Water pricing policy is believed to be an efficient method for addressing the water scarcity problem in China. The motivation of this study is to provide a better understanding of how reducing irrigation subsidies impacts farming sectors and rural households. We apply a Computable General Equilibrium model to simulate the irrigation water and irrigation subsidy in 16 provinces. The results show that reducing irrigation has great potential for resolving the water scarcity problem in China, especially for the provinces with high subsidy levels such as Guangdong, Shandong, and Jilin. The declines in farming outputs are significant, and then their producer prices and imports increase. Rural households suffer serious losses in food consumption, particularly for those in Jilin, Guangdong, and Shandong. As for policy recommendation, improving the mobility of cropland should be given greater attention for promoting water conservation, and improving the mobility of agricultural labor could mitigate the losses in the farming outputs and in the food consumption for rural households in most provinces. Reduction in irrigation subsidy as a policy option should be adopted gradually, and then increasing the water demand elasticity should be taken into account when the irrigation water price is close to the full-cost recovery level.

Introduction

Agriculture is the largest consumer of water in China and accounts for over 70% of total water use. Water stress is already affecting agricultural production in China (Qin et al., 2013). Furthermore, low efficiency in agricultural water use has also occurred due to inadequate farming skills, low access to advanced cultivation technology, poor irrigation management practices, and a lack of investment in infrastructure (Fan et al., 2012, 2014). Therefore, agriculture is a key sector for implementing water saving strategies, and many policy tools for water resource planning have been promoted by the central government to alleviate the severe water scarcity since 2000 (Liu et al., 2015).

Many factors can achieve greater water savings and higher water use efficiency in agricultural production. In general, these factors can be classified into three categories: technological advance (e.g. Blanke et al., 2007; Fan et al., 2012, 2014), management improvement (e.g. Huang et al., 2009; Xie, 2009; Chen & Hu, 2011; Li et al., 2016; Yu et al., 2016), and policy measures (e.g. Cai, 2008; Zhong & Mol, 2010; Qin et al., 2012; Aregay et al., 2013; Shi et al., 2014; Mamitimin et al., 2015). Water pricing policy plays a critical role in agricultural water savings. One of the main reasons for this is that irrigation water prices in China are heavily subsidized, which results in farmers having no incentive to save water and improve irrigation efficiency (Han & Zhao, 2007; Jiang, 2009; Wang et al., 2009; Cremades et al., 2015; Zhong et al., 2015b). However, strong resistance still exists against the water pricing policy due to food security and rising rural income acting as two crucial components of agricultural policies (Lohmar et al., 2003; Xiong et al., 2009; Liu et al., 2013). Understanding how water pricing policy affects agricultural production and rural households in a regional context would be helpful for policy makers and other stakeholders, such as farmers and agricultural companies.

Agricultural water conservation in China is not only a technological problem but is also a socio-economic problem related to national and regional food security and to domestic and international food markets. The Computable General Equilibrium (CGE) model is a comprehensive and systematic model that is widely used to analyze socio-economic policy and thus is very suitable for this issue (Tokunaga et al., 2003; Okiyama & Tokunaga, 2010; Ge et al., 2014). The CGE model has been applied to assess the efficiency of water allocation among different farming sectors with the assumption that water resources would be regulated by a price mechanism within the process of achieving market equilibrium, given the observation that farmers would reduce the irrigated area and change cropping patterns to higher-value crops as a response to the rising water price (Yang et al., 2003; Han & Zhao, 2007; Wang et al., 2009).

By applying a CGE model embodied with multi-provincial sectors, this study aims to evaluate a possible water pricing policy by measuring the effects of reducing irrigation subsidies on farming sectors and rural household income and consumption. In addition, we will determine which factor mobility limitation has a significant influence on farming sectors and the water savings potential, considering that water pricing reform depends significantly on the degree of factor mobility (Wittwer, 2012), which is represented by the different closure conditions of factor mobility among capital, labor, and natural resource (Wang et al., 2015). Moreover, the effects of policy on irrigation water demand, as well as income and consumption of rural households, will be collected from 16 provinces, with the concern that the spatio-temporal distribution of water resources is inconsistent with socio-economic demand for water (Zhang et al., 2015).

The next section presents a brief review of irrigation water pricing policy in China. Section 3 gives a simple description of the structure of the CGE model developed by a previous study, which also represents the direction of the effects caused by reduced irrigation subsidies. Section 4 describes four simulation scenarios with no constraints on irrigation water supply at the regional level. Section 5 provides a sensitivity analysis, where a comparative evaluation is performed using results from different degrees of reduction of irrigation subsidies and different settings of water demand elasticity (WDE). Finally, Section 6 concludes with several findings and policy recommendations.

