Impact of Nontariff Measures on Border Crossing Time and Costs: The Case of Perishable Goods Trade in the Central Asia Regional Economic Cooperation Region

I. Introduction

Perishable goods, including agricultural products and pharmaceuticals, are vulnerable to spoilage if improperly handled or preserved. This study explores the effects of nontariff measures (NTMs) on intraregional trade in 11 Central and West Asian countries alongside trade costs, trade duration, and their implications for perishable goods. The analysis utilizes bilateral trade flows, sanitary and phytosanitary (SPS) measures and technical barriers to trade (TBT) measures, and trade facilitation indicators (TFIs) during 2018–2021. By employing a structural gravity model and addressing empirical challenges—such as endogeneity, heteroskedasticity, and zero values—this research fills a gap in the literature by examining the distinct impact of behind-the-border SPS and TBT measures, as well as at-the-border measures, on time and costs at border crossing points (BCPs). Previous studies have investigated these trade policy variables separately, without considering their concurrent effect on perishable goods, which are particularly sensitive to transport duration and regulatory measures. By untangling the impacts of at-the-border and behind-the-border measures, policymakers can implement more targeted policies to enhance trade facilitation and improve market access for perishable goods.

In the 11 Central and West Asian economies, the average share of perishable goods in intraregional trade increased from 9.7% in 2018 to 14.2% in 2021, indicating the growing significance of imported perishable goods in the region’s food and pharmaceutical supply chains.¹ This trend presents opportunities for landlocked Central Asian economies such as Kazakhstan, the Kyrgyz Republic, Tajikistan, Turkmenistan, and Uzbekistan, which are geographically close to each other. This proximity facilitates prompt delivery of perishable goods and minimizes spoilage. Enhancing intraregional trade of perishable goods allows these economies to take advantage of shorter transport routes, reducing transit time and costs compared to long-distance international trade. Therefore, intraregional trade of perishable goods plays a critical role in ensuring food security and promoting public health in the region by mitigating potential supply chain disruptions in international trade.

In general, trade in perishable goods faces two main challenges: at-the-border and behind-the-border barriers. Business surveys highlight that the greatest complexity and uncertainty occur at-the-border during exporting and importing. Border clearance creates the longest time delays, and regulations are only as effective as their implementation on the ground. Inconsistent enforcement across transport modes and different BCPs, as well as heterogeneity in testing facility availability, add risk and uncertainty to business operations (APEC Business Advisory Council 2016).

Behind-the-border barriers such as NTMs, which include both SPS and TBT measures, are intended to ensure public health and safety, as well as product quality. From a consumer perspective, NTMs enhance consumer trust, lower transaction costs, and increase consumer demand by reducing information asymmetry and/or externalities. However, these measures can hinder trade by adding additional requirements and procedures. This is particularly challenging for landlocked developing countries with poor logistics, infrastructure, and governance. To illustrate, soft infrastructure in the Central Asia Regional Economic Cooperation (CAREC) region remains inadequate due to the absence of risk-based management systems, single windows, and green lanes for perishables, and the presence of cumbersome customs controls that still favor examinations and documentary checks. Inconsistent enforcement and lack of harmonization with international standards can further complicate the trade of perishable goods in the region (Asian Development Bank 2017).

The structure of the paper is as follows. Section II reviews the existing empirical literature. Section III discusses the methodology, data, and the estimation results. The final section concludes.

II. Review of Empirical Literature

Existing literature on the perishable goods trade and detailed studies examining tariffs and NTMs are limited so far. Evans and Harrigan (2005) show that products necessitating prompt delivery are likely to be imported from neighboring countries, despite higher costs. Liu and Yue (2013) find that extended delays at borders notably diminish both the quality and price of highly perishable agricultural goods, while streamlined and improved customs processes for such perishable agricultural items can lead to increased trade volumes and enhanced social well-being in importing countries. Using CAREC Corridor Performance Measurement and Monitoring (CPMM) trade facilitation data, Sharafeyeva (2023) demonstrates that time-sensitive, perishable agricultural products are confirmed to be strongly impacted by uncertainty in the time to export.²

There is a dearth of studies exploring the impact of time delays on trade in the region despite the challenges posed by limited soft and hard infrastructure in CAREC economies. Kim, Mariano, and Abesamis (2021) investigated this topic utilizing the CPMM TFIs. These indicators provide bilateral measures of time and cost associated with trade facilitation in the CAREC countries. The authors employed gravity model estimations and found that a 10% reduction in time at the inbound border leads to a 1%–2% increase in trade among CAREC countries. However, this study fails to account for variations in the impact of trade costs and duration on the perishable and nonperishable goods trade. The study also acknowledges that other trade policy variables, such as SPS and TBT measures, which are excluded from the CPMM TFIs, can exert influence on bilateral trade flows.

