Labor Endowment Change, Regional Difference, and Agricultural Production Location Adjustment: Evidence from China

Abstract

The outflow of the rural labor force has a great impact on the location distribution of agricultural production, which has not attracted enough attention in existing studies. This article describes the mechanism of agricultural production location adjustment and further studies the influence of a regional difference in labor endowment on agricultural production location adjustment. Based on commodity and province-level panel data calculation, the results show that the agricultural production location in China has been adjusted from the East to the Central and then to the West with accelerating transfer speed. Furthermore, it is found that the regional differences in labor endowment are the main driving force for the shift of low alternative crop production. The more labor-intensive the crop varieties are, the more obvious the influence of the regional differences in labor endowment is on the crop production location adjustment. Moreover, regional differences in labor endowment have little effect on the location adjustment of vegetable production, as expanding the vegetable market demand may offset the effect of labor supply shortage on the adjustment of vegetable production location. Therefore, it is necessary to formulate a regional industrial development plan consistent with the trend of agricultural production location adjustment, as well as promote the construction of agricultural circulation facilities and socialized services in less developed areas.

1. Introduction

The spatial distribution of agricultural production has been changing in China over the last two decades. Some labor-intensive agricultural industries (such as cotton, oil, sugar, small varieties of crops) are gradually withdrawing from the developed coastal areas in China, while these related agricultural industries are developing again in the inland underdeveloped areas. The above changes in agricultural production are manifested as agricultural production location adjustment.

Existing studies on the relationship between labor endowment and the location of agricultural production mainly focus on the following aspects: (1) the influence of changes in labor endowment on part-time and off-farm employment [1], on factor substitution [2], on farmland transfer [3], and on abandoned farmlands [4]; (2) the influence of changes in labor endowment on agricultural production efficiency [5,6,7]; (3) the influence of changes in labor endowment on agricultural production structure [8,9,10]; (4) the influence of changes in labor endowment on food security [11,12,13,14,15]; (5) the influence of changes in labor endowment on grain crop production.

For regional comparative advantages, the government of China has issued some policies to guide different crop varieties to the most suitable areas, improving the matching degree between agriculture and resources for building a specialized regional production pattern. However, due to the lack of systematic research on agricultural location adjustment, the formulation and implementation of relevant policies are restricted to a certain extent. Many scholars have discussed the above research topics, but there still exist some gaps, especially for cash crops. Grain crops mainly refer to the crops used as human staple food. Cash crops can provide raw materials for industry, which can be divided into fiber crops (such as cotton), oil crops, sugar crops, other crops (such as vegetables), etc. The production of grain crops has a high degree of standardization, while cash crop production has the characteristics of high regional requirements, intensive labor input, as well as high technical requirements. As a result, cash crop production requires more labor. In 2020, the labor input per hectare of oil, cotton, sugar, silkworms, and vegetables was predicted to be 1.55, 2.46, 2.45, 7.91, and 6.82 times that of grain crops, respectively. In addition, cash crops usually have a higher economic value than grain crops. In 2020, the average cash income per hectare of oil, cotton, sugar, silkworms, and vegetables was 1.51, 1.38, 1.72, 1.96, and 9.73 times that of grain crops, respectively, indicating the positive impacts of cash crop production on improving farmers’ welfare. There is a large difference between the production of cash crops and grain crops as cash crops are more sensitive to labor endowment change [16]. However, on the one hand, ensuring food security has always been the primary goal of China’s agricultural development, which will obviously affect the cultivation of cash crops. On the other hand, as incomes and consumption levels rise, so does the demand for high-quality cash crops. There is a lack of research on how to balance the above contradictions.

This article investigates the factors that influence the location adjustment of cash crop production within China using province-level panel data from 1990 to 2017. Specifically, this article analyzes the impact of labor endowment differences among Chinese provinces on the location adjustment of cash crop production. Due to the wide varieties and different agronomic characteristics of cash crop production, this article selects five kinds of labor-intensive cash crops, including cotton, oil, sugar, silkworm, and vegetables for detailed analysis based on the availability of statistical data. The results show that regional differences in labor resources constitute the driving force of the location adjustment of agricultural production. The more intensive the use of labor is in crop production, the stronger the effect of changes in labor endowment on production location. This article also finds that the location of vegetable production is less affected by regional differences in labor endowment.

This article is structured as follows: Section 2 includes the literature review and analysis logic. Section 3 provides the details of agricultural production location adjustment. Section 4 introduces the empirical methodology and describes the datasets. Section 5 presents the empirical results. Section 6 concludes this article.

