Maximum Residue Limits and Agricultural Trade: Evidence from China

Abstract

In recent years, maximum residue limit (MRL) standards regulated by governments have become increasingly stringent. MRL standards aim to protect human health and environmental sustainability by setting maximum allowable residue concentrations for chemical substances in imported products. In November 2020, the European Union published EU Regulation 2020/1085, imposing stricter MRL standards on imported agricultural products. As standard-setting leaders implement more restrictive MRLs and export-oriented countries adjust their chemical practices to align with these standards, it is essential to examine the emerging trade patterns associated with MRLs and exports. This paper investigates the impact of maximum residue limits (MRLs) on China’s agricultural exports, using disaggregated product-level data (HS eight-digit) for 237 products across 43 importing countries from 2006 to 2021. The results show that a 10% more stringent MRL at the mean in the previous year leads to a 6.6% decrease in the current-year export values of China’s agricultural products. Stricter MRLs for insecticides significantly reduce China’s agricultural exports, while MRLs for fungicides and herbicides have no significant effect. Additionally, I analyze the top 10 most frequently notified chemical substances in goods from China, as reported by the European Commission’s Rapid Alert System for Food and Feed (RASFF) over 2020–2024. More stringent MRLs for Chlorpyrifos, Diafenthiuron, and Carbendazim significantly impede China’s agricultural exports to the European Union. Conversely, China’s agricultural exports to USMCA, ASEAN, and BRI member countries are sensitive to more restrictive MRLs for Diafenthiuron, while benefiting from tougher MRLs for Tolfenpyrad, Dinotefuran, and Imidacloprid. The impacts of MRLs on the extensive and intensive margins of exports suggest that more restrictive MRLs do not significantly affect the probability of exporting, but they reduce export values conditional on the export decisions. The findings suggest that tightening MRL standards by importing countries do not necessarily impede agricultural trade; in some cases, they facilitate China’s agricultural exports to these markets. This study provides policymakers with a comprehensive analysis to formulate strategies for enhancing agricultural exports and highlights the role of food safety standards in sustainable development.

1. Introduction

The growing concern for human health and environmental protection has prompted policymakers to reform food safety standards, particularly those related to maximum residue limits (MRLs). MRLs for chemical substances are regulated by governments to ensure that imported agricultural products contain limited residues of specific substances, not exceeding the maximum allowable concentrations. By setting these limits on chemical substances such as fungicides, herbicides, and insecticides, MRL standards aim to protect human health and promote environmental sustainability.

Previous studies examining the impact of MRLs on trade have generated inconclusive results. On one hand, more stringent MRL standards may enhance consumer trust, boosting demand for imported agricultural products—this is referred to as the “demand-enhancing effect”. On the other hand, tighter MRLs increase compliance costs for exporters, potentially discouraging trade—referred to as the “trade-cost effect”. Xiong and Beghin [1] point out that stricter MRLs facilitate trade if the demand-enhancing effect outweighs the trade-cost effect, but they can impede trade when the trade-cost effect dominates.

In the empirical analysis, the trade effects of MRLs vary by importing country, product category, and the specific chemical substance under investigation. Chen et al. [2] found that a 10% increase in the stringency of MRLs for Chlorpyrifos (an insecticide) was associated with 2.8% decrease in the export values of China’s vegetables. Tran et al. [3] examined the impact of MRLs for Chloramphenicol (an antibiotic) on crustacean imports to the U.S., Canada, Japan, and EU countries, concluding that stricter MRLs impede crustacean trade. Xiong and Beghin [4] found no significant impact of MRLs for Aflatoxin on exports of edible groundnuts, groundnut oil, and shelled groundnut from 9 African countries to 14 European countries. Shingal et al. [5] adopted the random growth first difference model and find the positive effects of MRLs on the trade of vegetables and fruits. Similarly, Hejazi et al. [6] found positive trade effects for U.S. exports of fruits and vegetables to CPTPP member countries, attributing this to stricter MRLs for herbicides.

As MRL standards have become more stringent in recent years, particularly among standard-setting leaders, export-oriented countries have updated the chemical substances applied to agricultural products. In November 2020, the European Union published EU Regulation 2020/1085 (regarding maximum residue levels for chlorpyrifos and chlorpyrifos-methyl in or on certain products, Brussels, European Commission, 2020), which implemented stricter MRL standards to protect human health, environmental sustainability, and animal welfare. For instance, Wei et al. [7] used data on China’s tea exports from 1996 to 2009, and found that a 1% increase in the stringency of MRLs for endosulfan and fenvalerate led to a 0.06% and 0.16% decrease in tea exports, respectively. However, the European Commission’s Rapid Alert System for Food and Feed (RASFF) (RASFF website: https://webgate.ec.europa.eu/rasff-window/screen/search?event=SearchForm&cleanSearch=1 (accessed on 21 January 2025)) has not reported any alert notifications for residue limits of endosulfan in China’s tea shipments to the EU in the past five years (2020–2024), and only one alert has been issued for residue limits of fenvalerate in jasmine infusion. Specifically, on 20 December 2023, Belgium reported an alert notification on residue limits for 19 chemical substances in jasmine infusion sourced from China; fenvalerate was one of the substances detected.

In this study, I investigate the impacts of MRLs on China’s agricultural exports using disaggregated product-level data (HS eight-digit) for 237 products across 43 importing countries from 2006 to 2021. The chemical substances have been categorized as fungicides, herbicides, and insecticides. A detailed heterogeneity analysis is conducted by product category, type of chemical substance, and the Regional Trade Agreement (RTA) membership of importing countries. I find that a 10% more stringent MRL at the mean in the previous year leads to a 6.6% decrease in the current-year export values of China’s agricultural products. Stricter MRLs for insecticides significantly reduce China’s agricultural exports, while the restrictiveness of fungicides and herbicides has no significant effect on exports. Tightening MRLs significantly impedes exports to Association of Southeast Asian Nations (ASEAN), Belt and Road Initiative (BRI), European Union (EU), and Organization for Economic Co-operation and Development (OECD) countries, while facilitating exports to United States–Mexico–Canada Agreement (USMCA) countries.

Additionally, I collect all import rejections and alert notifications reported by the RASFF relating to exceeded residue limits in goods from China and identify the top 10 most frequently notified substances. The substances with the highest frequency of rejection at the EU border were selected, as they indicate potential weaknesses in food safety standards and regulatory conformity in China’s agricultural sector. Whether China diverts exports of agricultural products to destination markets with more lenient MRL standards or continues to serve countries with stricter MRL standards remains an empirical question.

A heterogeneity analysis is then performed to explore the impacts of MRLs for these ten chemical substances on China’s agricultural exports. The findings reveal several of the following key insights: (i) MRLs for Diafenthiuron and Pyraclostrobin reduce agricultural exports; (ii) MRLs for Tolfenpyrad and Imidacloprid promote agricultural exports; (iii) MRLs for Acetamiprid do not significantly affect the export values of agricultural products; (iv) MRLs for Chlorpyrifos, Lambda-cyhalothrin, Dinotefuran, Carbendazim, and Pyridaben have heterogeneous impacts on exports to various RTA member countries.

To further examine the impacts of MRLs on the extensive margin (the probability of exporting) and intensive margin (the export values conditional on the export decision), I employ the Heckman selection model. Drogué and DeMaria [8] found that larger dissimilarities in MRLs between country pairs significantly affected the intensive margin while not significantly influencing the extensive margin of trade for apples and pears. In contrast, Ferro et al. [9] reported the negative impact of more stringent MRLs on the probability of exporting vegetables and fruits, although they found no significant effects on the export intensity. In this analysis, I find that the MRL stringency does not significantly affect the probability of exporting, but significantly reduces the export value condition on the export decision. A higher number of MRLs that are stricter than China’s negatively impacts the extensive margin, with no significant effect on the intensive margin. Heterogeneity results for both the extensive and intensive margins of exports are examined by MRLs for fungicides, herbicides, and insecticides, as well as by the RTA membership of importing countries.

The objectives of this research are to examine the impacts of MRLs on China’s agricultural exports, providing policymakers with a detailed analysis to formulate strategies for facilitating agricultural exports in the face of increasingly restrictive MRL standards. The purpose of this study is to answer the following three empirical questions: (i) Does the restrictiveness of MRL standards regulated by importing countries impede China’s exports of agricultural products? (ii) As standard-setting leaders implement more stringent MRLs and export-oriented countries adjust their chemical practices to align with these standards, what is the new trade pattern that emerges? (iii) How do tightening MRL standards affect the probability of exporting and the average export values upon export decisions?

This study has three contributions. First, this study enhances the existing literature by utilizing comprehensive export data covering 237 products at the HS eight-digit level, 43 importing countries, and 399 chemical substances from 2006 to 2021. These highly disaggregated data provide a unique opportunity to examine the impacts of MRLs on agricultural exports, considering the RTA membership of importing countries, chemical substances, and product categories. Second, this study offers policymakers a detailed and updated analysis of the trade effects of MRLs, particularly examining the top ten most frequently notified chemical substances reported by the RASFF. The empirical findings can guide both standard-setting leaders and export-oriented countries, especially considering recent efforts by exporters to align MRL standards with those set by standard-setting leaders. Third, the extensive and intensive margins of exports are explored, shedding light on the trade patterns of exporters facing more restrictive MRL standards.

This article is structured as follows. Section 2 presents the literature review on the impacts of MRL standards on trade. Section 3 specifies the empirical model and introduces the data. Results are discussed in Section 4, along with a heterogeneous analysis, the effects of MRL stringency on the extensive and intensive margins of exports, and robustness checks. Section 5 concludes.

2. Literature Review

The impact of food safety standards on trade has been a subject of considerable investigation over the past decade. Many scholars have explored the role of MRLs in facilitating or impeding trade, with mixed conclusions. The empirical findings vary depending on the chemical substances involved, the agricultural products, and the country pairs examined.