Brief review of irrigation water pricing reform in China

Several traditional approaches, such as increasing water supply and extending water conservation technologies, were not able to overcome the water scarcity problem (Lohmar et al., 2003). One of the main reasons for this is that irrigation water prices in China are heavily subsidized and correspond exclusively to the costs incurred in the construction and maintenance of conveyance and storage facilities (Han & Zhao, 2007). Since the irrigation charges are sunk costs, farmers have no incentive to save water and improve irrigation efficiency (Jiang, 2009; Wang et al., 2009; Cremades et al., 2015). Aregay et al. (2013) indicated that irrigation water demand and fertilizer consumption primarily exhibit inelastic responses to water pricing. Mamitimin et al. (2015) found that under an increased water price, more than half of 128 interviewed farmers would not opt for decisions that would lead to improved water use efficiency or improve their crop production, and fruit farmers and farmers with less land and less cash income were particularly well represented in this group. This may be the reason for the low adoption rate (<20%) of water-saving technologies, such as plastic sheeting, sprinkler systems, drip irrigation, and other efficient techniques in water-strapped northern China (Deng et al., 2006; Blanke et al., 2007; Huang et al., 2009). Actually, deteriorating irrigation systems, unreliable water delivery, and poor cost recovery are common in many irrigation districts (Xie, 2007). Cai (2008) argued that there is a ‘dead lock’ within the following inter-related conflicts: (i) farmers are wasting water while there is increasing unsatisfied demand in industrial and municipal sectors; (ii) the society needs agricultural water conservation, but farmers cannot afford the costs of water conservation due to the low profits from crop production; and (iii) the national policy of food self-sufficiency requires high crop yields, even when the government may not maintain the required water for agriculture with the growing non-agricultural water demand.

Moreover, the institutional system for water resource management has become fragmented and ineffective over the past several decades (Jiang, 2009). The fragmented institutional framework and overlapping functions of various agencies has generated the water parallel pricing system, where the main division of water resources is between irrigation water and pipe water (Shen & Liu, 2008; Xie, 2009; Cheng & Hu, 2011; Nitikin et al., 2012; Zhong et al., 2015a). The water parallel pricing system results in a serious price distortion between irrigation water and pipe water, and water distribution between agricultural sectors and non-agricultural sectors becomes inefficient (Tsur et al., 2004; Dudu & Chumi, 2008). The pipe water users, which mainly include industrial and service sectors as well as households, must pay a set price to the water company, which is based on the marginal opportunity cost (MOC) with a volumetric pricing method. In agricultural settings, irrigation charges in many rural areas are still based on the number of irrigated areas rather than the actual amount of water used for irrigation (Nitikin et al., 2012; Zhong et al., 2015b).

Therefore, public attention has shifted to water valuation and improving water use efficiency via market mechanisms (Dudu & Chumi, 2008). Water pricing as a water-demand management tool is hypothesized to improve water allocation, water-use efficiency, equity, and sustainability (Veettil et al., 2011). For instance, a pricing mechanism has been implemented swiftly nationwide to encourage the efficient use of irrigated surface water and to alleviate the financial burden of operation and maintenance in the Bayi, Hunpu, and Tiejia irrigation districts (Han & Zhao, 2007). As a result, those traditional water policies are given lower priority compared to topics related to water pricing, such as setting the right prices for water resources, water rights, and tradable permits; full-cost recovery of water services; and cost control (Zhong & Mol, 2010).

A further reform on irrigation water price is to reduce the irrigation subsidy and thus to set the price as the full-cost recovery price of irrigation water supply (Yang et al., 2003; Tsur et al., 2004; Zhong & Mol, 2010; Cheng & Hu, 2011). In 1993, in Shenzhen, Guangdong Province, China began to promote water pricing reform to construct a water pricing system in which irrigation water and pipe water are volumetrically priced together according to the MOC (Shen & Liu, 2008; Xie, 2009; Nitikin et al., 2012). However, many regions have not implemented this reform due to complicated socio-economic and environmental affairs of water use, such as unclear responsibilities, poor collection rates and institutional capacities (Zhong & Mol, 2010). In regions in which integrated pricing reform has been implemented, the water price utilized by farmers for irrigation still contains irrigation subsidies (Han & Zhao, 2007; Huang et al., 2010; Zhang & Li, 2012). We conducted a field survey in the rural area of Jilin and Liaoning provinces, which provided a basic understanding of irrigation water pricing and management. According to the several interviews of 21 farmers in this survey, the irrigation cost paid by farmers is formulated according to the size of the irrigated area (CNY mu−1, 1 ha =15 mu) and is changed in relation to the weather and cultivated crops. Indeed, this irrigation cost only reflects the variable cost of the total irrigation cost. In contrast, the fixed cost of infrastructure is supported by the local government, acting as a subsidy for farmers to lower costs. More detailed discussion about this survey can be found in Zhong et al. (2015b, 2016).