2. Literature Review

With the accelerated consumption of the demographic dividend by economic development, the potential labor supply has been declining. Based on the normalization of “migrant worker shortage” in coastal areas and the rising wages, some scholars believe that China has entered an era of scarcity in labor endowment and the Chinese economy has reached the “Lewis Turning Point” [17,18,19]. However, in the process of China’s economic transformation, there are not only some characteristics consistent with the “Lewis Turning Point”, such as labor shortage and wage increases, but also some deviations from the “Lewis Turning Point”, such as the widening income gaps between urban and rural areas. Therefore, some studies have pointed out that the “Lewis Turning Point” has not yet arrived [20], and there is still a certain amount of surplus labor force in rural areas [21,22]. The main reasons for this divergence are related to the lack of statistical data and the observed samples. In fact, China is a country with great regional differences in the absolute and relative levels of labor-related factors, in which agricultural labor endowment varies from region to region. The employment rate and labor transfer rate of different regions also vary greatly. Some scholars have found that the change in labor supply shortage is not all-sided [23], and the arrival of shortcomings is characterized by gradual diffusion from the core of the economic circle to the periphery [24,25]. Thus, labor dissipation is a long-term and gradual process [26].

In traditional agriculture, intensive farming is the main characteristic of agricultural production, which determines that agricultural production relies heavily on labor input, and the regional difference formed by the advantage of labor endowment may constitute a necessary condition for the regional adjustment of agricultural production. However, the factor endowment theory reminds us that labor is not the only factor of production, as technological differences are also crucial to the formation of the industrial division [27]. On the one hand, the shortage of labor supply will inevitably influence the production pattern of smallholders, but on the other hand, the improvement in technical efficiency will help to transform traditional agriculture and ease the rigid constraint of the labor force [28]. It is assumed that agricultural production can improve productivity and reduce labor intensity using modern technology, which will help to improve the comparative advantages of agricultural production. A large number of studies have shown that in the context of increasing labor endowment constraints, farmers can adjust the factor input structure with the help of the factor market price, that is, replace expensive and scarce labor factors with cheap and abundant mechanical factors to promote the transformation of the labor-saving production mode [29,30]. Besides mechanical technology, biological breeding as well as chemical pesticide technology can also help to improve labor productivity [31]. Improving labor productivity by adopting advanced production techniques can help reduce agricultural production costs and alleviate or even offset the impact of labor shortage as well as rising labor prices.

The above analysis shows a potential logic of regional division under the expectation of successful transformation of agricultural modernization. However, the adjustment of factor input structures is inherent in the agronomic characteristics of crop varieties, and is also influenced by factors such as technology and natural conditions [32]. The production technology of cash crops is mainly biotechnology, that is, improving the standardization of crop production by selecting improved varieties or new pesticides, which can save labor within a certain range. However, cash crop production is often less standardized, and the technical details are numerous and scattered, which means that it is difficult to replace the labor force through machinery. It is worth noting that both biotechnology and mechanical technology are the main means of alleviating labor shortage. However, if the replacement of the labor force by machinery cannot be achieved, the effect of labor saving solely relying on biotechnology is limited. In China, the production of cotton, vegetables, and other cash crops still maintains the high-level labor input. If the labor-saving production mode cannot be applied, the negative impact of labor shortage will directly lead to the increase in costs and, thus, accelerate the transfer of agricultural production location to areas with lower production costs. In order to clearly observe the effect of labor endowment change on the location adjustment of cash crop production, this article defines two production scenarios as follows:

Scenario 1 (developed region A): With the constraints of agricultural labor supply, the intensive use of labor factors in crop production increases the opportunity costs of labor input for local farmers. In order to reduce labor costs, farmers may seek to increase other inputs, which improves production efficiency and maintains the current output. If the labor force cannot be replaced by modern factors for higher labor productivity, farmers will allocate labor factors to the agricultural sectors with lower labor intensity through the adjustment of planting structure, that is, to reduce the scale of crop production with high labor intensity and increase that with low labor intensity. Moreover, compared with the rapid growth of non-agricultural income, rational farmers either choose part-time production or simply allocate labor factors to non-agricultural sectors for profit maximization.

Scenario 2 (less-developed region B): As the labor transfer rate in less developed regions is lower than that in developed regions, agricultural labor resources are relatively abundant and, thus, their prices are relatively low. In the case of limited off-farm employment opportunities, farmers need to seek greater profit space in the agricultural sector. On the one hand, the high added value of cash crops helps to improve labor productivity and promotes the income level of farmers [31]. On the other hand, the intensive use of labor factors in cash crops can also solve the employment problem to a certain extent. Therefore, it is rational to expand the production scale of high-value-added crops under the condition of relatively abundant labor supply. The change in crop production scale is spatially manifested as the dynamic transfer of agricultural production location. This specific logic is shown in Figure 1.

3. Analysis of Agricultural Production Location Adjustment

3.1. Regional Differences in Labor Endowment

The rural labor force can be allocated between industries and regions to promote the development of industrialization and urbanization in China. Figure 2 shows the basic situation of rural–urban mobility of the rural labor force in China since 1979. With the continuous increase in labor demand in the industrial sector, the rising wages have induced millions of migrant workers to urban areas. At the national level, the total non-agricultural employment increased from 19.5 million to 275.8 million. The non-agricultural rate of the rural labor force reached 51.82% in 2017. From 1979 to 2005, with the relaxation of population mobility policies and the establishment of the market economy system, the non-agricultural employment scale of the rural labor force has shown an accelerating trend, while there has been a labor shortage along the coastal areas in China after 2006. Although the wages of migrant workers have accelerated since then, the growth rate of the non-agricultural employment rate of the rural labor force has slowed down, which means that the potential of the labor force in rural areas has begun to decline with the decreases in adjustment effect of labor factors on agricultural production.

