為什么新冠病毒對(duì)歐盟不同經(jīng)濟(jì)體的影響如此不同

1、Why has COVID-19 hit different European Union economies so differently?Andr SapirExecutive summaryAll European Union countries are undergoing severe output losses as a consequence of the COVID-19 crisis, but some have been hurt more than others. In response to the crisis, EU leaders have agreed on a

2、 Recovery and Resilience Fund (RFF), which will help all EU countries, but those hit hardest will benefit most.This Policy Contribution explores why some countries have been hit economically more than others by COVID-19. Using statistical techniques described in the technical appendices, several pot

3、ential explanations were examined: the severity of lockdown measures, the structure of national economies, the fiscal capacity of governments to counter the collapse in economic activity, and the quality of governance in different countries.We found that the strictness of lockdown measures, the shar

4、e of tourism in the economy and the quality of governance all play a significant role in explaining differences in economic losses in different EU countries. However, public indebtedness has not played a role, suggesting that that the European Central Banks pandemic emergency purchase programme has

5、been effective.preliminary draft .We used our results to explore why some southern EU countries have been more affected by the COVID-19 crisis than some northern countries. Depending on the pairs of countries or country groupings that we compared, we found that differences in GDP losses were between

6、 30 and 50 percent down to lockdown strictness, between 35 and 45 percent to the quality of governance and between 15 and 25 percent down to tourism.This could have implications for the allocation of the RRF between recovery and resilience expenditures. Supporting the recovery through a combination

7、of demand and supply initiatives is important to ensure that countries rebound as quickly as possible from the COVID-19 crisis, without leaving too much permanent damage to their economies. But in many countries, especially some of the southern countries hit hardest by the COVID-19 crisis, resilienc

8、e is a major sticking point. Too often, in some of these countries, the poor quality of governance has had a negative impact on their resilience, as the relatively large size of their GDP shocks has demonstrated. It is crucial therefore that RRF programmes devote sufficient attention (and resources)

9、 to improving the quality of governance in these countries.Recommended citationSapir, A. (2020) Why has COVID-19 hit different European Union economies so differently?,Policy Contribution 2020/18, Bruegel1 IntroductionUnderstanding why some countries have been hit harderby the COVID-19 shock could h

10、elp plan how countries should be helped by the Recovery and Resilience FacilityAfter five days and four nights of arduous negotiations, European Union leaders agreed on 21 July 2020 to set up a 750 billion fund to help EU countries recover from the COVID-19 crisis. All EU countries will benefit from

11、 the new fund, but those hit hardest will benefit the most1.This Policy Contribution explores first and foremost why some countries have been hit economically more than others by COVID-19. Several reasons for this have been put forward including the number of deaths per million inhabitants, the seve

12、rity of lockdown measures, the structure of the economy and the ability of the government to counter the collapse in economic activity. But there has been no systematic attempt so far that we are aware of to apportion the responsibility among various potential factors.Understanding why some countrie

13、s have been hit more than others could help in plan- ning how countries should be helped by the Recovery and Resilience Facility (RRF), the main instrument of the new fund2.To access money from the RRF, EU countries will have to prepare national recovery and resilience plans setting out their reform

14、 and investment agendas. Each plan will be assessed by the European Commission against various criteria, including “strengthening the growth potential, job creation and economic and social resilience” of the member state concerned.The Commissions assessments will be submitted to EU finance ministers

15、 for approval by a qualified majority. In addition, payments will be subject to the satisfactory fulfilment ofrelevant milestones and targets. In case one or several member states consider that there are “serious deviations from the satisfactory fulfilment of the relevant milestones and targets” by

16、another member state, payment may be suspended until a positive decision by EU leaders that the milestones and targets have been reached3.This strict mechanism was introduced at the insistence of the so-called frugal four countries Austria, Denmark, the Netherlands and Sweden whose economies have al

17、l been relatively less negatively affected by the COVID-19 crisis and will therefore benefit relatively little from the RRF. In addition to wanting to minimise their net contributions to the RRF, the frugal countries feared that some of the main beneficiaries such as Italy and other southern countri