Chen et al. [2] examined the impact of MRLs for Chlorpyrifos on China’s aquatic exports during the period 1992–2004. Their results showed that a 10% stricter MRL for Chlorpyrifos led to a 2.8% decrease in China’s export values of vegetables, and decreases of 3.2%, 2.1%, and 10.0% for garlic, onion, and spinach exports, respectively. Tran et al. [3] investigated the impact of MRLs for Chloramphenicol standards on crustacean imports to the U.S., Canada, Japan, and the EU15 from 2001 to 2008. They found that a 1 ppb (parts per billion) decrease in the MRL standard was associated with a 4% reduction in imports from the top thirty crustacean exporters. Wei et al. [7] analyzed the impact of MRLs for endosulfan, fenvalerate, and flucythrinate on China’s tea exports from 1996 to 2009. Their findings indicated that a 1% more restrictive MRL for endosulfan and fenvalerate would result in a 0.06% and 0.16% decrease in tea exports, respectively, whereas the effect of MRLs for flucythrinate was insignificant.

Some studies have found positive trade effects associated with MRL standards. Hejazi et al. [6] investigated the effects of MRL stringency on fruit and vegetable trade flows in 2013–2014. They found that an increase in the MRL index by 0.1 at the mean (about 10.4%) decreased fruit and vegetable exports by 8.8%. Additionally, MRLs for insecticides and fungicides negatively affected U.S. exports of fruits and vegetables to the EU, while MRLs for herbicides significantly increased these exports. Shingal et al. [5] utilized data from 53 importing countries on fruits and vegetables from 2005 to 2014, finding that a 1% more stringent MRL imposed by importers relative to exporters was associated with a 46.9% increase in bilateral trade.

The extensive and intensive margins of trade have also been empirically examined in the literature. Ferro et al. [9] analyzed the impact of MRLs on the intensive and extensive margins of trade using data on fruits and vegetables across 61 importing countries. Their findings suggested that MRLs had a negative and significant impact on the extensive margin of exports but did not significantly affect the intensive margin. DeMaria and Drogue [10] measured the impact of the MRL index on EU imports of baby food from 2008 to 2010, finding that more stringent MRL standards facilitated imports, increased the probability of trade, and reduced the volume of trade. Kareem et al. [11] examined the impact of MRL standards and EU entry price control on Africa’s exports of tomatoes, oranges, and limes and lemons to the EU from 2008 to 2013. They found that a 1% increase in the stringency of MRLs for insecticides would reduce the probability of exporting tomatoes to the EU by 6.0%, oranges by 4.0%, and limes and lemons by 6.9%. Fiankor et al. [12] used trade data from 2005 to 2014 for 145 products across 59 countries to measure the impact of MRLs on agri-food trade. Their results indicated that an increase in the MRL index by 0.1 units at the mean (about 9%) reduced the extensive margin by 0.7%, the intensive margin by 0.6%, and overall trade flows by 0.82%. Rabadan and Triguero [13] evaluated the impact of MRLs for aflatoxin on U.S. pistachio exports to Iran during the period 1996–2014. They found that more restrictive aflatoxin MRLs reduced the probability of exporting pistachio while increasing the export value of pistachio condition on exporting.

While some studies have found heterogeneous effects of MRLs on the extensive and intensive margins of trade, other research suggests positive or insignificant effects. Traoré and Tamini [14] analyzed data on 12 African countries that exported mangoes to 31 OECD countries in 2016, concluding that a 1% increase in the mean value of MRL stringency would generate an average increase of 0.05% in the likelihood of exporting and 1.39% increase in the average export values upon export decisions. Xiong and Beghin [4] examined the trade effects of the EU’s tightening of MRLs for aflatoxin in 2002 using data on edible groundnuts, groundnut oil, and shelled groundnuts traded between 14 European countries and 9 African countries. They found that the tightening of MRLs for aflatoxin had no significant impact on the extensive or intensive margins of exports.

Additionally, the similarity of MRL standards between exporting and importing countries has also been a focus of study, with several papers exploring its relationship with trade. Drogué and DeMaria [8] investigated the impact of the similarity of insecticide MRLs between importing and exporting countries on trade in apples, pears, and related processed products. They found that the harmonization of food safety standards had a significant effect on the intensive margin of trade but no significant effect on the extensive margin. Winchester et al. [15] examined the impact of MRL dissimilarities on trade between the EU and nine countries during 2009–2010, finding that an increase in the MRL index by 0.01 at the mean (about 2.6%) was associated with a 5% reduction in trade flows.

Several studies have also explored the determinants of MRL standards. Karemera et al. [16] found that countries with higher public health spending tend to adopt stricter MRLs, while countries with comparative advantages often set more lenient MRLs. Their findings also suggested that products subject to lower tariffs are associated with stricter MRLs. Fiankor et al. [17] concluded that stricter importer standards are trade restrictive, particularly for countries with smaller trade volumes.

In recent years, much attention has been paid to import rejections, especially in relation to the RASFF database. Beestermoller et al. [18] investigated Chinese firms exporting HS-4 agri-food products to the RASFF border and which were rejected in the past period on the Chinese firms’ entry, exit, and number of incumbent firms during the period 2000–2011. They found that rejection at the RASFF border in the last year would increase the probability of Chinese firms that export rejected HS-4 products exiting this market by 1.6%. Additionally, a 1% increase in the number of rejections at the RASFF border in the past was associated with a 25.3% decrease in the number of Chinese firms exporting to the RASFF market. Sun et al. [19] explored the impact of China’s import refusals on agricultural imports, finding that a 1% increase in the number of import refusals in the current period decreased the growth rate of the export value by 1.1%. The negative impact of import refusals was more pronounced in countries with lower food safety standards. Pigłowski and Niewczas-Dobrowolska [20] utilized the RASFF database prior to 2021 and found that the notifications primarily concerned pesticide residues in peppers from Turkey, okra, curry, and rice from India, as well as tea from China and India. Schaefer et al. [21] found that the negative impacts of RASFF fraud notifications on EU imports were even more severe following a food scandal.

The literature on RASFF data indicates that EU border rejections have a negative and significant impact on China’s agricultural exports. Furthermore, Chinese border refusals have been shown to reduce agricultural imports. Baylis et al. [22] pointed out that the number of notifications is positively associated with tariff rates, with a 1% decrease in tariffs leading to a 23% increase in the number of notifications. They also found that notifications banned at the border for low-hazard products were more influenced by tariff reductions than notifications for mid- or high-hazard products. He [23] concluded that market power affects countries’ export restriction decisions, suggesting that a one standard deviation increase in a commodity’s market power increases the probability of an export restriction by 5.5%.

Several pieces of work have also utilized firm-level data to examine the relationship between food safety standards and trade. Medin [24] used firm-level data on Norwegian seafood exports between 2002 and 2011, showing that SPS measures reduced the total exports of non-fresh seafood products by 47%, while increasing the exports of fresh seafood by 58%. Curzi et al. [25] analyzed firm-level customs data on Peruvian agri-food exports from 2000 to 2014, finding that Sanitary and Phytosanitary measures (SPS) and Technical Barriers to Trade (TBT) have positive impacts on both the EM and IM. Macedoni and Weinberger [26] found that strict standards would increase the likelihood of small firms exiting the market, while larger firms tend to benefit from resource transfers.

3. Estimation Strategy and Data

3.1. Empirical Model

Following Chen et al. [27], I employ a gravity model to estimate the impacts of MRLs on China’s agricultural exports, as follows:

$$\begin{array}{ll}& ln{Y}_{jkt}={\beta }_0+{\beta }_1{MR{L}_{\_restrictiveness}}_{jkt-1}+{\beta }_2\ln \left(1+{Number}_{jkt-1}\right)+{\beta }_3\ln \left(1+{Tariff}_{jkt}\right)+{\alpha }_{jk}+\\ & {\alpha }_{kt}+{\alpha }_{jt}+{\epsilon }_{jkt}\end{array}$$

(1)

where $ln{Y}_{jkt}$ is the logarithm of China’s annual exports of agricultural product $k$ at the HS 8-digit level to importing country $j$ in year $t$.

The main variable of interest, the restrictiveness index of MRLs, is defined as

$${MRL\_restrictiveness}_{jkt-1}={\displaystyle \frac{1}{N_s}}({\sum }_{sϵN_{kt-1}}\mathrm{e}\mathrm{x}\mathrm{p}\left({\displaystyle \frac{{MRL\_CHINA}_{\mathrm{s}kt-1}-{MRL\_IMPORTER}_{\mathrm{s}kt-1}}{{MRL\_CHINA}_{\mathrm{s}kt-1}}}\right))$$

(2)

where ${MRL\_CHINA}_{\mathrm{s}kt-1}$ is the MRL regulated by China for substance $s$ on product $k$ in year $t-1$; ${MRL\_IMPORTER}_{\mathrm{s}kt-1}$ is the MRL regulated by the importing country for substance $s$ on product $k$ in year $t-1$; $N_s$ is the number of chemical substances regulated for product $k$ in year $t-1$.

Following Li and Beghin [28], I employ a non-linear exponential index to capture the restrictiveness of MRLs. The main variable of interest, ${MRL\_restrictiveness}_{jkt-1}$, represents the average intensity of an importer’s MRL stringency relative to China’s MRLs for a specific product $k$ in year $t-1$. The index is defined on the domain $(0, e\approx 2.718]$, where a higher MRL restrictiveness index indicates that the importer regulates stricter MRLs than China for the specific product in the previous year. For instance, when the importer sets the MRLs at zero (the strictest level), the MRL restrictiveness index reaches its upper limit of $\mathrm{e}\mathrm{x}\mathrm{p}\left(1\right)\approx 2.718$. Conversely, when the importer imposes a very large (lax) MRL, the MRL restrictiveness index approaches its lower limit of $\mathrm{e}\mathrm{x}\mathrm{p}(-\mathrm{\infty })\approx 0$.

The parameter of the main variable of interest, ${\beta }_1$, implies the impacts of the average intensity of MRL stringency on China’s agricultural exports. If the estimate of ${\beta }_1$ is positive and significant, this indicates that more restrictive MRLs facilitate Chinas’ exports of agricultural products. In contrast, a negative and significant ${\beta }_1$ suggests that more stringent MRLs are associated with smaller export values of China’s agricultural products.

