile industry in eastern China.

As mentioned above, local governments use land as subsidies in the form of lower land prices in order to attract enterprises to invest and build factories in the local governments’ areas. This behavior, to a certain extent, will cause redundant construction and overcapacity. When enterprises receiving a “land subsidy” choose to enter the industry, other enterprises that might have hesitated to enter the industry often display an obvious “herd effect”. The capacity expansion of leading enterprises or incumbents will compel other enterprises to think that the supply of this industry is insufficient. Then, the latter will imitate or follow the behavior of “leading enterprises”, which in turn leads to an influx of many enterprises blindly entering the area [29]. The influx of many textile enterprises can increase the total output value of the industry, expand the number of textile products, and create a “prosperous scene” in which there appears to still be room for further development of the industry. Again, this leads to the influx of more enterprises. However, this behavior will eventually lead to the homogenization of industry investment and the formation of repetitive construction, which in turn will lead to serious overcapacity. Based on the above analysis, we propose:

Hypothesis 3 (H3).

The distorted price of industrial land leads to an influx of textile enterprises in eastern China, and the “herd effect” of the influx of enterprises leads to overcapacity of the textile industry.

3. Model Design and Data Description

3.1. Model Design

To verify hypothesis H1 proposed above, we use the data of prefecture level cities from 2007 to 2014 for panel regression and adopt the fixed effect model. The following model is constructed:

$$c{r}_{i,t}=\alpha +\beta distor{t}_{i,t}+\gamma X_{i,t}+{\lambda }_t+provinc{e}_i+{\epsilon }_{i,t}$$

(1)

where t is the year; i is the region (the whole region of cities at prefecture-level and above, excluding autonomous prefectures); λ_t is the year fixed effect; province i is the fixed effect of province; ε_i,t is the error term; and cr_i,t represents the capacity utilization rate of the textile industry in region i in t year. This represents the phenomenon of the textile industry’s overcapacity and is the explained variable of this paper. Next, distort_i,t represents the distortion degree of industrial land prices in region i in t, which is the core explanatory variable of this paper. Then, X_i,t represents regional control variables and industry control variables, α is a constant term, γ is the coefficient of the control variable, and β is the core coefficient with which this paper is concerned and which represents the direction and degree of the impact of industrial land price distortion on overcapacity in the textile industry. The method used to measure the main variables is explained below.

3.1.1. Price Distortion of Industrial Land (Distort)

The ratio of the difference between the minimum price standard of industrial land transfer and the actual price of industrial land transfer to the minimum price standard is adopted to measure the degree of the distortion of industrial land prices [17]. The larger this value is, the greater is the degree of distortion of the industrial land price. This is represented by:

$$distor{t}_{i,t}=(land{m}_{i,t}-lan{d}_{i,t})/land{m}_{i,t}$$

(2)

where landm_i,t stands for the lowest price standard for the transfer of industrial land, and land_i,t represents the actual price of the sold industrial land. The minimum price for the transfer of industrial land comes from the National Minimum Price Standard for Assigning Industrial Land, which was first implemented on 1 January 2007. The regulation defines the minimum price standard for the transfer of industrial land in each county and urban area in China, and also divides China’s industrial land into 15 different grades. The actual transfer price data of industrial land come from the transfer information of industrial land, which can be found on the China land market network (www.landchina.com, accessed on 27 February 2022). The transfer prices of each industrial land site in the same prefecture-level city are added up to obtain the actual transfer price of all the industrial land in the whole city. Divide each industrial land transaction in the China land market network according to grades, and then multiply the lowest unit price of the corresponding grades by area, in order to obtain the lowest transfer price of each piece of land. Then, the lowest selling prices of all the land in the same prefectural city are added up to get the lowest selling price of all the industrial land in the whole city. Finally, the degree of distortion of industrial land prices is calculated according to Equation (2).

3.1.2. Capacity Utilization (cr)

In Equation (1), cr_i,t represents the capacity utilization rate of the textile industry of prefecture-level city i in t. In this paper, the degree of overcapacity is measured by the capacity utilization rate. The lower the capacity utilization rate is, the more serious the degree of overcapacity will be.

