Public Disclosure Authorized WPS6857 Policy Research Working Paper 6857 Public Disclosure Authorized Technical Measures to Trade in Central America Incidence, Price Effects, and Consumer Welfare Sinéad Kelleher José-Daniel Reyes Public Disclosure Authorized Public Disclosure Authorized The World Bank Poverty Reduction and Economic Management Network International Trade Unit May 2014 Policy Research Working Paper 6857 Abstract Despite the widespread tariff reductions sparked by Honduran imports. The paper estimates the impact of the Dominican Republic-Central America Free Trade these technical measures on border prices and finds that Agreement, borders in the region remain thick, with the price impact of sanitary and phytosanitary measures many hurdles standing in the way of regional trade. is equivalent to an ad-valorem tariff of 11.6 percent. Although anecdotal evidence suggests that nontariff This price-rising effect is further investigated by looking measures raise trade costs and inhibit trade in the region, in detail at the impact of sanitary and phytosanitary little is known about the magnitude of these economic measures on the prices of beef, chicken meat, bread, and effects. This paper uses a newly collected data set to dairy products in Guatemala. The impact is estimated quantify the incidence of sanitary and phytosanitary to be equivalent to an ad-valorem tariff of 68.4 percent, measures and technical barriers to trade in the region 51.4 percent, 22.0 percent, and 5.0 percent, respectively. and benchmarks it with other parts of the world. The The paper shows that efforts to streamline key sanitary results indicate that the Central American region has and phytosanitary measures affecting these products the lowest prevalence of technical nontariff measures in by, for example, reducing the cost and time required the world. However, there is significant heterogeneity of to obtain sanitary registries, would likely reduce the trade-related regulations in Central America; for instance, Guatemalan urban extreme poverty rate from 5.07 48 percent of Salvadoran imports are subject to at least percent to 4.91 percent. one nontariff measure, compared with just 16 percent of This paper is a product of the International Trade Unit, Poverty Reduction and Economic Management Network. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at jreyes2@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Technical Measures to Trade in Central America: Incidence, Price Effects, and Consumer Welfare * Sinéad Kelleher † José-Daniel Reyes ‡ Keywords: Trade Policy; Nontariff Barriers; Empirical Studies of Trade; WTO; Economic Integration JEL Classification Codes: F13, F14, F15 Sector Board: Economic Policy (EPOL) * The authors thank Óscar Calvo, Jasmin Chakeri, Mariem Malouche, Michael Ferrantino, Sebastian Saez, Olivier Cadot, Magaly Clavijo, Mayra Del Carmen Alfaro De Moran, and Ana Paula Cusolito for helpful comments and discussions. We also thank participants at the workshop on Non-Tariff Measures in Central America, held in San José, Costa Rica on September 5th, 2013. All remaining errors are our responsibility. Financial support from the Trade Facilitation Facility Multi-Donor Trust Fund is gratefully acknowledged. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not represent the view of the World Bank, its Executive Directors, or the countries they represent. † University College Dublin, kellehsp@tcd.ie ‡ The World Bank, jreyes2@worldbank.org (corresponding author) 1. Introduction Broadly defined, Non-Tariff Measures (henceforth NTMs) are policy measures, other than ordinary customs tariffs, that may have an economic effect on international trade in goods by changing quantities traded, prices or both (UNCTAD, 2010). The fall in tariffs in recent decades has been accompanied by an increase in the importance of NTMs in the international trade agenda. NTMs allow countries to address legitimate non-trade objectives such as the protection of human, animal, and plant health. The World Trade Organization (WTO) recognizes that countries are entitled to use NTMs in this manner. However, there are concerns that the multilateral reduction of traditional trade barriers has tempted countries to replace tariffs with NTMs and to use them to protect domestic interests (Fischer and Serra, 2000, Aisbett and Pearson, 2012). The potential benefits of NTMs are evident, but it is important for policy makers to be cognizant of the costs associated with them. Unlike traditional trade policy instruments, such as tariffs, where the associated price increases are well understood, the costs associated with NTMs are more difficult to immediately quantify. Kono (2006), for instance, finds that democracy induces politicians to reduce tariffs but to increase the use of less transparent NTMs, and hypothesizes that this allows them to protect the domestic market but still to appeal to voters who prefer liberal trade policies. Although the outcome is the same, the price increases are less transparent to both policy makers and the public when NTMs are employed. Even when used legitimately and without protectionist intent, NTMs can have a major economic impact, potentially affecting regional integration, impacting domestic prices, diverting managerial attention, and penalizing small exporters. Countries imposing them may end up hurting their own poor (as NTMs may increase the price of key staples) and hampering the competitiveness of their own private sector (as NTMs may affect the price of imported inputs). The development an accurate metric of the costs of NTMs is thus important as it will inform policy makers of the minimum social benefit which is required to justify the presence of an NTM. This paper has a particular focus on Central America. Despite the widespread tariff reductions sparked by the Dominican Republic-Central America Free Trade Agreement (DR-CAFTA), Central American borders remain thick, with many hurdles standing in the way of regional trade (see Gordillo et al., 2010). Anecdotal evidence suggests that NTMs raise trade costs and inhibit trade in the region (see, for example, Cato et al (2005) for an analysis of the impact of Sanitary and Phytosanitary (SPS) measures on the shrimp industry in Nicaragua). However, there is a dearth of empirical papers quantifying the economic effects of these regulations, and, traditionally, very limited data availability. Seemingly simple questions such as ‘what policy measures are imposed by a country?’ or ‘what types of measures are faced by particular products?’ can often not be answered due to a lack of appropriate data. In order to fill this gap, and in response to the increased interest of both policy makers and researchers in the region, the World Bank, in collaboration with the United Nations Conference on Trade and Development (UNCTAD), initiated a new effort to collect NTM data in Central America. During the second half of 2012, NTM data was collected in Nicaragua, Guatemala, El Salvador, Honduras, and Costa Rica. This endeavor complemented a multiagency effort to gather NTM data worldwide (see Cadot and Malouche, 2012). To date, this program has resulted in the collection of NTM data in 48 countries. At the 2 time of writing, data on 33 countries was available on WITS 1 (the World Bank’s portal for trade data) with the remaining data still under verification by UNCTAD’s statistical division. This study is the first attempt to use this new database to provide information about the trade incidence of NTMs across countries in the Central American region. We specifically focus on the role of technical NTMs, namely Sanitary and Phytosanitary (SPS) measures and Technical Barriers to Trade (TBT). This assessment makes three key contributions to the literature on NTMs. Firstly, using this new and comprehensive dataset, we provide a descriptive analysis of the incidence of SPS and TBT measures in five Central American countries and benchmark Central America against other regions in the world. Secondly, we use econometric techniques to estimate the effect of NTMs on consumer prices, and determine how this effect varies across regions and type of NTMs. Thirdly, we explore the poverty impact of streamlining NTMs in Guatemala, one of the poorest countries in the region. Results indicate that the Central American region has the lowest prevalence of technical NTMs in the world. However, there is significant heterogeneity in the coverage of NTMs across countries. For instance, 16.2% of all imports to Honduras are covered by at least one NTM, compared to 48.2% of goods entering El Salvador. This variation is reflected in the implementation of TBTs; only 7.9% of imports to Honduras are subject to a TBT, compared to 42.3% of imports to El Salvador. By contrast, countries are quite consistent in their application of SPS measures with this NTM imposed on between 16.1% and 21.6% of trade in the region. Our analysis also indicates that SPS measures inhibit intra-regional integration. In particular, registration requirements for SPS reasons (the so-called sanitary registries) are identified as a possible obstacle to regional integration. When compared with other countries for which data are available, four of the five Central American countries are identified as heavy users of sanitary registries. We find that SPS measures have a significant impact on domestic prices in Central America. Across all products in the five countries, SPS measures have an average ad valorem equivalent of 11.6%. We investigate this price-raising effect in more detail for four key products in Guatemala: beef, bread and pastry, chicken meat, and dairy products. According to our estimation, the average import prices of these products rise by an amount equivalent to an ad-valorem tariff of 68.4%, 51.4%, 22.0%, and 5.0%, respectively. Finally, we estimate the impact of streamlining SPS measures in Guatemala, and find that rationalizing the use and processes related to SPS measures affecting beef, bread and pastries, chicken meat, and dairy products would reduce domestic prices. This would have the effect of making the basic consumption basket more affordable, thereby benefitting the poorest segment of the population. Our computations