In addition, the number of tougher-than-China MRLs and tariffs are controlled in the analysis. ${Number}_{jkt-1}$ is the number of chemical substances with tougher-than-China MRLs regulated on product $k$ by importing country $j$ in year $t-1$. Since multiple chemical substances may apply to a given product, variations in China’s agricultural exports may result from changes in the number of MRLs imposed by importing countries rather than the average intensity of the MRL stringency for a specific chemical substance. Including the number of tougher-than-China MRLs contributes to more effectively capturing the trade effects of the MRL restrictiveness on average. Furthermore, the effectively applied tariff, denoted as ${Tariff}_{jkt}$, is also controlled in the analysis.

To reduce concern of endogeneity due to reverse causality, I follow Fontagné et al. [29] by using a one-year lag for both the MRL restrictiveness variable and the number of substances stricter than China’s MRLs. This gravity framework is vulnerable to the endogeneity resulting from the omission of unobserved confounding factors that may be correlated with lagged MRL restrictiveness and export values. To address this, I include importer–product (${\alpha }_{jk}$), product–year (${\alpha }_{kt}$), and importer–year (${\alpha }_{jt}$) fixed effects, as suggested by Anderson and Yotov [30,31]. The product fixed effect is defined for each two-digit product category of $k$ to avoid multicollinearity. ${\epsilon }_{jkt}$ is the error term. The gravity estimations are implemented by the Poisson pseudo-maximum likelihood (PPML), where the disturbance term is assumed to follow a Poisson distribution. The PPML estimation leads to consistent estimation in the presence of heteroskedasticity and applies well to trade data with a considerable number of zeros [32,33].

3.2. Data

The MRL data are obtained from Homologa, a global regulatory database that compiles information on maximum residue limits from government sources. The MRL database includes information on 2692 chemical substances, 1459 products, and 75 countries and regions.

I manually match the MRLs with China’s annual export values of agricultural products at the importer–year–substance–product (HS 8-digit) level based on product descriptions. For instance, when the MRL data list onions, they are manually matched to the corresponding HS 8-digit code (e.g., 07031010). Matching the MRLs and trade data at the most disaggregated level (HS 8-digit) is crucial because multiple MRLs may apply within a single HS 6-digit product category. For example, both onions and shallots fall under the same HS 6-digit code (070310). However, Australia set an MRL of 0.2 ppm (parts per million) for onions and 0.01 ppm for shallots in 2015. Therefore, matching the MRL and trade data at the HS 8-digit level helps to precisely capture the trade effects of MRLs and offers a unique opportunity to explore the relationship between MRLs and exports.

Two rules are applied to address missing MRL standards. First, following Ferro et al. [9], observations are excluded if fewer than 30% of importing countries have regulated MRL standards for a specific chemical substance on product $k$ in year $t$. Second, missing MRLs for a specific substance are replaced with the most lenient MRL standard under the same HS 6-digit heading among all importing countries in year $t$. This approach for handling missing MRL values follows Drogué and DeMaria [8] and Fiankor et al. [12], who posit that substituting zeros for missing MRL values is inappropriate, as zero MRLs represent the strictest MRL standards. Finally, the data sample consists of 237 agricultural products, 43 importing countries, and 399 chemical substances (the list of agricultural products at the HS 2-digit heading is detailed in Appendix A 

Table A1. HS chapters.

HS2	Product Description	Number of HS 8-Digit Products Under 2-Digit Sector	Product Category
02	Meat and edible meat offal	41
04	Dairy produce; birds’ eggs; natural honey; edible products of animal origin not elsewhere specified or included	53
07	Edible vegetables and certain roots and tubers	10,954	Vegetables and fruits
08	Edible fruit and nuts; peel of citrus fruit or melons	6386
09	Coffee, tea, mate, and spices	2479	Coffee and tea
10	Cereals	993	Cereals and oilseeds
11	Products of the milling industry; malt; starches; inulin; wheat gluten	571
12	Oil seeds and oleaginous fruits; miscellaneous grains, seeds, and fruit; industrial or medicinal plants; straw and fodder	4763
14	Vegetable plaiting materials; vegetable products not elsewhere specified or included	282	Preparations and beverages
15	Animal or vegetable fats and oils and their cleavage products; prepared edible fats; animal or vegetable waxes	120
16	Preparations of meat, fish or crustaceans, or aquatic invertebrates	6
19	Preparations of cereals, flour, starch, or milk; pastrycooks’ products	367
20	Preparations of vegetables, fruit, nuts, or other parts of plants	4617
22	Beverages, spirits, and vinegar	360

). The sample at the importer–year–product level includes 31,992 observations, of which 66.7% are zero export values.

Tariff data are sourced from the UNCTAD Trade Analysis Information System (TRAINS) and the WTO’s Integrated Database (IDB).

Table 1. Summary statistics.

Variable	Description	Obs	Mean	SD	Min	Max
${MRL\_restrictiveness}_{jkt-1}$	The average intensity of the MRL stringency for all chemical substances regulated on product $k$ by country $j$ at time $t-1$	31,992	1.257	0.570	1 × 10−43	2.705
${Number}_{jkt-1}$	The number of substances with tougher-than-China MRLs regulated on product $k$ by country $j$ at time $t-1$	31,992	16.661	16.150	0	107
${Tariff}_{jkt}$	The effectively applied tariff imposed by country $j$ applied to product $k$ at time $t$ (%)	31,992	10.246	28.067	0	2001.667
${Fungicide}_{jkt-1}$	The average intensity of the MRL stringency for fungicides regulated on product $k$ by country $j$ at time $t-1$	25,698	0.250	0.221	2 × 10−19	2.586
${Herbicide}_{jkt-1}$	The average intensity of the MRL stringency for herbicides regulated on product $k$ by country $j$ at time $t-1$	24,734	0.118	0.169	5 × 10−45	2.460
${Insecticide}_{jkt-1}$	The average intensity of the MRL stringency for insecticides regulated on product $k$ by country $j$ at time $t-1$	30,925	0.998	0.525	3 × 10−29	2.705

shows the summary statistics. The mean value of the MRL restrictiveness index is 1.257, indicating that China’s agricultural exports are exposed to moderate MRL standards regulated by importing countries, since the index ranges from 0 to $e (\approx 2.718)$. The data include 100 fungicides, 113 herbicides, and 157 insecticides. The MRL index for fungicides, herbicides, and insecticides implies that the average intensity of the restrictiveness is the highest for insecticides, followed by fungicides and herbicides.

Figure 1 and Figure 2 show the MRL restrictiveness and the number of tougher-than-China MRLs regulated by RTA member countries during the period 2005–2020, respectively. The MRL restrictiveness of the EU is the most stringent, followed by OECD, CPTPP, ASEAN, BRI, and USMCA countries. The average MRL stringency of the EU and OECD countries was higher than that of China from 2006 to 2020. In contrast, the MRL restrictiveness of CPTPP, ASEAN, BRI, and USMCA countries was more lenient than China’s during the sample period. In addition, the number of tougher-than-China MRLs regulated by all RTA member countries grew rapidly during the following three periods: 2005–2007, 2012–2017, and 2019–2020.

Figure 3 and Figure 4 display the MRL restrictiveness and the number of tougher-than-China MRLs for fungicides, herbicides, and insecticides across 43 importing countries between 2005 and 2020. The average MRL stringency for insecticides is the highest, followed by fungicides and herbicides. The MRLs for insecticides regulated by importing countries were more restrictive than China’s between 2011 and 2013, and again from 2016 onward. However, the MRLs for fungicides and herbicides were less strict than China’s throughout the sample period. The number of tougher-than-China MRLs for insecticides peaked in 2017 at 20, declined to 13 between 2018 and 2019, and rose back to 20 in 2020. The number of tougher-than-China MRLs for fungicides and herbicides averaged four and two, respectively, during the sample period.

4. Results

4.1. Baseline Results

Table 2. Baseline results.

	All	Fungicide	Herbicide	Insecticide	USMCA	ASEAN	BRI	CPTPP	OECD	EU
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
${MRL\_restrictiveness}_{jkt-1}$	−0.522 ***	0.371	0.701	−0.877 ***	0.540 **	−1.361 ***	−0.992 **	−0.146	−0.180 *	−0.349 *
	(0.144)	(0.235)	(0.440)	(0.202)	(0.256)	(0.525)	(0.468)	(0.121)	(0.103)	(0.179)
$\ln \left(1+{Number}_{jkt-1}\right)$	−0.043	0.265 **	0.370 ***	−0.091	−0.386 **	0.519 **	0.482 ***	−0.357 ***	−0.408 ***	−0.422 ***
	(0.071)	(0.120)	(0.135)	(0.076)	(0.155)	(0.208)	(0.179)	(0.060)	(0.047)	(0.073)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.303 ***	−0.348 ***	−0.272 ***	−0.292 ***	−0.517 ***	−0.369 ***	−0.200 **	−0.479 ***	−0.347 ***	−0.158
	(0.052)	(0.059)	(0.059)	(0.054)	(0.127)	(0.119)	(0.097)	(0.100)	(0.068)	(0.107)
Constant	16.628 ***	15.875 ***	15.888 ***	16.858 ***	15.991 ***	16.585 ***	16.249 ***	17.461 ***	16.941 ***	16.239 ***
	(0.155)	(0.130)	(0.096)	(0.111)	(0.437)	(0.279)	(0.277)	(0.231)	(0.181)	(0.336)
Observations	27,279	21,651	20,796	26,570	1995	4286	4777	4882	20,500	13,074
R-squared	0.469	0.499	0.484	0.480	0.337	0.453	0.415	0.478	0.501	0.429
Importer–product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Importer–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Notes: Standard errors are in parentheses and are clustered by importer–product–year. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively.

shows the estimation results for the model specified in Equation (1), assessing the impacts of the MRL restrictiveness on China’s agricultural exports. Column (1) reports the PPML estimation for all of the chemical substances and importing countries. Columns (2)–(4) present results for fungicides, herbicides, and insecticides, respectively. Columns (5)–(10) show the impacts of the MRL restrictiveness on China’s agricultural exports to the following RTA member countries: United States–Mexico–Canada Agreement (USMCA), Association of Southeast Asian Nations (ASEAN), Belt and Road Initiative (BRI), Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP), Organization for Economic Co-operation and Development (OECD), and the European Union (EU). The list of RTA member countries is detailed in Appendix A 

Table A2. Member countries of RTAs.

