《COVID-19通过全球价值链对发展中国家的影 响.docx》由会员分享，可在线阅读，更多相关《COVID-19通过全球价值链对发展中国家的影 响.docx（34页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、Check forupdatesCling Together, Swing Together:The Contagious Effects of COVID-19 on DevelopingCountries Through Global Value ChainsStefan Pahl1, Clara Brandi2, Jakob Schwab2, Frederik Stender2struct:is paper aims at estimating the economic vulnerability of developing countries to disruptions in )ba
2、l value chains (GVCs) due to the COVID-19 pandemic. It uses trade in value added data for a ample of 12 developing countries in sub-Saharan Africa, Asia and Latin America to assess their dependence on demand and supply from the three main hubs China, Europe, and North America.ing first estimates on
3、CO VID-19-induced changes in final demand and production, we obtain an |ly projection of the GDP effect during the lockdowns that runs through trade in GVCs. Our stimates reveal that adverse demand-side effects reduce GDP up to 5.4 percent, and that collapsing leign supply puts an even larger share
4、of countries9 GDP at risk. Overall, we confirm conjecture at the countries most affected are those highly integrated in GVCs (South-East Asian countries). 卜 argue, however, that these countries also benefit from a well-diversified portfolio of foreign uppliers and demand destinations, possibly leadi
5、ng to a cushioning of economic downswing ause COVID-19 stroke major hubs at different times.y words: COVID-19, global value chains, input-output analysis, international trade, supply and rnand side dependency, shock spillover.onflict of interest: This is to acknowledge that no financial or other ben
6、efit, nor any competing erest has arisen from our research.iis article has been accepted for publication and undergone full peer review but has not been ough the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Recor
7、d. Please cite this article as doi: 10.1111/twec.l 3094 latter twos comparatively larger share of imports of final goods. Unsurprisingly, with close trade 、s in mind, a Chinese final demand shock generally affects the output of its Asian neighbors ongest in negative terms, outpacing economic turmoil
8、 caused in Europe and the United States, more depth, however, the authors demonstrate that demand shocks specifically to investment in imary and secondary sectors, that is, allegedly import-intensive sectors of China, have an even onger spillover elasticity than broad-based final demand shocks. The
9、latter finding attracts rticular attention in view of Chinas gradual transition of replacing investment and exports as ivers of its economic growth by consumption. Most recently, Cao et al. (2020) compare the lllover effects of import fluctuations in China and the United States. While the authors fi
10、nd that Europe and Asia would be most negatively affected through import declines in China and the lited States, that is, by implication revealing less impact on Latin America or Africa, export anges on GDP were felt more severely when caused by falling US imports.It is well beyond question, that th
11、e ongoing pandemic already poses the greatest challenge to p world community since the 1930s Great Depression. It is therefore all the more to be seen whether the well-documented insights from the latest, allegedly moderate, global financial crisis a decade ago can be transferred to the current cont
12、ext. As the nature of the COVID-19 crisis is ferent insofar as sectors of household consumption are affected very heterogeneously due to cial distancing measures, and that some supply chains may be interrupted fully for certain pods of time, the expected hazardous effects from this inflicted on deve
13、loping countries remain n empirical puzzle.3. Method and Datady GVC-induced demand and supply shocks on developing countries, we map individual untrys value added to demand for and production of specific value chains. Following Los et al. bl5) and Timmer et al. (2013), we define a value chain by the
14、 finalized products, that is, by final dustry grouping and country of completion, for example, textiles finalized in Bangladesh. To Ice each countrys contribution to this value chain, we need to find the output and value added sociated with the production of textiles in Bangladesh. In the last stage
15、 of production, output and |ue added is by definition generated in the textiles sector in Bangladesh. Yet, textiles production refluires intermediate inputs, such as cotton from the agricultural sector. This will generate outputd value added in the agricultural sector, which can be in Bangladesh or
16、in any other country from which the intermediate is imported. Those first-tier agricultural intermediates may require iWermediate inputs themselves, which again generates output and value added in a specific untry-industry, depending on the source of those second-tier intermediates. To get a complet