suggest that streamlining the most significant SPS measures affecting these products could reduce the urban poverty rate from 5.07% to 4.91%. We identify selected issues for policy dialogue at the regional level aimed at streamlining the NTM agenda. The manner in which countries’ sanitary registries work is a hindrance to deeper regional integration. A key issue is that while there is mutual recognition of registries across countries, all exporters must produce paperwork for each individual country they wish to enter. A solution that could make great strides in this area is the formation of a regional, electronic database that centralizes 1 http://wits.worldbank.org/wits/ 3 documentation about the credentials that each firm submits to meet the sanitary registration requirements. This would lessen the time and cost burden on exporters, who currently have to gather all the documentation each time they apply for a mutual recognition certificate. More general issues for policy dialogue include considering setting up a regional secretariat to improve national and regional competitiveness by identifying and streamlining NTMs that prevent further regional integration. This committee should include the main regional agencies that regulate trade as well as representatives from the business chambers and other key stakeholders. It is also important to disseminate information on the current laws and regulations on import and export procedures to the wider public, particularly the business community. In this regard, grouping and regularly updating information in a regional portal would alleviate informational asymmetries which increase the cost of doing business in the region. This paper is divided into six sections. This first section is a general introduction. The second provides an overview of the current literature on NTMs, details their classification and reports estimated frequency and coverage ratios. The third section assesses the price-raising effect of NTMs. The fourth section presents a preliminary analysis of the impact on poverty rates of streamlining key NTMs in Guatemala. Compiling all the information presented in the previous sections, part five highlights suggested topics for policy dialogue in the region aimed at facilitating international trade and regional integration. The final section presents some concluding remarks. 4 2. NTM Data and Stylized Facts a. Conceptual Issues In 2006, the Secretary-General of UNCTAD established the Group of Eminent Persons on Non-Tariff Barriers (GNTB). Their objective was to discuss the definition, classification, collection and quantification of NTMs. The GNTB established a Multi-Agency Support Team (MAST) which comprised representatives from several international organizations, including the Food and Agriculture Organization of the United Nations (FAO), the International Monetary Fund (IMF), the International Trade Centre UNCTAD/WTO (ITC), the Organization for Economic Co-operation and Development (OECD), the United Nations Industrial Development Organization (UNIDO), the World Bank and the World Trade Organization (WTO). Following a series of meetings and consultations, MAST proposed the following definition of NTMs: “Non-tariff measures (NTMs) are policy measures, other than ordinary customs tariffs, that can potentially have an economic effect on international trade in goods, changing quantities traded, or prices or both.” (UNCTAD, 2010) NTMs are a very diverse set of regulations in terms of type, effect and incidence (Gourdon and Nicita, 2012). Governments have a large degree of autonomy regarding the form and objectives of NTMs, and there is a high level of variation in how they are implemented. In addition, the desired level of NTMs varies across countries depending on perceived levels of risk, public demand and governmental priorities. With the reduction of traditional trade barriers, NTMs have become the new frontier of trade policy. This section briefly reviews the key empirical literature attempting to measure and quantify their effect on international trade. A range of papers find that the global significance of NTMs is large, with Henson et al (2000) asserting that “SPS measures are currently one of the foremost issues affecting exports of agricultural and food products from developing countries”. Andriamananjara et al (2004) report that the global welfare gains of the removal of certain NTMs would be in the order of US$90-92 billion, while Walkenhorst and Yasui (2009) find that a reduction in transaction costs of 1 percent of the value of world trade would yield aggregate welfare gains of US$40 billion. Estimates of tariff equivalents of NTMs tend to be quite high; Bradford (2003) focuses on final good trade and finds that the tariff equivalent of Japan’s NTMs is 57%. The tariff equivalent for European countries’ NTMs is slightly lower, ranging from 48% to 55%, and much lower for the United States at just 12%. Calvin and Krissoff (1998) use a price gap approach and show that the tariff equivalent of NTMs on US apples exported to Japan is 27.2%. This estimate significantly exceeds the average tariff rate of 19.3%. Chemingui and Dessus (2008) find that NTMs increase the domestic price of imported goods by an average of 17% across all product categories in Syria. By contrast, the average tariff is just 8%. Kee et al (2009) find that NTMs add, on average, an additional 87% to the level of trade restrictiveness imposed by tariffs, and note that for 34 countries (of a sample of 78) the contribution of NTMs to the overall level of restrictiveness is higher than the contribution of tariffs. 5 Henson et al (2000) report the results of a survey which asked respondents from low and middle income countries to indicate the significance of factors impeding their ability to export agricultural and food goods to the EU. They find that NTMs, specifically “SPS requirements”, “other technical requirements” (for example labeling regulations or compositional standards) and “transport and other direct export costs”, emerged as the top three issues, ahead of tariffs. Cadot and Gourdon (2012) find that SPS measures raise the price of African foodstuffs by 14%, with peaks for certain products, such as rice and other cereals, some types of meat, and edible oils. Combining this information with Kenyan household survey data, they find that NTMs act as a regressive tax, a result arising from the fact that foodstuffs make up a large share of poor households’ budgets and the major impact of NTMs, especially SPS measures, on these goods. However, NTMs cannot be simply dismissed as trade barriers. Results from gravity type models including Disdier et al (2008), Fontagné et al (2005), Portugal-Perez el al (2010), and Reyes (2011) indicate that while NTMs are generally found to be impediments to trade, they can also be trade promoting. This is because certain NTMs, such as labeling and content requirements, can provide consumers with information thereby lowering transaction costs, facilitating comparison and reducing uncertainty. Thus, NTMs can eliminate a market failure by reducing the cost of determining the quality of a product, thereby helping to prevent a ‘lemon’ problem (Akerlof, 1970). This point is also made by Crivelli and Gröschl (2012), Disidier et al (2008), and Ganslandt and Markusen (2001). Henson et al (2000) report that, to the extent that higher SPS standards increase costs of production, high SPS requirements offer low and middle income countries a source of competitive advantage over developed country suppliers. Chemnitz and Künkel (2006) study the heterogeneous effect of SPS measures across different countries noting that standards can “act like a catalyst and like a barrier at the same time”, and Fontagné et al (2005) find that “for sensitive products, enforcing a measure at the border may guarantee the existence of trade flows that would otherwise not be recorded at all”. However, research indicates that NTMs can have a particularly negative effect on small and developing countries. Reyes (2012) highlights the interplay between standards harmonization and market structure. Using firm-level data, the author finds that the international harmonization of European standards in the electronic sector increases intra-EU competition to the benefit of exporters from the US but to the disadvantage of exporters from the developing world, who lose market share. Ganslandt and Markusen’s 2001 general equilibrium model finds that incompatible standards are especially harmful for a small country, which can never win a “standards’ war”, where two countries impose cost-increasing standards on imports. In the small country, both producers and consumers may lose. Finally, Disdier et al (2008) find that developing and least developed countries’ exports to the OECD are significantly reduced by SPS or TBT measures. By contrast, these regulations do not have any effect on exports from other OECD states. b. Quantifying NTMs: The MAST Nomenclature The NTM nomenclature, developed by the Multi-Agency Support Team (MAST), comprises a tree/branch structure. Measures with similar objectives are categorized into 16 “branches” (chapters) denoted by alphabetical letters A to P. Chapters A to O specify the requirements the importing country imposes on imports, while Chapter P details those regulations they place on their exports. Each of these 16 “branches” is further disaggregated into “sub-branches” (1-digit), “twigs” (2-digits) and “leafs” (3 6 digits) allowing for a finer classification of NTMs. The structure of the NTM nomenclature at the highest degree of aggregation is in Table 1. Table 1. Non-Tariff Measures Classification A. Sanitary and Phytosanitary Measures (SPS) Technical B. Technical Barriers to Trade (TBT) Measures C. Pre-Shipment Inspection and other Formalities (PSI) D. Contingent Trade-Protective Measures E. Non-Automatic Licensing, Quotas and other quantitative controls (QC) F. Price Controls (PC) Imports G. Financial Measures H. Measures affecting Competition Non technical I.  Trade-Related Investment Measures Measures J.  