Impact of COVID-19 on the trade of goods and services in Spain

The COVID-19 crisis has led to the sharpest collapse in the Spanish trade of goods and services in recent decades. The containment measures adopted to arrest the spread of the virus have caused an especially intense fall of trade in services. Spain's export specialization in transport equipment, capital and outdoor goods, and services that rely on the movement of people has made the COVID-19 trade crisis more intense in Spain than in the rest of the European Union. However, the nature of the collapse suggests that trade in goods can recover swiftly when the health crisis ends. On the other hand, COVID-19 may have a long-term negative impact on the trade of services that rely on the movement of people.


Introduction
There is little doubt that 2020 will be remembered as the  year. The virus that emerged in China has, thus far, infected more than 51 million and killed almost 1. share of transport equipment, capital goods, products that are consumed outdoors (i.e., outdoor goods), and tourism in Spanish exports, and the fact that demand for these products and services decreased notably during the COVID-19 crisis, the pandemic has had a larger negative effect in Spain than in the rest of the European Union (EU).
The paper begins with a chronology of the trade collapse generated by  Spain. Although the first news about the virus emerged in January 2020, the trade crisis did not begin until confinement measures were introduced in Spain and its main trading partners in March. The most severe phase of the crisis occurred between the second half of March and May 2020, when very strict containment measures were implemented. However, the year-to-year decreases in trade flows began to soften from June 2020 onward. On the other hand, as this article goes to press, the latest available monthly trade figures of August were still lower than those at the same point in 2019.
Second, to understand the severity of the trade collapse, I compare the drop in exports during the COVID-19 crisis with the aforementioned Great Recession of [2008][2009]. I show that exports dropped faster during the COVID-19 crisis, but that the total drop was lower during the COVID-19 crisis. Regarding services, the first three months of the COVID-19 crisis were enough to reach the same overall collapse. In both cases, Spanish firms continued to sell to their foreign customers, but at lower amounts. Because Spanish firms maintained their relationship with foreign customers, it is reasonable to expect a rapid recovery of good exports after the health crisis ends.
Third, I show that COVID-19 had a larger negative impact on Spain's exports than on those of the rest of the EU. I further find that the country distribution of Spanish exports contributed to this differential impact. I also illustrate that COVID-19 had an heterogeneous impact on Spanish regions' export of goods.
Fourth, I analyze whether Spanish firms' dependence on imported goods contributed to the COVID-19 trade crisis. I show that a large share of Spanish imports was sourced by importers that only had a single supplier country. Despite this dependence, Spanish exporters did not identify disruptions in the supply chain as a major factor that drove the decrease in trade flows. However, the large increase in demand for medical materials and equipment, and the fact that Spain already depended on imports before the COVID-19 crisis led to severe sourcing problems for these products during the first stages of the health crisis.
Fifth, I speculate whether the COVID-19 crisis will cause permanent changes to Spanish trade. I argue that it may lead to changes in the composition of service exports, have a negative impact on the growth of Spanish exports in the medium term, and increase the concentration of Spanish exports in EU countries. The last section concludes this paper.

Chronology of a trade crisis
The first significant negative impact of COVID-19 on Spanish trade occurred on February 12, 2020, when the organizers of the World Mobile Congress, a major event contributing to service exports, announced its cancellation because of health risks. That month, firms began to worry that a lack of intermediate inputs manufactured in China would lead to production shutdowns (Expansión, 2020). These concerns further increased when some northern Italian regions that hosted factories that exported components to Spain were put under quarantine at the end of February (Vozpópuli, 2020). However, it was in March, in response to the declaration of the state of alarm in Spain and the introduction of confinement measures in many other countries, when COVID-19 began to have a significantly negative impact on trade flows.