At the regional level, the non-agricultural employment scale of rural labor in Western China was the largest with a ratio of 61.32%, followed by the Central regions (48.63%) and the Western regions (44.62%) in 2017. Since 2006, the growth rate of non-agricultural employment in the rural labor force among regions has slowed down, but the growth rate in the Western region is faster than those in the Eastern and Central regions. Therefore, the rural labor force has gradually transferred from the Eastern and Central regions to the West. The Eastern region has the strongest labor endowment constraint as well as the lowest comparative advantage of labor input in agricultural production.

3.2. Agricultural Technology Selection

Faced with the change in agricultural labor supply shortage, farmers usually adjust the factor input structure and replace labor factors with relatively abundant and cheap production factors (e.g., mechanical factors). However, due to the difficulty in integrating agricultural machinery and agronomy, the degree of mechanical substitution varies across different crops. Among them, the degree of mechanical substitution in grain production is highest, while that of cash crops is generally low. In 2017, the mechanical labor ratio of grain was 9.86, much higher than those of sugar (4.43), cotton (3.86), oil (3.58), vegetables (3.09), and silkworm (0.74). (Data from the Compiled Materials of Costs and Profits of Agricultural Products of China. The mechanical labor ratio refers to the ratio of machinery input to labor input per unit area of cultivated land. Mechanical inputs per mu include machinery operating costs, irrigation costs, energy costs, fuel costs, and depreciation costs of fixed assets (at a 1990’s constant price). As a result, the land–labor ratio of grain production is continuously decreasing, while that of cash crops is still maintained at a high level (in 2017, labor demand per mu of grain, oil, sugar, cotton, vegetable, and silkworm was 5.03, 7.53, 12.82, 15.62, 27.88, and 43.24, respectively). Due to the complex production process, there are often obstacles to replacing the labor force through mechanical technology, which is not conducive to the improvement of labor productivity. Therefore, it is necessary to develop agricultural biotechnology based on special agricultural products, such as using high-quality chemical pesticides to reduce labor input [31], or improving crop traits for higher labor productivity [33], which is the key to enhancing the market competitiveness of cash crops. However, due to the lack of market competitiveness, intensive use of labor factors will inevitably be impacted by the rising wages when facing rapid changes in factor markets.

The output elasticity is then introduced to explain the reflection of agricultural output on the change in factor input structure. Output elasticity reflects the sensitivity of the output change to the input change and, thus, can be used to evaluate the conversion effect of the factor input. Following Wang et al. [34],

Table 1. The estimated results of production factor output elasticity.

Crops	Output Elasticity			Substitution Elasticity
	Labor	Mechanical Input	Biological Input
Grain	0.183 *** (5.69)	0.417 *** (7.27)	0.191 *** (3.81)	1.243
Oil	0.247 *** (4.20)	0.154 *** (6.33)	0.052 (1.11)	0.883
Cotton	0.206 *** (3.43)	0.095 *** (4.34)	0.209 *** (4.24)	0.719
Sugar	0.132 * (2.48)	0.087 ** (2.73)	0.179 ** (2.80)	0.538
Silkworm	0.421 *** (10.28)	0.032 * (2.23)	0.312 *** (5.74)	0.645

Data source: The Compiled Materials of Costs and Profits of Agricultural Products of China. Notes: (1) The output elasticity is estimated by the Cobb–Douglas production function, and the substitution elasticity is estimated by the CES production function. (2) Mechanical inputs per mu include machinery operating costs, irrigation costs, energy costs, fuel costs, and depreciation costs of fixed assets. Biological inputs per mu include seed costs, fertilizer costs, and pesticide costs. (3) *** p < 0.01, ** p < 0.05, * p < 0.1, t-values are listed in parentheses. (4) The estimates of vegetable production are not included, due to severe data loss.

shows the estimated output elasticities of different crops. For grain crops, the elasticity with respect to mechanical input is largest, while that for labor input is smallest, indicating that grain production is more sensitive to the change in mechanical input. Among different input factors, for every 1% increase in machinery input, output increases by 0.42%, and for every 1% decrease in labor input, output decreases by 0.18%, indicating that machinery input can realize an effective substitution of labor.

For cash crops, the mechanical output elasticity of oil is relatively high, but it is still lower than that of labor. Other cash crops are more obvious, especially silkworm production. The labor output elasticity of silkworm is 0.42, which is much higher than that with respect to machinery. According to Table 1, when labor input increases by 1%, silkworm output increases by 0.42%, and when machinery input increases by 1%, silkworm output only increases by 0.03%, indicating that labor input has contributed the most to the output growth of silkworm production. Compared with the output elasticity of biological input, the mechanical output elasticity of cash crops is significantly lower (except for oil), which indicates that the technical selection direction of cash crops is mainly based on biochemical technology.