Distribution Restrictions K. Restrictions on Post-Sales Services L. Subsidies (Excluding Export Subsidies under P7) M. Government Procurement Restrictions N. Intellectual Property O. Rules of Origin Exports P. Export-Related Measures Source: UNCTAD 2012 Data on NTMs is traditionally quite fragmented and incomplete. The World Trade Report 2012 finds that “existing data sources are compromised by large gaps in country coverage, intermittent data collection and a lack of shared terminology”. This applies even to data collected by the WTO; the WTO framework itself is incomplete despite containing more than 200 different legal notification requirements. Specifically, notification requirements do not cover three of the 16 categories of international classification measures (finance measures, distribution restrictions and restrictions on post-sales services). The exclusion of these is not surprising; the notification system is not designed to create a comprehensive database. For example, the SPS and TBT notification requirement is intended so that early drafts of regulations can be viewed by partner countries at the stage when amendments are still possible. Countries are therefore not required to report regulations that predate the introduction of the SPS or TBT Agreements, or new regulations in their final form (WTO, 2012). In addition, because notifications are voluntary with no sanction mechanism, governments may simply choose not to comply with notification requirements, either due to logistical difficulties (Bacchetta et al, 2012) or concern that notification will expose their NTMs to scrutiny or objections. Finally, the quality of notifications varies widely. In many cases, even when NTMs are reported, precise information on certain dimensions of the measures, such as product coverage or the time period during which the measure will remain in place is not provided. Alterative inventories of NTMs have been obtained through concerns or complaints registered by traders, with sources including direct surveys and WTO Trade Policy Reviews. However, these databases are not comprehensive and tend to be biased as countries vary widely in terms of the administrative procedures required to file complaints against third parties (Beghin and Bureau, 2001). 7 In a response to this, an effort to collect comprehensive NTM data is currently being undertaken jointly by the World Bank, UNCTAD and the African Development Bank. Since 2009, local consultants have been tasked with drawing up NTM inventories in collaboration with ministries and agencies in 48 countries. One of the challenges related to data collection is that most countries do not have a national repository of NTM data. This is because laws and regulations affecting trade are often developed by different government agencies and regulatory bodies (for instance, in Lao, 14 organizations oversee the 27,975 NTMs, and in Tunisia, there are 10 authorities responsible for the 3,361 regulations). In practice, the data has to be carefully scrutinized for possible duplications, omissions, or any other problems in order to minimize inaccuracies. Once the gathered data are verified and formatted into a common, consistent nomenclature by UNCTAD and also verified by the relevant government, it is uploaded onto WITS, the World Bank’s portal for trade data, where it is publicly available. Variables in the WITS database include reporter name, NTM code, HS code, the year in which the measure was implemented, legal sources, enforcing agencies, and a list of affected partner countries. c. Data Availability At the time of writing, NTM data has been collected for 48 countries, with 33 of the data files publically available on the WITS website. The remaining data files are in the process of verification by UNCTAD. Table 2 shows the availability of data across each of the six World Bank regions, along with the year in which the data was collected. The table separates out the five Central American countries that are the subject of our analysis. However, they are officially classified under the Latin America and Caribbean region by the World Bank. Table 2. NTM Data Availability (by World Regions) Rest of Middle East Sub-Saharan East Asia Europe and Central High Income Latin America and North South Asia Africa and Pacific Central Asia America Countries and Caribbean Africa Costa Burkina Faso* Cambodia Kazakhstan* Argentina* Egypt* Afghanistan* EU* Rica* (2010) (2011) (2012) (2008) (2012) (2011) (2012) (2010) Kenya China* Russia Bolivia* El Salvador Lebanon* Bangladesh Japan* (2011) (2012) (2009) (2008) (2012) (2011) (2011) (2009) Madagascar* Indonesia Turkey Brazil* Guatemala* Morocco* India* (2011) (2011) (2010) (2008) (2012) (2011) (2012) Syrian Arab Lao PDR* Chile* Honduras Nepal* Malawi Republic (2011) (2011) (2008) (2012) (2011) (2012) Mauritius* Philippines Colombia * Nicaragua Tunisia* Pakistan* (2011) (2008) (2008) (2012) (2011) (2012) Namibia* Ecuador* Sri Lanka* (2011) (2008) (2012) Rwanda Mexico* (2011) (2008) Senegal* Paraguay * 8 (2011) (2008) South Africa Peru* (2011) (2008) Tanzania* Uruguay* (2011) (2008) Uganda Venezuela, RB* (2011) (2008) Note: This table shows the countries used in this study. The year in which NTM data was collected is in parentheses. An asterisk indicates those countries for which the NTM data are publically available in WITS as of June 2013. Source: Authors’ computation For the purposes of our analysis, a standardized methodology was developed in order to harmonize each country’s NTM data file. Firstly, we only used in the analysis those NTMs that applied to each of a country’s trading partners. In those cases where the database did not specify the group of countries to which an NTM applied, it was assumed that it applied to all trading partners. The second step in the harmonization was to standardize the product codes to the six digit level. In some cases only a two digit chapter heading was provided in the database. In these cases, it was assumed that the NTM applied to all products at the six digit level under that chapter. When a NTM was stated as affecting a product at an eight or ten digit level, the product code was trimmed to six digits to allow a consistent cross-country comparison exercise. In all cases, it was assumed that the NTM affected all goods under the six digit classification. Thirdly, we noted that four different NTM nomenclatures were used in the collection of the NTM data; each of the databases was converted to the 2012 version using a concordance across nomenclatures. Finally, the data was checked to ensure no duplications at the country-product-NTM code level, and we dropped from the analysis SPS and TBT measures that were applied to all products as these were assumed to be data entry mistakes. For each six digit code, two indicators were developed; a dummy indicating that at least one NTM applied to that good, and a count variable showing the total number of NTMs applying to that product. The same indictors are also created at the NTM chapter level; a dummy indicating the application of at least one regulation from a particular NTM chapter to a product, the second indicator showing the actual number of NTMs from each chapter applying to each good. In this analysis we investigate only SPS and TBT measures. d. The Incidence of NTMs in Central America In this section, two simple inventory indicators are used to measure the prevalence of NTMs: the frequency ratio (percentage of product lines exposed to NTMs) and the coverage ratio (share of total imports exposed to NTMs). The frequency ratio accounts only for the presence or absence of an NTM, and indicates the percentage of imported products to which one or more NTMs are applied. In more formal terms, the frequency ratio of NTMs imposed by country ( ) is calculated as: 9 ∑ = � ∑ � . 100 [1] where is a dummy variable indicating the presence of one or more NTMs and (also a dummy variable) indicates whether there are imports of good . Because products have all equal weights, this measure tends to over-emphasize products with very low import value. A measure of the importance of NTMs in terms of overall import value is given by the coverage ratio. This indicator measures the percentage of imports to country subject to at least one NTM. In formal terms the coverage ratio ( ) is given by: ∑ = � ∑ � . 100 [2] where, is defined as before, and Vi is the value of imports of product . 2 Unlike the previous indicator, the coverage ratio does not overemphasize products with a low import value, but because products affected by trade restricting measures are imported in lower quantities and therefore get a lower weight in the calculation, it tends to under-estimate overall restrictiveness (endogeneity bias). Neither the frequency ratio nor the coverage ratio make any distinction between NTMs with relatively little impact on trade and those with major, even prohibitive, trade effects. This is a significant issue as restrictiveness varies widely across type of NTM; Beghin and Bureau (2001) note that there is not even a clear correlation between the number of NTMs in place and their trade effect. Also, it should be noted that frequency and coverage ratios do not give any indication of the level of enforcement of regulations. Because neither of these standard measures is perfect, we present both. Despite their shortcomings, they provide useful information about what is most often observed in a country, what is likely to be important, and thus what may warrant deeper analysis. This paper focuses on SPS (chapter A) and TBT (chapter B) measures, both of which are technical NTMs. Central America is the region where these measures are least prevalent (Figure 1). On average, approximately 22.3% of active tariff lines in the region are affected by a least one NTM, and one third of imports are subject to an NTM. While the Middle East and North Africa has the highest frequency ratio, Europe and Central Asia has the highest coverage ratio indicating differences in the relative composition of their import baskets – Europe and Central Asia’s imports must be weighed more towards regulated goods. Figure 1: Technical NTMs: Frequency and Coverage ratios, by World Regions (Simple Averages, Percentages) 2 We use the average total import value by product (6-digit HS codes) in the period 2008-2010. 