(a) Measurement method 1: The fixed asset turnover (the ratio of operating revenue to fixed assets) is taken as a proxy variable of the capacity utilization rate, denoted as cr_fx.

(b) Measurement method 2: The production function method is used to measure the capacity utilization index [40,41]. The production function is used to measure the capacity level of the maximum output that can be realized when production factors (such as capital and labor) are fully utilized. Then, the ratio of actual output to theoretical output is used as the capacity utilization rate. This paper adopts the Cobb–Douglas production function to establish a production function model, which is in the following form:

$$Y_{i,t}=f(K_{i,t},L_{i,t})=A_i{K}_{i,t}^{\alpha }{L}_{i,t}^{\beta }{e}^{-\mu }$$

(3)

where Y_i,t is the actual output value, expressed by the annual operating income of the urban textile industry; K_i,t is the fixed capital stock, represented by the annual fixed assets of the urban textile industry. Next, L_i,t is the labor input, expressed by the number of employees in the urban textile industry at the end of the year, and A_i is the technical level, usually used as a fixed constant. The textile industry data of each city are summed from the textile enterprise data in the Database of Chinese Industrial Enterprises. The parameters α and β represent the output elasticity of fixed capital stock and labor input, respectively; the return to scale is assumed to be constant. The boundary production function deduced by estimation is:

$${\widehat{Y}}_{i,t}=e^{\widehat{\delta }}{K}_{i,t}^{\widehat{\alpha }}{L}_{i,t}^{\widehat{\beta }}$$

(4)

where ${\widehat{Y}}_{i,t}$, is the theoretical output value, the capacity utilization can be expressed as follows:

$$cr\_pf=Y_{i,t}/{\widehat{Y}}_{i,t}$$

(5)

3.1.3. Correlated Control Variables

The selection of control variables in this paper mainly refers to existing research literature that examines the factors that affect overcapacity. The omission variable bias in regression is taken as the selection criterion. Control variables are as follows: (1) Fiscal expenditure logarithm (lnfinexp) is used to control the influence of government subsidies on enterprise investment. (2) The annual population logarithm (lnpop) is used to control the regional market demand disturbance caused by population agglomeration. (3) The balance logarithm of various loans of financial institutions at the end of the year (lnloan) is used to control the impact of loan availability on enterprise investment. (4) The logarithm of GDP (lngdp) is used to control the influence of the urban economic development level on enterprise investment. (5) The year-end balance logarithm of urban and rural household savings (lnsav) is used to control the impact of household savings on enterprise capacity. (6) The logarithm of actual utilized foreign capital of the year (lnfori) is used to control the impact of accelerated foreign direct investment influx on enterprise investment, after the reduction made due to foreign trade policy uncertainty. (7) Industry growth opportunity (growth) is used to control the impact of the textile industry’s own development on overcapacity. To comprehensively reflect the growth opportunity of the textile industry, the growth rate of the textile industry is taken as a balance index.

3.2. Data Source and Description

The data sets used for regression in this paper are as follows: (1) China land market network (www.landchina.com, accessed on 14 August 2021). The land transfers of each prefecture level city from 2007 to 2014 were obtained from this website. (2) Database of Chinese Industrial Enterprises (2007–2014): In this paper, the annual data of enterprises are screened out according to the two-digit industry category codes of textile enterprises. The total sample size at enterprise level is 122,899, with an average of 17,557 per year. For this paper, data from the textile industry in prefecture-level cities are required, so the relevant variables of all textile enterprises in the same prefecture level city are aggregated. Finally, 588 samples of prefecture-level cities in the eastern region were obtained. In 2011, the standard for large enterprises was changed from 5 million yuan or more per year in main business revenue to 20 million yuan or more. In order to unify the statistical caliber, any enterprise whose main business income was less than 20 million yuan per year before 2011 was excluded. In addition, due to the poor quality of the database in 2010, complete industry variables could not be obtained, so 2010 is excluded from the sample period used in this paper. In addition, it should be noted that the database only counts until 2014 at present, so the research can only stop in 2014. This is one of the limitations of this paper. (3) Other control variables of prefecture-level cities were derived from the China Stock Market and Accounting Research Database, as well as the statistical yearbooks of prefecture-level cities. In addition, to eliminate the interference of price factors on the regression results, before the empirical analysis, the consumer price index (based on 2007 prices) was used to conduct price adjustments for all nominal variables to obtain the real value.