Following Henningsen and Henningsen [35], this article also estimates the factor substitution elasticity through the CES function. The substitution elasticity of grain production is significantly greater than 1, while that of cash crops is generally less than 1, indicating that the substitution ability of grain crops is stronger than that of cash crops, and it is not conducive to the realization of modernization transformation of the production mode. Due to the difficulty in changing the production mode, most cash crops are still characterized by intensive use of labor factors, which is, in turn, not conducive to alleviating the negative impact of insufficient labor resources and rising labor prices.

In general, the output elasticity with respect to machinery in grain production is much higher than those with respect to the labor force and biological input, while those of the labor force and biological input in cash production is generally higher than that of machinery. When the output elasticity of an input factor is significantly higher, crop production is usually supplemented by intensive use of this factor. With the increase in rigid labor constraints, it is easier for grain crops to relax the labor shortage constraint and to stabilize the original production location by mechanical substitution. However, the logic seems different for cash crops. Once the labor supply changes, the intensive use of labor factors is likely to induce the production of cash crops in areas with more abundant labor.

3.3. Agricultural Production Location Adjustment

In developed countries, industrial relocation is defined as the location change of an enterprise to study the absolute relocation [36]. Due to the lack of statistical data, relative indicators such as output value share, the proportion of employees, and industrial concentration are often used to identify industrial relocation in China [37]. This article adopts a similar approach to measure the situation of location adjustment using the crop-cultivated area. The equation is as follows:

$$\Delta Y=Y_{it}^k-Y_{it-1}^k=\frac{S_{it}^k}{{{\displaystyle \sum }}_i{S}_{it}^k}-\frac{S_{it-1}^k}{{{\displaystyle \sum }}_i{S}_{it-1}^k}$$

(1)

where $\Delta Y$ is the production location adjustment of crop k in region i. $\Delta Y$ is negative for outside-transfer and positive for inside-transfer. $Y_{it}^k$ and $Y_{it-1}^k$ represent the proportion of cultivated area in region i in the whole country for crop k in year t and year t − 1, respectively. $S_{it}^k$ and $S_{it-1}^k$ represent the cultivated area of region i for crop k in year t and year t − 1, respectively.

Table 2. The production location adjustment of different crops in China (1990–2017).

Time Period		Grain				Cotton
		Eastern	Central	Western		Eastern	Central	Western
1990–1995	$\Delta Y$	−0.21	−0.23	0.44	$\Delta Y$	−15.97	10.15	5.82
1996–2000		−0.75	−0.10	0.86		−1.58	−5.01	6.59
2001–2005		−1.14	2.11	−0.97		3.16	−2.20	−0.97
2006–2010		−1.13	1.11	0.02		−2.89	−2.68	5.57
2011–2017		−0.90	2.49	−1.59		−10.30	−14.95	25.24
Total		−5.52	8.35	−2.83		−30.45	−17.36	47.81
		Oil				Sugar
		Eastern	Central	Western		Eastern	Central	Western
1990–1995	$\Delta Y$	−2.95	5.00	−2.05	$\Delta Y$	−6.50	−3.39	9.88
1996–2000		0.16	−0.79	0.63		−1.91	−9.31	11.22
2001–2005		−2.78	0.46	2.32		−3.58	−8.43	12.01
2006–2010		−3.05	1.95	1.10		−3.13	−2.13	5.26
2011–2017		−4.87	−0.31	5.18		−0.40	−4.03	4.43
Total		−11.98	4.63	7.34		−15.52	−27.29	42.80
		Silkworm				Vegetable
		Eastern	Central	Western		Eastern	Central	Western
1990–1995	$\Delta Y$	−2.54	6.16	−3.62	$\Delta Y$	2.21	−1.03	−1.18
1996–2000		−1.92	1.71	0.21		0.12	0.16	−0.28
2001–2005		−2.13	1.28	0.85		−1.59	−0.88	2.47
2006–2010		−4.52	−0.18	4.70		−1.98	−0.47	2.45
2011–2017		−9.05	−1.60	10.65		−5.35	−0.96	6.32
Total		−20.16	7.37	12.79		−6.56	−3.92	10.48

Data source: China Rural Statistical Yearbook.

shows the identification results of different crop production location adjustments. In terms of grain crops, the proportion of grain production in the Eastern and Western regions decreased. Moreover, with the support of grain policies, grain production gradually concentrated in the Central regions based on stabilizing the original production location. In 2017, the 13 major grain producing areas accounted for 75.21% of the total grain cultivated area in China, with the Central regions accounting for nearly half (the 13 major grain-producing areas are Heilongjiang, Jilin, Liaoning, Inner Mongolia, Hebei, Henan, Shandong, Jiangsu, Anhui, Jiangxi, Hubei, Hunan, and Sichuan, among which Heilongjiang, Jilin, Jiangxi, Henan, Hubei, and Hunan are in the central region).

