1. Introduction
According to the IPCC Sixth Assessment Report (AR6), global surface temperature has increased faster since 1970 than in any other 50-year period over a least the last 2000 years, leading to an increase in the intensity and frequency of extreme weather events, which have severe harmful effects on agricultural production. In the agricultural sector, the overuse of fertilizer has been found to be a significant contributor to global warming [
1,
2,
3]. In addition, the overuse of fertilizer has caused a series of soil problems, such as non-point source pollution and soil nutrient imbalance, hindering the sustainable development of agriculture [
4]. Therefore, it is necessary to examine the factors affecting farmers’ fertilization behavior and then design corresponding policies to reduce fertilizer use.
In China, farmers’ fertilizer application rates are far higher than the internationally accepted safety limit of 225 kg/ha [
5,
6]. The Chinese government has introduced a series of measures to reduce farmers’ fertilizer application rates. For example, in 2015, the Chinese Ministry of Agriculture introduced the “Action to Achieve Zero Growth of Chemical Fertilizer Use by 2020”. Since then, the “No. 1 Document” of the central government has repeatedly emphasized the issue of fertilizer reduction for six consecutive years. However, the National Farm Product Cost-benefit Survey Data show that fertilizer application rates of different crops continued to increase slowly until 2020 (see
Table A1 in the
Appendix A), which means that considerable reductions in fertilizer use still need to be made.
Many studies have examined the factors influencing farmers’ fertilization behavior in China. First, production incentive policies and fertilizer price control policies formulated by the Chinese government have resulted in a high ratio of grain to fertilizer prices, which leads to deviations in the marginal output from the actual price of fertilizer [
7,
8,
9,
10]. Second, the scarcity of arable land, migration of the agricultural labor force, and rise of labor costs have led to the substitution of fertilizer for the above factors [
11,
12]. Third, characteristics of farmers, such as small-scale operations [
10,
13], high-volume fertilization habits [
14], and risk aversion [
15], have led to the high fertilizer application rates. Fourth, information asymmetry, such as fertilizer product updates that exceed farmers’ ability to assimilate the information [
16], fertilizer nutrient contents on packaging that are higher than the actual contents [
17], and asymmetry in soil quality information [
18], is another reason for the high fertilizer application rates by Chinese farmers.
However, studies on the determinants of fertilizer used by Chinese farmers have ignored the risk factors related to heavy rainfall, including fertilizer losses and the possible yield risk [
10,
14,
15,
16]. Heavy rainfall in China is mainly concentrated in the summer when crops are growing and their demand for fertilizer is high. From the time it is applied into soil to its uptake by the crop, fertilizer can easily be washed away by heavy rainfall, resulting in leakage and losses, which greatly reduces fertilizer use efficiency [
19]. Meanwhile, heavy rainfall may cause waterlogging in farmland (affecting the respiration of crop roots) and a high-humidity environment (more likely to lead to the occurrence of pests and diseases), thereby causing a crop yield reduction [
20]. Increasing fertilizer use can supplement the loss of fertilizer and reduce the adverse impact of heavy rainfall on crop yield. Furthermore, heavy rainfall patterns are uncertain, and the interannual variability in heavy rainfall may cause an interannual change in yield. Therefore, farmers cannot accurately predict rainfall levels for the year when applying fertilizer, and farmers in areas with high rainfall variability may apply more fertilizer to avoid yield risk (ensuring stable income).
Due to the synergistic relationship between soil nutrient uptake and water effectiveness, previous studies have mainly focused on examining the effect of rainfall on fertilizer use and found that the relationship is uncertain, depending on the level of economic development or agricultural infrastructure in different regions [
21,
22,
23,
24,
25,
26]. Only a few studies have examined farmers’ fertilization behavior under rainfall variability and found that rainfall abundance in the previous year increases fertilizer application rates by increasing yield and relaxing the liquidity limits of farmers, while rainfall variability decreases fertilizer application rates because it increases the possibility of harvest failure [
27,
28,
29].
However, the findings of these studies may not be applicable to China. First, these studies mainly focus on promoting farmers’ fertilizer application rates in semi-arid and rain-fed areas with high interannual rainfall variabilities, such as Ethiopia and Tanzania, where fertilizer application rates are low and economic conditions are poor (farmers face high liquidity constraints) [
27,
28,
29]. Given the high fertilizer application rates in China, which is different from these countries, the Chinese government mainly focuses on reducing farmers’ fertilizer application rates while maintaining yield [
5,
14]. Second, previous studies do not distinguish between rainfall intensity, which is problematic because normal rainfall and heavy rainfall have different effects on crop production (normal rainfall can improve yield, while heavy rainfall is unfavorable to production). This study defines daily rainfall exceeding 25 mm as heavy rainfall and examines the effect of heavy rainfall variability on farmers’ fertilizer use in China and its possible mechanisms.