Table 1. Descriptive statistics.

Variable		Observations	Mean	Standard Deviation	Minimum	Maximum
Explained variables	cr_fx	588	5.061	2.159	0.251	27.022
	cr_pf	588	8.262	2.695	0.769	21.646
Explanatory variable	distort	588	0.334	0.473	−1.836	0.950
Intervening variables	entr1	588	1.438	1.872	0	11.600
	entr2	588	1.326	3.211	0	56.953
Control variables	lnfinexp	588	14.588	0.899	12.797	17.463
	lngdp	588	7.602	0.886	5.902	9.885
	lnpop	588	6.110	0.569	4.577	7.238
	lnloan	588	16.567	1.186	14.120	19.715
	lnsav	588	16.334	0.949	14.459	19.092
	growth	588	0.165	0.384	−0.725	1.985
	lnfori	588	15.533	1.582	12.154	19.45
Instrumental variables	ln[slope∗(1/i)]	588	−1.870	1.532	−5.182	0.242
	ln(area∗year)	588	16.619	0.696	14.877	18.193

shows the descriptive statistics of the main variables.

4. Empirical Analysis

4.1. Baseline Regression Result

This paper first conducted an empirical test on the econometric model (1); the obtained benchmark regression results are reported in

Table 2. Baseline regression.

	(1) cr_fx	(2) cr_pf
distort	−0.4018 *** (0.1011)	−0.2496 ** (0.0963)
lnfinexp	−0.2458 (0.1957)	−0.3022 (0.1363)
lngdp	1.2556 *** (0.2205)	1.0188 *** (0.1803)
lnpop	0.1641 (0.1254)	0.0405 (0.1178)
lnloan	−0.6260 *** (0.1626)	−0.2924 (0.1298)
lnsav	−0.5300 ** (0.2464)	−0.4545 * (0.1860)
growth	0.7160 *** (0.3570)	1.0598 *** (0.6295)
lnfori	0.0699 (0.0531)	0.0145 (0.0427)
Constant	10.9969 *** (2.1178)	8.3645 *** (1.4662)
Year and province	Yes	Yes
N	588	588
Adj-R2	0.2307	0.2650

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

(heteroscedasticity robust standard error in brackets). It should be noted that all regression results in this paper were implemented using stata15 software. According to Hypothesis H2, the distortion of industrial land price causes overcapacity of the textile industry in eastern China. The higher the distortion is, the lower is the utilization rate of production capacity and the more serious is overcapacity. Therefore, the expected distort coefficient should be negative. As shown in Table 2, the explained variables in Column (1) take the fixed asset turnover rate as the proxy variable of the capacity utilization rate to represent the phenomenon of overcapacity in the textile industry (cr_fx). The interpreted quantity in Column (2) is the capacity utilization rate measured by the production function method (cr_pf). In general, no matter which measurement method is adopted, the regression coefficient of the core explanatory variable (distort) is negative. Therefore, the more distorted the price of industrial land in eastern China is, the lower the capacity utilization rate of the textile industry will be, and the more prominent overcapacity will be. This finding verifies Hypothesis H2. Moreover, the difference in the regression results between the two different measurement methods is also within the acceptable range, with little difference in coefficients. This finding further indicates that the benchmark regression results of this paper have internal validity. Therefore, the fundamental means to solve the overcapacity in traditional industries are to optimize the price system of industrial land and correct the distortion of industrial land prices.

4.2. Robustness Test

4.2.1. Change the Sample Interval

The sample scope taken from the China Industrial Enterprise Database is all state-owned industrial enterprises and non-state-owned industrial enterprises above the designated size. Moreover, before 2011, being “above scale” requires that the annual main business income (sales) of an enterprise was 5 million yuan per year or more. From 2011 onwards, the standard changed to 20 million yuan per year or more. In the benchmark regression results, in order to unify the standard, we excluded the enterprises whose main business income was less than 20 million yuan per year before 2011. However, this approach may lead to biased regression results, due to the elimination of a large number of samples. Therefore, in the robustness test, the sample interval was reset to 2011–2014; the obtained regression results are shown in

Table 3. Change the sample interval.

