Impact of Government Policies on Seed Innovation in China

Abstract

Seed innovation is of great importance for more sustainable agriculture and food systems. Using data on nationally approved rice varieties and farmers’ adopted varieties, this study examined rice varietal trait changes in China over the past four decades and explored the impact of national crop varietal approval policies on approved rice traits as well as the effect of seed subsidies on adopted rice trait changes. The results showed that the yield of approved varieties and adopted varieties showed an upward trend over the past decades, and the yield of approved varieties was slightly higher than that of adopted varieties in most years. The rice quality of approved rice varieties showed a trend of continuous improvement but the adopted varieties showed a downward trend. The disease resistance of the approved varieties failed to show an increasing trend overall while the adopted varieties remained unchanged. National crop variety approval policies seemed to exert a significant positive impact on approved rice yield traits but exert a negative influence on disease resistance. Subsidies for superior seed varieties significantly increased adopted rice quality but decreased yield. The results suggest that national crop variety approval policies are the gatekeeper of improved rice, so the government can improve the policies to more meet farmers’ and consumers’ needs.

1. Introduction

The progress of science and technology is a main driving force of the growth of agricultural total factor productivity [1], to which seed innovation contributes the most [2]. Since the foundation of the People’s Republic of China, improvement of adopted crop varietal traits has played a critical role in increasing crop productivity [3,4,5]. Studies show that during 1985–1994, the improvement of maize seed traits contributed 35.5% to maize yield [4].

Establishment of a crop varietal approval system is a gatekeeper for crop variety improvement [6]. The “National Crop Variety Approval Trial Regulations” promulgated in 1982 established national and provincial two-level crop variety approval systems. In 1989, the “Regulations of the People’s Republic of China on Seed Management” established a crop variety approval system in the form of regulations. The “Seed Law of the People’s Republic of China, hereinafter referred to as the “Seed Law”, released in 2000 established a crop variety approval system in legal form for the first time. These laws and regulations led the direction of crop production and variety improvement in China and have greatly promoted the rapid development of the seed industry. Moreover, the government has implemented a series of agricultural subsidy policies to encourage farmers to engage in grain planting and ensure stable and increased grain production. Subsidy for growing superior seed varieties is a special funding established by the central government to encourage farmers’ adoption of superior crop varieties, improve the quality and yield of agricultural products, and mobilize farmers’ enthusiasm for grain production in China. In 2002, the central government allocated 100 million yuan ($16.7 million) to subsidize the adoption of superior soybean varieties in Northeast China, which officially opened a precedent for agricultural subsidies. In 2004, the central government arranged special subsidy funding to support the adoption of superior rice varieties in seven provinces of Hunan, Hubei, Jiangxi, Anhui, Heilongjiang, Jilin, and Liaoning to stimulate farmers in the main rice-producing areas to plant rice.

Rice is the most important staple food crop in China. Adoption of superior new varieties may not only increase its yield potential but also improve its quality and disease resistance. With the wide adoption of new varieties by farmers, China’s rice yield has achieved substantial growth (Figure 1). Based on the sources or types of new rice varieties, the adoption of rice varieties in China can be divided into four stages. Before the mid to late 1950s, it was often called the period of selection and identification of farm seeds. At that time, most of the adopted seeds were unapproved farmer seeds. From the mid-to-late 1950s to the late 1960s, the first dwarf rice variety was bred and adopted, and compared to original high-stalk lodging varieties, the yield of dwarf rice varieties increased by approximately 30%. The third stage was from the early 1970s to the mid-1980s, during which three-line hybrid rice breeding was approved. Its yield was 34.5% higher than that of dwarf rice [7]. The fourth stage was from the end of the 1980s to the present. In this stage, the adoption of two-line hybrid rice was the third major breakthrough in the rice history of China, and rice yield increased by 111% compared to that in the 1950s.

Previous studies show that there is a disconnect between the adoption of agricultural technology and the needs of farmers in China, which leads to low efficiency in the adoption of new technologies by farmers and the transformation of scientific and technological achievements [8,9]. Research institutes and firms in China have approved a large number of superior new rice varieties for farmers. However, not all advanced technologies could be easily accepted and adopted by farmers due to a low level of relevant knowledge and information asymmetry [10]. The gaps between approved and adopted new rice varieties in China are of great importance for the government to identify farmers’ needs and promote the transformation of agricultural, scientific, and technological achievements into real agricultural productivity. Therefore, the objective of this study was to investigate trait changes and gaps between approved and adopted new rice varieties and examine the effect of national crop varietal approval policies on approved rice and subsidies for superior seed varieties on adopted rice trait changes.

There are two main contributions in this paper. First, this study employed national and provincial approval data and farmers’ adoption data to investigate the trait changes of approved and adopted rice varieties and their differences over the past four decades, which could identify the gaps between seed supply and demand and provide empirical evidence for the government to implement appropriate seed policies. Second, previous studies mainly used survey data to explore the impact of superior seed subsidies on farmers’ adoption behavior. This study was based on provincial-level data to examine whether government policies significantly influence approved and adopted rice traits, which would be helpful for the government to adjust seed industry policy in the future. Although superior seed subsidies are now combined with direct subsidies for grain and comprehensive subsidies for agricultural materials in China, called agricultural support and protection subsidies, the results of this study are still meaningful to capture the influence of government policies on researchers’ R&D behaviors.

2. Literature Review

2.1. Rice Variety Approval System and Changes in China

The crop varietal approval process in China is that the Varietal Approval Committee conducts regional and production tests on newly bred varieties in accordance with prescribed procedures and comprehensively reviews the adoption value and suitable scope of the variety in terms of yield, growth period, quality, and resistance. It is a key process from the selection of a new crop variety to adoption [11], which is of great significance to the safety, stability, and sustainable development of agricultural production.