RTA	Full Name of RTA	Country
USMCA	United States–Mexico–Canada Agreement	United States, Canada, Mexico
ASEAN	Association of Southeast Asian Nations	Argentina, Chile, India, Malaysia, Singapore, Thailand, Vietnam
BRI	Belt and Road Initiative	Argentina, Brazil, Colombia, India, Malaysia, Russia, Ukraine, Vietnam, South Africa
CPTPP	Comprehensive and Progressive Agreement for Trans-Pacific Partnership	Australia, Canada, Chile, Japan, South Korea, Mexico, New Zealand
OECD	Organization for Economic Co-operation and Development	Australia, Austria, Belgium, Canada, Switzerland, Czech Republic, Germany, Denmark, Spain, Estonia, France, United Kingdom, Greece, Hungary, Ireland, Israel, Italy, Japan, South Korea, Luxembourg, Mexico, Malta, Netherlands, Norway, New Zealand, Poland, Portugal, Slovakia, Sweden, Turkey, United States
EU	European Union	Austria, Belgium, Czech Republic, Germany, Denmark, Spain, Estonia, France, United Kingdom, Greece, Hungary, Ireland, Italy, Luxembourg, Malta, Netherlands, Poland, Portugal, Slovakia, Sweden

.

The results in column (1) suggest that a 10% increase in the importer’s MRL restrictiveness at the mean in the previous year leads to a 6.6% (${\displaystyle \frac{52.2\%}{(1/1.257)\ast 100\%}}\ast 10\%$) decrease in the current-year export values of China’s agricultural products. The calculation follows Chen et al. [27], and the interpretations of the 10% increase in the MRL restrictiveness index is provided in their footnote 30. In addition, tariffs significantly impede exports, while the number of chemical substances imposed by the importer on a specific product in the previous year does not significantly influence exports.

In columns (2)–(4), the results imply that stricter MRLs for insecticides significantly reduce the export values of China’s agricultural products, while the MRL restrictiveness of fungicides and herbicides has no significant effect on export values. In contrast, the number of fungicides and herbicides is positively associated with China’s agricultural exports, while the number of insecticides does not significantly affect exports. Previous literature, such as Hejazi et al. [6], found that the MRL restrictiveness for insecticides and herbicides impeded U.S. exports of fruits and vegetables, while MRLs for fungicides did not significantly impact U.S. exports.

Stricter MRL standards significantly impede China’s agricultural exports to ASEAN, BRI, the EU, and OECD countries. A novel finding is that more restrictive MRL standards facilitate the exports of agricultural products to USMCA member countries. On the other hand, the number of chemical substances regulated on agricultural products reduces China’s exports to USMCA, CPTPP, OECD, and EU countries, while encouraging larger exports to ASEAN and BRI member countries.

Sun et al. [19] conclude that the negative impacts of food safety standards on imports are greater for exporting countries with a lower food safety index. According to the 2015 Global Food Safety Index, China ranked 49th ([19], Table A2). Most EU, CPTPP, OECD, and USMCA member countries have a higher food safety index ranking than China, while some ASEAN and BRI member countries have a lower ranking relative to China (e.g., South Africa, Colombia, Vietnam, and India). This suggests a potential trade pattern in which China may divert its agricultural exports to destination markets with relatively lower food safety index rankings (e.g., ASEAN and BRI member countries), while reducing exports to RTA member countries with higher food safety index rankings (e.g., EU, CPTPP, OECD, and USMCA member countries). The results in Table 2, regarding the number of substances with tougher MRLs than China, support this trade pattern. In contrast, China’s agricultural products are sensitive to MRL stringency regulated by all RTA member countries, except for the USMCA countries.

4.2. Estimates by Product Category

Table 3. Heterogeneity results by product category.

	Vegetables and Fruits			Coffee and Tea			Cereals and Oilseeds			Preparations and Beverages
	Fungicide	Herbicide	Insecticide	Fungicide	Herbicide	Insecticide	Fungicide	Herbicide	Insecticide	Fungicide	Herbicide	Insecticide
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
${MRL\_restrictiveness}_{jkt-1}$	0.285	1.331 ***	−1.683 ***	−3.456	−35.902	0.476 *	0.865 ***	−0.477	1.400 ***	0.549 **	−0.448	−0.744 **
	(0.329)	(0.353)	(0.389)	(3.233)	(23.483)	(0.276)	(0.282)	(0.522)	(0.414)	(0.255)	(0.526)	(0.335)
$\ln \left(1+{Number}_{jkt-1}\right)$	0.250 *	0.704 ***	0.037	1.818 *	5.521 **	−0.145	0.288 *	−0.455 **	−0.371 ***	0.442 ***	−0.206 *	−0.145
	(0.145)	(0.169)	(0.133)	(0.952)	(2.322)	(0.092)	(0.175)	(0.225)	(0.084)	(0.113)	(0.118)	(0.177)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.427 ***	−0.344 ***	−0.349 ***	0.757	−0.144	0.332	−0.035	−0.166 **	−0.112	−0.297 ***	−0.270 ***	−0.385 ***
	(0.067)	(0.070)	(0.060)	(0.687)	(0.334)	(0.203)	(0.101)	(0.082)	(0.094)	(0.087)	(0.102)	(0.149)
Constant	16.240 ***	15.997 ***	17.403 ***	15.047 ***	14.884 ***	14.885 ***	14.124 ***	15.233 ***	14.329 ***	14.791 ***	15.559 ***	16.596 ***
	(0.147)	(0.128)	(0.112)	(0.628)	(0.238)	(0.395)	(0.295)	(0.116)	(0.326)	(0.229)	(0.190)	(0.582)
Observations	11,684	11,489	14,530	1424	1032	2121	2162	2391	2782	4141	4081	4823
R-squared	0.467	0.462	0.455	0.713	0.719	0.638	0.427	0.413	0.412	0.567	0.594	0.613
Importer–product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Importer–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Notes: Standard errors are in parentheses and are clustered by importer–product–year. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively.

presents the heterogeneity analysis by product categories, including vegetables and fruits (columns 1–3), coffee and tea (columns 4–6), cereals and oilseeds (columns 7–9), and preparations and beverages (columns 10–12). The PPML estimations for each product category are implemented for subgroups of fungicides, herbicides, and insecticides.

The results of columns (1)–(3) show that China’s exports of vegetables and fruits are vulnerable to stringent MRLs for insecticides, while stricter MRLs for herbicides facilitate the exports of vegetables and fruits. A greater number of tougher-than-China MRLs for fungicides and herbicides significantly facilitate China’s exports of vegetables and fruits.

In columns (4)–(6), the results indicate that MRLs for fungicides and herbicides do not significantly affect China’s exports of coffee and tea. However, more restrictive MRLs for insecticides significantly increase exports in this category. Conversely, the number of tougher-than-China MRLs for insecticides has no significant effect on the export values of coffee and tea, while the number of tougher-than-China MRLs for fungicides and herbicides is positively associated with coffee and tea exports.

Columns (7)–(9) present the results for China’s exports of cereals and oilseeds. The findings suggest that export values of cereals and oilseeds benefit from tougher MRLs for both fungicides and insecticides. The number of tougher-than-China MRLs for herbicides and insecticides negatively affects cereals and oilseeds exports, while the number of MRLs stricter than China’s for fungicides is positively associated with export values of these products.

In columns (10)–(12), the results report the export effects of MRLs for fungicides, herbicides, and insecticides on preparations and beverages. While stricter MRLs for insecticides reduce the export values of preparations and beverages, more restrictive MRLs for fungicides promote China’s exports of these products. A greater number of tougher-than-China MRLs for fungicides promote the exports of preparations and beverages, while a larger number of MRLs stricter than China’s for herbicides undermines the export values in this category.

4.3. Estimates by Chemical Substance

To further explore the export effects of specific chemical substances, the priority is to identify which substances have been widely applied by China and have frequently triggered import rejections over the past five years. I collected all of the EU notifications related to exceeding residue limits and resulting in border rejections or alert notifications for agricultural products sourced from China between 2020 and 2024. The data come from the RASFF, which includes 110 notifications related to MRLs. A summary of the chemical substances notified in the RASFF reports is provided in Appendix A 

Table A3. Frequency of chemical substance alerts reported in the RASFF for goods sourced from China, 2020–2024.

Chemical Substance	Frequency
CHLORPYRIFOS	40
TOLFENPYRAD	30
LAMBDA-CYHALOTHRIN	24
DINOTEFURAN	20
ACETAMIPRID	18
CARBENDAZIM	9
DIAFENTHIURON	9
IMIDACLOPRID	8
PYRACLOSTROBIN	7
PYRIDABEN	7
CARBOFURAN	6
THIAMETHOXAM	5
TEBUCONAZOLE	5
ISOPROCARB	5
DIFENOCONAZOLE	5
PROPAMOCARB	4
ANTHRAQUINONE	4
CLOTHIANIDIN	4
TRICYCLAZOLE	4
PROPICONAZOLE	4
PROCHLORAZ	3
FOLPET	3
HEXACONAZOLE	3
ISOCARBOPHOS	2
HEXAFLUMURON	2
BUPROFEZIN	2
CHLORFENAPYR	2
DIMETHOMORPH	2
FLONICAMID	2
TRIAZOPHOS	2
MATRINE	2
IMAZALIL	2
THIOPHANATE-METHYL	2
EMAMECTIN	1
SPIRODICLOFEN	1
CHLORPYRIFOS-METHYL	1
CHLORPYRIFOS-ETHYL	1
FLUOPICOLIDE	1
CHLORANTRANILIPROLE	1
HEXAZINONE	1
BIPHENYL	1
AMETRYN	1
PROPARGITE	1
BIFENTHRIN	1
FENVALERATE	1
DITHIOCARBAMATES	1
FENOBUCARB	1
AZOXYSTROBIN	1
CYPERMETHRIN	1
IPRODIONE	1
HCHGAMMA	1
ABAMECTIN	1
MINERALOIL	1
HCHBETA	1
METHAMIDOPHOS	1
PYRROLIZIDINEALKALOIDS	1
ISOFENPHOS-METHYL	1
TETRAMETHRIN	1
FIPRONIL	1
PROCYMIDONE	1
LUFENURON	1
HCH	1

.