	(1) cr_fx	(2) cr_pf
distort	−0.5866 *** (0.1454)	−0.4295 *** (0.1513)
Controls	Yes	Yes
Constant	12.3918 *** (3.1947)	11.6406 *** (2.4098)
Year and province	Yes	Yes
N	336	336
Adj-R2	0.2321	0.3345

Note: *** p < 0.01.

. The coefficient of distort remains significantly negative, with little variation. One can see that the regression results in this paper are less affected by sample elimination, and the results are relatively robust.

4.2.2. Substitution Explanatory Variable

The core explanatory variable used in the benchmark regression in this paper is the degree of distortion of all industrial land prices. This approach can fully reflect the impact of the price distortion of industrial land in the whole prefecture-level city on overcapacity in the textile industry. However, to verify the accuracy and robustness of the regression results, industrial land price distortion should be even further focused on the textile industry. That is, in this paper, only the degree of distortion of the industrial land price of the textile industry is used as the core explanatory variable, in order to eliminate the measurement error of the explanatory variable. Therefore, in the robustness test, the land transfer information of the textile industry was screened according to the industry classification of land transfer in the China land market network. In addition, distort_fz was used to calculate the distortion of the price of industrial land in the textile industry, according to the above method.

Table 4. Substitution explanatory variable.

	(1) cr_fx	(2) cr_pf
distort_fz	−0.2683 *** (0.0840)	−0.2560 *** (0.0901)
Controls	Yes	Yes
Constant	10.7729 *** (2.6286)	14.2100 *** (2.4987)
Year and province	Yes	Yes
N	512	512
Adj-R2	0.2079	0.2741

Note: *** p < 0.01.

reports the regression results. The coefficient of distort_fz is significantly negative, with little change, again verifying the robustness of the conclusions in this paper.

4.3. Instrumental Variable Estimation

Endogeneity is an unavoidable topic when discussing the impact of industrial land price distortion on overcapacity. The main sources of endogeneity are reverse causality, omitted variables, and measurement error problems. (1) Reverse causality: The reverse causality of industrial land price distortion in overcapacity is reflected in two aspects. On the one hand, when the eastern region offers low industrial land prices and thereby attracts a large influx of enterprises and excessive investment, more enterprises and employees tend to move in. This will serve to attract even more enterprises and greater population agglomeration, thus increasing the demand for (and thereby raising the price of) industrial land. On the other hand, when local governments realize that the degree of overcapacity is too serious, they may raise the price of industrial land by intervening in the scale and means used to transfer industrial land. These steps would be taken to prevent the negative impact on the long-term regional economy caused by the further expansion of enterprise investment. (2) Omitted variables: Although a series of control variables were added to the empirical test in this paper to avoid omission bias, it is impossible to control for all the factors that affect overcapacity in the eastern region. (3) Measurement error: Although this paper adopts different methods to measure the degree of distortion of industrial land prices via benchmark regression and a robustness test, there may still be endogeneity problems caused by measurement error.

Therefore, this study attempts to find appropriate instrumental variables to solve the endogeneity problem. Considering the instrumental variable requirements and the particularity of China’s land system, the logarithm of the product of the geographical average slope and the reciprocal loan interest rate (ln[slope∗(1/i)]), and the logarithm of the product of the administrative area and time of the prefecture-level cities (ln(area∗year)) were selected as instrumental variables. First of all, the correlation of the first instrumental variable is that the price of land factors depends on both the supply and demand of land factors. From the supply side, the steeper the land slope in a certain area is, the higher the housing price will be [42]. However, due to the substitution relationship between commercial residential land and industrial land, the area of transferred industrial land in Chinese cities will increase in line with the increase of the proportion of land with a slope lower than 15 degrees. The increased supply leads to a decline in the price of industrial land; that is, the average slope and the degree of distortion are inversely related. From the demand side, the increase in bank loan interest rates will increase the financing constraints of enterprises, reduce the funds that enterprises can use to purchase land, reduce the demand for industrial land, and thus lead to a decline in industrial land prices, thereby increasing the degree of distortion. At the same time, terrain, as a natural geographical condition, does not have other impacts on overcapacity in traditional industries. The benchmark loan interest rate at the national level also has no systemic impact on overcapacity, which is in line with the exogenous conditions. Therefore, the interaction between the two (terrain and the benchmark loan interest rate) is taken as the tool variable of this paper. Secondly, the correlation between the administrative division area and the logarithm of time product is that the larger the administrative division area is, the more industrial land the government can sell. Specifically, the greater the land supply is, the more local governments are able to use land as an attractor with lower land prices. Similarly, as a natural geographical condition, land area itself will not cause other impacts on overcapacity of traditional industries. This is in line with the exogenous conditions.