18、es that were badly hit by the crisis may not otherwise direct sufficiently the new EU funds to improving their economic resilience.The remainder of the paper is divided into four sections. Section 2 discusses how to measure the economic impact of the COVID-19 crisis. Section 3 explains the differenc

19、es between countries in terms of the economic impact of the crisis and summarises the results of our analysis of the relative contribution of several variables to the difference in GDP hit experienced by countries. Section 4 makes policy recommendations, including on the design of the RRF. In two te

20、chnical appendices we provide details of our statistical analysis.See Darvas (2020b) for a careful analysis of the allocation of the 750 billion fund between instruments and EU countries.The size of the RRF is 672.5 billion.All quotations in this paragraph are from European Council (2020).2 Measurin

21、g the economic shock suffered by each countryHow should we measure the economic impact of the COVID-19 pandemic on EU countries? Most reports use the fall or expected fall in the real GDP growth rate in 2020. For instance,in its summer 2020 economic forecast released a couple of weeks ahead of the J

22、uly 2020 European Council meeting, the European Commission (2020b) stated that “The EU economy will experience a deep recession in 2020 due to the coronavirus pandemic The EU economy is forecast to contract by 8.7 percent”.Figure 1 shows the July 2020 Commission forecast for GDP growth of the 27 EU

23、countries, with the countries arranged according to the size of the GDP decline.Figure 1: Forecast GDP decline in 2020: July 2020 Commission forecast (percentage points)Italy Spain Croatia France Portugal Greece Slovakia Belgium Ireland Czech. Estonia Cyprus Lithuania Austria Bulgaira Latvia Sloveni

24、a Hungary Nlands Germany Finland Luxemg Malta Romania Sweden Denmark Poland0%-2%-4%-6%-8%-10%-12%Source: Bruegel based on European Commission (2020b).There are major differences between the 27 countries. The five most-affected countries are Italy (-11.2 percent), Spain (-10.9 percent), Croatia (-10.

25、8 percent), France (-10.6 percent) and Portugal (-9.8 percent), with Greece ranking sixth (-9 percent). The five least-affected countries are Poland (-4.6 percent), Denmark (-5.2 percent), Sweden (-5.3 percent), Romania (-6 percent) and Malta (-6 percent), a group that includes two of the frugal fou

26、r (F4). The (unweighted) average decline for the F4 is 6.1 percent. This contrasts with a decline of 10.2 percent for the group of four southern countries (S4) (Greece, Italy, Portugal, and Spain), to which we will compare the F4 countries throughout this Policy Contribution.We prefer however to fol

27、low Darvas (2020a) and use the Commissions downward revision of GDP growth forecasts between winter 2020 and summer 2020 to measure the severity of the pandemics economic shock4. In practical terms this means subtracting the February 20204 Darvas (2020a) used the autumn 2019 (published in November 2

28、019) and spring 2020 (published in May 2020) forecasts. We use instead the winter 2020 and summer 2020 forecasts. The winter 2020 forecasts were published on 13 February 2020, well before the coronavirus reached epidemic level in Europe and more than a month before it was declared a global pandemic

29、by the World Health Organisation. In its winter forecasts, the European Commission (2020a) simply regarded the coronavirus as “a downside risk for the EU economy” so much so that those economic forecasts were basically the same as the autumn 2019 forecasts published before the outbreak of the corona

30、virus in China. We also use the summer instead of the spring 2020 forecasts because they are the most recent at the time of writing.forecast from the July forecast shown in Figure 1. To understand the reason for this adjustment and the difference it makes, consider the examples of Austria and Bulgar

31、ia, two countries that will see the same GDP decline of 7.1 percent in 2020 according the summer 2020 forecasts.Clearly, because of its much lower GDP per capita, Bulgaria was on a higher growth trajectory than Austria before the COVID-19 crisis. This was reflected by the fact that, in its winter 20

32、20 forecast, the Commission predicted that GDP would grow in 2020 by 2.9 percent in Bulgaria but only 1.3 percent in Austria. Taking account of this correction, we now obtain a GDP shock of 10 percent (7.1 + 2.9 percent) in Bulgaria and 8.4 percent (7.1 + 1.3 percent) in Austria.Figure 2 shows the G