Methods

In order to assess the systemic general equilibrium effects of reducing irrigation subsidy at the multi-province level, we use a CGE model that incorporates multi-provincial irrigation water inputs, irrigation subsidies, and the water parallel pricing system. A detailed discussion of model construction, the entire set of equations, and data sources can be found in Zhong et al. (2015b). In short, this model covers 34 production sectors and 16 provinces of cropland, irrigation water, irrigation subsidy, agricultural labor, and rural households. These production sectors are divided into two categories: (i) farming sectors, which this study focuses on, includes paddy, wheat, corn, vegetable, fruit, oil seed, sugarcane, potato, sorghum, and other crops; and (ii) other sectors, which include the non-farming agricultural, industrial, and service sectors. The selected 16 provinces are Guangdong, Jiangxi, Hainan, Yunnan, Guangxi, Henan, Jilin, Anhui, Heilongjiang, Hebei, Hubei, Chongqing, Sichuan, Inner Mongolia, Shandong, and an ‘Other provinces’ category covering the remaining provinces. The discussion involving irrigation water use in these 16 provinces can be found in Zhong et al. (2016). There are five factors in the model, which utilizes a multi-level nested constant elasticity of substitution production function. The agricultural labor of the 16 provinces are employed in both farming and non-farming agricultural sectors but not in industrial or service sectors, and non-agricultural labor and capital are employed by all production sectors. Cropland and irrigation water in the 16 provinces are the only inputs for the farming sector, and the provincial cropland area is combined with irrigation water use for each province to produce the composite land-water inputs at each multi-provincial level. All factor inputs that generate the farming outputs are based on the Cobb-Douglas assumption. Following a small-country assumption, farming outputs are separated into domestic and export products through a constant elasticity of transformation function. The domestic food consumption of the rural households in the 16 provinces is composed of domestic and import products with the Armington assumption. The consumption of rural households is determined by their income, which comes from the factor returns including the returns from capital and cropland as well as the wages from non-agricultural labor and agricultural labor. The irrigation water in the 16 provinces is regulated by the government, and thus the government collects the total payment of irrigation supply and irrigation subsidy. Figure 1 depicts the directions of the possible effects induced from the change in irrigation subsidy for other sectors such as farming sectors and rural households. Figure 1 also shows the set-up of the water parallel pricing system within the dotted box.
Fig. 1.

Paths of influence from irrigation subsidy to other sectors.

Fig. 1.

Paths of influence from irrigation subsidy to other sectors.

The water parallel pricing system is defined in the model framework, which refers to the study of Shi et al. (2014). We assume that the MOC level of irrigation water is equal to that of pipe water. The supply curve of irrigation water is vertical to represent the government's direct regulation. Because of irrigation subsidy, the actual price of irrigation water is lower than the pipe water price, and the price of irrigation water is represented as the full-cost recovery level of irrigation. The irrigation subsidy rate imposed on the irrigation water input is recognized as the difference between the irrigation water price and pipe water price, which assumes that water quality is the same for both irrigation water and pipe water. This assumption is acceptable because of the following two reasons. First, the difference between the actual prices of irrigation water and pipe water is obviously significant due to the remarkable diversities in supply costs and demand factors. Hence, we do not propose that irrigation water and pipe water should be priced at the same level; instead, we support that both of their prices should be determined by the volumetric pricing method according to the MOC level. The setting of irrigation subsidies is a convenient calculation given the data limitation of information about regional water quality of sectoral irrigation inputs. Secondly, according to findings from previous studies, impacts caused by reducing irrigation subsidies would significantly raise the price of irrigation water input, and then the production costs of farming sectors would be increased. Thus, farming sectors would have to reduce demand for irrigation water and decrease their output, and thus the prices of farming products would be higher, farming exports would be lower and farming imports would be higher. We used a CGE model to simulate this process, and aimed to investigate the relative changes in the given variables. Water price variance itself due to water quality might insignificantly affect the values of simulation results, but the relative changes and their trends in different scenarios would not be affected.

The detailed results of the irrigation subsidy rates for the 16 provinces can be found in Zhong et al. (2015b). The irrigation subsidy rates range from 0.48 (Henan) to 0.97 (Guangdong), which means that each unit of irrigation water price covers a percentage of the subsidy varying from 48% to 98% across the 16 provinces. The 16 provincial irrigation water inputs and their irrigation subsidies are introduced into the social accounting matrix (SAM), which usually serves as the dataset and simulation benchmark in the CGE model (Zhong & Tokunaga, 2014). A detailed discussion of the SAM and other parameter values are provided in Zhong et al. (2014, 2015b). Our approach differed in that we set the value of substitution elasticity between irrigation water input and cropland input in each province, which also acts as the WDE. We set this parameter equal to 0.2 to account for the fact that irrigation water demand in China is inelastic to water pricing policy (e.g. Aregay et al., 2013; Mamitimin et al., 2015).

Simulation

Simulation design and closure condition

This study focuses on the effects on farming sectors, and the changes in other sectors are not discussed. Simulation with multiple scenarios is carried out to analyze how reducing the irrigation subsidy by 5% in the selected 16 provinces impacts farming sectors and the rural households. According to our surveys of literature and the farmer interviews, there is no agreement on how much an irrigation subsidy should be reduced due to the complex situation of regional disparities, even though many studies support this policy. And also many regions did not fully implement water pricing reform. Therefore, we consider a reduction of 5% as a minor change in irrigation subsidies, which can be a policy experiment and may be accepted by local governments and farmers.