In terms of cash crops, first, as for the time change, the scale of outside-transfer among different crops is small in the Eastern region, and the transfer speed is relatively steady before 2005. However, the overall performance of the Eastern and Central regions becomes large outside transfers with the increasing scale and accelerating speed after 2005.

Secondly, as for the spatial order, the crop production in the Eastern regions transfers outside first, and then the Central regions later, while that in the Western regions generally transfers inside. To be more specific, the Eastern regions have basically turned into a trend of outside-transfers since 1990, and the Central regions have shown the same trend since 2000. Except for sugar crops, the transfer intensity of the Eastern regions is generally higher than that of the Central regions, which means that crop production in the Eastern and Central regions is faced with increasing production constraints, and the comparative advantage of production is declining with an increasing speed.

Thirdly, as for the various orders, crop varieties with weak mechanical substitution ability are the first to transfer outside with a larger scale and faster speed. Since 2005, the outside-transfer scales of silkworm and cotton in the Eastern regions have been 13.57% and 13.19%, respectively, significantly higher than those of oil (7.92%) and sugar (3.53%). With the higher labor intensity of silkworm and cotton production, it is difficult to achieve mechanical substitution, and, thus, the inelasticity of labor supply in the Eastern region restricts these two crop varieties severely. However, there are some exceptions for vegetables. The production of vegetable crops in the Eastern and Central regions shows a certain trend of outside-transfer, but the scale is relatively small compared to other cash crops. With the improvement in consumption levels, the market capacity of vegetables is constantly expanding, which helps to stabilize the original production location of vegetables.

4. Model and Data

4.1. Empirical Model

In the empirical analysis, Equation (1) can be used as an identification method for industrial transfer, but the change in outside-transfer and inside-transfer makes the identification result $\Delta Y$ in some provinces negative, which may cause a lot of sample losses when taking the logarithm. Following Wen [38], this article selects the proportion of cultivated area in the whole country as a variable to measure agricultural production location adjustment. The increase in the share of cultivated area implies inside-transfer, while the decrease represents outside-transfer. The following benchmark regression model is constructed:

$$Y_{it}=\alpha +\beta Labo{r}_{it}+\theta X_{it}+{\mu }_i+{\epsilon }_{it}$$

(2)

where $Y_{it}$ refers to the proportion of cultivated area in the whole country.$Labo{r}_{it}$ refers to the relative status of labor endowment as the key explanatory variable and is measured by labor abundance (i.e., rural surplus labor) and agricultural wages. $X_{it}$ is a vector of control variables. ${\mu }_i$ is the non-observed region fixed effect. ${\epsilon }_{it}$ is the random error term.

This article considers three groups of control variables in Equation (2). The first group is production variables, including: (1) the total power of agricultural machinery is measured by the ratio in each province to that in the whole country, and refers to the total power of various machinery mainly used in agriculture, calculated in watts of power; (2) the cultivated area per capita is measured by the ratio in each region to that in the whole country; (3) the degree of the aging population refers to the proportion of the population aged 65 or above in rural areas; (4) education level is measured by years of schooling.

The second group is economic and geographical variables, including: (4) transportation level is measured by the average proportion of passenger volume and cargo volume in each province to that in the whole country; (5) market demand. The new economic geography points out that enterprises tend to locate their industries near the consumer market [39], and a rise in demand may induce these enterprises to move toward a large market scale. This article uses the proportion of non-grain expenditure in food consumption between urban and rural residents to measure the purchasing level in each region, and then uses the ratio in each region to that in the whole country to measure consumer demand.

Furthermore, it is worth noting that economic policies play an important role in the process of industrial transfer, and, thus, a third group of control variables is introduced: (6) the agricultural participation of local government is measured by expenditure on agriculture, forestry, and water resources. Taking the differences of agricultural development among provinces into account, the ratio between local government and the whole country after standardization is taken to measure the relative status of local government’s participation.

4.2. Data

This article uses province-level panel data from 1990 to 2017. The data are mainly collected from official statistics, including the Compilation of Agricultural Statistical Data of the 30 Years of Reform and Opening Up, the 40 Years of Reform and Opening Up, China Statistical Yearbook, China Population & Employment Statistical Yearbook, China Yearbook of Household Survey, China Rural Statistical Yearbook, China Compendium of Silkworm Production Statistical 1949–2008, China’s National Income: 1952–1995, Data of Gross Domestic Product of China (editions of 1952–1995, 1996–2002, 1952–2004), Statistical Yearbook of the Chinese Investment in Fixed Assets, The Compiled Materials of Costs and Profits of Agricultural Products of China, and the provincial-level statistical yearbooks.