Crop varietal approval in China can be traced back to the 1960s and 1970s. With the improvement of breeding technology and farmers’ need to introduce new seeds, some provinces such as Heilongjiang, Liaoning, and Guizhou have successively carried out varietal trials, production tests, variety approval, registration, or approval of rice and other crops. In 1982, the former Ministry of Agriculture, Animal Husbandry and Fisheries promulgated the “National Trial Regulations on Crop Variety Approval”. Then, the national- and provincial-level crop variety examination and approval commissions were established. A series of regulations such as the “Regulations of the People’s Republic of China on Seed Management”, “Regulations on the National Crop Variety Approval”, and “Measures for the National Crop Variety Approval” were issued and implemented. The task of crop variety approval has gradually been standardized and national and provincial two-level variety review systems have been established. The process has greatly promoted the selection and adoption of crop varieties in China [12]. The “Seed Law” promulgated and implemented in 2000 clearly stipulated that five major crop varieties including rice, maize, wheat, cotton, and soybean must pass national or provincial approval before being adopted. Then, provincial governments formulated local seed laws and regulations. This implies that the varietal approval system of major crops such as rice has been legalized in China [13].

Since the establishment of national- and provincial-level systems for variety approval (SVA) in 1982, variety approval standards have varied among different periods, provinces, and crops [11,14]. Taking national rice variety approval as an example, the first National Crop Variety Examination and Approval Committee (NCVEAC) formulated the following two approval standards for rice varieties from 1983 to 1988. First, the tested varieties in regional trials and production trials must increase their yields by more than 10% compared with the control varieties or the yield growth must be statistically significant. Second, if the yield of the tested varieties is the same as that of the control varieties, at least one of the traits such as quality, growth period, and resistance must have outstanding performance [11].

From 1989 to 1996, the second NCVEAC revised rice varieties’ approval standards. It emphasized that if national-level approved varieties were adopted across provinces, the reviewed variety must be through provincial approval in two provinces or approval in one province and performed outstandingly in national-level regional trials and production trials [15].

From 1997 to 2001, the requirements proposed by the third NCVEAC emphasized further renewal of national-level approval varieties. The specific requirements were as follows. First, for the varieties that have been through two provincial-level approval processes, one of the approvals must be completed within two years. Second, clear requirements for yield, rice quality, and resistance levels of new varieties in national regional trials and production trials were proposed. Third, if trial varieties’ resistance and rice quality were the same as those of the control, its yield must increase 5% more than the control or an increase in yield must be statistically significant; or if its resistance level is the same as the control group, major indicators of rice quality must reach high-quality level two or above and its yield can be less than 5% lower than the control; or if yield and rice quality are the same as the control, it can be resistant to at one major pest [11].

In 2002, the first National Crop Variety Approval Committee (NCVAC) was established by the former Ministry of Agriculture. Given the poor quality of rice varieties and high demand for high-quality rice from consumers, the first NCVAC adjusted the approval requirements of rice varieties between 2002–2006. In addition to not relaxing the examination and approval of high-yield and resistant varieties, for varieties with quality up to the grade one, two, and three national standards, the yield could be 15%, 10%, and 5% less than that of the control, respectively. Under this standard, a large number of high-quality rice varieties emerged [16].

After the establishment of the second National Crop Variety Approval Committee in 2007, the government further relaxed the yield requirements for high-quality rice varieties. For the rice varieties whose quality level reached national standard of high-quality one, two, or three (or one, two, and three levels better than the control), its yield index was relaxed to no more than 10%, 5%, 0% than the control, respectively. Moreover, a one-vote veto system was implemented for variety resistance, which required that the comprehensive index of rice blast resistance be less than seven while the highest panicle blast loss rate of the northern rice area and the Wuling mountainous area is less than seven grades, and the highest panicle blast loss rate of the varieties in the southwestern rice region is not greater than seven grades [17].

In sum, the crop varietal approval system in China has experienced significant changes over the past decades. It potentially has a significant impact on the traits of newly bred rice varieties.

2.2. Subsidy Policies for Superior Rice Seed Varieties in China

Seed subsidy for superior variety (SSV) is a special subsidy policy for improving the competitiveness of China’s grain industry. Its main purpose is to encourage Chinese farmers to adopt new varieties or technologies in production and finally improve the quality and yield of grain [18]. The subsidies are mainly for farmers (including farm employees) to use superior crop seeds in production.

The Chinese government started to organize and implement superior seed adoption subsidies in 2002, and the central government invested 100 million yuan ($15.4 million) to subsidize superior soybean seed varieties. In 2003, wheat was included in the scope of subsidies for superior seed varieties due to a sharp reduction in national grain production. In March 2004, the Ministry of Finance and the former Ministry of Agriculture jointly formulated the “Interim Measures for the Management of Funds for the Promotion of Improved Varieties of Crops” that aimed to strengthen the management and supervision of the use of project funds. In April 2004, the “Interim Measures for the Management of Subsidy Funds for the Use of Improved Rice Varieties” was promulgated and implemented. The central government arranged special subsidy funds to support the adoption of superior rice seed varieties in seven provinces including Hunan, Hubei, Jiangxi, Anhui, Heilongjiang, Jilin, and Liaoning. The subsidy standard was 150 yuan/hm² for early rice, 225 yuan/hm² for medium rice and japonica rice, and 105 yuan/hm² for late rice. The subsidy area was often based on the actual planting area. In 2007, Sichuan, Guangxi, and Chongqing were incorporated as provinces subsidizing superior rice seed varieties. In 2008, 29.33 million hectares of rice farmland in China implemented the full-coverage subsidy. At the same time, the subsidy standard for late rice was increased from 105 yuan/hm² to 225 yuan/hm², the subsidy for early rice became 150 yuan/hm², and the subsidy for medium rice, japonica rice and late rice became 225 yuan/hm².