In this model, we fix the non-subsidized irrigation water price at the full-cost level to match the current pricing policy in most provinces, which means that reducing the irrigation subsidy would raise the actual irrigation water price. Meanwhile, regional irrigation supply is set equal to the sum of sectoral irrigation water inputs. Regional irrigation supply is initially fixed to derive the regional irrigation water price and is then set to be a flexible endogenous variable to represent the regional water demand. Thus, we can investigate the capacity of water conservation by examining the decline in water demand. Moreover, although we only consider one simulation in this study, four scenarios were designed with varying settings of factor mobility, which are represented as four closure conditions within this simulation. The goal of these scenarios is to determine, when reduced subsidies are implemented, the improvement for which factor mobility would contribute to greater water savings with less losses in farming outputs and limit the negative effects on rural household income and consumption. These settings are shown in Table 1, where the first column of Table 1 lists the factor inputs in farming sectors of the 16 provinces, including irrigation water, crop land, agricultural labor, non-agricultural labor and capital. The second column indicates the closure condition setting of factor mobility, defined as two conditions of ‘Mobile’ and ‘Fixed’ regarding whether factor inputs can freely flow across among ten farming production sectors. Indeed, ‘Mobile’ means that factor mobility is able to move from one sectoral production process to another sectoral production process under the assumption of laissez-faire; in contrast, ‘Fixed’ means that sectoral factor inputs are fixed, and thus the simulation would make factor returns diverge across different sectors. Accordingly, we conduct a comparative analysis between Scenario 1 (S1), Scenario 2 (S2), Scenario 3 (S3), and Scenario 4 (S4) regarding the assumption of improving factor mobility. For example, S1 displays the strictest situation, where the cropland and agricultural labor used to cultivate one crop cannot be used to cultivate any other crops. S2 and S3 set cropland and agricultural labor as constants, respectively. S4 is designed to test ‘free-will’ factor mobility, which is assumed as the first rationale to study market efficiency, and so change per unit of exchange cost in the free market is assumed to be a close approximation of reality. Indeed, the economic assumption of ‘free-will’ reflects a fact that the changes per unit of exchange cost act as the driving forces making the producers change their input strategies with concern for different factor costs, but various constraints would affect this process in some case, such as imperfect market mechanism and/or policy failures, which is beyond the scope of this study. The main reason for designing the other three scenarios (S1, S2 and S3) against the ‘free-will’ scenario (S4) is to provide a comparative analysis as a further discussion.

Table 1.

Closure conditions of factor mobility in five scenarios.

Farming factors Closure conditions Scenario 1 (S1) Scenario 2 (S2) Scenario 3 (S3) Scenario 4 (S4) 
16 provinces' irrigation water Mobile √ √ √ √ 
Pipe water Mobile √ √ √ √ 
16 provinces' crop land Mobile   √ √ 
Fixed √ √   
16 provinces' agricultural labor Mobile  √  √ 
Fixed √  √  
Non-agricultural labor Mobile √ √ √ √ 
Capital Mobile √ √ √ √ 
Farming factors Closure conditions Scenario 1 (S1) Scenario 2 (S2) Scenario 3 (S3) Scenario 4 (S4) 
16 provinces' irrigation water Mobile √ √ √ √ 
Pipe water Mobile √ √ √ √ 
16 provinces' crop land Mobile   √ √ 
Fixed √ √   
16 provinces' agricultural labor Mobile  √  √ 
Fixed √  √  
Non-agricultural labor Mobile √ √ √ √ 
Capital Mobile √ √ √ √ 

Results and analysis

Our results regarding changes in irrigation water demand at the multi-provincial level prove that the water pricing policy of reducing irrigation subsidy has great potential for water conservation in all provinces. Figure 2 shows that when the irrigation subsidy is reduced by 5%, the decline of irrigation water demand in the 16 provinces would be significant and various. Higher levels of initial irrigation water subsidy were markedly related to greater levels of water conservation. The most significant decline in water demand was found in Guangdong and Shandong. Jilin, Jiangxi, and Heilongjiang exhibited an intermediate level of decline in water demand. The provinces with the lowest declines in irrigation water demand are Henan and Chongqing. The results derived from different scenarios varied notably. We found that for almost all provinces, with Henan as the only exception, the reductions in irrigation water demand in S3 and S4 were marginally more significant than those in S1 and S2. This implies that improving the mobility of cropland would contribute to water conservation. However, a similar finding cannot be fully guaranteed for improving the mobility of agricultural labor when we compare S3 and S4.
Fig. 2.

Changes in irrigation water demand.

Fig. 2.

Changes in irrigation water demand.

Gross domestic outputs with real value (real gross domestic product [GDP]), consumer price index, and exchange rate, which are the main macroeconomic indicators used in this study, are precursors to the scale effect. The results indicate that reducing the irrigation subsidy by 5% does not have significant effects on the macro-economy. For the farming sector as a whole, however, the effects are significant and vary between scenarios. S3 is the worst case in terms of the farming sector effects (decreased by 0.247%) and the best case is S2 (decreased by 0.169%), and thus the highest increase in producer price was in S3 (increased by 0.583%) and the lowest one was in S2 (increased by 0.343%). Therefore, improving the mobility of agricultural labor would mitigate losses in farming output and limit the rise in producer price, but improving mobility of cropland would result in opposite effects. S4 exhibits the smallest increase in farming imports among all scenarios, and thus improving the mobility of both cropland and agricultural labor would produce a better outcome given the food self-sufficiency policy. Moreover, the income sources of rural households employed in farming sectors also changed significantly: the cropland returns decreased, and the agricultural labor wages and the capital returns increased. The results given by the different scenarios indicate that improving the mobility of cropland would narrow the loss in cropland returns from approximately 24% (S1: −24.128% and S2: −24.661%) to approximately 17% (S3: −16.972% and S4: −17.722%), but improving the mobility of agricultural labor would reduce the increase in agricultural labor wages from approximately 0.25% (S1 and S3) to 0.06% (S2) or 0.09% (S4). Furthermore, S2 has the lowest additional return of capital derived among all the scenarios (Table 2).