The estimated results of the instrumental variables are shown in

Table 5. Instrumental variable regression.

Phase I
	(1)distort	(2)distort
ln[slope*(1/i)]	−0.0519 *** (0.0103)	−0.0441 *** (0.0107)
ln(area*year)	0.1973 *** (0.0317)	0.2026 *** (0.0328)
Kleibergen–Paap rk LM statistic	38.4620 [0.0000]	36.9310 [0.0000]
Kleibergen–Paap rk Wald F statistic	23.6220	21.2590
Sargan–Hansen test	0.8290 [0.3624]	2.5300 [0.1117]
Phase II
	cr_fx	cr_pf
distort	−1.4218 *** (0.4659)	−0.6293 ** (0.2636)
Controls	Yes	Yes
Year and province	Yes	Yes

Note: *** p < 0.01, ** p < 0.05.

, in which Columns (1) and (2), respectively, report the regression results of the two different methods used to measure capacity utilization in traditional industries. The estimation results of the first stage showed that the Kleibergen–Paap rk LM test statistics of the two instrumental variables at 38.4620 and 36.9310, respectively, with p values of 0.000. All of these passed the significance level test of 1%, indicating that no problem existed with regard to the insufficient recognition of instrumental variables. (The Kleibergen–Paap rk LM test statistic is used for the under-identification test. One of the prerequisites for using the instrumental variable method is that the rank condition is established, that is, rank[E(z_ix_i’)] = K (full column rank), where zi ≡ (z_i1 ... z_iL)’ (L instrumental variables), x_i ≡ (x_i1 ... x_iK)’ (K explanatory variables), z_i and x_i may have overlapping elements, and L ≧ K (order condition). If the column rank of matrix E(z_ix_i’) is less than K, it is not recognizable. For whether the rank condition is established, the “under-identification test” can be carried out. The null hypothesis is “H₀:rank[E(z_ix_i’)] = K − 1”, and the alternative hypothesis is “H₁:rank[E(z_ix_i’)] = K”. If heteroskedasticity is allowed, the “Kleibergen–Paap rk LM statistic” is used, with an asymptotic distribution of χ²(L − K + 1). If the p-value of the Kleibergen–Paap rk LM statistic is significant, the under-identification null hypothesis is strongly rejected. Use the “ivreghdfe” command in stata15 to perform instrumental variable regression, and the software will report the Kleibergen–Paap rk LM statistic, the Kleibergen–Paap Wald rk F statistic, and the Stock and Yogo critical values. The p-value of the Kleibergen–Paap rk LM statistic in this paper is 0.0000, strongly rejecting the under-identification null hypothesis). The Kleibergen–Paap rk Wald F statistics were 23.6220 and 21.2590, both of which were significantly greater than the critical value of 19.93 given by Stock and Yogo at the 10% level. This result indicates that the problem of weak instrumental variables did not exist here. In this paper, the Sargan–Hansen test was also used to test the externality of the tool variables. The resulting p values were all greater than 0.1, indicating that the tool variables selected in this paper were all exogenous. Therefore, the regression results of the first stage proved that the instrumental variables selected in this paper meet the correlation and exogeneity assumptions. In addition, the estimation results of the second stage, shown in Table 5, revealed that the estimation coefficients of industrial land price distortion on the capacity utilization of traditional industries passed the significance tests of 1% and 5%, respectively. In addition, the sign and basic regression were consistent, indicating that the conclusion of this paper was still robust after the potential endogeneity was alleviated.

