Non-Tariff Measures, Pesticide Use, and Fruit Export: an Application of ASEAN

Abstract

Sanitary and Phytosanitary (SPS) measures are among the most frequent nontariff measures (NTMs) confronting agricultural trade. To ensure food and environmental safety, governments and international institutions have established Maximum Residue Limitations (MRLs) for regulating pesticide use. However, the national regulation of MRLs varies from country to country. This has become a barrier to the agricultural trade, making it difficult to measure its impact on agricultural trade.

In this study, a nonlinear aggregation index, the Bilateral Heterogeneity Index (BHI), was employed to investigate the impact of MRL regulatory heterogeneity on ASEAN’s fruit exports. According to the results, the regulatory heterogeneity of MRL between trading partners leads to a decrease in ASEAN fruit exports. The impact of regulatory heterogeneity acts differently for different entities; countries with small trading scales suffer much more negative impacts than larger exporters.

1.  Introduction

As tariffs become less of a market-entry barrier for developing countries, Non-Tariff Measures (NTMs) are becoming increasingly important. Sanitary and Phytosanitary (SPS), as a technical measure, is one of the most frequent NTMs agricultural trade is confronting.

The growing use of pesticides significantly increased the global food productivity over the past decades. Meanwhile, it has always been criticized for the linkage to some safety issues. To ensure food safety and environmental safety, governments and some international institutions established Maximum Residue Limitations (MRLs) to regulate the use of pesticides. MRL refers to the highest level of the pesticide residue that is legally tolerated in or on food or feed when pesticides are applied correctly in accordance with Good Agricultural Practice (GAP). MRL regulations make pesticide use easier and more explicit to grasp.

However, international harmonization of MRLs does not exist at the global level (Drogué and DeMaria 2012). Even though Codex Alimentarius offers an international standard, World Trade Organization (WTO) allows member countries to establish national MRL standards. Therefore, these legal limits can vary across countries, and have led to the regulatory heterogeneity of MRL globally.

The research questions raise: (i) how to measure the regulatory heterogeneity of MRL between trading partners? (ii) what is the impact of regulatory heterogeneity of MRL on agricultural trade flow? (iii) how is the trade impact on countries with different entities?

This study investigates the impact of regulatory heterogeneity of MRL on ASEAN’s fruit export, as measured by Bilateral Heterogeneity Index (BHI). BHI captures the regulation gap of MRL policies between ASEAN and the importing countries. Both the number of registered pesticides and difference in MRL regulation of each pesticide have been taken into consideration.

Special concern is given to ASEAN’s fruit export industry. ASEAN is the third largest fruit exporter after EU and US, accounting for about 10% of global total fruit export; for ASEAN, fruit is the largest export category in agricultural sector, this industry provides works and relatively higher income for many rural farmers, it is quite important for ASEAN’s economic development.

There are two MRL-related obstacles to ASEAN’s fruit export. Firstly, to deal with a wide variety of insects and diseases tropical crops are suffered from, farmers use several kinds of pesticides to ensure the harvest, which makes the products more difficult to pass the MRL tests by import customs. Secondly, tropical fruits are considered as minor crops because of the small cropping area when comparing with the major crops such as rice and soybean. In some countries, there is no MRL specification or relatively less specific regulation for some tropical fruits, because they are not grown in this country. The empty area may create an uncertain impact on ASEAN’s fruit export.

This study contributes to literature in two lines. Firstly, instead of focusing on single or several chemicals, this study takes all the registered pesticide into consideration to adequately measure the regulatory heterogeneity between ASEAN and the trading partners, and the result supports the debate of “standards as barriers to trade”. Secondly, most studies measure the impact of regulatory heterogeneity of MRL for all countries, the potential heterogeneity between specific country groups might be neglected. Considering that the different trading scale might impose different impact on exporting countries, this study investigates the heterogeneity of the impact of BHI on countries with different trading scale, the result suggests that countries with small trading scale suffer more than those large exporters.