Table 2.

Changes in macro indices, farming sector, and factor returns.

Unit: % Reducing irrigation subsidy by 5%
 
Scenario 1 (S1) Scenario 2 (S2) Scenario 3 (S3) Scenario 4 (S4) 
Macro indices 
 Real GDP −0.006 −0.007 −0.005 −0.007 
 Consumer price index 0.054 0.047 0.066 0.060 
 Exchange rate 0.015 0.014 0.016 0.018 
Farming sector as a whole 
 Output −0.213 −0.169 −0.247 −0.192 
 Producer price 0.529 0.343 0.583 0.433 
 Import 0.062 0.075 0.054 0.045 
Factor returns 
 Cropland −24.128a −24.661a −16.972b −17.722b 
 Agricultural labor 0.258c 0.063b 0.266c 0.092b 
 Capital 0.015 0.013 0.019 0.017 
Unit: % Reducing irrigation subsidy by 5%
 
Scenario 1 (S1) Scenario 2 (S2) Scenario 3 (S3) Scenario 4 (S4) 
Macro indices 
 Real GDP −0.006 −0.007 −0.005 −0.007 
 Consumer price index 0.054 0.047 0.066 0.060 
 Exchange rate 0.015 0.014 0.016 0.018 
Farming sector as a whole 
 Output −0.213 −0.169 −0.247 −0.192 
 Producer price 0.529 0.343 0.583 0.433 
 Import 0.062 0.075 0.054 0.045 
Factor returns 
 Cropland −24.128a −24.661a −16.972b −17.722b 
 Agricultural labor 0.258c 0.063b 0.266c 0.092b 
 Capital 0.015 0.013 0.019 0.017 

aIndicates the average value across farming sectors and 16 provinces.

bSignifies the average value across 16 provinces.

cIndicates the average value across agricultural sectors and 16 provinces; the minus values indicate the reduction level.

It is quite clear that farmers would respond to the rising water prices via a modification in their cropping patterns. Consequently, the changes in the outputs, producer prices, and imports of different farming sectors adopt different values, and these changes obviously depend on the mobility of cropland and agricultural labor. For example, sorghum has the most severe loss in output in all scenarios, but its producer price and import level only exhibits the highest increase in S2 and S4. This implies that the mobility of agricultural labor plays an important role in preventing losses in output for some farming sectors, including paddy, wheat, corn, and fruit. Therefore, the increases in the producer prices and imports of these sectors are more limited than the increases in other farming sectors. For instance, paddy farming suffers the most significant loss of output in S3 with decreases of 1.475%, and then its producer price and import increase by 1.620% and 7.020%, respectively. In S4, the decrease in paddy output is limited to 0.839%, and thus the increases in its producer price and import are limited to 0.913% and 3.860%, respectively. When considering the improvement of mobility of cropland, however, the situation becomes different. A comparison between the changes in S2 and S4 shows that only oil seed, sugarcane, and potato are improved with fewer losses of outputs and lower rises in their producer prices and imports (see Figure 3).
Fig. 3.

Changes in the outputs, producer prices, and imports of different farming sectors.

Fig. 3.

Changes in the outputs, producer prices, and imports of different farming sectors.

Figure 4 shows that the change in cropping structure improves the income of some rural households based on the increasing factor returns including those from agricultural labor and capital, and the highest increase occurs in Yunnan's rural households. Rural households in some provinces suffer losses to their income, such as in Guangdong, Shandong, Hainan, Jiangxi, and Guangxi. The worst declines in food consumption of rural households are found in Jilin, Guangdong, and Shandong. We also observed that the improvements in the mobility of both cropland and agricultural labor do not play a significant role in income changes for rural households, as the differences among the four scenarios only ranged from −0.03 to 0.07. Furthermore, the assumption of mobile agricultural labor results in reduced declines in food consumption for rural households in all of the provinces except for Heilongjiang. However, improving cropland mobility is generally not helpful in alleviating the losses in food consumption for most rural households.
Fig. 4.

Changes in rural households' income and food consumption.

Fig. 4.

Changes in rural households' income and food consumption.