In particular, the relevant variables are calculated as follows: (1) Followed by Hu [40], the degree of labor transfer can be measured by the proportion of agricultural labor force in the total labor force. The larger the proportion, the smaller the degree of labor transfer, leading to greater surplus labor force. To observe the situation of labor resources in the agricultural sector, this article measures the ratio of rural surplus labor force with the proportion of agricultural labor in rural employees. To reflect the relative level of agricultural labor resources as well as improve the comparability of indicators among different regions, the ratio of rural surplus labor in each province is divided by the national ratio of rural surplus labor to measure the labor abundance; the ratio of rural surplus labor force is measured by the ratio of workers in agricultural, forestry, animal husbandry and fishery, and rural workers; (2) as there are no existing data on the agricultural wages, this article estimates labor price. The labor remuneration in the agricultural sector at the province level has been reported in GDP accounting data since 1978, and, thus, labor price is the ratio of labor remuneration to the labor force number in this article (as this accounting method has been adjusted twice in 2004 and 2008, this article makes a consistent adjustment of the index comparability, which is not reported here, due to the space limitation).

Table 3. Data summary.

Variables	Definitions	Mean	Std. Dev.
Cotton area share	Cotton planting area by province/total cotton planting area (%)	6.13	6.59
Oil area share	Oil planting area by province/total oil planting area (%)	4.23	3.48
Sugar area share	Sugar planting area by province/total sugar planting area (%)	7.00	8.27
Silkworm area share	Silkworm planting area by province/total silkworm planting area (%)	5.76	5.12
Vegetable area share	Vegetable planting area by province/total vegetable planting area (%)	3.23	2.55
Labor abundance index	Surplus rural labor force by province/national surplus rural labor force	1.02	0.26
Labor price	Rural labor price by province/national rural labor price	1.18	0.58
Agricultural machinery	Power of agricultural machinery by province/national total power of agricultural machinery	3.23	3.05
Cultivated area per capita	Arable land per capita by province/national arable land per capita	1.21	1.03
Aging population	Population aged 65 and above by province/national population aged 65 and above	0.96	0.27
Education year	Education years of rural labor by province/national education years of rural labor	0.99	0.13
Market demand	Proportion of non-food consumption by province/national proportion of non-food consumption	0.97	0.12
Transportation condition	Total passenger and freight traffic by province/national total passenger and freight traffic	3.17	2.14
Local government participation	Agricultural expenditure as a share of GDP by province/national agricultural expenditure as a share of GDP	1.52	1.31

shows that crop production shares differ greatly in different regions. Meanwhile, labor resource abundance and labor price also show a great imbalance, indicating that crop production location is correlated with labor resource distribution.

5. Results

5.1. Basic Regression Results

As panel data are used in this article, the Hausman test is conducted on the model first, and the fixed-effect model is then selected for estimation based on the Hausman test.

Table 4. The estimated results of regional differences in labor endowment on agricultural production location adjustment using labor abundance index.

	Share of Cultivated Area
	Cotton	Oil	Sugar	Silkworm	Vegetable
Labor abundance	0.728 **	0.552 ***	0.628 **	2.182 ***	0.026
	(0.345)	(0.135)	(0.277)	(0.348)	(0.010)
Agri-machinery	0.589 ***	0.782 ***	−0.115	0.372 **	0.174 ***
	(0.170)	(0.058)	(0.106)	(0.176)	(0.051)
Area per capita	−0.037	0.084	−0.078	0.109	−0.153 **
	(0.261)	(0.068)	(0.130)	(0.111)	(0.059)
Aging population	0.092	0.126 *	−0.029	−0.257	0.0296
	(0.197)	(0.076)	(0.146)	(0.198)	(0.058)
Education level	4.054 ***	−1.710 ***	6.379 ***	−0.889	1.838 ***
	(0.969)	(0.353)	(0.661)	(1.080)	(0.278)
Market demand	−0.935	0.964 ***	−4.420 ***	−0.894 **	0.226 *
	(0.591)	(0.185)	(0.423)	(0.427)	(0.133)
Transportation	0.116 ***	−0.117 ***	0.200 ***	−0.019	0.065 ***
	(0.035)	(0.017)	(0.036)	(0.037)	(0.012)
Local government	0.638 ***	−0.231 ***	−0.016	−0.499 ***	−0.030
	(0.149)	(0.048)	(0.075)	(0.078)	(0.034)
Region fixed effect	Y	Y	Y	Y	Y
Year fixed effect	Y	Y	Y	Y	Y
R2	0.369	0.406	0.583	0.150	0.174
F-statistic	6.57 ***	11.38 ***	18.00 ***	2.14 ***	3.82 ***
Observations	444	640	503	476	700

Notes: Standard errors are listed in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the estimated results based on Equation (2) using the labor abundance index as the key explanatory variable. All the regression equations are significant at the 1% statistical level based on the F test value. The results show that there is a significant positive correlation between the labor abundance index and crop cultivated area in cotton, oil, sugar, and silkworm equations. To be more specific, for every 1% increase in the labor abundance, the cultivated area share of cotton increases by 0.728%, oil crops by 0.552%, sugar crops by 0.628%, and silkworm by 2.182%. The above results indicate that areas with relatively abundant agricultural labor supply may continuously induce crop production to be transferred inside; otherwise, outside-transfer will be promoted. Therefore, the regional difference in agricultural labor supply level has constituted an important force of agricultural production location adjustment for cash crops. Moreover, the regression result of the labor abundance index in the vegetable equation is positive, but not significant, which means that the regional difference in labor supply level is not the main factor for the location adjustment of vegetable production.