10 Central America 22.3 33.3 South Asia 32.7 46.6 East Asia and Pacific 34.8 38.0 Latin America and the Caribbean 40.5 49.4 Sub-Saharan Africa 41.6 51.3 Europe and Central Asia 43.3 55.9 Middle East and North Africa 47.0 54.7 60 40 20 0 20 40 60 80 Frequency Ratio Coverage Ratio Note: This figure presents the within regions simple average of frequency and coverage ratios. Only SPS and TBT measures are considered. The countries considered for this analysis are listed in Table 2. Source: Authors’ computation Central American countries differ greatly in their use of NTMs which indicates that there is not a common approach to trade regulation in the region (Figure 2). The more developed countries appear to utilize NTMs more intensely than their less developed neighbors. El Salvador applies NTMs to the highest proportion of import lines (32.2%), which is more than double the proportion of imported goods subject to an NTM in Honduras (15.0%). This heterogeneity may be an indication of differences in governmental priorities and approaches towards trade regulation, composition of import baskets, or the influence of import-competing sectors lobbying for protection. Figure 2: Central America: Frequency and Coverage ratios, by Country Honduras 15.0 16.2 Nicaragua 16.0 29.6 Guatemala 24.0 43.3 Costa Rica 24.4 29.4 El Salvador 32.2 48.2 40 30 20 10 0 10 20 30 40 50 60 Frequency Ratio Coverage Ratio 11 Note: This figure presents the frequency and coverage ratios for all Central American countries. Only SPS and TBT measures are considered. Source: Authors’ computation. Table 3 displays the coverage and frequency ratios for SPS regulations and TBT regulations separately. SPS regulations affect between 13.6% and 17.4% of products and cover not more than 22% of import value in the region. There is wider variation in the use of TBT, with only 5% of product lines imported to Honduras covered by a TBT regulation compared to 28.2% of goods imported to El Salvador. For each of the countries, the coverage ratio for TBT regulations is significantly larger than the frequency ratio (2.9 times bigger in the case of Nicaragua). This suggests that these countries import large volumes of products from sectors that use TBTs more extensively. Table 3. Frequency and Coverage Ratios by Type of NTM Costa Rica El Salvador Guatemala Honduras Nicaragua NTM Chapter FR CR FR CR FR CR FR CR FR CR A. Sanitary and Phytosanitary Measures (SPS) 17.4 16.2 15.7 21.6 16.2 19.4 14.5 16.1 13.6 16.6 B. Technical Barriers to Trade (TBT) 14.4 25.9 28.2 42.3 21.0 40.4 5.0 7.9 8.7 25.0 Note: This table shows the frequency and coverage ratios by type of NTM in each country in the Central American region. These indicators are computed for all active product lines (6-digit HS codes). Source: Authors’ computation Analyzing in more detail the type of NTMs in effect in each country provides more information about the nature of trade legislation in the Central American region. Table 4 reports coverage ratios for SPS and TBT measures across economic sectors in each of the five countries under analysis. Each cell represents the import value affected by at least one type of NTM as a share of total sectoral import value. This analysis reveals that SPS measures are primarily concentrated in animal, vegetable, and foodstuff sectors, while TBTs are especially prevalent in trade in animals and chemical goods, but are also used intensively in the vegetable and foodstuffs sectors. These two results are comparable with international practices. 12 Table 4. Pervasiveness of NTMs within Economic Sectors. (Coverage Ratios, Percentages) Type A: SPS B: TBT Sector . CRI SLV GTM HND NIC CRI SLV GTM HND NIC 01-05 Animal 78.8 100.0 100.0 100.0 90.5 58.8 97.5 95.4 57.5 46.8 06-15 Vegetable 68.8 98.4 99.8 85.9 99.8 45.0 32.1 78.8 24.2 91.1 16-24 Foodstuffs 83.3 87.5 98.7 78.5 77.7 35.5 85.9 81.8 21.5 34.6 25-26 Minerals 93.3 11.8 9.1 2.9 2.8 51.1 52.7 12.5 -- 82.2 27 Oil Minerals 44.3 16.8 -- 2.9 0.0 44.3 97.7 95.2 -- 4.6 28-38 Chemicals 36.4 9.0 40.8 19.5 13.8 81.4 98.5 74.0 4.5 63.7 39-40 Plastic / Rubber 0.2 0.8 -- -- -- 1.6 44.1 1.4 -- -- 41-43 Hides, Skins 44.1 42.1 41.2 -- -- -- 28.5 -- -- -- 44-49 Wood -- 3.8 1.6 -- -- 3.8 0.0 -- -- -- 50-63 Textiles, Clothing 0.3 5.1 4.0 -- -- 0.1 -- -- -- -- 64-67 Footwear -- 0.5 0.0 -- -- -- 2.4 -- -- -- 68-71 Stone / Glass -- -- -- 4.0 2.1 0.3 -- -- -- 2.1 72-83 Metals -- -- -- 0.0 0.0 0.0 0.3 0.3 -- 0.0 84-85 Mach/Elec 1.0 -- -- -- -- 9.2 0.0 2.9 -- -- Note: This table reports coverage ratios for SPS and TBT measures across economic sectors in each country in the region. Each cell represents the import value affected by at least one NTM as a share of total sectoral import value. Each country is labeled by its 3-digit ISO code. Source: Authors’ computation Table 5 shows the most important NTMs by their coverage ratio in each country, and shows that registration requirements are particularly common NTMs in Central America – these are highlighted in grey. These requirements are applied not only on importer firms (codes B150 and A150) but also on selected imported products (codes B810 and A810). As agricultural products are important for regional trade, we compare the use of registration requirements for SPS reasons (code A150) – the so called sanitary registries – in Central America to other countries for which data are available (Figure 3). It appears that each of the Central American countries, with the exception of Costa Rica, is a frequent user of this type of regulation. 13 Table 5. Most Prevalent NTMs in Central America (Coverage Ratios) Costa Rica El Salvador Guatemala Honduras Nicaragua Rank NTM CR NTM CR NTM CR NTM CR NTM CR B150 Registration requirement B140 Authorization requirement for A150 Registration requirements for A640 Storage and transport 1 21.5 B310 Labelling requirements 39.5 32.1 13.7 14.3 for importers for TBT reasons TBT reasons importers conditions B140 Authorization requirement for B150 Registration requirement for A140 Special Authorization 2 14.4 B810 Product registration requirement 33.0 27.3 8.7 B310 Labelling requirements 12.5 TBT reasons importers for TBT reasons requirement for SPS reasons B420 TBT regulations on transport 3 B110 Prohibition for TBT reasons 12.7 B830 Certification requirement 31.0 25.7 B310 Labelling requirements 6.8 B330 Packaging requirements 11.3 and storage A140 Special Authorization A640 Storage and transport B100 Prohibitions/restrictions of 4 11.8 21.5 B310 Labelling requirements 17.3 A820 Testing requirement 6.7 10.1 requirement for SPS reasons conditions imports for objectives … A150 Registration requirements for A140 Special Authorization 5 B810 Product registration requirement 9.5 A820 Testing requirement 21.3 16.3 A830 Certification requirement 6.4 6.4 importers requirement for SPS reasons B190 Prohibitions/restrictions of B200 Tolerance limits for residues and B700 Product quality or performance 6 9.3 A810 Product registration requirement 21.3 A810 Product registration requirement 16.3 4.5 5.7 imports restricted use of substances requirement A140 Special Authorization 7 A840 Inspection requirement 8.9 A310 Labelling requirements 15.6 14.1 B320 Marking requirements 4.5 A310 Labelling requirements 5.7 requirement for SPS reasons A150 Registration requirements for 8 B840 Inspection requirement 8.9 A830 Certification requirement 15.6 A830 Certification requirement 13.9 B330 Packaging requirements 4.5 5.6 importers A140 Special Authorization B700 Product quality or performance B490 Production or Post-Production A190 Prohibitions/restrictions of 9 B310 Labelling requirements 7.9 15.3 11.2 4.5 4.3 requirement for SPS reasons requirement requirements n.e.s. importsfor SPS reasons n.e.s. A410 Microbiological criteria of the 10 A810 Product registration requirement 5.0 13.3 A310 Labelling requirements 9.5 B600 Product identity requirement 4.5 A850 Traceability requirements 4.2 final product Note: This table shows coverage ratios for the most important NTMs in each country in the region. Registration requirements for both SPS and TBT reasons are highlighted in grey. Source: Authors’ computation 14 Figure 3: Registration Requirements for Importers due to SPS Reasons (A150) (Coverage Ratios, Percentages) 30.0 26.7 25.0 20.0 17.6 16.3 15.0 13.7 12.0 10.0 10.0 5.6 5.0 5.0 2.3 1.5 1.4 0.1 0.1 0.1 - India South Africa Mauritius Lao PDR Nicaragua Europe Union Tunisia Kenya Guatemala Honduras El Salvador Costa Rica Uganda Namibia Note: This figure shows the coverage ratio of registration requirements of importers due to SPS reasons across all countries for which data are available. Source: Authors’ computation SPS measures disproportionally affect intra-regional trade in Central America. Figure 4 compares total coverage ratios in the region, computed using total imports (X-axis), versus regional coverage ratios, computed using regional imports (Y-axis), for SPS and TBT measures. Each dot represents a NTM code, at the highest level of disaggregation. An NTM located above the 45 degree line indicates that it affects more imports from regional trading partners than imports from all countries in the world. This analysis shows that while TBT measures are relatively balanced across trading partners, SPS measures have a profound impact on intra-regional trade. Figure 4: Incidence of NTMs in Intra-regional Trade (Coverage Ratios) 15 Note: This scatter plot compares total coverage ratios, computed using total imports (X-axis), versus regional coverage ratios, computed using regional imports (Y-axis), for SPS and TBT measures. Each dot represents a NTM code, at the highest level of disaggregation. Data are plotted for Costa Rica, El Salvador, Guatemala, Honduras and Nicaragua. Source: Authors’ computation 16 3. Price Rising Effects a. Price Data: The Trade Unit Value Database Trade unit values from the French Research Center in International Economics (CEPII) are used to approximate trade prices. 