Sensitivity analysis

For the sensitivity analysis, we design three cases with different values of WDE, i.e. 0.2, 0.5, and 0.8, respectively. Furthermore, we consider several levels of irrigation subsidy reductions (i.e. 10%, 30%, 50%, and 80%). The total number of scenarios considered in this section is 12 (= 4 reduction degrees × 3 settings of WDE). These cases are based on the ‘free-will’ factor mobility of S4, which acts as a base scenario for the first efficient market solution. Overall, the sensitivity analysis shows that the empirical results are robust under different cases. It is worth noting that reducing irrigation subsidies should be gradual, and so it would be unpractical to set excessive reductions without any other improvements in technology and management and compensation to rural households. That is why this simulation is only considered a sensitivity analysis to test the robustness of the result and to measure the effects of changes in WDE as a further policy consideration. If the results could demonstrate that higher WDE would contribute more to water savings, the next research step would be to discover which specific improvements lead to rising WDE.

Figure 5 shows changes in irrigation water demand caused by different levels of irrigation subsidy reduction and different settings of WDE. We found that the irrigation water demand in all provinces decreases with increasing reductions in irrigation subsidy. Moreover, setting a higher WDE value would result in a more significant decline in irrigation water demand because of a higher water price. For different settings of WDE, the relative ranks of each province are very similar. Guangdong and Shandong are still the provinces with the lowest declines in irrigation water demand, followed by Jiangxi and Jilin. Other provinces, Heilongjiang, and Hainan experience slight changes in their relative ordering. Yunnan, Inner Mongolia, Guangxi, Hebei, Hubei, and Anhui had intermediate declines and retained their relative ranks. The fewest declines in irrigation water demand are still found in Henan, Chongqing, and Sichuan.
Fig. 5.

Changes in irrigation water demand under different cases of sensitivity analysis.

Fig. 5.

Changes in irrigation water demand under different cases of sensitivity analysis.

More remarkable decreases in irrigation water demand caused by a higher irrigation water price reduce the farming output and thereby increase the producer price and imports. For all cases, the relative ranks for each crop are unchanged, and thus the most significant output losses and producer price and import increases are still from sorghum. Under different levels of irrigation subsidy reduction, however, the same WDE setting results in different trends. When the value of reduced irrigation subsidy is lower than 50%, higher WDE settings result in decreased farming outputs, and higher increases in producer prices and imports. When the reduced irrigation subsidy is more than 50%, higher WDEs result in mitigated output losses for the farming sectors, and thus the increases in their producer prices and imports are limited to a lower level (see Figure 6).
Fig. 6.

Changes in farming sectors under different cases of sensitivity analysis.

Fig. 6.

Changes in farming sectors under different cases of sensitivity analysis.

Figure 7 shows that greater reductions in the irrigation subsidy result in higher incomes for rural households, and higher WDEs result in higher income increases. The relative ranks of changes in rural household income for different cases are similar, and the rural households in Yunnan and Chongqing still exhibit the highest improvements in income. On the food consumption side, the relative ranks of changes in food consumption for different cases are also similar, and thus the rural households in Jiangxi, Hainan, and Guangxi still suffer the most severe losses in food consumption. It is interesting that when the reduced irrigation subsidy is lower than 30%, higher WDE results in decreased rural household food consumption, but when the reduced irrigation subsidy is higher than 30%, higher WDE levels increasingly mitigate losses in food consumption.
Fig. 7.

Changes in the situation of rural households under different cases of the sensitivity analysis.

Fig. 7.

Changes in the situation of rural households under different cases of the sensitivity analysis.

Conclusion, policy recommendation and future research needs

This study utilizes a CGE model to examine the changes in water conservation capacity that result from reducing irrigation subsidy by 5%. The simulation results showed that reducing irrigation has significant potential for resolving the water scarcity problem in China, as the current cost of water is highly subsidized and far below the full-cost recovery level of water in the 16 provinces. Guangdong, Shandong, and Jilin were the three provinces with the highest capacity for water conservation, and these were followed by the provinces of Jilin, Jiangxi, and Heilongjiang. In contrast, Henan and Chongqing were recognized as the provinces with lowest capacity for water conservation.

The results derived from the different scenarios of factor mobility show that improving the mobility of cropland would contribute to water conservation in all provinces, but improving mobility of agricultural labor does not make the same contribution. Significant losses were found in all farming sectors. However, we also found that improving the mobility of agricultural labor would mitigate losses in total farming output and limit the rise in producer price, but improving the mobility of cropland would produce the opposite results. In detail, the mobility of agricultural labor helps prevent output losses of paddy, wheat, corn, and fruit, and thus the increases in their producer prices and imports are limited. On the other hand, improving the mobility of cropland results in superior outcomes for oil seed, sugarcane, and potato with lower output losses and reductions in the increase of their producer prices and imports.

The scenarios indicate that many provincial rural households would benefit from higher income, especially those in Yunnan, but the rural households in Guangdong, Shandong, Hainan, Jiangxi, and Guangxi would suffer losses in their income. According to the comparative evaluation for the four scenarios of factor mobility conditions, it was found that improving the mobility of agricultural labor would mitigate the negative effects for the rural households with decreasing income. However, a significant loss in food consumption was found for rural households in all 16 provinces, especially for those in Jilin, Guangdong, and Shandong. Moreover, taking agricultural labor as mobile lowered the decline in food consumption for rural households in nearly all of provinces. In contrast, making cropland mobile does not produce this effect.