This article further compares the transfer degree among different cash crops. The influence of labor abundance on the location adjustment of cash crop production is silkworm, cotton, sugar, and oil in descending order. As mentioned above, in 2017, the labor demand per mu of silkworm, cotton, sugar, and oil is 43.24, 15.62, 12.82, and 7.53, respectively. This means that crop varieties with a more intensive use of labor factors will be more affected by regional differences in labor endowment in the process of production location adjustment.

Although the labor intensity of vegetables is also high, the production location adjustment is less affected by regional differences in labor endowments. One possible reason is that with rising consumption levels, the scale of vegetable market demand keeps expanding [16], especially in developed regions (where labor supply is scarce), and then offsets the impact of labor supply shortage on the location adjustment of vegetable production. On the other hand, it can be explained by market demand. The market demand index is significantly negative in the sugar and silkworm equation, but significantly positive in the vegetable equation. Generally, primary agricultural products such as sugar and silkworm usually do not enter the consumer market directly, and they provide raw materials for downstream industrial enterprises. Therefore, the location selection of such agricultural products is mainly about production costs. Moreover, there are some substitutes for silkworm and sugar, such as artificial fiber and sweeteners, which further reduces the market dependence on these agricultural products.

The estimated effects of other conventional factors are generally as expected. First, the improvement of transportation conditions promotes the location adjustment of cash crop production, but it is worth noting that the long-distance transportation efficiency of vegetables in China is still low. In 2020, the proportion of cold chain transportation of vegetables in China was about 30%, and the circulation decay rate exceeded 20 percent, with the vegetable circulation cost accounting for about half of the total cost. There are three main reasons for the above situation. One is the shortage of special technical equipment for vegetable logistics, coupled with long chains and large loss. Secondly, the lack of a professional vegetable logistics center as well as cold storage cannot meet the needs of vegetable circulation. Thirdly, due to the lag of vegetable information construction in rural areas, the links between vegetable production and circulation seem disconnected, leading to the serious product backlog. In order to meet the growing consumption demand and stabilize the original vegetable production location, the main trend of vegetable production is to expand the scale in the suburban areas. Secondly, education levels help to promote the location adjustment of cash crop production. Thirdly, the increased participation of local government guarantees the location adjustment of cotton production. Based on differentiated regional developments, great changes have taken place in the spatial distribution of productivity. Hence, it is of significance to guide the orderly transfer of the agricultural industry and promote agricultural supply-side structural reform.

5.2. Robustness Check

Changes in labor supply usually respond to market prices and, thus, affect agricultural production. On this basis, the labor price index is used to replace the labor abundance index in this section, and to estimate the regional difference of labor endowment on agricultural production location adjustment.

Table 5. The estimated results of regional differences in labor endowment on agricultural production location adjustment using labor price.

	Share of Cultivated Area
	Cotton	Oil	Sugar	Silkworm	Vegetable
Labor abundance	−0.562 ***	−0.221 **	−0.330 **	−1.112 ***	0.303 ***
	(0.210)	(0.093)	(0.155)	(0.203)	(0.061)
Agri-machinery	0.612 ***	0.749 ***	−0.148	0.233	0.229 ***
	(0.168)	(0.058)	(0.107)	(0.163)	(0.049)
Area per capita	−0.137	0.231 ***	−0.022	0.233	−0.182 ***
	(0.256)	(0.041)	(0.131)	(0.186)	(0.058)
Aging population	0.101	0.090	−0.049	−0.371 *	0.028
	(0.196)	(0.074)	(0.145)	(0.194)	(0.057)
Education level	4.196 ***	−1.707 ***	6.167 ***	−2.302 **	1.785 ***
	(0.967)	(0.344)	(0.652)	(1.077)	(0.273)
Market demand	−0.630	0.909 ***	−4.398 ***	−0.615	0.307 **
	(0.537)	(0.178)	(0.422)	(0.398)	(0.124)
Transportation	0.114 ***	−0.119 ***	0.198 ***	−0.015	0.065 ***
	(0.035)	(0.017)	(0.036)	(0.036)	(0.011)
Local government	0.682 ***	−0.196 ***	−0.001	−1.166 ***	−0.019
	(0.144)	(0.049)	(0.075)	(0.148)	(0.033)
Region fixed effect	Y	Y	Y	Y	Y
Year fixed effect	Y	Y	Y	Y	Y
R2	0.374	0.426	0.583	0.189	0.203
F-statistic	6.70 ***	12.34 ***	17.96 ***	2.83 ***	4.66 ***
Observations	444	640	503	476	700

Notes: Standard errors are listed in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the estimated results of Equation (2) using the labor price index as the key explanatory variable. It is shown that the labor price coefficients in the cotton, oil, sugar, and silkworm equation are −0.562, −0.221, −0.330, and −1.112, respectively. The rising labor price has a significant negative impact on the cultivated area share of these cash crops, that is, regions with high labor costs would reduce the production scale and then induce more of these crops to regions with lower labor costs.