Panel A1 of Figure 1 shows the evolution of good exports in 2019 and 2020. 2 Goods represented two-thirds of all Spanish export values in 2019. The blue dash and the red solid lines are monthly exports in 2019 and 2020, respectively. Exports in January and February were 2.7% and 4.2% higher than during the same months of 2019. With the introduction of containment measures in many EU countries (the most important destinations of Spanish exports), the value of exports decreased by 14. 5%,39.3%,and 34.3% in March,April,and May, respectively, compared to the previous year. Coinciding with the relaxation of containment measures at the destinations of Spanish exports, the decreases began to smoothen: exports "only" dropped by 9.2%, 7.5%, and 9.1% in June, July, and August, respectively.   The negative impact of COVID-19 on trade was greater for services than for goods (Panels B1 and B2 of Figure 1). 5 Many services require the movement of the supplier or the customer, and the confinement measures made such movement across national boundaries almost impossible. Export of services decreased by more than 60% between April and June, and almost by half in July and August relative to the same months of 2019. However, the drop in imports, although large, was smaller than in exports.
Tourism, which represented 49.4% and 32.7% of service exports and imports in 2019, respectively, was the main contributor to the massive reduction in the trade of services (Panels C1 and C2 of Figure 1). In April and May 2020, tourism receipts and payments were zero. In June, when hotels were allowed to reopen in Spain, and border controls were lifted in the rest of Europe, tourism receipts and payments were still 98.0% and 94.1% lower, 4 According to INE and Eurostat data, during the second quarter of 2020, Spanish GDP decreased by 17.8% relative to the first quarter of 2020. In the EU, the most important Spanish trading partner, the drop was of 11.5%. 5 Data on service exports and imports were obtained from the Bank of Spain, Balance of Payments and International Investment Position database, available at https://www.bde.es/webbde/en/estadis/ infoest/temas/sb_extbppii.html. respectively, than during the same period of 2019. Although tourism receipts and payments recovered in July and August, they remained around three-quarters lower than those of the same period of 2019.
The decrease in non-tourism services trade, comprising transport, finance, communications, and business services, was much lower than that of tourism. Between March and August 2020 exports and imports of non-tourism services decreased by 12.6% and 16. Overall, during the March-August period of 2020, exports and imports of goods decreased by 19.5% and 24.1%, respectively, relative to the same period of 2019. Service exports and imports decreased by 52.5% and 37.8%, respectively. If we add exports to imports and merchandises to services, total Spanish trade decreased by 28.9%. This was the largest biannual drop in Spanish trade in decades. COVID-19 also had a negative impact on the number of firms exporting and importing goods ( Figure 2). 7 During March and April, there was a decrease in the number of traders, especially exporters. However, from May onward, there was a very fast recovery in that number. In August 2020, the number of exporters was 6.6% lower than that during the same month of 2019. Contrarily, the number of importers was already higher than that of 2019.
The drop in the number of exporters would be more worrisome if it had affected regular

COVID-19 vs. the Great Recession
To for each year and month, I compute a moving window totaling the value of exports in the actual month and the previous eleven months: 8 Regular exporters are defined as firms that exported without interruption in the last four years. 9 There were many regular exporters that exported a very small amount per year. I recalculated the number of regular exporters removing all export transactions with a value lower than 1,500 euros. The number of regular exporters decreased from 37,873 in January to 37,519 in August, which represented a 1% fall.
where X m,t is the value of exports in the month m in year t , and t ∈ (t − 1, t). I define the beginning of a crisis when the moving window is lower than that of the previous month, but is higher than that of the following month (W indow m−1,t > W indow m,t > W indow m+1,t ).
I give the value of 100 to W indow m,t when m is the month prior to the beginning of the crisis. Using this definition, the Great Recession began in October 2008 for goods and one month later for services. The COVID-19 crisis began in March 2020 for both goods and services. when a firm continues to export a product to a destination and sales are reduced. The extensive margin comprises two sub-margins. Sub-margin (a), denoted by extensive -prod-ucts&destinations, is the reduction in exports when a firm ceases to sell a product in a destination. Sub-margin (b), denoted by extensive -firms, is the reduction in exports when a firm ceases to export. The definition of margins is equivalent for imports.