Table 4. Top 10 most frequently reported substances in the RASFF notifications for goods from China, 2020–2024.

Name of Substance	Type of Substance	Notification	Border Rejection		Alert Notification		Serious Risk
		Frequency	Frequency	%	Frequency	%	Frequency	%
Chlorpyrifos	insecticide	40	21	52.5	19	47.5	22	55
Tolfenpyrad	insecticide	30	29	96.7	1	3.3	7	23.3
Lambda-cyhalothrin	insecticide	24	23	95.8	1	4.2	4	16.7
Dinotefuran	insecticide	20	20	100	0	0	6	30
Acetamiprid	insecticide	18	11	61.1	7	38.9	9	50
Diafenthiuron	insecticide	9	8	88.9	1	11.1	3	33.3
Carbendazim	fungicide	9	2	22.2	7	77.8	7	77.8
Imidacloprid	insecticide	8	3	37.5	5	62.5	5	62.5
Pyridaben	insecticide	7	6	85.7	1	14.3	6	85.7
Pyraclostrobin	fungicide	7	4	57.1	3	42.9	3	42.9

shows the top 10 most frequently notified substances reported in the RASFF for goods from China. The most frequently notified substance is Chlorpyrifos, with 52.5% of notifications for Chlorpyrifos leading to border rejections. The second, third, and fourth most frequently notified substances are Tolfenpyrad, Lambda-cyhalothrin, and Dinotefuran, each with more than 95% of notifications for these substances resulting in border rejections. In terms of risk classifications, more than 50% of notifications for Acetamiprid, Carbendazim, Imidacloprid, and Pyridaben are identified as serious risks.

The specification for evaluating the impacts of the top 10 most frequently notified chemical substances in the RASFF on China’s export values of agricultural products is given as follows:

$$\begin{array}{ll}ln{Y}_{jkt}={\beta }_0+& {\beta }_1{MRL\_restrictiveness}_{jkt-1}+{\beta }_2{Other\_restrictiveness}_{jkt-1}\\ & +{\beta }_3\ln \left(1+{Number}_{jkt-1}\right)\\ & +{ \beta }_4\ln \left(1+{Tariff}_{jkt}\right)+{ \beta }_5{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{U}\mathrm{S}\mathrm{M}\mathrm{C}\mathrm{A}\\ & +{\beta }_6{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{A}\mathrm{S}\mathrm{E}\mathrm{A}\mathrm{N}+{\beta }_7{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{B}\mathrm{R}\mathrm{I}\\ & +{ \beta }_8{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{C}\mathrm{P}\mathrm{T}\mathrm{P}\mathrm{P}+{\beta }_9{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{O}\mathrm{E}\mathrm{C}\mathrm{D}\\ & +{\beta }_{10}{MRL\_restrictiveness}_{jkt-1}\ast \mathrm{E}\mathrm{U}+{\alpha }_{jk}+{\alpha }_{kt}+{\alpha }_{jt}+{\epsilon }_{jkt}\end{array}$$

(3)

where ${MRL\_restrictiveness}_{jkt-1}$ is the MRL restrictiveness for each specific substance under investigation; ${Other\_restrictiveness}_{jkt-1}$ represents the average MRL stringency across substances other than the specific substance under investigation. The total number of tougher-than-China MRLs regulated by importer $j$ on product $k$ in year $t-1$ is denoted as ${Number}_{jkt-1}$. This analysis includes six interaction terms between the MRL restrictiveness index of the specific substance under investigation and indicators for RTA member countries.

Table 5. Heterogeneity results by chemical substance: top 10 substances notified by the RASFF border, 2006–2021.

Name of Chemical Substance	Chlorpyrifos	Tolfenpyrad	Lambda-Cyhalothrin	Dinotefuran	Acetamiprid	Diafenthiuron	Carbendazim	Imidacloprid	Pyridaben	Pyraclostrobin
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
${MRL\_restrictiveness}_{jkt-1}$	0.648	6.270 ***	0.928 ***	−0.135	−0.256	−1.141 ***	0.619	2.196 ***	−1.486	−0.926 *
	(0.429)	(1.941)	(0.308)	(0.102)	(0.437)	(0.335)	(0.513)	(0.401)	(1.667)	(0.480)
${Other\_restrictiveness}_{jkt-1}$	−3.766 ***	7.718	−1.214 ***	0.273	−1.654 ***	5.127 *	−1.947 ***	−1.730 ***	8.974 **	−0.037
	(1.128)	(5.558)	(0.329)	(2.117)	(0.605)	(2.973)	(0.321)	(0.521)	(3.964)	(0.628)
$\ln \left(1+{Number}_{jkt-1}\right)$	1.569 ***	−1.338	0.268 *	−1.183 **	1.279 ***	−4.331 ***	0.472 ***	0.105	−3.823 ***	0.370
	(0.399)	(1.770)	(0.148)	(0.498)	(0.429)	(1.655)	(0.132)	(0.244)	(1.441)	(0.313)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.229 *	0.813 ***	−0.246 ***	0.187	−0.395 ***	0.482	−0.209 ***	−0.012	−0.423	−0.451 **
	(0.133)	(0.295)	(0.069)	(0.178)	(0.142)	(0.485)	(0.062)	(0.082)	(0.318)	(0.202)
${MRL\_restrictiveness}_{jkt-1}$ $\ast USMCA$	−0.071	(omitted)	0.595 ***	1.351 ***	0.215	(omitted)	0.414	1.076 ***	−15.679 ***	0.541
	(0.331)		(0.204)	(0.130)	(0.271)		(0.289)	(0.399)	(2.145)	(0.352)
${MRL\_restrictiveness}_{jkt-1}$ $\ast ASEAN$	−0.867 **	(omitted)	0.012	0.655 ***	−0.303	(omitted)	0.381	−1.041 ***	0.524	0.241
	(0.368)		(0.241)	(0.057)	(0.227)		(0.430)	(0.253)	(1.279)	(0.304)
${MRL\_restrictiveness}_{jkt-1}$ $\ast BRI$	0.127	(omitted)	−0.067	0.338 ***	0.300	(omitted)	−0.419	−1.075 ***	−0.957	0.847 **
	(0.361)		(0.268)	(0.112)	(0.418)		(0.489)	(0.336)	(1.499)	(0.423)
${MRL\_restrictiveness}_{jkt-1}$ $\ast CPTPP$	−1.174 ***	(omitted)	−0.003	0.903 ***	−0.016	0.633	−0.168	0.268	10.920 **	0.482
	(0.398)		(0.182)	(0.324)	(0.304)	(0.537)	(0.315)	(0.342)	(4.617)	(0.314)
${MRL\_restrictiveness}_{jkt-1}$ $\ast OECD$	0.822	(omitted)	−1.334 ***	−0.545 *	0.063	(omitted)	−0.045	−1.956 ***	4.925	0.487
	(0.620)		(0.373)	(0.321)	(0.505)		(0.579)	(0.502)	(4.172)	(0.534)
${MRL\_restrictiveness}_{jkt-1}$ $\ast EU$	−1.158 **	(omitted)	0.067	0.365	−0.346	(omitted)	−0.840 **	0.250	−4.142	0.456
	(0.516)		(0.255)	(0.544)	(0.405)		(0.348)	(0.477)	(4.134)	(0.407)
Constant	15.193 ***	10.224 ***	16.442 ***	18.772 ***	14.622 ***	24.907 ***	16.532 ***	17.497 ***	23.843 ***	16.038 ***
	(0.850)	(2.945)	(0.251)	(0.700)	(0.916)	(2.075)	(0.273)	(0.588)	(1.591)	(0.638)
Observations	2987	149	12,008	404	7114	587	11,889	3885	1287	2570
R-squared	0.645	0.796	0.498	0.665	0.537	0.912	0.545	0.680	0.884	0.593
Importer–product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Importer–year FE	Yes	No	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes

Notes: Standard errors are in parentheses and are clustered by importer–product–year. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively. To avoid multicollinearity, the importer–year fixed effect is excluded from the analysis reported in columns (2) and (6) due to the limited number of product categories involved in these specific regressions. For example, the regression for Tolfenpyrad includes only HS chapters 08 and 09, and the regression for Diafenthiuron includes only HS chapters 07 and 09.

reports the heterogeneity analysis of the top 10 most frequently notified chemical substances in the RASFF during the period 2020–2024. These ten substances ranked in the top 10 during the period 2020–2024, as the EU publicly released data for this timeframe. It is imperative to note that the regressions in Table 5 are based on export data from 2006 to 2021. The substances with the highest frequency of rejection at the EU border were chosen, as they indicate potential weaknesses in food safety standards and conformity in China’s agricultural sector. The empirical findings presented in this analysis examine the impacts of the MRLs for these substances on China’s agricultural exports, offering policymakers guidance on how to adjust their chemical practices to better align with international standards.

The estimates of the interaction terms reveal novel findings. Tightening the MRL for Chlorpyrifos reduces the export values of agricultural products shipped to ASEAN, CPTPP, and EU member countries. A more restrictive MRL for Lambda-cyhalothrin significantly facilitates China’s agricultural exports to USMCA countries while impeding exports to OECD countries. Stricter MRLs for Dinotefuran increase the export values to USMCA, ASEAN, BRI, and CPTPP countries but decrease the agricultural exports to OECD countries. The restrictiveness of Acetamiprid MRLs does not significantly affect trade, while a tougher MRL for Carbendazim generates smaller export values to the EU. More stringent MRLs for Tolfenpyrad and Imidacloprid facilitate China’s agricultural exports to all RTA member countries. In contrast, tightening the MRL for Pyridaben significantly decreases the export values to USMCA member countries while enhancing exports to CPTPP member countries. Finally, a tougher MRL for Diafenthiuron or Pyraclostrobin is negatively associated with the export values of agricultural products.