The remaining part of this paper proceeds as follows: Section 2 is the literature review on the measurement and trade impact of regulatory heterogeneity of MRL. Section 3 describes the methodology and data used in this study. Section 4 gives the discussion of empirical results of econometric analysis. Finally, in Section 5, I draw conclusion and give several policy implications.

2.  Literature review

To measure the regulatory heterogeneity of MRL between trading partners, early publications focus on some specific chemicals (Wilson et al., 2003, Wilson and Otsuki 2004, Xiong and Beghin 2012). In practice, there are often numerous MRLs applied on one product, the analysis of one chemical may overstate its impact on agricultural trade flow (Hejazi et al., 2022), thus these applications somehow appear to be ex parte.

In addition, recent studies turn to construct an MRL-based aggregation index (Drogué and DeMaria 2012; Winchester et al., 2012; Peterson et al., 2013; Ferro et al., 2015). But either of them does not compare the regulatory heterogeneity of MRL between trading partners or does not assign larger cost for the larger regulatory heterogeneity. Li and Beghin (2014) provide a better solution that overcomes those drawbacks to compare the regulatory heterogeneity between individual countries and the international standard.

The preceding literature also attempts to investigate the impact of regulatory heterogeneity of MRL on agricultural trade flow. There is a well-known debate—“standards as barriers to trade”. The mainstream studies conclude that standards generally reduce trade flow, especially developing countries’ agricultural export to developed countries (Wilson and Otsuki 2004; Disdier et al., 2008; Ferro et al., 2015; Fiankor et al., 2021; Hejazi et al., 2022). Some studies also point out that stricter standards increase compliance cost (Wilson and Otsuki 2004; Henson and Jaffee 2008; Ferro et al., 2015), sometimes they may shut off trade as products get rejected by the import custom (Xiong and Beghin 2012).

Nevertheless, there are some different voices. Some studies point out that regulatory heterogeneity is not wide enough to act as a trade impediment, sometimes it could even become trade-prompting (Drogué and DeMaria 2012; Winchester et al., 2012; Shepherd and Wilson 2013). Especially, Peterson et al. (2013) note that the negative effect of SPS measures on fresh fruit and vegetable export diminishes as exporters accumulate treatment experience, and it vanishes when exporters reach a certain threshold.

The discussion has never led to a consensus, but following Li and Beghin (2012), “the literature shows a wide range of estimated effects from significantly impeding trade to significantly promoting it”.

3.  Methodology and data

(1)  BHI calculation

In this study, I revised Li and Beghin (2014)’s method to construct a non-linear aggregation index—Bilateral Heterogeneity Index (BHI), to measure the regulatory heterogeneity between ASEAN and other trading partners. ${\mathit{x}}_{\mathit{i}\mathit{n}\mathit{t}\mathit{l},\mathit{p}\mathit{t}}^{\mathit{k}}$ (MRL in CODEX standard) in the original form is replaced by ${\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}$ (MRL in exporting country i) to make it able to compare the difference between different trade partners.

There are two dimensions that need to be taken into consideration when constructing the index: the number of registered pesticides (n) in one country and the MRL of each pesticide (x). I assume that if the importer’s MRL requirement is stricter than the exporter’s regulation ( ${\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}-{\mathit{x}}_{\mathit{j}\mathit{p}\mathit{t}}^{\mathit{k}}>0$ ), then it would become a trade barrier.

  

$${\mathit{B}\mathit{H}\mathit{I}}_{\mathit{i}\mathit{j}\mathit{t}}^{\mathit{k}}=\frac{1}{\mathit{n}}\sum _{\mathit{p}=1}^{\mathit{n}}\mathit{e}\mathit{x}\mathit{p}\left(\frac{{\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}-{\mathit{x}}_{\mathit{j}\mathit{p}\mathit{t}}^{\mathit{k}}}{{\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}}\right)$$（1）

　

As shown in Equation (1), i refers to exporting countries, and j refers to importing countries. ${\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}$ is the MRL of pesticide p on product k in country i and year t. The exponential form is applied in accordance with Li and Beghin (2014) to penalize larger regulatory heterogeneity of MRL regulation between country i and country j, mapping the score from 0 to 2.72.