Sensitivity analysis via considering a series of modifications to WDE showed that the simulation results were robust. Setting a higher value for WDE would improve the capacity of water conservation, and the relative provincial ranks for different levels of reducing irrigation subsidy were similar. It is interesting that in the context of increasing reductions in irrigation subsidy, the same setting of WDE resulted in different trends for farming sectors. When the reduction in irrigation subsidy was lower than 50%, higher WDE would worsen the decline in farming outputs, and thus their producer prices and imports were more significantly increased. In contrast, when irrigation subsidy was reduced by more than 50%, higher WDE would reduce the losses in farming outputs and lower the rise in their producer prices and imports.

Several policy recommendations can be proposed according to the simulation results. First, improving the mobility of cropland is a functional option to increase water conservation; improving mobility of agricultural labor should be considered as an option to reduce food consumption losses for rural households in most provinces. Additionally, reducing irrigation subsidy as an option of water pricing policy should be gradually implemented until the irrigation cost is more in line with the full-cost recovery level, and additional policies for preventing the significant losses in farming outputs should be incorporated into this schedule. Finally, when the irrigation cost is close to the full-cost recovery level of irrigation water price, policy efforts should be made to increase WDE, such as expanding new irrigation techniques; developing alternate sources such as rainfall harvest by small reservoirs, groundwater, and reuse of return flow; as well as crop pattern change.

This research has several limitations that need to be addressed in future studies. One of them is the model assumptions, which provided a perfect market solution for the situation of reducing irrigation subsidy, and thus the potential of water conversation is likely excessively optimistic. The actual potential might be limited by several conditions, such as the food security policy, factors hindering market mechanisms, and imbalanced economic development along with uneven distribution of water resources at the multi-provincial level. Therefore, this study can be further improved by a more realistic development of model assumptions and parameters, which requires additional field surveys and data collection. Additional field surveys and in-depth data are still needed for future studies to address specific problems occurring in water pricing reform, such as ineffective water management and barriers to water-saving technology, which will be considered during the next model extension to provide new findings for policy decisions.

It should be noted that the objective of this study is not to examine efficiency of water use, which could be completed by a Data Envelopment Analysis (DEA) or other methods. The basic purpose of the CGE model is to provide an ‘evaluation’ with an emphasis on market equilibrium, especially on pricing issues. In a CGE model, equilibrium occurs at a set of prices at which all supply and demand are balanced, where producers have maximized profits by setting the optimal input and output levels and consumers have maximized utility by purchasing the most satisfying bundle of products. It is assumed that nothing will change without external influences, such as drought, additional investment or a water pricing policy like a reduced irrigation subsidy as in this study. As a result, the CGE model is widely used for policy evaluation, with market mechanisms playing a critical role within the simulation. A DEA can be seen as a partial equilibrium model, which aims to find an optimal ‘solution’ at given input or output levels or at other given objectives. Therefore, in the next step of this study when we could properly set some specific objectives of the water policy, such as food security or a water-savings plan, we would seek to employ a DEA to investigate an optimal solution to achieve these objectives and also to alleviate the damage derived from the CGE analysis as shown in this study.

Acknowledgements

The authors appreciate the extensive support of Professor Suminori Tokunaga and Professor Mitsuru Okiyama from Reitaku University, Chiba-ken, Japan, as well as Professor Jinghua Sha and Professor Jingjing Yan from China University of Geosciences, Beijing, China. The authors appreciate the support of three anonymous reviewers and associate editors for their valuable comments and suggestions for improving the research. This work was financially supported by the National Natural Science Foundation of China (Grant No. 41271547, No. 41501604 and No. 41271546), and the China Postdoctoral Science Foundation (Grant No. 2015M571109). The authors thank Accdon for its linguistic assistance during the preparation of this manuscript.