A great number of studies have investigated the effect of superior seed subsidy policy on crop production but have obtained mixed results. Some scholars believe that superior seed subsidy policies are helpful for increasing grain yield [19,20,21,22], improving quality [21], decreasing prices [21], increasing sown area [22] and total factor productivity [23], and encouraging farmers to engage in grain planting [19]. Moreover, subsidy policy was helpful for improving farmers’ technical efficiency [24], promoting the adoption of new crop varieties [25,26], and narrowing gender differences in variety adoption behavior [27]. There are also some scholars that argued that superior seed subsidy policy did not exert any significant impact on farmers’ production. Guan [28], for example, analyzed the production efficiency of major cotton-producing provinces in China before and after the implementation of cotton seed subsidies. The results showed that China’s current cotton seed subsidy policy had a limited effect on improving cotton production efficiency.

2.3. Previous Literature on Crop Variety Selection and Adoption

The purpose of China’s variety approval system is to encourage seed-breeding units to create new germplasm and breed new varieties. Therefore, relaxing or strengthening crop varieties’ approval standards can significantly influence the R&D behavior of researchers from the public or private sector, especially firms and individuals. This can further influence the traits of new crop varieties developed by researchers. However, until now, there have been few studies examining the impact of seed industry regulations and policies on approved crop variety traits in China.

Previous studies have shown that the yield potential, quality, and disease resistance of rice varieties that had been obtained through national or provincial approval systems have been improving, but different regions present different trends. Zhao et al. [29], for example, analyzed rice regional trials for provincial-level approval in Zhejiang Province from 1991 to 1996 and found that the growth rate of rice yield remained at approximately 5% in general. Que et al. [30], based on the data of approval rice varieties in Jiangsu Province from 1986 to 2005, indicated that rice’s yield potential, quality, and single resistance had improved, although its comprehensive resistance level had declined. Based on approval rice variety data in Sichuan Province from 1984 to 2014, Deng [31] found that rice quality and resistance had improved. Based on the trait data of mid-indica late-maturing rice varieties approved in Sichuan Province from 2001 to 2010, Zeng et al. [32] showed that the plant height, panicle length, grain number, and thousand-grain weight of rice had increased but the growth period, effective panicle, seed setting rate, and resistance grade of rice blast showed a downward trend.

The role of new variety adoption could often increase crop yield, improve agricultural product quality, adjust agricultural structure, and strengthen disease resistance [33]. Decision making for new variety adoption has been extensively examined by many scholars. The research has focused on two aspects: the role of new variety adoption and determinants of farmers’ decisions on adopting new varieties. Studies have shown that the adoption of new varieties increases crop yields [34,35,36], farmers’ incomes [34,35,36,37,38,39,40], and consumer expenditures [41] and reduces poverty [39,41] and pesticide use [36].

The main factors that influence crop new variety adoption are farmland scale [34], policy reforms [42], cost and output of new varieties [34,43,44,45], and information [46,47]. Based on corn adoption data from farmers in Minnesota and Wisconsin in the United States, for example, Useche et al. [48] found that farmers’ preference for variety traits was influenced by producer and regional heterogeneity. Negatu et al. [49] employed data from 96 wheat farms in Ethiopia and found that farmers’ decision to adopt new varieties was affected by factors such as farm size, farmer income, and soil type. Mariano et al. [50] found that farmers’ adoption of new varieties in rice production in the Philippines was affected by factors such as ownership of machinery and irrigation water supply. Using survey data from 200 corn farmers in Mexico, Sanchez et al. [42] found that the 1994 North American Free Trade Agreement Mexico’s agricultural policy reform had a significant impact on the adoption rate of superior seeds.

In recent years, some scholars have also examined major factors affecting Chinese farmers’ decision to adopt new varieties, which were mainly divided into economic factors and noneconomic factors. Economic factors mainly include farmers’ income level [51,52,53,54] and seed prices [53,55,56], while noneconomic factors mainly include mechanical input [57], factor endowments [57], and policy rewards [52,53,56]. Zhu et al. [52] conducted a survey of 289 farmers in impoverished mountainous areas in western Hubei and found that government subsidies promoted farmers’ adoption of new technologies. Based on survey data of 196 farmer households in four provinces, Song et al. [58] found that farmers’ decision making regarding new variety selection was affected by farmers’ income level, arable land area, and education level. In addition, some studies have shown that the maximum adoption area of superior rice varieties is mainly influenced by yield potential, quality, and disease resistance [59]. Based on the potential impact of superior seed subsidies on farmers’ choice of rice seeds, this study aimed to investigate whether subsidies for growing superior seeds promote farmers to choose high-yield, high-quality, and high-resistance-to-pest rice varieties.

3. Materials and Methods

3.1. Data

This study employed four sets or sources of data. Data set 1 involves major agronomic traits of approved rice varieties such as yield, growth period, plant height, rice quality indicators, and resistance to major diseases and insect pests. The information was mainly from the regional experiments on seed varieties published by the Ministry of Agriculture, the “National Crop Approved Varieties” and “Chinese Rice Varieties and Their Genealogy” published by China Agriculture Press over years, and the website of the National Rice Data Center.

Data set 2 involves major agronomic traits and planting areas of adopted rice varieties, which are those whose planting areas are more than 6666.7 hectares in at least one province in a certain year. It was from the “Statistical Table for the adoption of National Major Crop Varieties” and the “Chinese Rice Varieties History.”

Data set 3 involves economic factors that influence rice adoption such as per capita disposable income and per capita arable land area. These data were from the National Bureau of Statistics, the “China Statistical Yearbook”, and the “China Rural Statistical Yearbook”.