Recovery from a crisis can be faster if the intensive margin is the main contributor to the decrease in exports and imports. In this margin, the relationship between the exporter (impoter) and its foreign customer (supplier) remains alive, and exports (imports) recover quickly when the slump ends. Contrarily, if the adjustment comes from the extensive margin, many trade relationships will be lost. Firms will then need to invest in recovery efforts and create new export (import) relationships. Because these investments are costly, some firms cannot return to their foreign markets, even when demand recovers. new exporters was larger than the reduction of exports by firms that ceased to export. 13 The intensive margin was also the main contributor to the decrease in imports during the COVID-19 crisis: 93.3%. The reduction in the destination and product portfolio was 7.9%. Interestingly, new importers smoothed the decrease in imports by 1.2%. The intensive margin was also the main contributor to the drop in imports during the Great Recession, but with a lower percentage: 82.1%. The drop in imports by ongoing importers was larger during the Great Recession, 12.5%. As opposed to the COVID-19 crisis, imports by new importers did not compensate the loss of imports by firms that ceased to import.
The large share of the intensive margin in the decrease of exports and imports, therefore, leads to the anticipation of a rapid recovery of trade flows when the health crisis ends.
A feature that can further support a rapid recovery is a larger drop in exports among top exporters than non-top exporters. I define top exporters as the top 1% of firms that exported France, Germany, and Italy. 14 I compare total exports of March-August 2020 with that of the same period of 2019. 15 The decrease in exports in Spain, 31.0%, was three-quarters larger than that of the EU22 (17.6%). Exports also fell more in Spain than in Germany (16.6%), France (26.0%), and Italy (23.1%). These figures indicate that COVID-19 had a particularly negative effect on Spanish exports.
Panel (B) of Figure 5 shows the decrease in the export of goods. The decrease in Spain (19.5%) was also larger than that of the EU22 (16.6%), but the difference was lower than that shown in panel (A). The export of goods decreased more in Spain than in Germany (16.3%), was similar to that of Italy (19.9%), and less than that of France (28.0%). Panel (C) shows that the differential impact of COVID-19 on Spanish exports can be explained by the large decrease in service exports (52.5%), which is more than twice the decrease in the EU22 countries (20.3%). Italy also experienced a very large reduction in service exports (36.0%).
To understand the larger decrease in the export of goods in Spain than in EU22, Figure 6 presents Contrarily, exports increased in food-related chapters and pharmaceutical products. The specialization in transport equipment, capital goods, and outdoor goods explains the larger 14 The EU22 comprises Belgium, Bulgaria, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Italy, Latvia, Lithuania, Luxembourg, Malta, The Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, and Sweden.
15 Monthly data on export of goods and services for EU countries were obtained from Eurostat's Balance of payments by country -monthly data (BPM6) database (https://appsso.eurostat.ec.europa.eu/nui/ show.do?dataset=bop_c6_m&lang=en). Spain's data was not available in this dataset. Therefore, I used the Customs and Bank of Spain's databases mentioned previously.