NTMs can function as trade barriers (or facilitators) by decreasing (or increasing) domestic imports. For instance, the elimination of SPS measures would result in a rise in Australian imports of apples from New Zealand (Yue and Beghin 2009), as well as the increased trade of meat between the European Union (EU) and the United States (US) (Beckman and Arita 2016). Similarly, Arita, Beckman, and Mitchell (2017) discovered that SPS measures significantly decreased EU–US trade in meat, fruit, vegetables, cereals, and oilseeds. Furthermore, Webb, Gibson, and Strutt (2018) demonstrate that SPS compliance measures reduce the number of countries exporting agrifood products to the US by 35%. SPS measures also impede trade in fruit exports between developing countries (Melo et al. 2014).

Crivelli and Groeschl (2015) examined the effects of different types of SPS measures on the trade of agricultural and food products, categorizing them into conformity assessment (e.g., certificate requirements, testing, inspection, and approval procedures) and product characteristics (e.g., quarantine treatment, pesticide residue levels, labeling, and packaging). The findings indicate that conformity assessment related to SPS regulations hinders market entry, while measures related to product characteristics facilitate trade once exporters meet stringent standards. Conformity assessment regulations increase fixed costs due to the burdensome certification, testing, and inspection procedures in various export destinations. In contrast, SPS measures related to product characteristics enhance consumer trust by providing safety guidelines for imported agricultural goods.

Considering the limitation and variability in research findings, this study aims to fill the void by differentiating between at-the-border and behind-the-border trade measures. Perishable goods requiring specialized handling, such as cold chain logistics, typically incur higher transport costs and are highly sensitive to travel duration. Therefore, understanding how the trade of perishable goods affects intraregional trade in landlocked economies like those in the CAREC region is crucial.

III. Methodology

A. Data

1. Corridor Performance Measurement and Monitoring Trade Facilitation Indicators

This paper uses the detailed TFIs made available for 11 Central and West Asian countries, published as the CAREC CPMM trade facilitation data, which allows us to differentiate between at-the-border and behind-the-border trade measures for the 11 countries included in CAREC. The CPMM TFIs are used to assess and monitor the performance of the six transport corridors within the CAREC region in terms of time and cost. These indicators are derived from the “time/cost–distance framework” developed by the United Nations (UN) Economic and Social Commission for Asia and the Pacific. The TFIs encompass four measurements, which can be expressed in hours or days, US dollars, and kilometers required for trade along the corridors (Table 1).

One advantage of the CPMM TFIs is that they are collected directly from truck drivers, who are responsible for monitoring the time and costs involved in transporting goods from their origin to their destination. This approach mitigates the measurement issues encountered by the World Bank’s Doing Business report.³ A key attribute of the CPMM TFIs is their comprehensive coverage of trade facilitation aspects specific to the CAREC region. Since most of CAREC’s intraregional trade occurs within its transport corridors, the CPMM TFIs provide a more representative view of trade facilitation than other indicators.

Figure 1 presents the average quarterly times and costs for road-based clearances at BCPs. Before the coronavirus disease (COVID-19) pandemic, there were noticeable differences in clearance times across BCPs in the CAREC region. However, the overall average time has decreased since 2017. A significant change occurred in 2017 when trucks took an average of 16.7 hours for border clearances, a 48% increase from 11.3 hours in 2016. This increase was mainly due to challenges at the Afghanistan and Pakistan BCPs. Factors such as the sudden border closure in early 2017, stricter border checks, and less efficient procedures contributed to the longer wait times. On the financial side, the cost of transporting goods across borders has remained relatively stable since 2014. However, the pandemic led to an increase in both time and cost due to stricter travel restrictions, reduced efficiency in customs, and occasional border closures (Kim, Mariano, and Abesamis 2021; Kim, Abesamis, and Ardaniel 2022; Sharafeyeva 2023).