3 These data provide reliable and comparable unit values across countries. Following Cadot and Gourdon (2012), we use Cost of Insurance and Freight (CIF) trade unit values (reported by the importer country) to estimate any potential price-raising effects of NTMs. The CIF price was deemed most appropriate for the analysis as it includes the Freight on Board (FOB) price of the product along with the cost of compliance with SPS and TBT regulations, and any other expenses from pre-shipment inspection, licensing, or any other sort of NTMs. b. Methodology The estimation below can be thought of as a simple treatment-effect approach where prices are “treated” by NTMs. We assume that prices are determined by Equation 3. Here, i indexes countries, k indexes products, p∗k is the world price of product k, pik is its price in country i, t ik the tariff imposed by country i on product k, and λi a cost-of-living adjustment factor for country i depending on its level of income and a number of other characteristics such as landlocked status, remoteness, the quality of the infrastructure, and the regulatory environment. δijk is a dummy variable that equals one if country i imposes an NTM of type j on product k, and zero otherwise. Let aj be the cost of complying with NTM of type j. We assume that this cost is the same across countries and products. Assuming that tariffs and NTM compliance costs are fully passed through to domestic prices, the basic price determination equation is: = ∗ ( + )( + )∏� + � [] Logging Equation 3 yields: ∗ = + (1 + ) + (1 + )+ ∑ �1 + � [4] Due to our assumption of perfect tariff pass-through, we can redefine our dependent variable as: � = − (1 + ) [5] Also note that: ln�1 + � = 1 [6] ln�1 + � = � 0 = 0 so ln�1 + aj δijk � = δijk ln�1 + aj �. 3 This database provides a world-level dataset reporting reliable unit values at a high level of disaggregation. One shortcoming is important to note: despite the treatment of missing unit values in the database, there is still a relatively large share of missing data. Berthou and Emlinger (2011) estimate that the database covered around 80% of world total trade values in 2008. 17 Let k = ln p∗ k and γi = ln(1 + λi ). Using this result, and letting be the error term, the basic estimation equation is: � = + + � + [7] where k and i are product and country fixed effects, respectively. Note that the algebraic interpretation �ȷ is: of the estimated coefficient β �ȷ = ln(1 + a β �ȷ ) or � �ȷ = eβȷ − 1 a [8] �ȷ is the estimated ad-valorem equivalent (AVE) of NTM type j. where a Note that since we do not observe the compliance costs of NTMs (we only know if an NTM exists and the number of such measures), the pass-through parameter for NTMs cannot be identified. Thus, the estimated AVE for NTMs, a �ȷ , is the fraction of the compliance cost that is ‘passed through’, e.g. if a �ȷ = 0.2, an NTM imposed on product k translates, on average, into a 20% rise in its domestic price across products and countries in the sample. A second series of regressions are also run, using counts rather than dummy variables to capture the presence of NTMs. c. Results NTMs are positively and significantly related to domestic prices across our sample of countries. Table 6 shows the baseline regressions results. The first and third specifications use a dummy variable to indicate the presence of an NTM, while the second and fourth specifications use the count of NTMs applied at the product level. Standard gravity type controls are used. Bilateral data on distance between two countries, common language, contiguity and common colonizer since 1945 are sourced from CEPII, and GDP figures come from the World Bank’s World Development Indicators database. Three variables on endowments are used, physical capital stock per worker, arable land per worker and education per worker. These figures are sourced from Shirotori et al (2010). After controlling for systematic cross-country cost- of-living differences, and partner and product specific unobservables, we find that the presence of at least one NTM increases domestic prices by 10.6% (exp(0.101)-1) on average across all 33 countries for which data are available. If we interpret the coefficient with the number of NTMs as the increase in domestic prices of adding one additional NTM (regardless of type), then an additional NTM increases domestic prices by an average of 2.2% (exp(0.022)-1). 18 Table 6. Price Rising Effect of NTMs Across all Countries (CIF Unit Values) Specification NTM dummy Log Number of NTMs NTM dummy Log Number of NTMs Variables [1] [2] [3] [4] NTM 0.101*** 0.022*** (0.008) (0.001) SPS 0.194*** 0.023*** (0.011) (0.002) TBT -0.015* 0.019*** (0.009) (0.003) Log (GDP) -0.082*** -0.087*** -0.083*** -0.087*** (0.003) (0.003) (0.003) (0.003) Log (Distance) 0.155*** 0.161*** 0.158*** 0.161*** (0.004) (0.004) (0.004) (0.004) Log (H/L) 0.713*** 0.704*** 0.694*** 0.705*** (0.013) (0.013) (0.013) (0.013) Log (T/L) -0.160*** -0.151*** -0.156*** -0.151*** (0.004) (0.004) (0.004) (0.004) Log (K/L) 0.367*** 0.375*** 0.369*** 0.375*** (0.005) (0.005) (0.005) (0.005) Contiguity -0.071*** -0.065*** -0.068*** -0.065*** (0.007) (0.007) (0.007) (0.007) Common Official Language 0.120*** 0.115*** 0.119*** 0.114*** (0.008) (0.008) (0.008) (0.008) Common Ethnical Language -0.081*** -0.071*** -0.072*** -0.071*** (0.008) (0.008) (0.008) (0.008) Common Colonizer 0.007 0.006 0.006 0.007 (0.010) (0.010) (0.010) (0.009) Constant 3.088*** 3.123*** 3.104*** 3.127*** (0.099) (0.100) (0.100) (0.099) Exporter FE Yes Yes Yes Yes Product FE Yes Yes Yes Yes Observations 1,922,527 1,922,527 1,922,527 1,922,527 R-squared 0.587 0.588 0.589 0.589 Note: Robust standard errors adjusted for clustering at the product (6-digit HS codes) level are in parentheses. All regressions include exporter and product fixed effects. *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level. Source: Authors’ computation Our results also indicate that SPS measures are positively and significantly related to an increase in border prices. The price increase arising from SPS measures is equivalent to an ad-valorem tariff of 21.4% across all countries and products in the dataset (column 3). By contrast, TBT measures are negatively related to border price with an estimated AVE of -1.5%. The coefficient for TBT is significant only at the 10% level. A negative AVE indicates a net trade-facilitating effect, which may be the result of positive externalities arising from the TBT outweighing its compliance costs (see Beghin et al, 2013). The signs and significance of the control variables are generally as expected. In order to explore the impact of SPS and TBT measures on border prices across geographic regions, we use specification 3 and interact the SPS and TBT dummies variables with a dummy for each world region. 19 Results are presented in Appendix A and graphically depicted in Figure 5 and Figure 6 for SPS and TBT measures respectively. We find that SPS regulations in Central America are positively and significantly related to an increase in border prices, and have a price effect equivalent to a tariff of 12%. This AVE, however, is small compared to that experienced in the Asian and the Sub-Saharan African regions. Figure 5 also shows that the price increase experienced by the Central American countries is close to what is observed for the rest of the Latin America and the Caribbean region. Figure 5: Price-Rising Effect of SPS Measures, by World Bank Region 0.60 0.40 0.43 0.35 0.25 0.20 Coefficient 0.11 0.10 0.07 0.00 -0.20 -0.32 -0.40 -0.60 East Asia and Sub-Saharan South Asia Central Rest of Latin Middle East Europe and Pacific Africa America America and and North Central Asia the Caribbean Africa Note: This figure shows the 95% confidence intervals of the price raising impact of SPS regulations by region. The underlying coefficients are computed as the sum of the free-standing coefficient and the respective interaction term. The level of significance is computed by a Wald test applied to this computation. The point estimate is denoted by an X when the coefficient is not statistically different to zero. Source: Authors’ computation. TBTs, on the other hand, do not have a significant effect on prices in the Central American region (see Figure 6). This is in contrast to the positive estimated impact computed for the rest of the Latin America and Caribbean region. We also find that TBTs have no impact on border prices in Sub-Saharan Africa whereas the South Asian region experiences a negative price effect that is equivalent to a subsidy of 32%. Figure 6: Price-Rising Effect of TBTs, by World Bank Regions 20 0.40 0.30 0.24 0.20 0.19 0.17 0.10 Coefficient 0.07 0.00 0.00 -0.02 -0.10 -0.20 -0.30 -0.28 -0.40 Middle East East Asia and Europe and Rest of Latin Sub-Saharan Central South Asia and North Pacific Central Asia America and Africa America Africa the Caribbean Note: This figure shows the 95% confidence intervals of the price raising impact of TBT regulations by regions. The underlying coefficients are computed as the sum of the free-standing coefficient and the respective interaction term. The level of significance is computed by a Wald test applied to this computation. The point estimate is denoted by an X when the coefficient is not statistically different to zero. Source: Authors’ calculations In order to decompose the impact of SPS measures on each country in the Central American region, we interact country specific dummies with SPS dummies in specification 3. Results are presented in Appendix B and graphically depicted in Figure 7. The impact of SPS measures on prices varies widely between the five countries in Central America. SPS measures are found to have no impact on prices in Honduras and Nicaragua and have the largest price effect in Guatemala with an AVE of 55%. The AVEs in El Salvador and Costa Rica are 23% and 6% respectively. Figure 7: Central America: Price-Rising Effect of SPS Measures, by Country 21 0.70 0.60 0.50 0.44 0.40 Coefficient 0.30 0.26 0.20 0.21 0.22 0.10 0.06 0.00 -0.10 -0.20 -0.30 Guatemala El Salvador Costa Rica Nicaragua Honduras Note: This figure shows the 95% confidence intervals of the price raising impact of SPS regulations by Central American country. The underlying coefficients are computed as the sum of the free-standing coefficient and the respective interaction term. The level of significance is computed by a Wald test applied to this computation. The point estimate is denoted by an X when the coefficient is not statistically different to zero. Source: Authors’ calculations 4. SPS Measures, Household Expenditure, and Consumer Welfare in Guatemala This section concentrates on the impact of SPS measures on domestic prices and, ultimately, on urban poverty in Guatemala. We focus our analysis on those products for which SPS measures in general, and registration requirements in particular, are especially challenging for traders in the region. These products are bread and pastries, beef, chicken meat, and dairy products. 