16 I exclude chapter 98 (Complete industrial plants) and 99 (Special combined nomenclature codes) from the analysis, because it is unknown the products that were included in them. negative effect of COVID-19 on Spanish export of goods than on that of EU22 countries. 17 The larger decrease in the export of services in Spain can be explained by the larger share of tourism in Spanish exports. According to the Organization for Economic Cooperation and Development data, personal travel accounted for 52.0% of Spanish service exports in 2018, whereas it was only 16.9% for the EU28. As shown in Section 2, tourism was the most 17 An outdoor good is a product that is mostly consumed outside the household. For example, ski-boots are outdoor goods, while slippers are not. The list of outdoor goods comes from de Lucio et al. (2020c). They showed that the export value of goods that are consumed outdoors decreased five times more than that of the remainder consumption goods.  vs 1.5%), and the lower growth of exports in some of these zones than in the remainders. 20

Impact of COVID-19 by region
The impact of the COVID-19 crisis varied across Spanish provinces according to their export specializations. 21 Figure 7 shows Darker colors represent provinces that specialized in vehicles, fuels, and apparel. Contrarily, provinces with lighter colors specialized in foodstuff, pharmaceuticals, and miscellaneous manufacturing. 23 These patterns are in line with Figure 6, which shows the HS chapter impact of COVID-19. Although there are no data on the regional distribution of service ex- The largest growth in exports occurred in the province of Salamanca. The product having the highest share of exports in that province in 2019 was meat. However, the growth in exports in 2020 can be explained by nuclear fuel.
24 Data from Exceltur. Available at: https://www.exceltur.org/pib-y-empleo-turistico-por-c-ca-a/ Concerns about the risks of relying on foreign suppliers were especially intense at the beginning of the COVID-19 crisis because they were related to the sourcing of medical materials and equipment (e.g., protective gear and medical ventilators), which were key fighting the pandemic. For example, to ensure that domestic production was dedicated to local needs, in March 2020, the EU introduced a regulation making exports of certain products subject to authorization. 26 In the scramble to acquire needed medical equipment, 25 In another survey, firms considered that a decrease in foreign demand would have a larger impact on their activity than would disruptions in the supply chain (Instituto Nacional de Estadística, 2020). IMF (2020) also contended that the decline in trade volumes largely reflected weak final demand from customers and firms. 26 The EU 2020/402 regulation came after France and Germany introduced barriers to the export of medical equipment that also affected other EU countries. In exchange for removing those barriers, the EU imposed Note: The ratio of imports to domestic production (imports/production) represents 2018. Import growth calculates changes in imports between March and August 2020 and the same period of 2019. The main supplier is the largest supplier in the category during the March-August 2020 period. The main single supplier was the country most repeated among importers having a single supplying country per product as of March-August. Share in total was calculated using imports by firms having only one supplier over total imports. Source: author's own elaboration using data from Prodcom and Customs.
some EU members temporarily seized shipments destined to other member states, were outbid by contenders when shipments were bound to take off from source countries, or purchased products that were defective or below standard (El País, 2020; BBC News, 2020).
Spain was especially vulnerable to the COVID-19 crisis because the domestic capacity could not respond to the increase in demand for medical materials and equipment. Table 1 shows the ratio of imports to domestic production (imports/production) before the COVID-19 crisis for the six categories identified as critical to combating the COVID-19 pandemic by the World Customs Organization (WCO) and the World Health Organization (WHO). 27 I sorted the categories, from top to bottom, by the growth in imports during the COVID-19 crisis, relative to the same period of 2019. The growth in imports/production ratio was very high for protective garments, oxygen-therapy equipment, and tests kits (the three categories that experienced the largest growth of imports during the pandemic). The growth regulations on exports to non-EU countries (Felbermayr and Görg, 2020). This regulation was applied for a six-weeks period, and was further extended for a period of 30 days by regulation 2020/568 (April 23rd, 2020). It ceased to apply on May 26th, 2020. 27 Production data come from the Eurostat Prodcom Database and correspond to 2018. The products included in each critical category and their HS six-digit codes were obtained from WCO and WHO (2020). I linked the HS codes with Prodcom using a correspondence table from the Reference and Management of Nomenclatures (https://ec.europa.eu/eurostat/ramon/relations/index.cfm?TargetUrl=LST_REL).
in protective garment imports was impressive and stark: 430.7%. Nevertheless, most of this increase can be explained by the surge in the price of imports. The kilograms of imported garments only rose by 93%. The dependence ratio was also large for other medical devices.