2. Data Description

To estimate the impact of at-the-border trade costs of perishable goods on intraregional trade, we employ bilateral CPMM TFI1 and TFI2 measures for road crossings at BCPs from 2018 to 2021.⁴ These trade facilitation measures capture the time and cost of moving goods at the Harmonized System two-digit level of product aggregation (HS2). Hence, there is no need to generate the average time and cost that HS2 goods spend at BCPs at the bilateral country level, as done by Kim, Mariano, and Abesamis (2021). The TFI1 and TFI2 measures employed in our study include the time and cost at inbound BCPs (i.e., importing countries) of the CAREC economies because imports are used to represent intraregional trade.

Table A1 in the Appendix lists the HS2-level products identified as perishable goods in the CPMM dataset. CPMM defines perishable goods as those products falling under HS codes 01 to 30 and/or requiring refrigeration while being transported. Bilateral import data for HS2-level products of the 11 CAREC economies are from UN Comtrade.⁵

Tables A2 and A3 in the Appendix present the top three perishable goods traded between CAREC economies that share BCPs. These include vegetables (HS 08), sugar confectionery (HS 17), cereals (HS 10), pharmaceutical products (HS 30), and fruits (HS 08). Perishable goods dominate trade between Afghanistan and Turkmenistan, with vegetables (48.5%), residues of waste from industry (33.0%), and sugar confectionery (18.4%) constituting more than 99.8% of Afghanistan’s exports to Turkmenistan. Sugar confectionery appears to be a major importing product of Turkmenistan; for example, it constitutes 74.6% of Azerbaijan’s exports to the country. Pharmaceutical products are among the People’s Republic of China’s (PRC) top exports to all CAREC economies with which it shares BCPs (i.e., Kazakhstan, the Kyrgyz Republic, Mongolia, Pakistan, and Tajikistan) (Figure 2). These economies—particularly Mongolia, Pakistan, and Tajikistan—saw their imports of pharmaceutical products from the PRC drastically increase in 2020 and 2021, making the PRC an important trading partner for the region during the COVID-19 pandemic.

Behind-the-border regulations, such as SPS and TBT measures, are derived from the UN Conference on Trade and Development Trade Analysis and Information System NTMs database.⁶ These cover NTMs that have been imposed or in effect from 2018 to the present. In terms of affected partner countries, SPS and TBT measures can affect all trading partners (i.e., multilateral) or specific partners (i.e., bilateral). SPS and TBT measures are also categorized according to their classifications: either conformity assessment or technical regulation. Conformity assessment measures include certificate requirements, testing, inspection, and approval procedures. Meanwhile, product characteristics cover requirements on quarantine treatment, pesticide residue levels, labeling, and packaging. Data on the existence of regional trade agreements between countries are drawn from the CEPII database.⁷

B. Estimating the Effects of Border Trade Costs

1. Model Specification

This paper attempts to untangle the separate and distinct impact of behind-the-border and at-the-border costs of the perishable goods trade on intraregional trade in the CAREC region using a gravity model following the Anderson and van Wincoop (2003) specification, which considers multilateral trade resistance factors. The structural gravity model is expressed in the logarithmic form :

where $Y_{ijt}^{\mathrm{\text{HS}}2}$ is the value of country $i^{\prime }s$ HS2-level imports from country j at time t, expressed in levels. $F_{ij}$ is a vector of reporter–partner fixed effects to control for bilateral time-invariant characteristics that might influence the trade flow between two countries (i.e., distance, cultural ties, common language, proximity, and historical relationships) without precluding the estimation of time-varying bilateral trade policy covariates, thereby reducing the problem of omitted variable bias. They also account for the unobservable relationship between the endogenous trade policy variables and the error term, which addresses the problem of endogeneity of trade policy variables (Baier and Bergstrand 2007, Santeramo and Lamonaca 2022). $F_i$ and $F_j$ are vectors of country fixed effects interacted with a year dummy variable t. They control for multilateral resistance in both importer i and exporter j, including the unobservable country-specific characteristics that vary over time (Yotov et al. 2016). $F_{{\mathrm{\text{BCP}}}_{ij}}$ is the vector of BCPs to control for trade intensities at each BCP. The term ${\mathrm{\text{rta}}}_{ijt}$ is a dummy variable representing the presence ($=1$) or absence ($=0$) of regional trade agreements between countries.