Given that, as food for humans and livestock, maize has the largest sown area in China (requiring a large amount of fertilizer), its fertilizer application rates rank second among the three major crops (rice, wheat, and maize) (see the China agricultural products cost-benefit compilation of information). Reducing fertilizer use for maize plays an important role in decreasing fertilizer application rates in China. Meanwhile, as maize grows in summer with its higher probability of heavy rainfall, heavy rainfall may have a greater effect on maize farmers’ fertilizer use. Therefore, this study uses maize farmers to examine the impact of heavy rainfall variability on fertilizer application rates.
This study makes the following contributions. First, this is the first study to empirically examine the impact of rainfall variability on maize farmers’ fertilizer application rates in China, which is conducive to providing a supplementary explanation for the high fertilizer application rates by Chinese farmers. Second, this study adopts heavy rainfall variability as the proxy variable of income risk and examines its impact on farmers’ fertilizer application rates, which has important reference value for the design of fertilizer reduction policies. Third, in light of similar existing literature, this study uses panel data (which are the largest longitudinal household survey in rural China, covering more than 350 villages every year) and daily rainfall data (which can accurately identify rainfall intensity) to explicitly examine farmers’ responses to rainfall variability.
The remainder of this study is organized as follows.
Section 2 discusses the theoretical framework.
Section 3 describes the data and methodology used.
Section 4 presents the empirical results and robustness checks.
Section 5 presents further discussion on potential channels. Conclusions are presented in
Section 6.
2. Theoretical Analysis and Hypothesis Proposal
2.1. Risk Characteristics of Fertilizer
Previous studies have generally defined the input that leads to greater or lower yield variance as “risk-increasing” or “risk-reducing” inputs, respectively [
30,
31]. Many studies have found that fertilizer can increase yield variance, which is a “risk-increasing” input, and risk-averse farmers should theoretically apply less fertilizer under risk [
30,
32,
33]. However, they prefer to apply more fertilizer in reality [
15,
34].
To explain the divergence between the reality and theory, two issues need to be clarified. First, do farmers apply more fertilizer to reduce variation or loss? Farmers do not avoid upward fluctuations in yield and income, and mainly avoid downward fluctuations. If applying more fertilizer can reduce the possibility of yield reduction, farmers have an incentive to apply more fertilizer. Second, is risk the cause or consequence of applying more fertilizer? If yield and income variability are thought to be caused mainly by the high fertilizer application rates, then risk-averse farmers will theoretically apply less fertilizer. However, owing to the uncertainty of agricultural production, external risk should be a more important factor affecting farmers’ fertilization behavior. Moreover, previous studies have found that farmers seem to have a different understanding of fertilizer from economists. Farmers generally argue that fertilizer does not increase risk, but instead they use more fertilizer as a means to reduce risk [
34,
35,
36,
37,
38]. Therefore, under risk conditions, farmers are inclined to apply more fertilizer to prevent low yield and income [
32,
36,
39].
2.2. Heavy Rainfall Variability and Risk-Averse Farmers’ Fertilizer Use
Agricultural production and management are risky, and farmers are risk-averse with the goal of utility maximization (pursuing the trade-off between profit maximization and income stability). Therefore, the strict marginal analysis framework, which assumes that farmers pursue profit maximization, is no longer suitable for analyzing farmers’ fertilization behavior, and the equilibrium condition of fertilization is no longer that the price of fertilizer is equal to the value of marginal products. Given that farmers generally consider that applying more fertilizer can mitigate the negative effect of uncertain production conditions, they tend to prevent unacceptable decreases in income by applying more fertilizer to maximize utility when there are production risks.
Therefore, applying more fertilizer will break the original equilibrium between the original marginal benefits and marginal costs, leading to the right shift of the equilibrium point, which seems to be “economically undesirable”. However, applying more fertilizer increases the income expected value and its distribution, ensuring that farmers’ income is not too low even in years of bad weather, thus maximizing their utility. Therefore, under risk conditions, farmers will weigh the costs and benefits of applying more fertilizer. The equilibrium point of fertilizer use remains at the second stage of production, where the yield has not yet reached the maximum and the marginal yield is greater than zero.