Sometimes ${\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}$ could be missing because pesticide p is unregistered in country i. To deal with those missing values, the national regulation of MRL deferrals is essential to decide the value of ${\mathit{x}}_{\mathit{i}\mathit{p}\mathit{t}}^{\mathit{k}}$ . Some countries explicitly prohibit the use of those unregistered pesticides, while others set a default value (usually 0.01 ppm), or they defer to CODEX standard or some other regional standard (such as ASEAN standard). There are still several countries who do not establish a sophisticated MRL regulation system, they have no limitation for the use of unregistered pesticides, thus I assign 50 ppm (a frequent maximum MRL value in the database) for this situation.

(2)  Model specification and data description

In order to investigate the impact of regulatory heterogeneity of MRL on ASEAN’s fruit export, the structural Gravity model with fixed effect (Yotov et al. 2016) is applied. The dependent variable $y_{ijt}^k$ , is country j’s import value of product k in year t from country i. Import data is obtained from the United Nation Commodity Trade Statistics Database (UN Comtrade), the unit is thousand US dollars.

As shown in Equation (2), there are four independent variables. dist_ij is the geographical distance between trading partners, it is regarded as a proxy of transportation cost. In the case of perishable fruit export, the trade flow is supposed to be more sensitive to transportation. The data is collected from CEPII. ${tariff}_{ijt}^k$ is the simple average tariff rate at HS-6 level (crop level), accessed from United Nations Conference on Trade and Development (UNCTAD). FTA_ijt is a dummy variable, it equals to 1 when importer j established Free Trade Agreement (FTA) with ASEAN in year t.

  

$$\begin{array}{cc}{y}_{ijt}^k=& {\beta }_0+{\beta }_1\log \left({dist}_{ij}\right)+\hfill \\ & {\beta }_2\log \left({1+tariff}_{ijt}^k\right)+{\beta }_3{BHI}_{ijt}^k+\hfill \\ & {\beta }_4{FTA}_{ijt}+{\mu }_{it}+v_t^k+{\epsilon }_{ijt}^k\hfill \end{array}$$（2）

　

The key variable, ${BHI}_{ijt}^k$ , measures the regulatory heterogeneity of MRL between ASEAN members and the importers. MRL data is accessed from HOMOLOGA (2022). The sign of ${BHI}_{ijt}^k$ is supposed to be negative because the higher BHI score, the stricter MRL requirements of importers, and it may cause higher compliance cost and then impede trade.

In order to verify the difference in impact between countries with different trading scales, in Equation (3), the interaction term D_middle and D_large are specified. D_middle represents the middle-exporter group, it equals to 1 when the import volume is larger than 10 tones but less than 100 tones. While D_large represents the large-exporter group, equals to 1 when the import volume is larger than 100 tones. Therefore, the coefficient β₃, measures trade impact on exporters with small trade scale (import volume less than 10 tones), (β₃+β₄) measures the impact on middle exporters, and (β₃+β₅) measures the impact on large exporters.

  

$$\begin{array}{cc}{y}_{ijt}^k=& {\beta }_0+{\beta }_1\log \left({dist}_{ij}\right)+\hfill \\ & {\beta }_2\log \left({1+tariff}_{ijt}^k\right)+{\beta }_3{BHI}_{ijt}^k+\hfill \\ & {\beta }_4{BHI}_{ijt}^k\times D_{middle}+{\beta }_5{BHI}_{ijt}^k\times \hfill \\ & D_{large}+{\beta }_6{FTA}_{ijt}+{\mu }_{it}+v_t^k+{\epsilon }_{ijt}^k\hfill \end{array}$$（3）

　

About the estimation method, the most common practice is Ordinary Least Squares (OLS) method. However, the standard practice of interpreting the parameters of log-linear models estimated by OLS as elasticities can be misleading in the existence of heteroskedasticity (Silva and Tenreyro 2006). Besides, there are inevitably too many zeros in the dataset, which log-linear OLS cannot deal with. Therefore, I apply Poisson Pseudo Maximum Likelihood (PPML) method as suggested in the previous studies. It is a better approach to incorporate zero observations and solve inherent heteroskedasticity issue. Besides, Country-time specific (μ_it) and crop-time specific ( $v_t^k$ ) fixed effect are included as multilateral resistance term as suggested by Yotov et al. (2016).