17、e aracterization of the value chain, we trace the entire chain of intermediate suppliers across untries and industries.To do so, we make use of the global system of input-output relationships. We define a column :tor F for final demand for finalized products grouped by industry (eg, textiles) and co
18、untry of npletion (eg, Bangladesh). With A being a matrix of intermediate input coefficients, we cantrace the contributions to the production of F making use of the well-known Leontief-inverseA) t. (I A) -1F then describes the output generated in any country-industry to produce the vector of final d
19、emand F.1 Multiplying by a matrix V of value added to gross output ratios for each Ountry-industry on the diagonal, we further obtain the value added by each country-industry (vector VA) generated in the production of F. That is,(1) VA = V(I - A)密 appropriate definition of F, we obtain value added g
20、enerated in a specific country-industry rSated to finalization of a specific product or to final demand by consumers in a specific region or rountry. For example, setting all elements to zero in F except those for Chinese consumers, we /tain value added generated in any country-industry associated w
21、ith final demand in China. In a rnilar vein, setting all elements to zero except those for final demand (anywhere in the world) for )ducts finalized in China, we obtain value added in any country-industry associated with oroduction for goods finalized in China.obtain each countrys demand dependencie
22、s by calculating that countrys share of GDP that js generated in the production for final goods consumed in the three world regions Europe (EU28 of 2014 plus Switzerland), North America (US and Canada) and China. We obtain each ountry,s supply dependency by calculating each countrys share of GDP tha
23、t is generated in VCs that are dependent on the three hubs. We define a GVC as dependent on a specific hub ifA) t is a geometric series. That is, (I A) -1F can be written as (AF + A2F + A3F + AF then describes first-tier intermediate use, A2F the second tier use and so on. Therefore (I A) -1F gives
24、the output in any ntry-industry in the system that participates in production of F. in For details, see Miller and Blair (2009).that hub generates at least five percent of value added m production of the respective final good jfined by country of completion and sector grouping).2The main advantage o
25、f our supply-side approach is that it accounts for upstream as well as wnstream dependencies. For example, let us assume that Ethiopia exports cotton to China where j cotton is processed and then exported to Europe as a textiles product. Ethiopian agriculture es not require any Chinese inputs in thi
26、s value chain, but its production is nonetheless dependent Chinese producers. By decomposing the value chain by its final product, we can trace all rticipants in that chain independent of their relative position.To estimate the size of the effects of the COVID-19 pandemic, we combine these demand an
27、d pply dependencies with first estimates on the demand and supply shock. For the demand shock, )combine the data on sectoral value added dependencies with estimations of the sectoral Terences in the downturn in final demand. For Europe, we obtain sectoral retail consumption ta from Eurostat (2020).
28、For the sectoral demand effects in the United States (which we use for )rth America), we rely on estimates by Coibion et al. (2020). These authors use householdsurvey data in the United States and exploit regional variation in the exposure to COVID-19 and asures to contain it to estimate the effect
29、on different categories of consumption goods by jseholds. The survey was held in April, and thus during the time when the US economy was hithardest by the pandemic.For China, we use data from the National Bureau of Statistics ChinaBS, 2020b). The sectoral classifications differ slightly between the
30、three sources. However, all、information can be straightforwardly mapped with the ISIC (International Standard Industrial assification of All Economic Activities) Rev. 4 categories in the TiVA (Trade in value-added)ta (see Table Al in the Appendix for the respective mappings and the estimated decline
31、s in al sectoral demand).3For the supply shock, we use data on the drop in percent of industrial production in the three bs Europe, North America, and China in the month of the largest drop in the first months of 20.4 The months with the largest drop was February for China (26.6 percent), and April
32、forhere are 50 countries of completion and 28 sectors, and therefore 50 % 28 = 1400 value chains. By definition, the |ue-added shares of all countries that participate in the GVC add up to expenditure for the respective final product, equation (1).Mostly for reasons of exposure, our analysis concent
33、rated on those ISIC Rev. 4 categories for which final demand in three hubs meaningfully contributed to GDP in the set of developing countries.)r Europe, we use data from the European Union, for North America from the United States. The data come from respective national statistical bureaus.Europe (2