Figure 1 shows the relationship between utility maximization and farmers’ fertilizer use. Specifically, it is known that farmers pursue utility maximization:
where
x denotes fertilizer application rates,
y denotes yield (
y ≤
f(
x),
f(
x) denotes the production function under certain technology), and
Ur denotes the utility of income stabilization from additional fertilizer (
Ur =
rx and production risk factor
r ≥ 0). Under certain risk,
Ur increases with fertilizer application rates in the second production stage.
When there is no production risk (r = 0), the slope of the iso-utility line (the iso-profit line) is k = px/py, the equilibrium point of fertilizer use is a, and the fertilizer application rates are xa. When there is production risk (r > 0), the slope of the iso-utility line is k’ = (px − r)/py, the equilibrium point of fertilizer use is b, and the fertilizer application rates are xb. xb > xa indicates that farmers apply more fertilizer to avoid production risk.
The characteristics of agricultural production and fertilizer have led farmers to apply more fertilizer under risk conditions. First, because there are many uncertainties in agricultural production, and fertilizer is a yield-enhancing input that can compensate for the loss of yield caused by an excess or shortage of other factors to a certain extent. Second, the process of fertilizer uptake by crops is special because fertilizer application and crop nutrient uptake are not synchronized. Nutrient uptake by crops is continuous, while fertilizer is applied once or in stages, with most fertilizers needing a period after application before they are absorbed. During this period, adverse weather may cause fertilizer losses, resulting in yield losses. When the above situations cause yield losses, farmers can increase crop yield and reduce yield risk by applying more fertilizer.
Therefore, it is necessary to apply additional fertilizer to restore plant growth and compensate for fertilizer losses owing to heavy rainfall. Heavy rainfall causes not only waterlogging in farmland, which affects crop metabolism, but also high humidity conditions, leading to the spread of diseases and multiplication of insect pests, thus causing crop yield reductions. If heavy rainfall occurs before the middle term of crop growth, foliar fertilizer can be sprayed to quickly restore crop growth by improving the photosynthesis of crop leaves. However, foliar fertilizer accounts for a small proportion of the total fertilizer use. The amount of fertilizer required to compensate for the fertilizer losses caused by heavy rainfall is relatively large. Meanwhile, heavy rainfall in China mainly occurs in summer when crops require more fertilizer and the concentration of fertilizer in the soil is relatively high. The scouring and leaching effect of heavy rainfall on fertilizer is relatively large, which causes the leakage and runoff of fertilizer. According to the barrel theory, crop yield depends on the input of the “shortest board”. If additional fertilizer is not applied after heavy rainfall, serious crop losses may occur owing to fertilizer runoff.
Figure 2 shows the impact of heavy rainfall variability on farmers’ fertilizer use.
Meanwhile, rainfall is random and heavy rainfall has significant interannual variations. Given that farmers cannot accurately predict rainfall levels for the year when making fertilizer decisions (fertilization is an ex ante action), they can only make rough estimates based on rainfall levels in previous years. Therefore, in areas with large rainfall levels and variability in previous years, farmers may apply more fertilizer to ensure that yield and income are not too low. In addition, poor detectability of soil nutrient contents, nutrient losses, and crop uptakes results in incomplete information on fertilizer use efficiency, exacerbating farmers’ uncertainty about fertilizer application rates. Farmers tend to apply more fertilizer to ensure sufficient nutrients for crop growth.
It should be noted that fertilization is mainly an ex ante decision rather than an ex post disaster relief, and the occurrence of risk is random. Avoiding risk requires more fertilizer to be systematically applied annually rather than just adding fertilizer after risks occur. The additional fertilizer applied each year can be seen as a fixed cost to achieve the expected goal. Moreover, although farmers seem to increase fertilizer use to avoid yield risk, they actually avoid income risk (applying more fertilizer to ensure that the probability and extent of yield reduction are not too large and income is not too low).
Thus, we propose the following two hypotheses:
- (1)
There is a positive relationship between heavy rainfall variability and farmers’ fertilizer application rates on maize.
- (2)
Yield fluctuations are a channel through which heavy rainfall variability affects farmers’ fertilizer application rates on maize.