4 Using the empirical framework developed above, this section estimates the price-raising effects of SPS measures for these products and then presents scenarios of their partial and total removal. Specifically, we offer a preliminary analysis of the direct impact of a reduction in domestic prices due to the streamlining of SPS measures on urban poverty levels. The principle behind our calculation of changes in poverty due to a reduction or elimination of NTMs is as follows. For urban households, price-raising NTMs increase the cost of goods purchased, thereby acting as a tax, which should be balanced by the reduction of some externality. Without experimental evidence on the willingness to pay for the reduction of this negative externality, we can only estimate the tax side (the cost, but not the benefit). This value constitutes a lower bound on the social benefit (for example a reduction in mortality or disease risks brought about by the measure) which would be required to justify the presence of the NTM. 4 Consultation with the private sector in the region was undertaken in 2012. The cost of complying with SPS regulations for these four products was identified as one of the main barriers to regional integration. 22 a. Price-Raising Effect of SPS Measures on Selected Products We estimate the price-raising effect of SPS measures by employing the empirical framework explained in Section 3. Specifically, we use interaction terms in our cross-country regression to estimate how different the effects of Guatemalan SPS measures, applied to our selected set of products, are from the average effect across countries and products. We then use Equation 8 to compute ad-valorem equivalents. Table 7 shows the results along with the list of the different NTMs (at the highest level of disaggregation) that importers need to comply with in order to import these products into the Guatemalan market. Table 7. SPS Measures Influencing Trade in Selected Products in Guatemala Beef Bread and Pastry Sanitary and Phytosanitary Measures Sanitary and Phytosanitary Measures A140 Special Authorization requirement for SPS reasons A140 Special Authorization requirement for SPS reasons A150 Registration requirements for importers A150 Registration requirements for importers A310 Labelling requirements A310 Labelling requirements A330 Packaging requirements A330 Packaging requirements A420 Hygienic practices during production A810 Product registration requirements A640 Storage and transport conditions A830 Certification requirement A810 Product registration requirements A820 Testing requirement A830 Certification requirement A840 Inspection requirement AVE 68.4% 51.4% Chicken Meat Dairy Products Sanitary and Phytosanitary Measures Sanitary and Phytosanitary Measures A140 Special Authorization requirement for SPS reasons A140 Special Authorization requirement for SPS reasons A150 Registration requirements for importers A150 Registration requirements for importers A310 Labelling requirements A310 Labelling requirements A330 Packaging requirements A330 Packaging requirements A420 Hygienic practices during production A420 Hygienic practices during production A640 Storage and transport conditions A640 Storage and transport conditions A810 Product registration requirements A810 Product registration requirements A820 Testing requirement A820 Testing requirement A830 Certification requirement A830 Certification requirement A840 Inspection requirement A840 Inspection requirement AVE 22.0% 5.0% Note: This table lists the SPS measures which apply to the selected products. SPS ad-valorem equivalents are statistically significant at least at the 10% level. The HS codes associated with each product are as follows. Beef: 0201, 0202, 0206, 1502. Bread and Pastry: 1905. Chicken meat: 0207. Dairy products: 0401, 0402, 0403, 0404, 0405. Source: Authors’ computation 23 The results confirm that registration requirements for SPS reasons (the so-called sanitary registries) are a measure affecting trade in Guatemala. As shown in Figure 3, there are only a few other countries that apply these types of measures as broadly as Guatemala. Other regulations, such as labeling and packaging requirements are also imposed on these goods. The impact of SPS measures varies greatly across these four products. According to our estimation, the average increase in the Guatemalan domestic price of beef due to SPS measures is equivalent to an ad-valorem tariff of 68.4%. This seems consistent with the perception of the Guatemalan beef market as quite closed to regional exporters, especially trade from Nicaragua. The AVE equivalents for the SPS measures facing bread and pastry, chicken meat and dairy products are 51.4%, 22.0% and 5.0% respectively. b. Household Expenditure on Selected Products We now explore the implications of our results for the cost of living across the income distribution in Guatemala. To do this, we combine our estimated AVEs with household expenditure data from the 2011 Guatemalan Household Survey (HHS) and calculate the budget share of these products by quintiles of the income distribution. Results are presented in Table 8. As expected, given the essential nature of these items, the percentage of total income spent on them decreases with household income. Therefore, any policy aimed at streamlining NTMs would mainly benefit the poorest sector of the population in Guatemala. Table 8. Guatemala: Households’ Budget Shares of Selected Products (%) (By Income Quintiles). Products Q1 Q2 Q3 Q4 Q5 Beef 1.18 0.90 0.71 0.57 0.35 Bread and Pastries 1.04 0.96 0.85 0.69 0.38 Chicken Meat 1.30 1.03 0.83 0.70 0.46 Dairy Products 0.59 0.50 0.41 0.36 0.24 Total 4.11 3.39 2.80 2.31 1.43 Note: This table shows households’ budget shares of the selected products across the quintiles of income distribution. Budget shares are the percentage of total annual income by household spent on the consumption of each good. Each type of product is mapped to a group of goods in the HHS as follows: Beef (carne de cerdo con y sin hueso and carne de res sin y con hueso), bread and pastries (pan de rodaja, pan dulce, pan frances), chicken meat (carne de pollo o gallina, viceras de pollo o gallina), and dairy products (leche liquida, queso fresco o duro). Source: Authors’ computation using the Guatemala 2011 HHS. c. Measuring the Impact of Streamlining SPS Measures on Poverty We now look at the hypothetical cases of the total and partial removal of SPS measures on our set of selected products and analyze the impact of these changes on poverty rates in Guatemala. We follow Porto (2010) in this analysis. Using the 2011 Guatemalan Household Survey, we calculate the head count ratio (HC), defined as the fraction of the population with an income below the poverty line z. That is, 24 1 = � 1� < � [9] where N is total population, is income, and 1{ } is an indicator function that takes value 1 if the < within brackets is true. Guatemala uses two poverty lines. The extreme poverty line is the cost of purchasing the poverty consumption basket, which includes food items that satisfy a minimum caloric and energetic intake. 5 This value corresponds to 4,380 quetzales per person, per year (around 560 US dollars). The general poverty line includes some expenditure for non-food items and is set to 9,030.93 quetzales per person, per year (around 1,150 US dollars). 6 The poverty analysis requires a comparison of the proportion of individuals in poverty before and after the policy scenarios. Given a poverty line z, the head count is given by F(z), where F(z) is the observed cumulative distribution function of income before the shock. The head count is defined as = � () [10] 0 where f(.) is the density of per equivalent adult income associated to F(.). Let t* be the policy parameter (partial or total removal of NTMs) that represents the level of protection. Differentiating with respect to t*, we get: = () +� () [11] ∗ ∗ 0 ∗ A policy shock that changes the ad-valorem equivalent (t*) has two effects on poverty: a change in the poverty line, as a variation in t* affects consumer prices and, therefore, the cost of purchasing the poverty bundle, and a change in the distribution of income. For the sake of simplicity, this study focuses on the first effect (the direct effect on the poverty line). The poverty line can also be defined as = ∑ , where is the price of good , and is the quantity determined in the construction of the poverty line. Holding quantities constant, the change in the (log) poverty line caused by a change in price of good is given by: ln = ln (t*) [12] where is the weight attached to good and ln (t*) is the price-raising effect of NTMs. We update the poverty line by summing equation [12] across all products affected by the policy shock. ̂ ln = � � ∗ ln ( ) [13] ∈ We estimate the weights, � , using the average budget share spent on food products by households in the second quintile of the distribution. 5 These values correspond to 2,246 calories in urban areas and 2,362 in rural areas. 6 Guatemala does not compute different poverty lines for rural and urban households. 