While Poisson regression gives a natural way to address zero-trade issue, it is criticized for the vulnerability of overdispersion and excessive zero flows (Burger et al., 2009). In the context of fruit export, some zeros reflect the absence in trade flow, others might be the result of importer’s strict MRL regulations (the “certain zeros”). Therefore, I also apply Zero Inflation Poisson (ZIP) model to deal with these issues.

The dataset includes seven exporters (Indonesia, Malaysia, Lao PDR, Myanmar, Philippines, Thailand and Vietnam), fifteen importers with import value more than 100 million dollars in 2020 (Australia, Canada, China, Germany, Hong Kong SAR, India, Japan, Korea, Rep., Malaysia, Netherlands, Singapore, Thailand, United Kingdom, USA and Vietnam), nine fruit crops (cashew nuts, durian, tamarind and others, mango and others, coconuts, plantain, banana, pineapple and watermelon), covering the period from 2008 to 2020. The number of total observations is 8522¹, 4886 zero observations, and 3636 non-zero observations.

4.  Result and discussion

(1)  BHI score

Generally, ASEAN members’ MRL regulations are far from importers’ requirements. MRL regulations in developed countries are stricter than in developing countries: they registered more pesticides, often require lower MRLs, and the measures of MRLs are more complicated (temporary MRLs, special trade concern, etc.).

Fig. 1 shows the simple average BHI score aggregated at exporter level. Apparently, there are clear differences between exporters, which are caused by the existence of MRL deferrals. Malaysia and Thailand maintain the same default value (0.01ppm) as EU, Japan and Korea, Rep., even though the registered pesticides in these countries are even less than half of importers’ requirements, they still receive very low BHI score (in 2020, the simple average BHI score of Malaysia is 1.09, and Thailand is 0.91). Philippines applies CODEX standard to improve the poor national MRL regulations, it really helps to narrow the disparity (1.83, in 2020). The rest ASEAN members do not apply any MRL deferrals, as a result, they all receive very high BHI scores (on average, BHI>2.2).

Fig. 1

Simple Average BHI score, 2008 to 2020

Note: ASEAN refers to Lao PDR and Myanmar, who apply ASEAN MRL standard.

(2)  Econometric analysis

The result of econometric analysis is presented in Table 1. Column (1) and Column (2) correspond to the result of Equation (2) and Equation (3) with PPML method explained in the preceding section respectively. Column (3) and Column (4) present the result based on ZIP estimation.

Table 1.

Econometric result

Estimation method	Column (1)	Column (2)	Column (3)	Column (4)
	PPML	PPML	ZIP	ZIP
Variable
log(distij)	0.10 (0.08)	0.42 (0.08)	0.17* (0.08)	0.44 (0.08)
$\log \left({1+tariff}_{ijt}^k\right)$	−0.59*** (0.10)	−0.41*** (0.10)	−0.49*** (0.11)	−0.41*** (0.10)
${BHI}_{ijt}^k$	−0.40** (0.17)	−29.64** (15.76)	−0.36** (0.16)	−9.21*** (1.00)
${BHI}_{ijt}^k\times D_{middle}$		24.73* (15.61)		4.13*** (0.90)
${BHI}_{ijt}^k\times D_{large}$		29.48* (16.00)		8.97*** (0.94)
FTAijt	0.36** (0.16)	0.67*** (0.18)	0.59*** (0.00)	0.64*** (0.00)
Inflation regression (probit)
${BHI}_{ijt}^k$			0.43*** (0.02)	−1.86 (0.18)
Fixed effect	Yes	Yes	Yes	Yes
Observation	8522	8522	8522	8522
Pseudo R-squared	0.44	0.67	0.37	0.56
AIC	2.35*108	1.41*108	2.00*108	1.39*108

Note: ***, ** and * show 1%, 5% and 10% significance level, with standard errors in parenthesis.