Vectors $Z_{{\mathrm{\text{BCP}}}_{ijt}}^{\mathrm{\text{HS}}2}$ and ${TP}_{ijt}^{\mathrm{\text{HS}}2}$ represent the time-varying bilateral trade costs where the former consists of trade facilitation measures representing at-the-border trade costs while the latter constitutes NTMs controlling for behind-the-border trade costs (i.e., SPS and TBT). The vectors of coefficients, $\lambda$ and $\gamma$, show the magnitudes of the partial impacts $Z_{{\mathrm{\text{BCP}}}_{ijt}}^{\mathrm{\text{HS}}2}$ and ${TP}_{ijt}^{\mathrm{\text{HS}}2}$ on bilateral trade flows.

Following Santeramo and Lamonaca (2022), we utilize a count variable, such as ${\mathrm{\text{SPSFrequency}}}_{ijt\mathrm{\text{HS}}2}$ or ${\mathrm{\text{TBTFrequency}}}_{ijt\mathrm{\text{HS}}2}$, which represents the total number of SPS or TBT measures in a given country pair at the HS2 product level for a given year to account for the extent of regulations. This indicator—also employed in Schlueter, Wieck, and Heckelei (2009)—enables us to assess the effects of introducing an additional SPS or TBT measure. The estimated coefficients of ${\mathrm{\text{SPSFrequency}}}_{ijt\mathrm{\text{HS}}2}$ or ${\mathrm{\text{TBTFrequency}}}_{ijt\mathrm{\text{HS}}2}$ represent the elasticity of the value of trade flows concerning the number of SPS and TBT measures calculated based on Table 2.

Heteroskedasticity and zero trade flows are two key challenges in the estimation of gravity models like Anderson and van Wincoop’s (2003). To address these concerns, the study employs the Poisson pseudo-maximum likelihood (PPML) estimation following Santos Silva and Tenreyro (2006). This method is robust to heteroskedastic error terms, a common occurrence in trade data, particularly with smaller and more remote economies such as the CAREC countries, where the conditional variance of the trade flow variable tends to approach zero. PPML’s robustness to heteroskedastic errors allows for estimation in levels with a multiplicative error term, assuming proportionality between the conditional variance and the conditional mean.

Furthermore, an additional challenge is zero trade flows—which may be either structural, due to inherently low trade volumes between small, distant countries with high transaction costs, or statistical, resulting from rounding errors or missing observations. Both types of zeroes are also more common in trade involving smaller, more remote economies. The PPML estimator, applied in a multiplicative rather than logarithmic form, accommodates these zero observations in the dependent variable.

C. Estimation Results

1. Trade Outcomes of Time and Costs at Border Crossing Points

We first investigate the effects of the average time and cost at the BCPs. The estimation results in Table 3 verify the importance of considering the product variation in examining the relationships. The baseline estimation results in Table 3 (columns 1–3) show that while the coefficients of the average time taken at the inbound BCPs (i.e., importing countries) are negative, they are not statistically significant for all goods and nonperishable goods when controlling for both SPS and TBT measures (columns 1 and 3 of Table 3). For perishable goods, the coefficients are positive and not statistically significant. These results are counterintuitive but consistent with the findings of Sharafeyeva (2023). A similar observation can be found for the average cost. The baseline estimation results in columns 1–3 of Table 3 appear to show that higher average cost at inbound BCPs is associated with more trade.

Introducing exporter-product-year fixed effects into the model clarifies these dynamics. Including a dummy variable for each exporter-product-year combination accounts for product-level variation, such as differences in production costs, technological advancements, and domestic policies. By controlling for these factors, the fixed effects allow the estimation to isolate the direct impact of time and cost at the border on imports, ensuring that product-specific variations are appropriately captured. When these fixed effects are included, the results show a negative impact of the time and cost at the border on trade flows, except for perishable goods, where the relationship is positive but not statistically significant. This outcome aligns with the theoretical expectation that longer times and higher costs at the border generally reduce imports by increasing uncertainty, risk of damage, and expenses, making imports more costly for businesses. Although not statistically significant, the exception for perishable goods likely reflects the specific nature of these products, which may require time-sensitive handling and consistent demand regardless of border delays.