Data set 4 involves agricultural input factors that may also affect farmers’ decisions regarding rice variety adoption such as the total power of agricultural machinery, flood disaster areas, drought disaster areas, and diseases and insect or pest areas. These data were from the National Bureau of Statistics, the “China Agricultural Statistical Yearbook”, and the “Fifty Years of China’s Plant Protection”.

3.2. Conceptual Framework

Breeding superior crop varieties is an effective measure to promote the progress of agricultural technology, while adopting new varieties in production and their performance are the standards to measure a country’s agricultural breeding technology [33]. This study attempted to investigate the trait changes of approved and adopted rice varieties from the perspectives of supply and demand.

In terms of the supply of new rice varieties, researchers in the public sector aim to select seed varieties that are likely to meet the country’s national development strategy, while firms give top priority to developing varieties that satisfy farmers’ needs and make more profits [60]. However, even when researchers have developed a new variety, it does not guarantee that it is a good variety in production. Since the difference in appearance characteristics among different varieties of the same crop is small, it is extremely difficult for farmers or consumers to distinguish the difference in its traits and quality. Such small differences can also give rise to counterfeiting when the quality assurance systems do not operate well, which has a great impact and a strong negative multiplier effect on production. Once fake and inferior seeds flow into the market, they can cause great losses to agriculture and form various “seed accidents”. Therefore, one of the major purposes of a country’s approval system for seed innovation is to choose superior seed varieties to avoid “seed accidents”. Accordingly, requirements of approval systems influence researchers’ R&D behavior and the products they develop. One of the aims of this study was to examine trait changes in newly bred rice varieties developed by research institutes or firms during the past four decades and the impact of crop variety approval standards.

From the standpoint of demand, previous studies show that farmers’ decisions on rice variety adoption are often influenced by varieties’ traits [59]. In other words, varieties adopted by farmers to some extent indicate what kind of new varieties are needed in real production. If subsidies for growing superior seed varieties can encourage more farmers to choose seeds with better traits such as high quality or resistance to disease, rice aggregate traits would change accordingly under the influence of subsidies. Another aim of this study was to investigate newly bred rice varieties’ trait changes adopted by farmers during the past four decades and the impact of subsidies for superior rice seed varieties.

Moreover, with the improvement of people’s living standards and the enhancement of other agricultural technologies, researchers must also adjust the traits of newly developed seed varieties. For example, with the improvement of fertilizer technology and increasing use of fertilizer, the yield potential of high-stem rice varieties is limited due to their susceptibility to lodging. Low stem and lodging resistance to high yield have become a new goal for breeders since the 1950s in China [33]. Similarly, a serious rice blast in Zhejiang Province occurred in 1975 and 1984. After that, most rice varieties used by farmers have had strong blast resistance [33]. Therefore, per capita disposable income, per capita cultivated land area, mechanization, and diseases as well as insect pests are expected to exert some influence on the trait changes of newly bred rice varieties. The specific theoretical framework is shown in Figure 2.

3.3. Model Specification

3.3.1. Variety Approval Model

Since crop variety approval standards involve many requirements on crop variety traits, changes in the approval standards would result in approved crop trait changes. In this study, three major agronomic traits of rice including yield-relevant traits (growth period, plant height, and yield), rice quality, and disease resistance were employed. The head rice rate was used as rice quality while the resistance of rice to blast and bacterial blight indicated rice disease resistance. It is worth noting that the trait value of approved rice varieties in this study was obtained by averaging all varieties’ traits in specific provinces.

For approved rice varieties, agronomic traits (AT) of a certain province in a certain year could be represented by the simple average value of agronomic traits of all rice varieties approved in that year in that province.

It is as follows:

$${\mathrm{MAT}}_{\mathrm{it}}=\frac{1}{{\mathrm{N}}_{\mathrm{it}}}{{\displaystyle \sum }}_{\mathrm{k}=1}^{{\mathrm{N}}_{\mathrm{it}}}{\mathrm{AT}}_{\mathrm{kit}}$$

where i represents the province and t represents the year; MAT_it represents the simple average traits of all approved rice varieties in province I in year t; k = 1, 2, ... N_it; N_it represents the total number of approved varieties in province i in year t; and AT_kit represents agronomic traits of the approved variety k in province I in year t.

Provincial-level varieties were calculated by directly averaging approved varieties. For national-level approved varieties, the first step was to adjust the approved province from “national” to 31 provinces and then average all the varieties in a specific province.

The specific model is as follows:

$${\mathrm{Y}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{Seedlaw}+{\mathsf{\beta }}_2\mathrm{STDdum}+{\mathsf{\beta }}_3{\mathrm{LnDPI}}_{\mathrm{it}}+{\mathsf{\beta }}_4{\mathrm{LnRjgdarea}}_{\mathrm{it}}+{\mathsf{\beta }}_5{\mathrm{LnMachine}}_{\mathrm{it}}+{\mathsf{\beta }}_6{\mathrm{Dry}}_{\mathrm{it}}+{\mathsf{\beta }}_7{\mathrm{Flood}}_{\mathrm{it}}+{\mathsf{\beta }}_8{\mathrm{Pest}}_{\mathrm{it}}+{\mathsf{\beta }}_9{\mathrm{Hybrid}}_{\mathrm{it}}+{\mathsf{\beta }}_{10}{\mathrm{Middle}}_{\mathrm{it}}+{\mathsf{\beta }}_{11}{\mathrm{Latter}}_{\mathrm{it}}+{\mathsf{\beta }}_{12}{\mathrm{Indica}}_{\mathrm{it}}+{\mathrm{a}}_{\mathrm{i}}+{\mathrm{u}}_{\mathrm{it}}$$