5. Further Analysis

Based on the above conclusions, distorted industrial land prices have a positive impact on overcapacity in the textile industry (the result shows a negative impact on the capacity utilization rate). To solve the overcapacity in traditional industry, optimizing the price system of industrial land needs to be paid attention to. Further, this study attempts to analyze the possible channels and paths of industrial land price distortion and how they affect overcapacity in the textile industry. In terms of a mechanism test, this paper analyzes how the distortion of industrial land prices leads to overcapacity in the textile industry, from the perspective of enterprise influx. To verify Hypotheses H1 and H3, this paper constructed the following test equation on the basis of the benchmark empirical model:

$$c{r}_{i,t}={\alpha }_1+{\beta }_{1}distor{t}_{i,t}+{\gamma }_1{X}_{i,t}+{\lambda }_t+provinc{e}_i+{\epsilon }_{i,t}$$

(6)

$$ent{r}_{i,t}={\alpha }_2+{\beta }_{2}distor{t}_{i,t}+{\gamma }_2{X}_{i,t}+{\lambda }_t+provinc{e}_i+{\epsilon }_{i,t}$$

(7)

$$c{r}_{i,t}={\alpha }_3+{\beta }_{3}distor{t}_{i,t}+{\beta }_{4}ent{r}_{i,t}+{\gamma }_4{X}_{i,t}+{\lambda }_t+provinc{e}_i+{\epsilon }_{i,t}$$

(8)

In the above equation, entr is the mediating variable ‘enterprise entry’; the mediating effect is β₂β₄. As for the measurement of enterprise entry, existing literature mostly measures the number of new enterprises and the gross industrial output value of new enterprises [43]. Based on existing studies, this paper measures the entry of enterprises in eastern China from two aspects, in order to ensure the robustness of the results. Here, entr1 represents the number of newly established enterprises, and entr2 represents the total industrial output value of those newly established enterprises. The term ‘newly established enterprises’ refers to enterprises established within and including two years. The gross industrial output value of newly entered enterprises is calculated by summarizing the gross industrial output value of textile enterprises by region. All data are from the China Industrial Enterprise Database.

Looking at the regression results of Columns (1) and (4) in

Table 6. Mechanism test.

	(1) entr1	(2) cr_fx	(3) cr_pf	(4) entr2	(5) cr_fx	(6) cr_pf
distort	0.5070 *** (0.1880)	−0.3886 *** (0.1001)	−0.2243 ** (0.0877)	0.9247 ** (0.3611)	−0.3739 *** (0.0996)	−0.2137 ** (0.0850)
entr1		−0.0260 (0.0219)	−0.0500 ** (0.0279)
entr2					−0.0302 ** (0.0126)	−0.0389 ** (0.0194)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−11.7523 *** (4.4375)	10.6918 *** (2.1505)	7.7772 *** (1.5138)	−13.4290 * (7.5155)	10.5919 *** (2.1291)	7.8426 *** (1.5055)
Year and province	Yes	Yes	Yes	Yes	Yes	Yes
N	588	588	588	588	588	588
Adj-R2	0.2889	0.2310	0.2699	0.1785	0.2371	0.2765

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

, it is not difficult to see that the enterprise entry measured by the two methods was positively affected by industrial land price distortion, which is the core explanatory variable. This finding shows that the strategy of attracting land investment with low land prices in the eastern region does attract an influx of textile enterprises. Columns (2), (3), (5), and (6) all indicated that the influx of textile enterprises has reduced the capacity utilization of the industry and has caused the textile industry’s overcapacity in the eastern region. In the empirical test, the regression results obtained were consistent, regardless of the different methods used to measure enterprise entry, or if the capacity utilization rate was measured by different methods. This finding shows that industrial land price distortion in eastern China leads to an influx of textile enterprises, which in turn leads to overcapacity. The path has significant internal effectiveness and robustness. Therefore, to solve the problem of overcapacity and overcapacity in traditional industries, we should start from the root of industrial land price.