34、7.0) and North America (16.6). China was the first country globally to be affected by M)vid-19, and implementing a drastic lockdown, with other countries around the globe following staggered sequence with the eruption of the disease, with the resulting effects on industrial Dduction. implement this
35、method to be applied to estimating both the demand and the supply shock, we ed information on the global system of input-output relationships (depicted in A), information d value added to gross output ratios (V) and a vector of final demand (F). In particular, to obtain one needs to turn to global i
36、nput-output tables, which describe the supply and use relationships etween producers within and across countries. Global input-output tables are constructed mbining a large amount of information on value added, gross output, trade flows (intermediates, lai goods), and final demand categories. As thi
37、s is a highly data-intensive exercise, a major ttleneck to studying the involvement of developing countries is the relatively poor coverage of less developed countries, in particular in Sub-Saharan Africa. Some attempts have been made to Mdge this gap. The construction of the EORA database (Lenzen e
38、t al., 2013) takes a global approach covering a large amount of countries since the 1990s, but naturally has to compromise ith respect to a clear anchoring in official statistics and simplifying assumptions. As a country- Jpcific alternative, we use the data constructed in Pahl et al. (2019). This c
39、onstruction closely ollows the approach laid out in the construction of the World Input-Output Database (WIOD; Bnmer et al., 2015; Dietzenbacher et aL, 2013), but adds seven new developing countries for the ime period 2000 to 2014. These are: Ethiopia, Kenya, Senegal, South Africa, Bangladesh, lalay
40、sia and Vietnam. We base the estimates on the final year in that dataset, that is, 2014.kThis country-specific approach (rather than a global one) allows for a number of improvements, are particularly important when studying the value added or income effects related to VCs. As is easy to see from eq
41、uation (1), value added to gross output ratios in V are crucial to in a countrys value added in global production. An advantage of using data from Pahl et al. 019) is the yearly variation in the input data in those ratios for each of the covered sectors and pastries. Secondly, the construction in Pa
42、hl et al. (2019) provides a careful treatment of trade ows (eg, re-exports, missing trade flows, classification by use category), which is paramount to picting the cross-country relationships in A. Moreover, to obtain the domestic supply and use ations in A, the data are built up from national suppl
43、y and use tables or official input-output )les, and as such are highly country-specific. Lastly, F is consistent with national accounts, andsplit between household consumption, government consumption, gross fixed capital formation andnentories. As this paper assesses the impact of a reduction in fin
44、al demand, mostly running reduced household consumption, this distinction proves useful.Using this dataset, we will analyze 12 developing and emerging economies: four countries inUb-Saharan Africa (Ethiopia, Kenya, Senegal, South Africa), six in East and South-East Asia angladesh, China, India, Indo
45、nesia, Malaysia, Vietnam), and two in Latin America (Brazil,xico).4. Demand-Side Vulnerabilitystudy demand-side related GDP effects for developing countries arising from the pandemic, we the value added trade data to compute how much of value added in each of the developing duntries in the sample de
46、pends on final demand in the different regions in the world. Table 1 presents these results aggregated across sectors, where rows show individual developing countries.Xiie first six columns list separate world regions, and the last column developing countries9 home rkets. The values then depict how
47、much of value added in each country depends on final mand in each of these regions.5As shown in Table 1, Vietnam and Malaysia are most strongly dependent on foreign demand, ith only around 50 percent of domestic value added dependent on final demand in their home rkets. For other countries, GDP depe
48、ndence on foreign demand ranges between 27 (South rica) to below 10 percent (Brazil). At the same time, for example, Bangladeshs GDP is jjely strongly dependent on demand from Europe, with 6.1 percent, and Mexico, unsurprisingly, on demand from North America, with 13.8 percent. We would expect these
49、 ntries to be most strongly affected by the economic downturn and plummeting demand in urope and the US. Considering regional differences, value added in Asian countries is on average re dependent on foreign final demand than that in African countries.6hese numbers are related to the export to GDP ratio, but not equal. Differences arise in different shares of domestic value added to gross exports across the countries.or Latin America, the sample is quite small and par