5. Further Discussion
In addition to examining the impact of heavy rainfall variability on farmers’ fertilizer application rates on maize, it is important to understand farmers’ perceptions of fertilizer use efficiency and their response measures when facing heavy rainfall. In this section, we use the Jiangsu Province Rural Fixed Observation Point Survey Data, collected by Nanjing Agricultural University in eight counties (Donghai, Sihong, Dafeng, Baoying, Danyang, Lishui, Haimen, and Kunshan) in January 2022 to examine farmers’ perceptions of fertilizer use efficiency and their response measures to heavy rainfall, testing the mechanism of the impact of heavy rainfall variability on maize farmers’ fertilizer application rates.
Table 7 reports the distribution of maize and rice farmers.
When farmers were asked “Would a 10% reduction in current fertilizer application rates have an impact on yield?”,
Table 8 shows that 72% of maize farmers and 84% of rice farmers argued that this scenario would have an impact on yield. When fertilizer application rates are assumed to decrease by 20%, the number of maize and rice farmers who argue that yield will be affected reaches 92% and 100%, respectively, indicating that most farmers argue that existing fertilizer application rates are appropriate and reducing fertilizer application rates will adversely affect yield. Perhaps because of this fertilization perception, it is difficult for farmers to voluntarily reduce fertilizer use without a change in technology, which is the main reason for the poor outcomes of fertilizer reduction work. Therefore, farmers’ perceptions of the impact of fertilizer reduction on crop yield may be a reason for the high fertilizer application rates by Chinese farmers.
When farmers were asked, “When applying base fertilizer, if 30% of the fertilizer is estimated to be lost due to waterlogging within half a month, will you apply more fertilizer?”,
Table 9 shows that 56% of maize farmers and 44% of rice farmers argued that they would increase base fertilizer (adding fertilizer before disasters represents a preventive fertilization practice, while adding fertilizer after disasters represents a remedial fertilization practice). When farmers were asked, “If waterlogging occurs and may reduce the yield by 30%, will you apply topdressing?”, the results show that 70% of maize farmers and 76% of rice farmers argued that they would apply additional fertilizer (remedial fertilization practice). Furthermore, we found that 45 maize farmers and 17 rice farmers answered that they would increase fertilizer use before and after disasters, indicating that more than half of the farmers would take both preventive and remedial actions. The above results suggest that farmers apply more fertilizer mainly to prevent and remedy the adverse effects of natural risk (heavy rainfall variability) on crop yield. Farmers view applying more fertilizer as a means to avoid risks, that is, by applying more fertilizer to ensure that crops can absorb sufficient nutrients and reduce the likelihood of yield reductions, which is in line with Rajsic et al. [
32], Babcock [
36], Yang et al. [
37], and Stuart et al. [
38].
6. Conclusions and Policy Implications
This study uses farmers’ responses to changes in rainfall levels to examine whether heavy rainfall variability is a reason for the high fertilizer application rates of Chinese farmers. Using the 2003–2018 National Rural Fixed Observation Point Survey Data, we found that farmers’ fertilizer application rates on maize increase with heavy rainfall variability, indicating that heavy rainfall variability is a reason for the high fertilizer application rates by Chinese farmers. Farmers’ fertilizer application rates will increase by 1.7 kg for a 0.1 increase in heavy rainfall variability. Furthermore, we investigated the heterogeneous effects of heavy rainfall variability on fertilizer use, and found that heavy rainfall variability has greater effects on fertilizer application rates in hills and mountainous areas and areas with good irrigation and economic conditions. Finally, we examined the potential channels behind the estimated effects and found that yield fluctuations are a channel through which heavy rainfall variability affects fertilizer application rates. The above results indicate that heavy rainfall variability increases the risk or uncertainty associated with income and yield and farmers apply more fertilizer to prevent yield and income risk.
The findings of this study indicate that yield uncertainty caused by heavy rainfall variability leads to the high fertilizer application rates by Chinese farmers. When making fertilizer decisions, farmers not only rely on the traditional marginal revenue analysis framework, but also consider the impact of natural risk. To reduce farmers’ fertilizer application rates, the Chinese government can promote corresponding fertilizer reduction technologies (deep loosening and deep application) and products (slow- and controlled-release fertilizers) through financial subsidies (e.g., agricultural machinery purchases and service subsidies, policy and financial support to enterprises to reduce the price of environmentally friendly fertilizers, and subsidies for farmers to purchase slow- and controlled-release fertilizers) and demonstration households, to reduce fertilizer losses caused by heavy rainfall and improve fertilizer use efficiency. Furthermore, the Chinese government can reduce fertilizer losses and yield reduction caused by heavy rainfall by strengthening agricultural infrastructure construction (e.g., promoting farmland reconstruction projects and implementing farmland protection measures according to local conditions).