2. Trade Outcomes of Sanitary and Phytosanitary Measures and Technical Barriers to Trade

The heterogeneous impacts of SPS and TBT measures across different products are evident in the baseline estimation results in columns 1–3 of Table 3. These results are consistent with findings in recent literature (see, for example, Santeramo and Lamonaca [2019, 2022]). When all goods are considered, SPS measures (modeled as a count to quantify the marginal impact of the introduction of an additional SPS measure) do not affect intraregional imports (column 1 of Table 3). For perishable goods, SPS measures positively impact intraregional trade, increasing imports by 38.7% (column 2 of Table 3). SPS measures quantifying the impact of SPS on nonperishable goods reduce intraregional imports of nonperishable products by 63.7% (column 3 of Table 3). These results remain robust even when exporter-product-year fixed effects are applied. The findings are also intuitive, as perishable goods, like food and pharmaceuticals, must meet strict standards to ensure safety and quality for human consumption. Therefore, regulations that build consumer trust can stimulate trade in these products.

Meanwhile, TBT measures positively impact intraregional imports of all goods (column 1 of Table 3) boosting trade by 27.2% (marginal impact of additional TBT measure) and increasing intraregional perishable imports by 5.3%. Using exporter-product-year fixed effects yields qualitatively the same results. While the impact of TBT measures on intraregional trade is positive when all goods are considered, its impact is higher on nonperishable goods (41.8%). However, these results are not robust to a different fixed effects structure.

TBT measures encompass a wide range of objectives, such as the safety of electrical appliances and the compatibility of telecommunication devices. While our findings suggest that these regulations may positively impact trade by ensuring imported goods meet certain standards, the lack of robustness in the results under alternative fixed-effects structures indicates that this relationship may be context dependent. Nevertheless, the potential trade-facilitating role of TBT measures can be explained through their ability to increase consumer confidence in product quality. By lowering informational barriers and market frictions, properly designed and implemented TBT measures can align product standards and reduce consumer uncertainty. However, the nuanced findings imply that the effectiveness of TBT measures in stimulating trade likely varies depending on factors such as product type, market conditions, and the regulatory environment.

The impact of SPS and TBT measures is not uniform; it varies not only across different product categories but also depending on the type of measure. To investigate this variation, we conduct a sensitivity analysis examining the effects of SPS and TBT measures on conformity assessment and product characteristics (Table 4). The findings reveal that, in general, SPS measures boost perishable goods trade (columns 2 and 5 of Table 3); specifically, SPS measures on product characteristics positively and significantly affect intraregional perishable goods trade (column 5 of Table 4). These results indicate that stringent safety and quality standards are important in facilitating trade for goods prone to spoilage, such as food and pharmaceuticals. Ensuring these products meet health and safety requirements can enhance consumer confidence and reduce health-related risks, stimulating trade. These results underscore the importance of regulatory standards that address product-specific vulnerabilities, suggesting that such measures can promote trade in sensitive product categories when properly designed.

However, SPS measures on conformity assessment harm intraregional trade in all goods, including both perishable and nonperishable products, when using an exporter-product-year fixed effects structure (column 2 of Table 4). This outcome can be attributed to several economic and logistical factors. First, conformity assessments, which involve inspections, testing, and certifications, introduce additional compliance costs for exporters. These procedures can significantly raise the cost and risk for perishable goods with limited shelf life and high sensitivity to delays. The extra steps impose direct financial burdens and lead to delays at the border, increasing the risk of spoilage and reducing the economic attractiveness of trading such goods.

While the spoilage risk is lower for nonperishable goods, compliance costs and delays can still create logistical hurdles. These assessments can slow down the movement of goods, increase storage costs, and lead to inventory inefficiencies. Moreover, the unpredictability of passing conformity assessments can create market access uncertainties, discouraging exporters from participating in regional trade or prompting them to seek markets with less stringent regulations.

The analysis isolates these specific effects by employing an exporter-product-year fixed effects structure. It reveals that while conformity assessments safeguard consumer health and safety, they can inadvertently function as trade barriers for a broad range of products. This suggests that although such regulations are essential, they must be designed to balance safety concerns with the need for smooth and cost-effective trade, particularly for goods sensitive to border delays.