(1)

where Y_it includes lnYieldmu_it, lnDuration_it, lnHeight_it, Headrice_it, and lnBacterbt_it. lnYieldmu_it is the variety’s trial (regional trial or production trial) yield in logarithm, lnDuration_it is the approved rice’s period of duration in logarithm, and period of duration refers to the time that a crop elapses from seeding to seed maturity (harvesting). lnHeight_it is the plant height, plant height refers to the distance between the base of the plant and the top of the main stem in logarithm, Headrice_it is the percentage of head rice to net rice sample mass, and lnBacterbt_it is the level of disease resistance (bacterial blight and rice blast) in logarithm. The variable Seedlaw represents a dummy variable for before or after 2001 when the Chinese government issued the first seed law in China. It equals 1 in or after 2001, otherwise 0. The variable STDdum represents a set of dummy variables of the variety approval standard. The crop variety approval reforms were divided into five stages. The first stage is 1982–1988, the second stage is 1989–1996, the third stage is 1997–2001, the fourth stage is 2002–2006, and the fifth stage is 2007. The first stage of 1982–1988 was taken as the baseline. lnMachine, lnDPI, and lnRjdgarea are the logarithms of the degree of mechanization, per capita disposable income, and per capita arable land area, respectively. The variables Flood, Dry, and Pest are a set of disaster variables that represent the proportion of flood disaster area, drought disaster area, and disease and insect damage area to the sown area of crops, respectively. Hybrid, Middle, Latter, and Indica represent the proportions of hybrid rice, middle rice, late rice, and indica rice to the total number of approved varieties, respectively.

A fixed-effect model was used in the model estimation. It is worth noting that due to inconsistent statistical caliber of disease resistance, it was set to five levels: the lowest value 1 represents high resistance of the variety, 3 represents resistance, 5 represents medium resistance, 7 represents sense, and the highest value 9 represents high sense. Since this study simply averaged the variety traits by province, the original five-level disease resistance became a continuous variable.

3.3.2. Variety Adoption Model

Consistent with the traits of the approved varieties, this study selected 6 agronomic traits of rice for regression analysis including rice yield, growth period, plant height, quality, and disease resistance. The traits of adopted rice varieties were obtained after weighing the average rice planting area by province. The specific formula is as follows:

$${\mathrm{WAT}}_{\mathrm{it}}={{\displaystyle \sum }}_{\mathrm{k}=1}^{{\mathrm{N}}_{\mathrm{it}}}\frac{{\mathrm{area}}_{\mathrm{kit}}}{{\mathrm{area}}_{\mathrm{it}}}·{\mathrm{AT}}_{\mathrm{kit}}$$

(2)

where WAT_it represents the weighted average of the traits of all rice varieties used in province i in year t; k = 1, 2, ... N_it; N_it represents the number of adopted varieties (actually planted) in province i in year t; and AT_kit represents the agronomic traits of variety k adopted (actually planted) in province i in year t.

As mentioned above, this study assumed that subsidy policies for superior varieties affect the traits of rice varieties adopted in China. To examine the contribution of subsidy policy to the changes in the traits of adopted rice varieties, this study employed the following model:

$${\mathrm{Y}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{Allowance}+{\mathsf{\beta }}_2\mathrm{Subsidy}+{\mathsf{\beta }}_3{\mathrm{LnDPI}}_{\mathrm{it}}+{\mathsf{\beta }}_4{\mathrm{LnRjgdarea}}_{\mathrm{it}}+{\mathsf{\beta }}_5{\mathrm{LnMachine}}_{\mathrm{it}}+{\mathsf{\beta }}_6{\mathrm{Dry}}_{\mathrm{it}}+{\mathsf{\beta }}_7{\mathrm{Flood}}_{\mathrm{it}}+{\mathsf{\beta }}_8{\mathrm{Pest}}_{\mathrm{it}}+{\mathsf{\beta }}_9{\mathrm{Hybrid}}_{\mathrm{it}}+{\mathsf{\beta }}_{10}{\mathrm{MiddleLatter}}_{\mathrm{it}}+{\mathsf{\beta }}_{11}{\mathrm{Indica}}_{\mathrm{it}}+{\mathrm{a}}_{\mathrm{i}}+{\mathrm{u}}_{\mathrm{it}}$$

(3)

where Y_it includes lnYieldmu_it, lnDuration_it, lnHeight_it, Headrice_it, and lnBacterbt_it. The variable Allowance represents a dummy variable of subsidy policy for improved rice varieties. It equals 1 for Hunan, Hubei, Jiangxi, Anhui, Heilongjiang, Jilin, Liaoning, 7 provinces from 2004, and all provinces after 2010; otherwise, it equals 0. The Subsidy variable represents the direct subsidy policy for grains. It equals 1 since 2004 for all provinces, otherwise 0. Other control variables have the same definition as Equation (1). Similarly, a fixed-effect model could be used to estimate Equation (3). To avoid pseudo-regression, this study also conducted a test of the stationarity of the data. The methods and results are provided in the Supplementary Materials (Tables S1 and S2).

Considering the influence of control variables on the changes in traits of approved varieties may have a lag. According to the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), the lag periods of the variables are provided in the Supplementary Materials (Table S3).

4. Results

4.1. Trait Changes of Approved and Adopted Rice Varieties

The results showed that the traits of approved and adopted rice varieties presented completely different trends in China. Regarding the yield, both approved and adopted varieties showed an upward trend. However, the yield of approved varieties was slightly higher than that of adopted varieties in most years (Figure 3). The yield of approved rice varieties increased from 402.63 kg/mu in 1982 to 599.12 kg/mu in 2016 with an average annual growth rate of 1.14%, while the yield of adopted varieties increased from 445.33 kg/mu in 1982 to 554.97 kg/mu in 2016 with an average annual growth rate of 0.63%, which was lower than the average annual growth rate of approved varieties. This suggests that the increase in the yield of approved rice varieties has encouraged farmers to choose rice varieties with high yield potential, which is conducive to promoting the improvement in the yield of rice varieties used in China.