As for the transfer degree, the influence of labor price on the location adjustment of cash crop production is silkworm, cotton, sugar, and oil in descending order. This means that the more intensively the crop varieties use labor input, the more deeply the regional difference of labor price affects production location adjustment, which is basically consistent with the regression results in Table 4.

However, the results are quite different for vegetables. For every 1% increase in labor price, the cultivated area share of vegetables increases by 0.303%, which is significantly positive at the 1% statistical level. This indicates that the relatively rising labor price is conducive to the expansion of the vegetable cultivated area. On the one hand, the rising labor price implies the increase in production cost, leading to a decrease in the planting willingness of farmers, and, thus, reduces the share of vegetable production. On the other hand, the rising labor price also generates income increases, which will expand the market capacity of vegetables. To sum up, the effect of labor price increases on vegetable production is influenced by these two factors. Generally, the income increase will significantly expand the market demand of vegetables, which determines that the promotion effect is greater than the inhibition effect on vegetable production. Combined with Table 4, it can be concluded that regional differences in labor endowment have little impacts on the location adjustment of vegetable production. The estimated effects of other conventional factors in Table 5 are basically consistent with Table 4.

To sum up, the regional differences in labor endowment have constituted the push–pull force of the production location adjustment of crops with low substitution elasticity, such as cotton, oil, sugar, and silkworm crops. For the outside-transferred provinces, the declining labor supply as well as rising labor prices have squeezed the profit margins of these crops, leading to a shift in crop production to areas with more profits. For the inside-transferred provinces, with the abundant labor supply, relatively low labor price, as well as the characteristics of intensive labor use for the above crops, surplus rural labor force can be used, resulting in the expansion in the crop production scale.

6. Conclusions and Discussion

This article analyzes the impact of regional differences in labor endowment on the adjustment of agricultural production location. The results are as follows:

First, the agricultural production location adjustment shows obvious characteristics of time, space, and variety order. In terms of time sequence, the overall performance of crop production has become outside-transfer with increasing scale and accelerating speed since 2005. In terms of spatial order, the Eastern region transfers outside first, the Central region transfers with a lag, while the Western region generally transfers inside. In terms of variety order, crops are basically adjusted based on the factor substitution elasticity from low to high.

Secondly, the regional differences in labor endowment have constituted the push–pull force of the production location adjustment of crops with low substitution elasticity. Moreover, the more intensively the crop varieties use the labor force, the more obviously the labor endowment affects the production location adjustment. As capital substitution cannot alleviate the impact of labor supply shortage on agricultural production, regional differences in labor endowment have already been formed.

Thirdly, regional differences in labor endowment have little impacts on the location adjustment of vegetable production. Due to the large elasticity of vegetable demand, regions with scarce labor endowments tend to have larger vegetable market capacity. The expansion of market demand helps to stabilize the original production location of vegetables, thus reducing the impact of regional differences in labor endowment.

In conclusion, the agricultural production location adjustment in China develops constantly toward the direction of regional division. Taking the spatial order and variety order into account, it is suggested to guide the orderly transfer of crop production to the most suitable regions through related policies. Meanwhile, it is suggested to develop the construction of agricultural products circulation facilities, and improve the cold chain transportation technology as well as storage capacity, etc. The distribution network of agricultural products should also be rationally distributed and optimized to build a perfect circulation system for agricultural location adjustment. On the one hand, traditional production locations with the advantages of heat, water, and other resources are facing the scarcity of the labor force. On the other hand, the agricultural production resources are relatively deficient in regions with the abundant labor forces. Therefore, how to balance the relationship between resource endowment and economic development as well as the reduction in the waste of agricultural resources is an important issue to be considered.

Furthermore, although one of the main transfer logics lies in the regional distribution pattern of the labor comparative advantage, it is worth noting that the agricultural production location adjustment among regions is in constant change. As the labor supply shortage will gradually spread from the coastal regions to the inland regions, the production locations with comparative advantages are likely to be transformed into inferior regions in the future. On this basis, the development of a production mode is key to improving the comparative advantages of crop production. On the one hand, it is suggested to speed up the development of small- and medium-sized agricultural machinery to save the labor input for cash crops. Varieties should also be improved as well as production processes simplified to make crop production more standardized and mechanical. On the other hand, the increase in unit output can improve labor productivity, which alleviates the impact of the labor price rise caused by labor shortage.