Basically, most of the independent variables receive significant result with expected sign, except the geographical distance. The coefficients of average tariff rate are negative and significant, implies that even though whittled down by a series of agreements, tariff still reduces the trade flow. FTA receive positive and significant result cross all specifications, which indicate the positive impact on ASEAN’s fruit export.

It is not surprising that the estimated coefficients of BHI are all negative and statistically significant under PPML estimation and ZIP estimation. This means, BHI has a negative impact on trade flow, importer’s strict MRL standards reduce ASEAN’s fruit export. Besides, in Column (3), I notice that the estimated coefficient of BHI in the probit model (as shown in the lower part, inflation regression) is positive and significant, which indicates that BHI increase the possibility of “certain zeros”, some exporters cannot afford the compliance cost to meet importer’s requirement, they must stop exporting and seek alternative buyers. This is exactly evidence that bilateral heterogeneity of MRL not only decreases the trade flow, but also threatens the market access.

In addition, I also evaluate trade impact of regulatory heterogeneity of MRL on exporters with different trading scales, the result is statistically significant under different estimation methods. In the PPML scenario in Column (2), the estimated coefficient for small exporters (β₃) is −29.64, for middle exporters it is −4.91 (−29.64+24.73), and for large exporter, the estimated coefficient is −0.16 (−29.64+29.48). The result under ZIP estimation, as shown in Column (4), is somehow different, with coefficient of −9.21 for small exporters, −5.08 (−9.21+4.13) for middle exporters, and −0.24 (−9.21+8.97) for large exporters. Both estimation methods show the same tendency—countries with small trading scale suffer more negative impact than those large exporters when confronting strict MRL regulations imposed by import customs.

The reason is intuitive. Large exporters are more experienced in adjusting to importers’ requirement. Besides, they are able to do so economically. It is more profitable to export instead of selling in the domestic market. Therefore, even though the strict MRL regulations imposed by importers lead to additional compliance cost, large exporters are willing to pay the it if the cost is lower than the expected revenue. The finding is consistent with Fiankor et al. (2021)’s conclusion.

5.  Conclusion

The relationship of MRL standard and agricultural trade flow is receiving increased attention. In this study, I employ a non-linear aggregation index—BHI to measure the impact of regulatory heterogeneity of MRL between ASEAN and the trading partners, and then add BHI into the Gravity model to investigate the trade impact. The estimation is performed on an unbalanced panel data including seven exporters, fifteen importers and nine fruit crops during the period from 2008 to 2020, and the result is statistically significant under PPML estimation and ZIP estimation. It is confirmed that the regulatory heterogeneity of MRL between ASEAN and trading partners leads to the decrease of ASEAN’s fruit export. Furthermore, the trade impact acts different for different entities: countries with small trading scale suffer much more negative impact than those large exporters.

It is highlighted that ASEAN members really need some technical and financial support to improve the poor ASEAN MRL database. It can not only specify the domestic production, but also help narrow the regulation gap between ASEAN and their importers. In addition, for those small exporters, who are vulnerable to the strict MRL regulations, they need more financial investment to improve the product quality to meet importer’ requirement.

One limitation of this study is that it does not examine the impact of regulatory heterogeneity of MRL on ASEAN’s fruit production, export price and farmer’s income. This will remain to be the future work.

Acknowledgments

I would like to appreciate HOMOLOGA for kindly providing MRL data for free.