Plant height presented similar changes to the trait yield (Figure 4), which is also in line with the fact that yield and plant height were highly related. Breeding researchers and farmers tend to increase yield by increasing plant height. The plant height of approved varieties increased from 87.45 cm in 1982 to 110.89 cm in 2016 with an average annual growth rate of 0.68%, while the plant height of adopted varieties increased from 89.96 cm in 1982 to 108.17 cm in 2016 with an average annual growth rate of 0.53%.

For the growth period, approved varieties first decreased in the early 1980s then increased in the 1990s, which is consistent with the goal of China’s rice variety improvement. Early maturity breeding in the late 1970s shortened the average growth period of rice. With the changes in breeding goals, increase in labor costs, and rising demand for single-season rice and mid-season rice, the average growth period has gradually increased. Figure 5 shows that the growth period of approved varieties has increased and decreased significantly and the annual growth rate has remained between −12.17% and 7.62%, while the adopted varieties have generally shown a relatively gentle growth trend.

The rice quality of approved rice varieties in China has been continuously improved to meet people’s urgent needs for high-quality rice. This study used the head rice rate to evaluate rice quality; that is, the higher the head rice rate, the better the rice quality (Figure 6). The head rice rate of approved varieties has shown an upward trend in general. It increased from 50.5% in 1982 to 59.34% in 2016 but the fluctuation range was large with an annual growth rate between −10.23% and 22.34%, which revealed that breeding research staff gradually regarded rice quality indicators as breeding targets. In contrast, the head rice rate of adopted varieties showed a downward trend, from 70.2% in 1982 to 60.05% in 2016 with an annual growth rate between −8.64% and 4.97%. This shows that farmers always consider the issue of rice quality in rice production. In recent years, they may have relaxed their rice quality indicators after comprehensively considering yield and other influencing factors. However, the head rice rate has always remained above 55%.

Compared with economic traits, the trend of disease resistance traits was different (Figure 7 and Figure 8). The bacterial blight resistance and rice blast resistance of the approved varieties failed to show an increasing trend in general with the smaller the resistance level, the stronger the resistance, which may be related to the fact that breeding researchers do not regard disease resistance as a breeding goal. Bacterial blight resistance and rice blast resistance of the adopted varieties did not change much as the whole. In particular, it was concentrated at the levels of five and seven, which suggests that farmers always regard disease resistance as an important criterion in the selection of varieties. For bacterial blight resistance, the resistance grade of the approved varieties was always maintained between four and nine, and the average annual growth rate was maintained between −35.28% and 56.28%. The resistance grade of the adopted varieties remained between six and seven. For rice blast resistance, the resistance grade of the approved varieties was maintained between two and nine, and the average annual growth rate was between −52.36% and 119.3%. The adopted varieties were maintained between four and seven, showing a gentle upward trend with an average annual growth rate of −12.63% to 12.16%. These results are consistent with that of a previous study [61].

4.2. Estimation Results of Approved Policies on Approved Rice Variety Traits

The estimation results of approved rice varieties’ trait changes are shown in

Table 1. Results of the impact of approval standards on the traits of approved rice varieties.

	Yield	Plant Height	Growth Period	Head Rice Rate	Bacterial Blight	Rice Blast
	(FE)	(FE)	(FE)	(FE)	(FE)	(FE)
Time trend	0.014 ***	0.005 ***	0.006 ***	−0.169 ***	−0.015 ***	0.012 ***
	(0.001)	(0.001)	(0.001)	(0.059)	(0.004)	(0.003)
Approval standard	−0.076 ***	−0.011	−0.034 ***	1.913 **	0.177 ***	−0.626 ***
(1988–1996)	(0.016)	(0.010)	(0.010)	(0.816)	(0.039)	(0.033)
Approval standard	−0.081 ***	−0.024 *	−0.021 *	3.015 ***	0.225 ***	−0.554 ***
(1997–2001)	(0.019)	(0.013)	(0.013)	(0.974)	(0.065)	(0.049)
Approval standard	−0.138 ***	−0.030 *	−0.023	7.195 ***	0.267 ***	−0.455 ***
(2002–2006)	(0.023)	(0.016)	(0.015)	(1.217)	(0.082)	(0.060)
Approval standard	−0.163 ***	−0.026	−0.060 ***	9.841 ***	0.347 ***	−0.474 ***
(After 2007)	(0.029)	(0.019)	(0.019)	(1.508)	(0.102)	(0.073)
Seed Law	0.008	−0.004	−0.063 ***	−4.947 ***	0.089 **	0.257 ***
	(0.011)	(0.007)	(0.007)	(0.555)	(0.039)	(0.032)
Per capita disposable income	0.003	0.000	−0.005 **	0.070	0.019 *	−0.009
	(0.003)	(0.002)	(0.002)	(0.169)	(0.011)	(0.008)
Machine	−0.008 **	−0.001	0.003	0.107	−0.003	0.042 ***
	(0.003)	(0.002)	(0.002)	(0.192)	(0.013)	(0.008)
Per capita arable land area	0.012	0.010	0.020 ***	0.250	−0.053 *	−0.118 ***
	(0.009)	(0.007)	(0.006)	(0.530)	(0.031)	(0.023)
Flood	0.001	0.001 **	0.000	0.041	0.001	−0.001
	(0.001)	(0.000)	(0.000)	(0.030)	(0.002)	(0.002)
Dry	−0.000	0.000 **	−0.000	0.018	0.000	−0.001
	(0.000)	(0.000)	(0.000)	(0.014)	(0.001)	(0.001)
Diseases and insect pests	0.000	0.000	0.001	−0.007	−0.001	−0.066 ***
	(0.001)	(0.001)	(0.001)	(0.074)	(0.005)	(0.009)
Hybrid rice	0.001 ***	0.001 ***	0.001 ***	0.006	0.001 *	−0.000
	(0.000)	(0.000)	(0.000)	(0.009)	(0.001)	(0.001)
Middle rice	0.000 ***	0.001 ***	0.001 ***	−0.004	−0.001	0.004 ***
	(0.000)	(0.000)	(0.000)	(0.007)	(0.000)	(0.000)
Late rice	0.001 ***	0.000 ***	0.000	0.036 ***	0.001	0.001 **
	(0.000)	(0.000)	(0.000)	(0.008)	(0.001)	(0.000)
Indica rice	−0.001 ***	−0.001 ***	−0.002 ***	−0.084 ***	0.001 ***	−0.003 ***
	(0.000)	(0.000)	(0.000)	(0.008)	(0.000)	(0.000)
R2	0.598	0.598	0.604	0.345	0.162	0.784