6. Conclusions and Policy Recommendations

This paper used the panel data of prefecture-level cities (from 2007 to 2014) to study the causes of excessive influx and overcapacity of textile industry in eastern China. In the benchmark regression, this study found that the greater the regional industrial land price distortion is, the more obvious overcapacity in the textile industry will be. This result was still significant after a series of robustness tests. In the mechanism test, this paper verified that the excessive influx of textile enterprises is itself the transmission mechanism between the distortion of industrial land price and overcapacity. To enjoy the land concessions offered by local governments, enterprises will choose to enter an area, even if the project is losing money. This behavior leads to waves of repeated construction and excess production capacity in the industry again. The phenomenon of the revival of overcapacity also appears.

Based on the above empirical research conclusions, this study has led to the following observations: the overcapacity in traditional industries in recent years needs further attention from both the government and academia. If overcapacity becomes too serious, it may lead to a sharp drop in the price of industry products, a decline in industry profits, or even negative growth. In addition, overcapacity can also damage the investment expectations of enterprises and the consumption expectations of residents, ultimately leading to downward pressure on the economy. Therefore, solving the problem of overcapacity is becoming increasingly urgent. Based on the theoretical analysis and empirical test in this paper, the following policy suggestions are suggested.

First, we will continue to optimize the price system of industrial land. Specifically, the management system for the transfer of land-use rights based on bidding, auction, and listing should be improved, the monopoly position of local governments over land should be broken, and the allocation of the supply and demand of land factors should be optimized through the market supply and demand mechanism. When the marketization degree of land transfer becomes high, and when the local government is no longer monopolistic, the location of enterprises will be constrained by the cost of land to a much greater degree. Even if the market is more developed in the east, the net return on land is likely to be lower. When this occurs, companies are more likely to move into the central and western regions, thus alleviating the problem of overcapacity in the east.

Second, the distortion of industrial land prices is mainly caused by regional competition. Under the background of a ‘promotion tournament’, with economic growth as the core, more favorable policies must be implemented to optimize the price system of industrial land and correct the distortion of industrial land prices. For example, the performance evaluation criteria of officials should be reformed; the focus should be shifted from quantity to quality of economic growth. As a result, local governments would tend to be more market-oriented in their efforts to attract efficient enterprises and would be less concerned with simply increasing the number of enterprises. In this way, the problems of the distortion of industrial land prices and the overcapacity of traditional industries can be effectively solved.

Third, local governments should pay more attention to the sustainable development of the textile industry. Most of China’s textile enterprises still produce products with primary processing and low added value, which is largely due to the improper intervention of local governments through land investment. The empirical results show that for the purpose of obtaining short-term economic performance, local governments often pay attention to the number of enterprises rather than quality, resulting in overcapacity. The problems of low land use efficiency and environmental pollution caused by many inefficient textile enterprises have damaged the long-term sustainable development of the textile industry. Therefore, local governments should reduce land investment or attract high-quality textile enterprises with low pollution and high value-added products, so as to improve the competitiveness of domestic textile industry. In addition, local governments should pay attention to guiding enterprises to implement the measures of environmental cost internalization to realize the long-term sustainable development of textile industry.

In addition, this study is of great significance for the developing countries. First, in the process of optimizing the price system of industrial land, the market should be the leading factor and the government should provide support, but government monopolies should be prevented. Secondly, in the process of land attracting investment, the quality and development prospects of the company should be paid attention to, instead of just seeking quantity for temporary gains. Thirdly, in the process of transferring traditional industries, it is necessary to focus on the combination of industrial transfer and industrial upgrading, eliminate backward production capacity during the transfer, and encourage enterprise innovation to enhance international competitiveness. Finally, countries that have or are about to experience excess capacity can help domestic companies seek profitable orders or reduce operating costs in international markets. For example, the Belt and Road Initiative is used reasonably. Some countries along the Belt and Road Initiative are still in the early stage of industrial economic development. There is an urgent need for machinery, equipment, and infrastructure construction. Industrial transfer to these countries would help the home country to relieve production capacity and help the host country to develop related industries so as to realize the sustainable development of the textile industry on a global scale.