However, imports increased very little in this category during the pandemic. Germany was the main supplier of imports during the COVID-19 crisis in the following three categories: test kits, other medical devices, and medical consumables. China was the main supplier of protective garments and oxygen therapy. The United States was the top supplier of disinfectants. However, if we analyze the most important source for importers that only relied on a single supplier, China was the most repeated country. The value of imports that were sourced by importers with a single source country per product over total imports was very high for protective garments and non-negligible for the remaining categories.

Will COVID-19 change Spanish trade?
Richard Baldwin argued that "things are the way they are for good reasons. If the reasons don't change permanently, the outcomes won't change permanently" (Baldwin, 2020). In this section, I speculate whether COVID-19 has altered some of those "goods reasons" and if it will lead to permanent changes in Spanish trade.
COVID-19's clearest impact was the reduction of people's mobility. Firms responded to this by moving many activities online. Domestic equipment and internet connections were improved, and workers quickly learned how to be productive remotely. If firms can learn how to manage remotely-provided day-to-day activities, they might decide to further outsource those jobs to personnel located in lower-wage countries (Baldwin, 2019). In parallel, the digital transformation accelerated by COVID-19 might cause firms to be more open to hiring business services from foreign countries. This could increase Spanish exports in services, such as market research or engineering, which can be provided online and where the share of exporters is large (Minondo, 2015).
Contrarily, services requiring the movement of people have been very negatively affected by COVID-19. Until a vaccine is distributed, restrictions to mobility will remain. This and people's ongoing concerns about health conditions in destination countries will continue to discourage tourism and similar provision of services that require the movement of people.
In any case, a vaccine notwithstanding, COVID-19 has shown that tourism is very fragile to pandemics. Firms will explore business models characterized by a lower dependence on long-range foreign tourists and overcrowding. This transition may be particularly difficult for Spain, where tourism represented 12.3% of its GDP and foreign tourists accounted for 75% of the total expenditure in this industry. 28 Restrictions to the movement of people could also have a negative impact on trade in the medium and long runs in Spain. de Lucio et al. (2011b) showed that as new and incumbent exporters' diversified into new destinations and products, they contributed by half to the growth of Spanish exports between 1997 and 2007. To grow along these extensive margins, Spanish firms must build trust-based relationships with new foreign customers.
These relationships rely on a social rapprochement that cannot be successfully substituted with video calls yet. 29 If COVID-19 continues to increase the costs of traveling, international business meetings and fairs will continue to diminish, reducing Spanish firms' opportunities for opening new trade relationships.
It is also likely that countries will rely less on foreign suppliers to source essential products.
To address future crises, countries can build strategic inventories of essential equipment, while keeping production in countries that have a comparative advantage in these products. 30 However, future health crises may demand new types of medical devices, medicines, or tests that can only be developed if local capacities are present. This dynamic perspective may lead 28 INE, Cuenta Satélite del Turismo en España, Encuesta de Turismo de Residentes, and Encuesta de Gasto Turístico. GDP data are from 2018; share of foreign tourists in total expenditure data are from 2019.
29 Despite the improvement in communication brought by the internet, business trips continued to rise (Cristea, 2011;Coscia et al., 2020). 30 Switzerland holds reserves of food and medicines to cope with unseen crises. During the COVID-19 crisis, confidence on the availability of those reserves precluded the panic purchases observed in other countries (Financial Times, 2020). countries to trade efficiency for domestic reaction capacity, nurturing domestic skills through contracts between the public health sector and local firms. 31 For example, on August 6th, 2020, the US President issued an executive order to ensure that essential medicines would be produced in the US. 32 Such nurturing of domestic capacities in health-related products is consistent with previous policies aiming to develop strategic sectors (e.g., electric cars, digital-intensive goods and services, and green technologies). For Spain, it is likely that the drive for higher domestic capacity will be framed under EU's open strategic autonomy framework, which could lead to an increase in the share of this economic area of Spanish trade (European Commission, 2020).