The results in Table 5 support the trade pattern proposed by Sun et al. [19]. For example, more stringent MRLs for Dinotefuran reduce the exports to OECD countries while increasing the exports to other RTA member countries. Additionally, stricter MRLs for Pyridaben lead to an increase in China’s exports to CPTPP member countries, while exports to USMCA member countries decrease, indicating that China diverges exports from USMCA countries to CPTPP countries. Similarly, more restrictive MRLs for Lambda-cyhalothrin have a positive impact on exports to USMCA countries and a negative effect on exports to OECD countries, suggesting that China shifts agricultural exports from OECD countries to USMCA countries. In most countries, agricultural product imports are sensitive to stricter MRL standards. For example, all importing countries are sensitive to MRLs for Diafenthiuron and Pyraclostrobin. EU member countries are particularly sensitive to MRLs for Chlorpyrifos and Carbendazim. The positive impact of MRLs for Tolfenpyrad and Imidacloprid on China’s agricultural exports to all RTA member countries implies that China’s food safety standards for these two substances are relatively stricter than those of other countries.

4.4. Extensive and Intensive Margins of Exports

In this section, I employ the Heckman selection model to examine the impacts of MRLs on the probability of export (extensive margin) and the volume of exports conditional on the export decision (intensive margin).

The selection equation, which estimates the impacts of MRLs on the extensive margin (EM) of exports, is specified as follows:

$$\begin{array}{ll}& P_{jkt}={\beta }_0+{\beta }_1{MRL\_restrictiveness}_{jkt-1}+{\beta }_2\ln \left(1+{Number}_{jkt-1}\right)+{\beta }_3\ln \left(1+{Tariff}_{jkt}\right)+\\ & {\beta }_4{lnGDP\_CHINA}_t+{\beta }_5{lnGDP}_{jt}+{\beta }_6{lnDist}_j+{\beta }_7{Border}_j+{\beta }_8{Language}_j+{\beta }_9{RTA}_{jt}+{\beta }_{10}{Religion}_j+\\ & {\delta }_k+{\epsilon }_{jkt}\end{array}$$

(4)

where $P_{jkt}$ is a dummy variable that equals one if positive export values of China’s agricultural products $k$ to importing country $j$ occurred in year $t$, and zero otherwise. Control variables include the logarithm of China’s GDP in year $t$ (${lnGDP\_CHINA}_t$), the logarithm of GDP in importing country $j$ and year $t$ (${lnGDP}_{jt}$), and the logarithm of distance between China and importer $j$ (${lnDist}_j$). Indicators for whether China and importer $j$ share a common border (${Border}_j$), common language (${Language}_j$), and are members of the same RTA in year $t$ (${RTA}_{jt}$) are incorporated in this analysis. The covariate ${Religion}_j$ represents the religious proximity index between China and the importer. The product fixed effect is denoted as ${\delta }_k$ and ${\epsilon }_{jkt}$ is the error term.

An appropriate exclusion restriction is required in this model. The exclusion restriction is assumed to influence the fixed cost of exports while not affecting the variable cost of exports. Previous studies, such as Helpman et al. [34], Crivelli and Groeschl [35], and Hejazi et al. [6], have used the common religion dummy as the exclusion variable. In line with these studies, this analysis adopts the religious proximity index between China and the importer as the exclusion restriction. However, in some cases, the exclusion variable may not meet the requirement of a valid exclusion restriction if it does not significantly affect the probability of exporting. For example, the religious proximity index does not significantly affect the probability of China exporting agricultural products to BRI member countries. Consequently, an alternative exclusion variable is employed in the analysis that measures the impacts of MRLs on the extensive and intensive margins of exports to BRI member countries. Following Ferro et al. [9], the indicator for a shared common language between China and the importer is adopted. Additionally, to assess the robustness of the exclusion variables used in the baseline regression, an analysis is conducted in Section 4.5.

The outcome equation measuring the impacts of MRLs on the intensive margin (IM) of exports is given as

$$\begin{array}{ll}& {lnM}_{jkt}={\beta }_0+{\beta }_1{MRL\_restrictiveness}_{jkt-1}+{\beta }_2\ln \left(1+{Number}_{jkt-1}\right)+{\beta }_3\ln \left(1+{Tariff}_{jkt}\right)+\\ & {\beta }_4{lnGDP\_CHINA}_t+{\beta }_5{lnGDP}_{jt}+{\beta }_6{lnDist}_j+{\beta }_7{Border}_j+{\beta }_8{Language}_j+{\beta }_9{RTA}_{jt}+\\ & {\beta }_{10}{IMR}_{jkt}+{\delta }_k+{\epsilon }_{jkt}\end{array}$$

(5)

where ${lnM}_{jkt}$ is the logarithm of the export value of agricultural product $k$ from China to importer $j$ in year $t$. The inverse mills ratio (${IMR}_{jkt}$) is computed from the selection equation and included in the outcome equation to correct for sample selection bias. The inverse Mills ratio is the ratio of the probability density function to the cumulative distribution function.

Table 6. Extensive and intensive margins of exports.

	All				ASEAN				BRI
	EM	IM	EM	IM	EM	IM	EM	IM	EM	IM	EM	IM
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
${MRL\_restrictiveness}_{jkt-1}$	−0.028	−0.159 *			−0.008	−0.049			0.128 **	−0.329
	(0.018)	(0.085)			(0.065)	(0.234)			(0.059)	(0.249)
$\ln \left(1+{Number}_{jkt-1}\right)$	−0.119 ***	−0.049			0.009	0.259 ***			−0.047 *	0.116
	(0.009)	(0.070)			(0.027)	(0.100)			(0.026)	(0.113)
${Fungicide}_{jkt-1}$			0.865 ***	2.448 ***			1.003 ***	−2.170			1.884 ***	0.935
			(0.111)	(0.803)			(0.377)	(2.158)			(0.352)	(1.558)
${Herbicide}_{jkt-1}$			−1.224 ***	−8.054 ***			−2.244 ***	4.508			−2.479 ***	−3.266
			(0.161)	(1.193)			(0.549)	(3.717)			(0.528)	(2.636)
${Insecticide}_{jkt-1}$			−0.260 ***	−1.215 ***			−0.642 ***	0.732			−0.128	−3.387 ***
			(0.046)	(0.287)			(0.182)	(1.133)			(0.172)	(0.755)
$\ln \left(1+{N\_Fungicide}_{jkt-1}\right)$			−0.072 ***	−0.241 *			0.054	−0.096			−0.153 **	0.053
			(0.027)	(0.141)			(0.075)	(0.366)			(0.067)	(0.286)
$\ln \left(1+{N\_Herbicide}_{jkt-1}\right)$			−0.005	0.750 ***			0.152 *	−0.134			0.192 **	1.199 ***
			(0.030)	(0.152)			(0.088)	(0.450)			(0.084)	(0.362)
$\ln \left(1+{N\_Insecticide}_{jkt-1}\right)$			−0.009	−0.198			0.098	0.210			−0.024	0.520 **
			(0.027)	(0.132)			(0.065)	(0.337)			(0.061)	(0.263)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.082 ***	−0.301 ***	−0.069 ***	−0.255 ***	−0.042 *	−0.040	0.005	−0.103	−0.029	−0.094	−0.027	−0.004
	(0.008)	(0.054)	(0.010)	(0.068)	(0.023)	(0.086)	(0.028)	(0.134)	(0.022)	(0.097)	(0.028)	(0.123)
$ln{GDP\_CHINA}_t$	0.162 ***	1.042 ***	0.226 ***	1.610 ***	0.451 ***	0.886 **	0.772 ***	−1.070	0.490 ***	0.706 ***	0.740 ***	0.168
	(0.020)	(0.116)	(0.030)	(0.209)	(0.069)	(0.397)	(0.107)	(1.037)	(0.054)	(0.262)	(0.080)	(0.378)
$ln{GDP}_{jt}$	0.298 ***	1.122 ***	0.282 ***	1.265 ***	−0.586 ***	−1.003 **	−0.826 ***	0.941	−0.184 ***	−0.606 ***	−0.337 ***	−0.371 *
	(0.007)	(0.145)	(0.009)	(0.197)	(0.082)	(0.488)	(0.139)	(1.085)	(0.038)	(0.168)	(0.050)	(0.215)
${lnDist}_j$	−0.250 ***	−1.262 ***	−0.247 ***	−1.536 ***	−0.665 ***	−0.113	−0.692 ***	1.512 *	0.905 ***	0.898 **	0.955 ***	0.707
	(0.016)	(0.131)	(0.019)	(0.182)	(0.069)	(0.508)	(0.087)	(0.901)	(0.096)	(0.396)	(0.120)	(0.500)
${Border}_j$	−0.011	0.476 ***	0.035	0.318	−0.198	0.897 **	0.239	0.169	1.707 ***	1.749 ***	2.019 ***	1.256 **
	(0.033)	(0.148)	(0.042)	(0.202)	(0.126)	(0.433)	(0.192)	(0.855)	(0.129)	(0.458)	(0.162)	(0.565)
${Language}_j$	0.733 ***	1.663 ***	0.865 ***	2.475 ***	−0.193 *	−0.500	−0.095	−0.572	1.716 ***		1.869 ***
	(0.044)	(0.340)	(0.054)	(0.541)	(0.117)	(0.335)	(0.142)	(0.559)	(0.160)		(0.193)
${RTA}_{jt}$	0.154 ***	0.614 ***	0.108 ***	0.366 *	−0.146	−0.105	−0.060	−0.229	0.860 ***	1.253 *	0.648 ***	0.537
	(0.027)	(0.157)	(0.034)	(0.200)	(0.128)	(0.567)	(0.175)	(0.927)	(0.137)	(0.713)	(0.168)	(0.782)
${Religion}_j$	−0.128 ***		−0.142 ***		0.415 ***		0.336 ***
	(0.024)		(0.030)		(0.111)		(0.118)
Inverse Mills Ratio		2.211 ***		4.027 ***		−2.133 *		−5.214 ***		−1.988 ***		−2.461 ***
		(0.709)		(0.983)		(1.090)		(1.899)		(0.529)		(0.577)
Observations	30,445	30,445	19,995	19,995	3945	3945	2599	2599	4505	4505	3033	3033
Product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Notes: The exclusion variable shown in columns (1)–(8) is the religious proximity index between China and the importer, and the exclusion variable shown in columns (9)–(12) is the indicator for sharing a common language between China and the importer. Standard errors are in parentheses. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively.