25 We explore two policy shocks associated with an attempt to streamline SPS measures in Guatemala. The first is the unrealistic case of a total elimination of the costs of complying with SPS measures for the selected products (i.e. AVE=0). The second is the case where efforts to streamline SPS measures reduce the AVE by half. Given that we abstract from the impact on production (the change in income distribution), we only consider the impact on poverty for urban households. The underlying assumption is that the price change brought about by the NTM reduction decreases prices but leaves income distribution unchanged because urban households are employed in economic activities outside the agricultural sector. Finally, since we are not allowing the non-tradable prices to adjust, we center the discussion on the impact of our policy scenarios on extreme poverty rates, although we also report the results for general poverty rates. Table 9 compares the proportion of the population living in poverty now (2011) and after streamlining SPS measures. Table 9. Poverty Impact of Streamlining SPS Measures in Guatemala (Urban Poverty Rates). Initial Ex-Post Poverty Poverty Poverty Rates Line Rate Total Partial Extreme 5.07 4.59 4.91 Total 34.97 32.66 33.98 Note: This table shows the poverty impact of total and partial elimination of SPS measures applied to beef, bread and pastries, chicken meat, and dairy products in Guatemala. Source: Authors’ computation using the 2011 Guatemalan HHS. The main finding is that an effort to streamline NTMs in Guatemala would likely reduce poverty. Specifically, working towards rationalizing the processes related to SPS measures affecting beef, bread and pastries, chicken meat, and dairy products would reduce domestic prices and, thus, make the basic consumption basket more affordable. Our computations suggest that streamlining the key SPS measures affecting these products by, for example, reducing the cost and time required to obtain sanitary registries, could reduce the urban poverty rate from 5.07% to 4.91%. This would mean lifting approximately 20 thousand people out of extreme poverty in urban areas in Guatemala. Figure 8 depicts the left shift in the poverty line due to a scenario of total elimination of the cost of complying with SPS measures in our products. The dotted line is the new extreme poverty line under total reduction of price-controls in Guatemala. Figure 8: Poverty Effect of Streamlining SPS Measures in Guatemala Shift in Urban Poverty Lines 26 .6 .4 Density .2 0 7 8 9 10 11 12 13 Log per Equivalent Adult income kernel = epanechnikov, bandwidth = 0.1028 Note: This figure shows the density function of per equivalent adult income and two poverty lines (the vertical lines). The dashed poverty line corresponds to the situation after the hypothetical elimination of SPS measures in Guatemala. Source: Authors’ computation using the 2011 Guatemalan HHS. 27 5. Suggested Topics for Policy Dialogue Central America should view the NTM agenda as a crucial issue in enhancing regional integration, increasing private sector competitiveness, and reducing poverty. Sanitary registrations for processed food and beverages are a hindrance to deeper regional integration. To enable goods to enter the market, companies need to obtain a sanitary registry. On average, it takes 70 days to obtain this document (INCAE 2012), and costs an average of US$400 the first time a product enters the market. Even though countries have already approved a regulation for mutual recognition of sanitary registries for food and beverages, the implementation process is stymied by regulatory and technical constraints. The first registry is done in the producer’s country, but the company needs to manually re-register the product in the other Central American countries and provide physical documentation. This process is costly, unharmonized, and bureaucratic. It takes 3-8 days to reregister a product by mutual recognition and costs an average of US$250 in each of the remaining countries. These lengthy and costly procedures, as well as the lack of a centralized system for mutual recognition of sanitary registries, increase the cost of doing business in the region. A solution that could make great strides in this arena is the formation of a regional, electronic database that centralizes documentation about the credentials that each firm submits to meet the sanitary registration requirements. This would lessen the time and cost burden on exporters, who currently have to gather all the documentation each time they apply for a sanitary registry. Other potential problems could be identified through a well-organized consultation process with the private sector. Technical solutions should be sought through careful analysis and public-private dialogue with the aim of reducing the burden of the regulatory environment on the economy. The underlying philosophy of streamlining NTMs is similar to what is known as “regulatory improvement” or “Regulatory Impact Assessment” (RIA). 7 However, whereas RIA is for ex-ante analysis of measures, a competitiveness agenda should include a review of existing measures in response to specific demands from countries struggling with legacies of complicated and penalizing regulations. Dealing with existing measures has the advantage of responding to an immediate need and focusing on measures whose effects are known. It can be thought of as the first stage of a process of regulatory improvement covering not just the stock of existing measures, but also the process of creating new ones. Central America can address the NTM agenda by setting up a regional secretariat to improve national and regional competitiveness and by convening a forum to consider setting up a public-private steering committee. This should include the main agencies that regulate trade as well as representatives from the business chambers and other key stakeholders such as consumer protection organizations from all member countries. Efforts to anchor NTM reforms in regional integration have led to various forms and levels of implementation of regional secretariats across the world, which mostly depend on the political economy, capacity of national and regional institutions, and the approach adopted. Despite these challenges, some regions have been more successful than others in effectively addressing and streamlining NTMs. The Association of Southeast Asian Nations (ASEAN) has adopted the most ambitious NTM agenda amongst developing countries. Low-income economies in Southern Africa, also determined to address NTMs regionally, have achieved effective results in terms of transparency. Finally, the EU offers the most comprehensive example of economic integration in NTMs. 7 See Cadot and Malouche (2012) for a complete overview of these processes. 28 Another important issue in the region is the need to disseminate information on the current laws and regulations on import and export procedures to the wider public, particularly the business community. In this regard, grouping and regularly updating information in a regional portal would alleviate informational asymmetries that increase the cost of doing business. 29 6. Concluding Remarks With the decline in traditional trade barriers initiated by the DR-CAFTA, Non-Tariff Measures (NTMs) are increasingly viewed as a key trade policy issue in Central America. Yet, despite anecdotal evidence that borders remain thick, there is little quantitative understanding of the incidence of NTMs in the region. This paper provides, to the best of our knowledge, the first attempt to look at incidence of technical NTMs (SPS and TBT regulations) and to estimate their impact on border prices in Central America. It also assesses the impact of SPS measures on the border prices of important consumption products in Guatemala and estimates the poverty impact of streamlining key SPS measures affecting them. We use data collected by the World Bank in Central America in 2012 and carry out three distinct analyses. Firstly, we provide a descriptive analysis of the incidence of technical NTMs in the Central American region. Our results indicate that, of the seven geographic regions of the World Bank, Central America has the lowest prevalence of these regulations, with 33.3% of imports covered by at least one technical NTM. By contrast, over half of all goods imported to Sub-Saharan Africa, Europe and Central Asia and the Middle East and North Africa are governed by at least one technical NTM. However, there is significant variation in the use of NTMs within the Central American region with El Salvador using them most intensively (coverage ratio of 48.2%), but Honduras availing of them quite sparingly (coverage ratio of 16.2%). We then use econometric techniques to estimate the impact of NTMs on domestic price and find that SPS measures increase prices by an average of 11.6% across the Central American region. However, the ad- valorem equivalent varies widely across the individual countries, ranging from 6% in Costa Rica to 55% in Guatemala, with NTMs having no significant price impact in either Nicaragua or Honduras. Private sector consultation undertaken in 2012 identified four products that are important for regional trade and for which SPS measures in general, and registration requirements in particular, are especially challenging: beef, bread and pastry, chicken meat, and dairy products. Using our empirical framework, we find that the ad-valorem equivalents of SPS measures for these products are 68.4%, 51.4%, 22.0%, and 5.0%, respectively. Finally, we explore the poverty impact of streamlining NTMs measures in Guatemala by merging the NTM data with Guatemalan household survey data. The main finding is that rationalizing the use and processes related to compliance with SPS measures would reduce domestic prices. This would make the basic consumption basket more affordable benefitting the poorest sector of the population. Our computations suggest that streamlining the main SPS measures affecting these products by, for example, enforcing harmonization and mutual recognition of sanitary registries, could reduce the urban poverty rate from 5.07% to 4.91%. We suggest three issues for policy dialogue at the regional level aimed at streamlining trade regulation and improving national and regional competitiveness. Our first suggestion is to better disseminate information on the current laws and regulations about import and export procedures to the wider public, particularly the business community. In this regard, grouping and regularly updating information in a regional portal would alleviate informational asymmetries that increase the cost of doing business in Central America. The second suggestion is considering setting up a regional secretariat by assembling a public-private steering committee that oversees and discusses the regional regulatory framework affecting trade. A comprehensive regulatory impact assessment for existing and proposed NTMs may be initiated 30 from this committee. The third suggestion is the establishment of a regional, electronic database that centralizes documentation about the credentials that each firm submits to meet the sanitary registration requirements in each country in the region. 