Note: The standard deviation in parentheses, “*”, “**” and “***” indicate significant at the level of 10%, 5%, and 1%, respectively.

. The results showed that approved rice varieties’ yield, plant height, and growth period in terms of yield traits increased by 1.4%, 0.5%, and 0.6%, while head rice rate decreased by 0.169% in terms of rice quality. From the perspective of resistance to disease, bacterial blight resistance decreased by 1.5% and rice blast resistance increased by 1.2% per year. This suggests that seed breeding technology has been increasing the yield of rice and improving the bacterial blight resistance as a whole.

Variety approval standards have a significant impact on the traits of approved rice varieties. The coefficients of head rice rate and bacterial blight resistance grade were significantly positive while the coefficients of yield, plant height, growth period, and rice blast resistance were significantly negative. This suggests that the head rice rate and bacterial blight resistance grade in the last four approval stages were generally higher than those in the first stage while the rice yield was generally lower than that in the first stage. The plant heights of the third and fourth stages decreased by 2.4% and 3%, respectively, compared with the first stage, while the growth periods of the second, third, and fifth stages were reduced by 3.4%, 2.1%, and 6%, respectively, compared with the first stage. The results are in line with the requirements for varieties in different stages of approval standards. The first stage focused on the yield of varieties and required a 10% increase in yield while the latter four stages relaxed the restrictions on the increase in yield, instead focusing on the quality and resistance of rice. Bacterial blight and rice blast resistance showed opposite trends at different approval stages. This may be because bacterial blight can be effectively controlled with pesticides, and breeders may gradually no longer regard bacterial blight resistance as the main breeding goal.

The coefficients of Seed Law on growth period and head rice rate were significantly negative and those on resistance grades of bacterial blight and rice blast were significantly positive, but on yield and plant height was not significant. Specifically, after the promulgation of the Seed Law, the growth period and head rice rate of the approved rice varieties decreased by 6.3% and 4.9%, and the resistance levels of bacterial blight and rice blast increased by 8.9% and 25.7%, respectively. There was no significant difference in rice yield or plant height before and after the promulgation of the Seed Law. This indicates that Seed Law had relatively little effect on trait improvement in Chinese rice-approved varieties and had failed to significantly improve rice yield potential, rice quality, and disease resistance.

From the perspective of rice types, the coefficients of the ratio of hybrid rice on yield, plant height, growth period, and bacterial blight resistance grade were significantly positive, which indicates that the higher the proportion of newly bred hybrid rice, the higher the yield, plant height, growth period, and white leaf blight. For every 1% increase in the proportion of hybrid rice, the yield, plant height, growth period, and bacterial blight resistance grade increased by 0.1%. The coefficients of the ratio of medium rice on yield, plant height, growth period, and rice blast resistance were all positively significant. For every 1% increase in the proportion of late rice, the yield, head rice rate, and rice blast resistance increased by 0.1%, 0.036%, and 0.1%, respectively. The coefficient of indica rice for bacterial blight grade was also significantly positive, and the coefficients for other explained variables were significantly negative. This implies that the higher the proportion of indica rice, the higher the bacterial blight resistance grade and the lower the yield, plant height, growth period, head rice rate, and rice blast resistance. The control variables also significantly influenced some traits of approved varieties.

4.3. Estimation Results of the Impact of Seed Subsidies on Adopted Rice Traits

The estimated results are shown in

Table 2. Results of the impact of the subsidy policy for improved varieties on the traits of adopted rice varieties.