presents the results for the extensive and intensive margins of exports. Columns (1)–(2) show the estimation results for the model specified in Equations (4) and (5) across all importing countries, respectively. In columns (5)–(6) and (7)–(8), the EM and IM results for agricultural exports from China to ASEAN and BRI member countries are presented. Additionally, the MRL restrictiveness for fungicides (${Fungicide}_{jkt-1}$), herbicides (${Herbicide}_{jkt-1}$), and insecticides (${Insecticide}_{jkt-1}$), as well as the number of tougher-than-China MRLs for fungicides (${N\_Fungicide}_{jkt-1}$), herbicides (${N\_Herbicide}_{jkt-1}$), and insecticides (${N\_Insecticide}_{jkt-1}$), are incorporated into the selection and outcome equations. The results for the extensive and intensive margins of exports, driven by the MRLs for these three types of substances across all importing countries, are displayed in columns (3)–(4). The impacts of the MRLs for fungicides, herbicides, and insecticides on the extensive and intensive margins between China–ASEAN and China–BRI member countries are reported in columns (7)–(8) and (11)–(12).

The results in columns (1)–(2) suggest that the MRL stringency does not significantly affect the probability of exporting, but it significantly reduces the export values conditional on the export decision. Conversely, a greater number of MRLs stricter than those of China decreases the likelihood of exporting, while having no significant effect on the intensity of exports. Tariffs and distances negatively influence both the extensive and intensive margins of exports. Higher GDPs in China and the importing country, shared common language, and RTA membership facilitate exports and enhance export opportunities. The positive coefficient of ${Border}_j$ in the outcome equation indicates that China significantly exports larger values of agricultural products to neighboring countries when exporting. The estimate of ${Religion}_j$ in the selection equation is negative and significant, implying that China is more likely to export agricultural products to destination markets with smaller religious proximity.

In columns (3)–(4), the positive and significant estimates of ${Fungicide}_{jkt-1}$ indicate that stricter MRLs for fungicides are associated with a higher likelihood of exporting and greater export values of agricultural products. In contrast, the coefficients for ${Herbicide}_{jkt-1}$ and ${Insecticide}_{jkt-1}$ are negative and significant, suggesting that more restrictive MRLs for herbicides and insecticides imposed by importers not only reduce the likelihood of trading but also decrease the export values. It is worth noting that the number of tougher-than-China MRLs for fungicides significantly impedes both the probability and intensity of exports. The greater number of MRLs stricter than China for herbicides does not significantly impact export decisions but increases the export value condition on exporting. However, the number of tougher-than-China MRLs for insecticides is not significantly associated with the extensive or intensive margins of exports.

Columns (5)–(6) present the results for the extensive and intensive margins of agricultural exports to ASEAN member countries. Tightening the MRL stringency does not significantly impact the export decision or export values to ASEAN member countries, but a greater number of tougher-than-China MRLs induces higher export values from China to ASEAN member countries conditional on the export decision. In columns (7)–(8), the findings reveal that the MRLs for fungicides, herbicides, and insecticides do not significantly affect the intensive margin. However, the MRL restrictiveness for herbicides and insecticides significantly reduces the probability of exporting to ASEAN member countries. More restrictive MRLs for fungicides increase the likelihood of exporting to ASEAN member countries. The greater number of tougher-than-China MRLs for herbicides improves the export decision to ASEAN member countries, while not significantly affecting the export intensity.

Columns (9)–(10) report the impacts of MRLs on the extensive and intensive margins of agricultural exports from China to BRI member countries. A more stringent MRL standard significantly increases the probability of exporting agricultural products to BRI member countries while not significantly affecting the export value condition on the export decision. The number of tougher-than-China MRLs impedes exports to BRI member countries by reducing the likelihood of exporting. The coefficients for ${Fungicide}_{jkt-1}$ in columns (11)–(12) indicate that the MRL stringency for fungicides is positively associated with the likelihood of exporting to BRI member countries, while not significantly impacting the export intensity. The estimates for ${Herbicide}_{jkt-1}$ suggest that the MRL stringency for herbicides discourages the decisions of exporting to BRI member countries, while not significantly affecting the export value condition on the export decisions. Alternatively, the coefficients for ${Insecticide}_{jkt-1}$ imply that more restrictive MRLs for insecticides do not significantly impact the extensive margin but negatively affect the intensive margin of exports to BRI member countries. Additionally, the number of tougher-than-China MRLs for herbicides improves exports to BRI member countries by increasing the probability of exporting and export values upon exporting. A greater number of tougher-than-China fungicides undermines export opportunities to BRI member countries while not affecting the export values. On the other hand, the number of tougher-than-China insecticides does not significantly influence the export decisions but leads to higher export values to BRI member countries conditional on the export decisions.

4.5. Robustness Checks

A series of additional analyses were conducted to assess the robustness of the baseline findings. Specifically, the robustness was examined by (i) implementing the Heckman selection model with the interaction term between the religious proximity index and HS 2-digit indicators as an exclusion restriction; (ii) replicating the baseline regression using an alternative method for imputing missing MRL values; (iii) replacing the one-year lagged MRL restrictiveness index and the one-year lagged number of tougher-than-China MRLs with their two-year lagged counterparts. The results of the robustness checks are presented in

Table 7. Robustness analysis.

	EM	IM	EM	IM	Alternative Missing MRL Imputation	Two-Year Lagged Restrictiveness
	(1)	(2)	(3)	(4)	(5)	(6)
${MRL\_restrictiveness}_{jkt-1}$	−0.026	−0.143 *			−1.763 ***
	(0.018)	(0.082)			(0.204)
$\ln \left(1+{Number}_{jkt-1}\right)$	−0.119 ***	0.028			0.590 ***
	(0.009)	(0.062)			(0.097)
${MRL\_restrictiveness}_{jkt-2}$						−0.566 ***
						(0.166)
$\ln \left(1+{Number}_{jkt-2}\right)$						−0.005
						(0.089)
${Fungicide}_{jkt-1}$			0.868 ***	0.792
			(0.111)	(0.621)
${Herbicide}_{jkt-1}$			−1.247 ***	−5.666 ***
			(0.161)	(0.936)
${Insecticide}_{jkt-1}$			−0.251 ***	−0.705 ***
			(0.046)	(0.231)
$\ln \left(1+{N\_Fungicide}_{jkt-1}\right)$			−0.072 ***	−0.120
			(0.027)	(0.118)
$\ln \left(1+{N\_Herbicide}_{jkt-1}\right)$			−0.001	0.762 ***
			(0.030)	(0.130)
$\ln \left(1+{N\_Insecticide}_{jkt-1}\right)$			−0.012	−0.178
			(0.027)	(0.112)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.082 ***	−0.242 ***	−0.071 ***	−0.114 **	−0.217 ***	−0.333 ***
	(0.008)	(0.048)	(0.010)	(0.052)	(0.039)	(0.056)
$ln{GDP\_CHINA}_t$	0.162 ***	0.943 ***	0.227 ***	1.193 ***
	(0.020)	(0.107)	(0.030)	(0.161)
$ln{GDP}_{jt}$	0.298 ***	0.925 ***	0.283 ***	0.710 ***
	(0.007)	(0.121)	(0.009)	(0.135)
${lnDist}_j$	−0.251 ***	−1.102 ***	−0.248 ***	−1.069 ***
	(0.016)	(0.113)	(0.019)	(0.130)
${Border}_j$	−0.010	0.482 ***	0.037	0.250
	(0.034)	(0.143)	(0.043)	(0.170)
${Language}_j$	0.735 ***	1.257 ***	0.871 ***	1.056 ***
	(0.045)	(0.296)	(0.055)	(0.384)
${RTA}_{jt}$	0.154 ***	0.496 ***	0.107 ***	0.100
	(0.027)	(0.146)	(0.034)	(0.168)
Inverse Mills Ratio		1.208 **		1.131 *
		(0.590)		(0.669)
Observations	30,445	30,445	19,995	19,995	42,215	23,733
R-squared					0.461	0.477
Product FE	Yes	Yes	Yes	Yes	No	No
Importer–product FE	No	No	No	No	Yes	Yes
Product–year FE	No	No	No	No	Yes	Yes
Importer–year FE	No	No	No	No	Yes	Yes

Notes: Columns (1)–(4) present the results from the Heckman selection model. The interaction term between the religious proximity index and HS 2-digit indicators serves as the exclusion restriction to account for product heterogeneity. For brevity, the estimates of these interaction terms are not reported. In column (5), the missing MRLs for a specific substance are replaced with the most lenient MRL standard under the same HS 4-digit heading (instead of the HS 6-digit heading) among all importing countries in year $t$. Column (6) examines the impact of the two-year lagged MRL restrictiveness and the number of tougher-than-China MRLs on China’s agricultural exports. Standard errors are in parentheses. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively.

.

First, following Xiong and Beghin [1] and Shingal et al. [5], the interaction term between the religious proximity index and HS 2-digit indicators is used as the exclusion restriction to account for product heterogeneity. The results in columns (1)–(2) are consistent with the baseline regression, indicating that the MRL stringency does not significantly affect the probability of exporting while it significantly reduces the export values conditional on the export decision. In contrast, a higher number of MRLs stricter than those of China negatively impacts the extensive margin, with no significant effect on the intensive margin.