References

References
Blanke
A.
Rozelle
S.
Lohmar
B.
Wang
J.
Huang
J.
(
2007
).
Water saving technology and saving water in China
.
Agricultural Water Management
87
(
2
),
139
150
.
Cremades
R.
Wang
J.
Morris
J.
(
2015
).
Economic incentives and the adoption of modern irrigation technology in China
.
Earth System Dynamics
6
,
399
410
.
Deng
X. P.
Shan
L.
Zhang
H.
Turner
N. C.
(
2006
).
Improving agricultural water use efficiency in arid and semiarid areas of China
.
Agricultural Water Management
80
(
1–3
),
23
40
.
Dudu
H.
Chumi
S.
(
2008
).
Economics of irrigation water management: A literature survey with focus on partial and general equilibrium models
.
Policy Research Working Paper, Series No. WPS4556
.
The World Bank
,
Washington, DC, USA
. .
Fan
M.
Shen
J.
Yuan
L.
Jiang
R.
Chen
X.
Davies
W. J.
Zhang
F.
(
2012
).
Improving crop productivity and resource use efficiency to ensure food security and environmental quality in China
.
Journal of Experimental Botany
63
(
1
),
13
24
.
Huang
Q.
Rozelle
S.
Wang
J.
Huang
J.
(
2009
).
Water management institutional reform: a representative look at northern China
.
Agricultural Water Management
96
(
2
),
215
225
.
Huang
Q. Q.
Rozelle
S.
Howitt
R.
Wang
J. X.
Huang
J. K.
(
2010
).
Irrigation water demand and implications for water pricing policy in rural China
.
Environment and Development Economics
15
,
293
319
.
Jiang
Y.
(
2009
).
China's water scarcity
.
Journal of Environmental Management
90
(
11
),
3185
3196
.
Liu
J. G.
Zang
C. F.
Tian
S. Y.
Liu
J. G.
Yang
H.
Jia
S. F.
You
L. Z.
Liu
B.
Zhang
M.
(
2013
).
Water conservancy projects in China: achievements, challenges and way forward
.
Global Environment Change
23
,
633
643
.
Lohmar
B.
Wang
J.
Rozelle
S.
Huang
J.
Dawe
D.
(
2003
).
China's Agricultural Water Policy Reforms: Increasing Investment, Resolving Conflicts, and Revising Incentives
.
Agriculture Information Bulletin No. 782
.
Economic Research Service, US Department of Agriculture
,
Washington, DC
,
USA
.
Mamitimin
Y.
Feike
T.
Seifert
I.
Doluschitz
R.
(
2015
).
Irrigation in the Tarim Basin, China: farmers’ response to changes in water pricing practices
.
Environmental Earth Sciences
73
(
2
),
559
569
.
Nitikin
D.
Shen
C. L.
Wang
J.
Zou
H. F.
(
2012
).
Water service delivery reform in China: safeguarding the interests of the poor
.
Annals of Economics and Finance
13
,
463
487
.
Qin
C. B.
Jia
Y. W.
Su
Z.
Bressers
H. T. A.
Wang
H.
(
2012
).
The economic impact of water tax charges in China: a static computable general equilibrium analysis
.
Water International
37
(
3
),
279
292
.
Tsur
Y.
Roe
T.
Doukkali
R.
Dinar
A.
(
2004
).
Pricing Irrigation Water: Principles and Cases From Developing Countries
.
Resources for the Future
,
Washington, DC
,
USA
.
Veettil
P. C.
Speelman
S.
Frija
A.
Buysse
J.
Mondelaers
K.
van Huylenbroeck
G.
(
2011
).
Price sensitivity of farmer preferences for irrigation water-pricing method: evidence from a choice model analysis in Krishna River Basin, India
.
Journal of Water Resources Planning and Management
137
(
2
),
205
214
.
Wang
J. X.
Huang
J. K.
Rozelle
S.
Huang
Q. Q.
Zhang
L. J.
(
2009
).
Understanding the water crisis in Northern China: what the government and farmers are doing
.
International Journal of Water Resources Development
25
(
1
),
141
158
.
Wang
Z.
Deng
X.
Li
X.
Zhou
Q.
Yan
H.
(
2015
).
Impact analysis of government investment on water projects in the arid Gansu Province of China
.
Physics and Chemistry of the Earth
79–82
,
54
66
.
Wittwer
G.
(
2012
).
Economic Modeling of Water: The Australian CGE Experience
.
Springer
,
Dordrecht, Heidelberg, New York
,
London
.
Xie
M.
(
2007
).
Global development of farmer water user associations
. In:
Proceedings of the Regional Workshop on WUAs Development: Water Users’ Associations Development in Southeastern European Countries: Bucharest, Romania
.
Hussain
I.
Naseer
Z.
(eds).
Xie
J.
(
2009
).
Addressing China's water scarcity: recommendations for selected water resource management issues. World Bank-free PDF, No. 47111
.
The World Bank
,
Washington, DC, USA
. .
Xiong
W.
Conway
D.
Lin
E.
Xu
Y. L.
Ju
H.
Jiang
J. H.
Holman
I.
Li
Y.
(
2009
).
Future cereal production in China: the interaction of climate change, water availability and socio-economic scenarios
.
Global Environment Change
19
,
34
44
.
Yang
H.
Zhang
X.
Zehnder
A. J. B.
(
2003
).
Water scarcity, pricing mechanism and institutional reform in northern China irrigated agriculture
.
Agricultural Water Management
61
(
2
),
143
161
.
Yu
G. M.
Yang
Y. M.
Tu
Z. F.
Jie
Y.
Yu
Q. W.
Hu
X. Y.
Yu
H. L.
Zhou
R. R.
Chen
X. X.
Wang
H. Z.
(
2016
).
Modeling the water-satisfied degree for production of the main food crops in China
.
Science of the Total Environment
547
,
215
225
.
Zhang
X.
Li
H.
(
2012
).
Water cost evaluation and agricultural production policy recommendations of northwest of China
.
Research of Agricultural Modernization
33
,
82
85
.
Zhong
L.
Mol
A. P.
(
2010
).
Water price reforms in China: policy-making and implementation
.
Water Resources Management
24
(
2
),
377
396
.
Zhong
S.
Sha
J. H.
Shen
L.
Okiyama
M.
Tokunaga
S.
Yan
J. J.
Liu
L. T.
(
2016
).
Measuring drought based on a CGE Model with multi-regional irrigation water
.
Water Policy
18
(
4
),
877
891
.