	Yield	Plant Height	Growth Period	Head Rice Rate	Bacterial Blight	Rice Blast
	(FE)	(FE)	(FE)	(FE)	(FE)	(FE)
Time trend	0.009 ***	0.003 ***	0.002 ***	−0.110 **	0.005 *	0.081 ***
	(0.001)	(0.000)	(0.000)	(0.050)	(0.002)	(0.010)
Subsidy policy for improved varieties	−0.040 ***	0.003	−0.008	1.214 **	−0.025	−0.463 ***
	(0.010)	(0.005)	(0.005)	(0.599)	(0.035)	(0.128)
Direct subsidy policy for planting grains	−0.013	−0.002	−0.013 ***	2.028 ***	0.091 **	−0.075
	(0.011)	(0.006)	(0.005)	(0.670)	(0.037)	(0.132)
Per capita disposable income	−0.013 ***	0.000	−0.001	0.122	0.003	0.059
	(0.005)	(0.002)	(0.002)	(0.305)	(0.016)	(0.066)
Machine	0.015 ***	−0.001	−0.000	−0.247	−0.000	−0.082
	(0.005)	(0.003)	(0.003)	(0.350)	(0.018)	(0.078)
Per capita arable land area	0.017	0.003	0.017 ***	−1.033	−0.149 ***	0.105
	(0.011)	(0.006)	(0.005)	(0.934)	(0.047)	(0.161)
Flood	0.000	0.001 **	0.001 **	−0.045	−0.006 **	−0.001
	(0.001)	(0.000)	(0.000)	(0.037)	(0.002)	(0.007)
Dry	0.000	0.000 **	0.000	0.028	−0.000	0.003
	(0.000)	(0.000)	(0.000)	(0.021)	(0.001)	(0.004)
Diseases and insect pests	0.000	0.000	0.001	−0.716 ***	0.003	−0.004
	(0.002)	(0.001)	(0.001)	(0.272)	(0.006)	(0.020)
Hybrid rice	−0.000 ***	0.000 ***	0.000	−0.019 ***	−0.001 ***	0.004 ***
	(0.000)	(0.000)	(0.000)	(0.006)	(0.000)	(0.001)
Mid-late rice	−0.000	0.002 ***	0.001 ***	0.112 **	−0.004 **	0.029 ***
	(0.001)	(0.000)	(0.000)	(0.052)	(0.002)	(0.006)
Indica rice	0.000	−0.000 **	−0.000 ***	−0.036 ***	0.000	−0.002
	(0.000)	(0.000)	(0.000)	(0.010)	(0.000)	(0.002)
R2	0.425	0.526	0.172	0.120	0.122	0.453

Note: The standard deviation in parentheses, “*”, “**” and “***” indicate significant at the level of 10%, 5%, and 1%, respectively.

. In all adopted rice variety models, most of the coefficients of the time trend variables were significant, which shows that within the study time range (1982 to 2016), there were significant changes in the five agronomic traits. The annual growth rates of the adopted rice varieties’ yield, plant height, growth period, bacterial blight resistance, and rice blast resistance were 0.9%, 0.3%, 0.2%, 0.5%, and 8.1%, respectively, while the annual decreasing rate of the head rice rate was 0.11%. Meanwhile, due to the lag in the adoption of new varieties by farmers, this study assumed that the farmers’ decision to adopt new varieties in that year was affected by the economic and noneconomic factors of the previous year. Therefore, we considered the first-order lag period for the control variables of yield, plant height, growth period, head rice rate, bacterial blight, and rice blast resistance.

The subsidy policy for improved varieties had a significantly positive coefficient for head rice rate, significantly negative coefficients for yield and rice blast resistance, and a nonsignificant coefficient for plant height, growth period, and bacterial blight resistance grade. This implies that after the implementation of the subsidy policy for improved varieties, the head rice rate increased by 1.2% compared with before, while the yield was reduced by 4% and the resistance level of rice blast was 46.3% lower than that before the implementation of the subsidy policy. In other words, the resistance level of rice blast increased by 46.3%. This suggests that the subsidy policy for improved varieties was conducive to the improvement of the traits of farmer adopted rice varieties in China. The possible reason is that when making decisions on the adoption of new varieties taking the yield potential as the only decision variable, farmers take into account first the rice quality which is associated with price and disease resistance, which is then linked to yield stability of rice.

The impact of the direct subsidy policy for planting grains on the adopted rice varieties has been shown to improve the quality of rice but it has failed to significantly increase the yield potential and disease resistance.

Rice variety types also had significant effects on the traits of adopted rice varieties. For every 1% increase in the proportion of hybrid rice, the rice blast resistance grade increased by 0.4% and the head rice rate and bacterial blight resistance grade decreased by 0.02% and 0.1%, respectively. For every 1% increase in the proportion of mid-late rice, plant height, growth period, head rice rate, and rice blast resistance increased by 0.2%, 0.1%, 0.1%, and 2.9%, respectively, and bacterial blight resistance grade decreased by 0.4%. The coefficients of the proportion of indica rice on plant height, growth period, and head rice rate were all significantly negative.

5. Conclusions

This study employed data on national-level and provincial-level approved rice varieties and adopted rice varieties over the past four decades to investigate the trait changes and differences of approved and adopted rice varieties in China. It also analyzed the influence of variety approval standards on the traits of approved rice varieties and the influence of national subsidies for superior seed varieties on the traits of adopted rice varieties.

The results showed that the yield, plant height, growth period, and head rice rate traits of the approved and adopted varieties were significantly improved while the bacterial blight and rice blast resistance increased in the 1980s but showed a downward trend after the 1990s. Moreover, the yield gap between approved varieties and adopted varieties became larger while the differences of plant height and growth period remained unchanged. Traits of rice quality and rice disease resistance presented a convergence trend between approved varieties and adopted varieties, which indicates a smaller gap between approved varieties and adopted varieties.

Crop variety approval standards play an important leading role in the direction of rice variety breeding and significantly improve the rice quality and resistance of approved varieties. The improvement of variety approval standards is conducive to promoting the selection of rice varieties and guiding the direction of rice variety trait improvement.

To a certain extent, the subsidy policy for improved varieties has significantly improved the quality of the adopted varieties but has failed to effectively promote the improvement of traits such as yield, plant height, and growth period. The head rice rate and rice blast resistance have increased by 1.2% and 46.3%, respectively, compared with before the implementation of the policy, which shows that the subsidy policy has been beneficial to the improvement of the traits of rice varieties in China. Therefore, improving the rice subsidy policy should help encourage farmers to adopt new varieties, increase farmers’ willingness to grow grain, convert the potential productivity of new varieties into actual productivity, and ensure the increase and stability of rice production.