note

1  This is an unbalanced dataset, some data is missing due to the lack of MRL data at crop level as well as year level.

References

•  Burger,  M.,  F.  Van Oort, and  G. J.  Linders (2009) On the specification of the gravity model of trade: zeros, excess zeros and zero-inflated estimation, Spatial Economic Analysis 4(2): 167–190. https://doi.org/10.1080/17421770902834327.
•  Disdier,  A. C.,  L.  Fontagné, and  M.  Mimouni (2008) The impact of regulations on agricultural trade: evidence from the SPS and TBT agreements, American Journal of Agricultural Economics 90(2): 336–350. https://doi.org/10.1111/j.1467-8276.2007.01127.x.
•  Drogué,  S. and  F.  DeMaria (2012) Pesticide residues and trade, the apple of discord?, Food Policy 37(6): 641–649. https://doi.org/10.1016/j.foodpol.2012.06.007.
•  Ferro,  E.,  T.  Otsuki, and  J. S.  Wilson (2015) The effect of product standards on agricultural exports, Food Policy 50: 68–79. https://doi.org/10.1016/j.foodpol.2014.10.016.
•  Fiankor,  D. D. D.,  O. K.  Haase, and  B.  Brümmer (2021) The heterogeneous effects of standards on agricultural trade flows, Journal of Agricultural Economics 72(1): 25–46. https://doi.org/10.1111/1477-9552.12405.
• Hejazi, M., J. H. Grant, and E. Peterson (2022) Trade impact of maximum residue limits in fresh fruits and vegetables, Food Policy 106: 102203. https://doi.org/10.1016/j.foodpol.2021.102203.
•  Henson,  S. and  S.  Jaffee (2008) Understanding developing country strategic responses to the enhancement of food safety standards, World Economy 31(4): 548–568. https://doi.org/10.1111/j.1467-9701.2007.01034.x.
• HOMOLOGA, the global crop production database (2022) MRLs. Retrieved from https://homologa.com/solutions/mrls/.
•  Li,  Y. and  J. C.  Beghin (2012) A meta-analysis of estimates of the impact of technical barriers to trade, Journal of Policy Modeling 34(3): 497–511. https://doi.org/10.1016/j.jpolmod.2011.11.001.
•  Li,  Y. and  J. C.  Beghin (2014) Protectionism indices for non-tariff measures: An application to maximum residue levels, Food Policy 45: 57–68. https://doi.org/10.1016/j.foodpol.2013.12.005.
•  Peterson,  E.,  J.  Grant,  D.  Roberts, and  V.  Karov (2013) Evaluating the trade restrictiveness of phytosanitary measures on US fresh fruit and vegetable imports, American Journal of Agricultural Economics 95(4): 842–858. https://doi.org/10.1093/ajae/aat015.
•  Shepherd,  B. and  N. L.  Wilson (2013) Product standards and developing country agricultural exports: The case of the European Union, Food Policy 42: 1–10. https://doi.org/10.1016/j.foodpol.2013.06.003.
•  Silva,  J. S. and  S.  Tenreyro (2006) The log of gravity, The Review of Economics and Statistics 88(4): 641–658. https://doi.org/10.1162/rest.88.4.641.
•  Wilson,  J. S. and  T.  Otsuki (2004) To spray or not to spray: pesticides, banana exports, and food safety, Food Policy 29(2): 131–146. https://doi.org/10.1016/j.foodpol.2004.02.003.
•  Wilson,  J. S.,  T.  Otsuki, and  B.  Majumdsar (2003) Balancing food safety and risk: do drug residue limits affect international trade in beef?, Journal of International Trade & Economic Development 12(4): 377–402. https://doi.org/10.1080/0963819032000154810.
•  Winchester,  N.,  M. L.  Rau,  C.  Goetz,  B.  Larue,  T.  Otsuki,  K.  Shutes,  C.  Wieck,  Heloisa L.  Burnquist,  Maurício J.  Pinto de Souza, and Nunes de Faria, R. (2012) The impact of regulatory heterogeneity on agri-food trade, The World Economy 35(8): 973–993. https://doi.org/10.1111/j.1467-9701.2012.01457.x.
•  Xiong,  B. and  J.  Beghin (2012) Does European aflatoxin regulation hurt groundnut exporters from Africa?, European Review of Agricultural Economics 39(4): 589–609. https://doi.org/10.1093/erae/jbr062.
• Yotov, Y. V., R. Piermartini, and M. Larch (2016) Partial equilibrium trade policy analysis with structural gravity, An advanced guide to trade policy analysis: The structural gravity model. Genève: World Trade Organization, 11–17. https://doi.org/10.30875/abc0167e-en.