Under the exporter-product-year fixed effects structure, TBT measures on conformity assessment reduce intraregional trade across all goods, including perishable and nonperishable goods (columns 1–3 of Table 4). This implies that the costs and delays associated with conformity assessments may outweigh the potential benefits of standardization, thus acting as a barrier to trade. Interestingly, when using the same fixed effects’ structure, TBT measures on product characteristics positively impact intraregional trade in perishable goods (column 5 of Table 4). This finding suggests that regulations focusing on product-specific attributes, such as safety and quality standards tailored to perishables, can enhance trade by addressing consumer concerns and market demands, even amid the stringent checks that typically accompany cross-border trade.

IV. Conclusion

By utilizing a novel dataset of CPMM trade facilitation measures covering CAREC member countries, this paper investigates the distinct impacts of at-the-border and behind-the-border costs on intraregional trade, focusing on perishable goods within the CAREC region. The analysis employs a gravity model using HS2-level bilateral data on imports, SPS and TBT measures, as well as time and costs at BCPs in the region.

Estimating the gravity model using PPML yields the following key results and implications:

(i)	The analysis confirms the importance of considering the variation across different product types. Introducing exporter-product-year fixed effects clarifies that both time and costs significantly and negatively impact trade, suggesting that delays and higher border costs directly reduce import volumes. This effect is particularly critical for perishable goods, where time-sensitive processing at the border can significantly affect product quality and market value.
(ii)	SPS and TBT measures related to conformity assessment are found to reduce intraregional trade across all goods, including both perishable and nonperishable goods, when using an exporter-product-year fixed effects structure. This indicates that the added costs, delays, and complexities introduced by these conformity assessments can act as trade barriers, outweighing their intended benefits of standardization and safety.
(iii)	Conversely, SPS measures related to product characteristics continue to boost intraregional trade in perishable goods, suggesting that targeted regulations focusing on specific product attributes, such as safety and quality standards tailored to perishables, can still have a positive trade-facilitating effect.

These findings highlight a critical policy challenge: While regulations like SPS and TBT measures on conformity assessment are designed to ensure product safety and quality, their current implementation appears to impose burdensome compliance costs and delays, thereby hindering trade. This is especially problematic for perishable goods, which are highly sensitive to time at the border.

Streamlining conformity assessment procedures should be considered to mitigate their negative impact on trade. This could involve simplifying inspection processes, introducing electronic certification systems, or adopting mutual recognition agreements to reduce redundant checks. For perishable goods, implementing “green lanes” for expedited processing at border crossings could help alleviate delays, preserving product quality and enhancing trade flows.

Additionally, a shift toward more risk-based inspections—focusing on high-risk products while allowing low-risk goods to pass more quickly—can help balance the dual goals of regulation and trade facilitation. By refining these measures, countries can maintain safety standards without unnecessarily restricting trade, thereby supporting economic growth and the livelihoods of small farmers and businesses that rely on cross-border trade.

ORCID

Dorothea M. Ramizo  https://orcid.org/0000-0003-4493-2152

Akiko Terada-Hagiwara  https://orcid.org/0000-0002-0249-357X

Notes

¹ The 11 economies covered in this study include Afghanistan, Azerbaijan, the People’s Republic of China, Georgia, Kazakhstan, the Kyrgyz Republic, Mongolia, Pakistan, Tajikistan, Turkmenistan, and Uzbekistan. ADB placed its regular assistance to Afghanistan on hold effective 15 August 2021.

² Asian Development Bank (ADB). CAREC Corridor Performance Measurement and Monitoring. https://cpmm.carecprogram.org/data/ (accessed 1 February 2023).

³ For instance, data are transcribed by individuals not directly involved in trade and logistics. See World Bank. Doing Business Archive. https://archive.doingbusiness.org/en/doingbusiness (accessed 30 August 2024).

⁴ The data for rail transport time and cost are excluded because of numerous missing data points.

⁵ United Nations Statistics Division. UN Comtrade Database. https://comtradeplus.un.org (accessed 1 February 2023).

⁶ United Nations Conference on Trade and Development (UNCTAD). Trade Analysis and Information System (TRAINS) Online. https://trainsonline.unctad.org/home (accessed 1 February 2023).

⁷ Database is available at Centre d’Etudes Prospectives et d’Informations Internationales (CEPII). The CEPII Gravity Database. https://www.cepii.fr/cepii/en/bdd_modele/bdd_modele.asp (accessed 1 July 2022).