The regionalization of Spanish trade might also be favored by the changes that COVID-19 might cause in global value chains. The Tohoku earthquake and tsunami in Japan in 2011 and the floods in Thailand during the same year already revealed many of the vulnerabilities of global value chains to natural disasters (Boehm et al., 2019;Haraguchi and Lall, 2015).
However, the global scope of COVID-19 rendered the potential risk of disruption much more apparent to firms. 33  argued that firms could trade off cost efficiencies for supply chain shock resilience by keeping larger inventories, investing in technologies that closely monitor supplier situation, and diversifying their pool of suppliers. Furthermore, they argued that firms might want to locate their supply chains geographically closer. First, difference of labor costs between China and developed countries has decreased, and increased automation could make these differences even less relevant. Second, geographically-close suppliers would allow firms to react quickly to changes in demand. Third, policies to avert climate change could eventually lead firms to save on transportation costs. Finally, firms will likely choose suppliers belonging to the same regional trade agreements, because they are less likely to 31 See Guinea and Forsthuber (2020) for a critical view on this strategy. 32 https://www.whitehouse.gov/presidential-actions/executive-order-ensuring-essentialmedicines-medical-countermeasures-critical-inputs-made-united-states/ 33 Kilic and Marin (2020) argued that firms could respond to potential disruptions in the supply chain by substituting workers with COVID-19-free machines. change their trade policies unilaterally.

Díaz
Gandoy-Juste and Díaz-Mora (2020) argued that Spain could benefit from this supply chain regionalization process because of its competitive labor costs, relative to other EU countries, good infrastructures, and labor skills. However, the regionalization of supply chains might harm Spanish firms that supply industrial equipment to customers outside Europe. Potential customers might doubt that Spanish firms will be able to supply their goods owing to pandemic-related disruptions. This concern will be especially acute for largematuration projects, which require the movement of personnel to fine-tune equipment at the customer's premises.
In any case, finding new suppliers is not an easy task. Firms invest heavily in identifying suppliers that will meet their specifications, quality standards, and shipment schedule and will have the capacity to respond rapidly to changes in production planning. Suppliers often must make specific investments to adapt their product to customers' needs. Furthermore, there are many elements in the supplier-customer relationships that cannot be written into contracts or enforced outright. Thus, firms must be confident that other parties will not take advantages. All these reasons lead to sticky supplier-customer relationships (The World Bank, 2020). Moreover, the swift recovery in the trade of goods anticipated in Section 3 of this document relies on the stickiness of the customer-supplier relationships.

Conclusion
COVID-19 has caused the steepest decrease in Spanish trade in decades. Trade fell for both goods and services. However, the containment measures adopted to arrest the spread of the virus made the decrease especially intense for services. Spain's export specialization in products that facilitate the movement of people and goods, capital goods, outdoor products, and services that rely on the movement of people has made the negative impact of COVID-19 on trade greater than in other large EU countries. The nature of the trade collapse in goods, which concentrated on the intensive margin and particularly affected top exporters, suggests a rapid recovery of trade flows when health risks disappear. However, the recovery of tourism might take longer because it depends on the perceptions of foreign customers and their governments on the health situation in Spain.
In addition to the short-term massive reduction in trade flows, COVID-19 might also have longer-term impacts. Spain must develop strategies to reduce its dependence on foreign tourism receipts and exploit opportunities to expand business service exports. In the medium and long terms, Spanish trade might grow less owing to the higher costs of establishing new trade relationships. Furthermore, the development of domestic capacity in strategic sectors and the regionalization of value chains could lead to further increases in the shares of EU countries hold in Spanish trade.
Finally, the difficulties of obtaining medical materials and equipment to fight COVID-19 have triggered a general mistrust in the capacity of trade to adequately respond to countries' needs in the face of emergencies. This concern will exacerbate if countries do not manage the future distribution of the COVID-19 vaccine wisely. Failures in that case would quickly reinforce isolationist positions that would surely damage future trading relationships.