In columns (3)–(4), the results represent the impacts of MRLs for fungicides, herbicides, and insecticides on the extensive and intensive margins of exports. The signs and significance levels of the estimates are similar to those in Table 6. One exception is the significance level of the coefficients for ${Fungicide}_{jkt-1}$ and $\ln \left(1+{N\_Fungicide}_{jkt-1}\right)$ in the outcome equation. In the robustness analysis, the MRL restrictiveness for fungicides and the number of tougher-than-China MRLs for fungicides do not significantly impact the export values conditional on the export decision. The widespread use of fungicides across almost all product categories (HS-2 headings) may help to explain the insignificant impacts on the intensive margin of exports when product heterogeneity is incorporated.

Column (5) shows the results of replicating the baseline regression using an alternative method for imputing the missing MRL values. Specifically, missing MRLs for a specific substance are replaced with the most lenient MRL standard under the same HS 4-digit heading (instead of the HS six-digit heading) among all importing countries in year $t$. This adjustment leads to a 1.5-fold increase in the number of observations. The negative and statistically significant estimate for the MRL stringency is consistent with the baseline regression results, though its magnitude is greater. This finding aligns with the imputation method, where the missing MRLs are replaced with more lenient standards compared to those used in the baseline regression. Interestingly, the coefficient of $\ln \left(1+{Number}_{jkt-1}\right)$ is positive and significant, indicating that more substances with tougher MRLs than China imposed by the importing countries facilitate China’s agricultural exports. This trade pattern mirrors the results observed for fungicides and herbicides in the baseline regression, where the number of tougher-than-China MRLs positively correlates with exports.

Column (6) reports results where the main variables of interest are replaced with the two-year lagged MRL restrictiveness index and the number of tougher-than-China MRLs. The coefficient of ${MRL\_restrictiveness}_{jkt-2}$ is negative and significant, with a magnitude slightly larger than the baseline estimate. The number of substances with stricter MRLs than China’s does not significantly impact exports, which is consistent with the findings in Table 2. This analysis suggests that stricter MRLs impede China’s agricultural exports in the subsequent period, with the negative effects becoming greater two years later.

There is concern regarding the time inconsistency between the RASFF database (2020–2024) and the main dataset (2006–2021). This issue is partly addressed in Section 4.3, which highlights that substances with the highest frequency of rejection at the EU border might indicate poor food safety standards or inadequate regulatory conformity in China’s agricultural sector during 2006–2021. However, the time inconsistency still needs to be further addressed.

To address this, an additional heterogeneity analysis is conducted on the top 10 chemical substances regulated by importing countries between 2006 and 2021. The results are presented in

Table 8. Heterogeneity results by chemical substance: top 10 substances regulated by importing countries, 2006–2021.

Name of Chemical Substance	Permethrin	Acephate	Chlordane	Phoxim	Phorate	Methamidophos	Carbofuran	Dieldrin	Fenamiphos	Aldrin
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
${MRL\_restrictiveness}_{jkt-1}$	−0.011	−0.725 ***	1.857 ***	−1.990 ***	−0.458	−1.130 **	0.365	0.494 *	−3.710 ***	0.447
	(0.235)	(0.227)	(0.452)	(0.570)	(0.810)	(0.492)	(0.534)	(0.259)	(0.999)	(0.303)
${Other\_restrictiveness}_{jkt-1}$	−1.868 ***	−1.726 ***	−2.453 ***	−1.322 ***	−2.331 ***	−0.895 **	−2.128 ***	−2.907 ***	−1.526 ***	−2.802 ***
	(0.338)	(0.373)	(0.355)	(0.422)	(0.413)	(0.426)	(0.462)	(0.419)	(0.386)	(0.440)
$\ln \left(1+{Number}_{jkt-1}\right)$	0.748 ***	1.005 ***	0.546 ***	0.933 ***	1.127 ***	0.901 ***	0.915 ***	1.266 ***	0.902 ***	1.191 ***
	(0.175)	(0.180)	(0.156)	(0.247)	(0.209)	(0.185)	(0.224)	(0.202)	(0.177)	(0.209)
$\ln \left(1+{Tariff}_{jkt}\right)$	−0.254 ***	−0.272 ***	−0.335 ***	−0.242 ***	−0.286 ***	−0.272 ***	−0.277 ***	−0.291 ***	−0.271 ***	−0.295 ***
	(0.061)	(0.060)	(0.063)	(0.065)	(0.065)	(0.063)	(0.067)	(0.068)	(0.065)	(0.069)
${MRL\_restrictiveness}_{jkt-1}$ $\ast USMCA$	0.538 ***	−0.405 ***	−1.650 ***	−0.365	0.822 **	0.016	0.233	0.399	−0.636	0.630 *
	(0.160)	(0.137)	(0.310)	(0.258)	(0.348)	(0.207)	(0.368)	(0.371)	(0.403)	(0.359)
${MRL\_restrictiveness}_{jkt-1}$ $\ast ASEAN$	0.139	0.511 **	−1.283 ***	1.265 ***	1.046	0.444	0.095	0.262	2.119 **	0.188
	(0.192)	(0.216)	(0.384)	(0.381)	(0.756)	(0.472)	(0.486)	(0.193)	(0.937)	(0.208)
${MRL\_restrictiveness}_{jkt-1}$ $\ast BRI$	−0.201	−0.288	−1.766 ***	−0.297	−0.612	−0.474	−0.878 ***	−0.054	−0.560	0.071
	(0.165)	(0.177)	(0.391)	(0.441)	(0.522)	(0.306)	(0.316)	(0.226)	(0.422)	(0.272)
${MRL\_restrictiveness}_{jkt-1}$ $\ast CPTPP$	−0.547 ***	−0.220 *	−0.122	−0.252	−0.262	0.373	−0.622 *	−0.391	0.042	−0.452
	(0.152)	(0.131)	(0.317)	(0.309)	(0.329)	(0.231)	(0.330)	(0.277)	(0.456)	(0.310)
${MRL\_restrictiveness}_{jkt-1}$ $\ast OECD$	0.283	1.114 ***	−0.182	1.783 ***	−0.380	0.545	0.098	−0.423	4.017 ***	−0.372
	(0.276)	(0.274)	(0.553)	(0.604)	(0.862)	(0.524)	(0.602)	(0.331)	(1.071)	(0.388)
${MRL\_restrictiveness}_{jkt-1}$ $\ast EU$	0.661 **	−0.162	−1.654 ***	0.491	−0.103	0.290	−0.554	−0.742 **	0.367	0.304
	(0.295)	(0.376)	(0.471)	(0.355)	(0.378)	(0.300)	(0.389)	(0.327)	(0.541)	(0.348)
Constant	15.699 ***	15.339 ***	16.973 ***	16.132 ***	15.584 ***	15.077 ***	15.868 ***	15.730 ***	15.506 ***	15.531 ***
	(0.292)	(0.394)	(0.244)	(0.471)	(0.454)	(0.409)	(0.428)	(0.404)	(0.400)	(0.393)
Observations	21,763	17,641	17,101	16,866	16,625	16,165	15,999	15,825	15,877	15,708
R-squared	0.516	0.533	0.485	0.548	0.516	0.534	0.507	0.520	0.550	0.519
Importer–product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Importer–year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Notes: This analysis examines the top 10 chemical substances regulated by importing countries between 2006 and 2021 on China’s agricultural exports. All of the top 10 substances are insecticides. Standard errors are in parentheses and are clustered by importer–product-year. *, **, and *** denote p < 0.10, p < 0.05, and p < 0.01, respectively.

. All of the top 10 substances are insecticides: Permethrin, Acephate, Chlordane, Phoxim, Phorate, Methamidophos, Carbofuran, Dieldrin, Fenamiphos, and Aldrin. Among the coefficients for ${MRL\_restrictiveness}_{jkt-1}$ in Table 8, four are not statistically significant, suggesting that the MRL stringency for Permethrin, Phorate, Carbofuran, and Aldrin does not have a significant impact on China’s agricultural exports. Conversely, stricter MRLs for Acephate, Phoxim, Methamidophos, and Fenamiphos significantly reduce China’s agricultural exports. Additionally, more stringent MRLs for Chlordane and Dieldrin are associated with a significant increase in exports.

5. Conclusions

Using China’s agricultural export data from 2006 to 2021, I find that a 10% more stringent MRL at the mean in the previous year leads to a 6.6% decrease in the current-year export values of China’s agricultural products. Stricter MRLs for insecticides significantly reduce China’s agricultural exports, while the restrictiveness of fungicides and herbicides has no significant effect on exports. Tightening MRLs significantly impedes exports to ASEAN, BRI, the EU, and OECD countries while facilitating exports to USMCA countries. Exports of vegetables and fruits, as well as preparations and beverages, suffer from more restrictive MRLs for insecticides. However, exports of coffee and tea, as well as cereals and oilseeds, benefit from tougher MRLs for insecticides. More stringent MRLs for fungicides facilitate China’s exports of cereals and oilseeds, and preparations and beverages, while tightening MRLs for herbicides increases the export values of vegetables and fruits.

Although the EU frequently issues alert notifications on the ten chemical substances investigated in this study, stricter MRLs for these substances did not significantly reduce China’s agricultural exports to the EU, except for Chlorpyrifos, Diafenthiuron, and Carbendazim. Tightening MRLs for Tolfenpyrad, Dinotefuran, and Imidacloprid encourages exports to USMCA, ASEAN, and BRI member countries. Conversely, China’s agricultural exports to USMCA, ASEAN, and BRI member countries are sensitive to more stringent MRLs for Diafenthiuron. CPTPP member countries import more agricultural products from China when they regulate tougher MRLs for Tolfenpyrad, Dinotefuran, and Pyridaben, while importing less from China when the MRLs for Chlorpyrifos and Diafenthiuron are more restrictive.

The results of the extensive and intensive margin of exports in this study suggest that the MRL stringency negatively impacts the export values conditional on the export decisions, but does not significantly affect the probability of exporting. In contrast, the number of tougher-than-China MRLs decreases the likelihood of exporting but does not significantly affect the intensity of exports.

This study contributes to providing policymakers with a detailed and updated analysis of the trade effects of MRLs. Tightening MRL standards by importing countries do not necessarily impede agricultural trade; in some cases, they facilitate China’s agricultural exports to these markets.