31 7. References Aisbett, E. and L. Pearson (2012), “Environmental and Health Protections, or New Protectionism? Determinants of SPS Notifications by WTO Members”, Crawford School Research Paper No. 12-13, December Akerlof, G. (1970), “The Market for “Lemons”: Quality Uncertainty and the Market Mechanism”, The Quarterly Journal of Economics, 84 (3), pp 488-500 Andriamananjara, S., J. M. Dean, R. Feinberg, M. Ferrantino, R. Ludema, M. Tsigas (2004), “The Effects of Non-Tariff Measures on Prices, Trade, and Welfare: CGE Implementation of Policy-Based Price Comparisons”, USITC Economics Working Paper No. 2004-04-A Bacchetta, M., J. Richtering and R. Santana (2012), “How Much Light Do WTO Notifications Shed on NTMs?” in Non-Tariff Measures. A Fresh Look at Trade Policy’s New Frontier (eds) Cadot O. and M. Malouche, The World Bank Beghin, J. and J-C Bureau (2001), “Quantitative Policy Analysis of Sanitary, Phytosanitary and Technical Barriers to Trade”, Economic Internationale 87, pp 107-130 Beghin, J., A-C Disdier, and S. Marette (2013), “Trade Restrictiveness Indices in Presence of Externalities: An Application to Non-Tariff Measures”. Working Paper No. 12023, Department of Economics, Iowa State UniversityBerthou, A. and C. Emlinger (2011), “The Trade Unit Values Database”, CEPII Working Paper 2011-10, CEPII Research Center Bradford, S. (2003), “Paying the Price: Final Goods Protection in OECD Countries”, The Review of Economics and Statistics, 85(1), pp 24–37 Cadot, O. and J. Gourdon (2012), “Assessing the Price-Raising Effect of Non-Tariff Measures in Africa”, CEPII Working Paper 2012-16, CEPII Research Center Cadot, O. and M. Malouche (eds), (2012), Non-Tariff Measures. A Fresh Look at Trade Policy’s New Frontier, The World Bank Calvin, L. and B. Krissoff (1998), “Technical Barriers to Trade: A Case Study of Phytosanitary Barriers and U.S.-Japanese Apple Trade”, Journal of Agricultural and Resource Economics, 23(2), pp 351-366 Cato J. C., W. S. Otwell, and A. S. Coze (2005), “Nicaragua’s Shrimp Subsector: Developing a Production Capacity and Export Market during Rapidly Changing Worldwide Safety and Quality Regulations”, Agriculture and Rural Development Discussion Paper, The World Bank Chemingui, M. and S. Dessus, (2008), “Assessing Non‐Tariff Barriers in Syria”, Journal of Policy Modeling, 30(5), pp 917-928 Chemnitz, C. and N. Künkel (2006), “Standards, a Catalyst for the Winners – a Barrier for the Losers? An Empirical Analysis of the Impact of Higher SPS Measures on the Trade Performance of Developing Countries”, Paper Prepared for Presentation at the 98th EAAE Seminar ‘Marketing Dynamics within the Global Trading System: New Perspectives’, Chania, Crete, Greece on 29 June – 2 July, 2006 32 Crivelli, P. and J. Gröschl (2012), “The Impact of Sanitary and Phytosanitary Measures on Market Entry and Trade Flows”, Ifo Working Paper No. 136 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), pp 336-350 Fischer, R. and Serra, P. (2000), “Standards and Protection”, Journal of International Economics, 52(2), pp 377–400 Fontagné, L., M. Mimouni and J-M Pasteels (2005), “Estimating the Impact of Environmental SPS and TBT on International Trade”, Integration and Trade Journal, 22, pp 7-37 Ganslandt, M. and J. Markusen (2001), “Standards and Related Regulations in International Trade: A Modeling Approach”, NBER Working Paper 8346, National Bureau of Economic Research Gordillo D. M., A. Stokenberga, and J. Schwartz (2010), “Understanding the Benefits of Regional Integration to Trade: The Application of a Gravity Model to the Case of Central America”, World Bank Policy Research Working Paper 5506 Gourdon, J. and A. Nicita (2012), “NTMs: Interpreting the New Data”, in Non-Tariff Measures. A Fresh Look at Trade Policy’s New Frontier (eds) Cadot O. and M. Malouche, The World Bank Henson S., R. Loader, A. Swinbank, M. Bredahl, and N. Lux (2000), “Impact of Sanitary and Phytosanitary Measures on Developing Countries”, Centre for Food Economics Research, University of Reading, April INCAE Business School (2013), “Registros Sanitarios de Alimentos y Bebidas en Centroamérica” mimeo Kee, H., A. Nicita and M. Olarreaga (2009), “Estimating Trade Restrictiveness Indices”, Economic Journal, Royal Economic Society, 119(534), pp 172-199, 01 Kono, D. (2006), “Optimal Obfuscation: Democracy and Trade Policy Transparency”, American Political Science Review, 100(3), pp 369-384 Porto, G. (2010), “International Market Access and Poverty in Argentina”, Review of International Economics, 18(2), pp 396-407 Portugal-Perez, A. J-D Reyes and J. Wilson (2010), “Beyond the Information Technology Agreement: Harmonization of Standards and Trade in Electronics”, The World Economy, 33(12), pp. 1870-1897 Reyes, J-D (2011), “International Harmonization of Product Standards and Firm Heterogeneity in International Trade” World Bank Working Paper 5677 Reyes, J-D (2012), “The Pro-Competitive Effect of International Harmonization of Product Standards” in Non-Tariff Measures. A Fresh Look at Trade Policy’s New Frontier (eds) Cadot, O. and M. Malouche, The World Bank 33 Shirotori, M., B. Tumurchudur, and O. Cadot (2010) “Revealed Factor Intensity Indices at the Product Level”, Policy Issues in International Trade and Commodities Study Series No. 44, UNCTAD UNCTAD (2010), “Non-Tariff Measures: Evidence from Selected Developing Countries and Future Research Agenda”, Developing Countries in International Trade Studies, United Nations Conference on Trade and Development, New York and Geneva, ISSN: 1817-1214 UNCTAD (2012), “Classification of Non-Tariff Measures”, Geneva, February Walkenhorst, P. and T. Yasui (2009), “Quantitative Assessment of the Benefits of Trade Facilitation”, in Overcoming Border Bottlenecks: The Costs and Benefits of Trade Facilitation, OECD, Paris WTO (2012), “World Trade Report, 2012, Trade and Public Policies: A Closer Look at Non-Tariff Measures in the 21st Century”, Geneva 34 Appendix A: Region Central Rest of Latin Sub-Saharan East Asia Europe and Middle East and South Variables America America and the Caribbean Africa and Pacific Central Asia North Africa Asia SPS 0.093*** 0.123*** 0.258*** 0.142*** 0.222*** 0.181*** 0.203*** (0.013) (0.011) (0.011) (0.012) (0.011) (0.011) (0.012) X Region Dummy 0.014 -0.053** 0.096*** 0.287*** -0.542*** -0.083 0.051 (0.020) (0.021) (0.029) (0.039) (0.049) (0.059) (0.035) TBT 0.007 -0.032*** -0.065*** -0.050*** -0.022** -0.038*** 0.029*** (0.009) (0.011) (0.010) (0.009) (0.009) (0.009) (0.010) X Region Dummy -0.026 0.097*** 0.064*** 0.239*** 0.189*** 0.283*** -0.312*** (0.034) (0.017) (0.020) (0.038) (0.056) (0.028) (0.027) Region Dummy 0.544*** -0.675*** 0.754*** -0.385*** 0.818*** -0.324*** -0.087*** (0.020) (0.011) (0.015) (0.016) (0.031) (0.013) (0.014) Log (GDP) -0.036*** -0.079*** -0.055*** -0.090*** -0.086*** -0.088*** -0.084*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Log (Distance) 0.144*** 0.279*** 0.097*** 0.146*** 0.186*** 0.111*** 0.154*** (0.004) (0.003) (0.004) (0.004) (0.003) (0.003) (0.004) Log (H/L) 0.900*** 0.958*** 0.503*** 0.785*** 0.530*** 0.603*** 0.634*** (0.014) (0.014) (0.013) (0.014) (0.013) (0.013) (0.014) Log (T/L) -0.135*** -0.097*** -0.172*** -0.165*** -0.214*** -0.150*** -0.173*** (0.004) (0.004) (0.004) (0.004) (0.005) (0.004) (0.004) Log (K/L) 0.291*** 0.391*** 0.480*** 0.345*** 0.381*** 0.389*** 0.370*** (0.006) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) Contiguity -0.063*** 0.147*** -0.051*** -0.089*** -0.046*** -0.138*** -0.060*** (0.007) (0.006) (0.007) (0.007) (0.007) (0.006) (0.007) Common Official Language 0.152*** 0.028*** -0.038*** 0.121*** 0.144*** 0.117*** 0.103*** (0.008) (0.008) (0.007) (0.008) (0.008) (0.008) (0.008) Common Ethnical Language -0.099*** 0.012 -0.073*** -0.069*** -0.070*** -0.086*** -0.085*** (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) Common Colonizer 0.043*** -0.016* -0.064*** -0.023** -0.036*** -0.042*** 0.056*** (0.009) (0.010) (0.009) (0.009) (0.010) (0.010) (0.009) Constant 2.365*** 1.301*** 1.656*** 3.479*** 3.100*** 3.679*** 3.259*** (0.096) (0.095) (0.098) (0.102) (0.101) (0.100) (0.099) Exporter FE Yes Yes Yes Yes Yes Yes Yes Product FE Yes Yes Yes Yes Yes Yes Yes Observations 1,922,527 1,922,527 1,922,527 1,922,527 1,922,527 1,922,527 1,922,527 R-squared 0.591 0.600 0.598 0.589 0.590 0.589 0.589 35 Appendix B: Region Costa El Salvador Guatemala Nicaragua Honduras Variables Rica SPS 0.183*** 0.198*** 0.183*** 0.128*** 0.176*** (0.00978) (0.00977) (0.00962) (0.0102) (0.0101) X Country Dummy -0.124*** 0.041*** 0.257*** 0.090 0.083 (0.0382) (0.0486) (0.0491) (0.200) (0.187) Country Dummy -0.420*** 0.456*** 1.172*** 0.507** 0.0608 (0.0163) (0.0161) (0.0197) (0.199) (0.186) Log (GDP) -0.0949*** -0.0775*** -0.0633*** -0.0738*** -0.0818*** (0.00308) (0.00311) (0.00302) (0.00314) (0.00314) Log (Distance) 0.159*** 0.160*** 0.129*** 0.156*** 0.158*** (0.00352) (0.00354) (0.00352) (0.00352) (0.00353) Log (H/L) 0.702*** 0.681*** 1.042*** 0.738*** 0.698*** (0.0131) (0.0132) (0.0152) (0.0131) (0.0130) Log (T/L) -0.175*** -0.152*** -0.151*** -0.165*** -0.156*** (0.00442) (0.00421) (0.00419) (0.00428) (0.00422) Log (K/L) 0.384*** 0.369*** 0.292*** 0.356*** 0.368*** (0.00503) (0.00499) (0.00531) (0.00499) (0.00502) Contiguity -0.0738*** -0.0609*** -0.119*** -0.0574*** -0.0679*** (0.00666) (0.00668) (0.00682) (0.00668) (0.00666) Common Official Language 0.119*** 0.126*** 0.130*** 0.131*** 0.121*** (0.00799) (0.00804) (0.00805) (0.00797) (0.00799) Common Ethnical Language -0.0726*** -0.0772*** -0.0771*** -0.0823*** -0.0744*** (0.00778) (0.00781) (0.00780) (0.00778) (0.00777) Common Colonizer 0.00137 0.0125 0.0204** 0.0155 0.00704 (0.00954) (0.00953) (0.00955) (0.00952) (0.00955) Constant 3.240*** 2.960*** 2.916*** 2.923*** 3.076*** (0.0987) (0.0994) (0.0987) (0.0998) (0.0993) Exporter FE Yes Yes Yes Yes Yes Product FE Yes Yes Yes Yes Yes Observations 1,922,527 1,922,527 1,922,527 1,922,527 1,922,527 R-squared 0.589 0.589 0.592 0.589 0.588 36