33、DP shock in 2020 on the basis of this approach, with the countries arranged according to the size of the economic shock.Figure 2: GDP shock in 2020: difference between the July and February 2020 Commission forecasts (percentage points)Croatia Spain Ireland France Italy Portugal Greece Slovakia Cypru

34、s Hungary Belgium Malta Bulgaria Estonia Czech. Romania Lithuania Slovenia Latvia Luxemg Austria Nlands Poland Finland Germany Denmark Sweden0%-2%-4%-6%-8%-10%-12%-14%-16%Source: Bruegel based on European Commission (2020a,b).As in Figure 1, Figure 2 shows major differences in the GDP shocks felt by

35、 the 27 countries, but there are also some significant changes in the country rankings. The five most-affected countries are now Croatia (-13.4 percent), Spain (-12.5 percent), Ireland (-12.1 percent), France (-11.7 percent) and Italy (-11.5 percent), a group that includes again three of the S4 coun

36、tries, although Portugal ranks a close sixth. The five least-affected countries are now Sweden (-6.5 percent), Denmark (-6.7 percent), Germany (-7.4 percent), Finland (-7.8 per- cent) and Poland (-7.9 percent), a group that includes two of the F4 countries, but with the Netherlands and Austria ranki

37、ng sixth and seventh. The (unweighted) average negative shock to the S4 countries is now 11.7 percent and 7.4 percent for the F4 countries.Notably, the shocks to Croatia and Ireland increase substantially when considering their high growth rates anticipated before the crisis, of respectively 2.6 per

38、cent and 3.6 percent. By contrast the negative shock to Italy appears relatively less severe when considering that before the COVID-19 crisis its expected growth rate for 2020 was only 0.3 percent. In the remainder of this paper, we focus entirely on the economic shocks reported in Figure 2.3 Explai

39、ning differences between countriesWhat could explain the major differences between countries in the economic shocks they are undergoing in the wake of the COVID-19 crisis?The first factor that comes to mind is the strictness of lockdown measures implemented by national, regional or local authorities

40、 to cope with the coronavirus pandemic5. We use the stringency index provided by the Oxford COVID-19 Government Response Tracker (OxCGRT, Oxford University, 2020). This index, which varies between 0 and 100 is available on a daily basis for most countries, including 26 of the 27 EU countries6. We co

41、nstructed an aggregate index, by computing the average of the Oxford stringency index over the first semester of 2020. Our index of strictness, therefore, reflects both the daily intensity of the measures and their duration. This aggregate index averages 42 for the 26 EU countries (using an unweight

42、ed average), and ranged from 25 in Sweden to 55 in Italy. It averaged 48 for the S4 and 38 for the F4 countries, meaning that in the F4 countries, lockdowns were less strict.A second possible factor is the economic structure of countries. Measures taken to combat the spread of the coronavirus have a

43、ffected economic activities differently, dependingmainly on whether they involve proximity between consumers and producers, which is often the case with services. Among services, those viewed by consumers (or sometimes also public authorities) as less essential, such as leisure and tourism, have bee

44、n more affected than activities such as shopping, which also involve some proximity between consumers and producers, although less so with the development of e-commerce. We use the share of tourism in GDP as a proxy for the impact of a countrys economic structure on the economic shock due to sanitar

45、y measures to combat coronavirus. In 2019, tourism averaged 4.8 percent of GDP in the 26 EU countries (using a simple rather than a weighted average) and ranged from 1.1 percent in Germany to 19.9 percent in Croatia. It was twice as high for the S4 countries (5.9 percent) as for the F4 countries (2.

46、9 percent).A third potential factor is the level of public indebtedness. It has been argued that some countries with high COVID-19 fatalities (for example Italy, with 590 deaths per million inhabitants as of 18 September 2020) would be unable to borrow money (or unable to do so at a relatively low c

47、ost) to meet the costs of the COVID-19 crisis because of their high public debt levels (135 percent of GDP in 2019 in Italy), and that this explains why they were unable to implement enough of a fiscal response to prevent a severe economic contraction. In fact, this line of reasoning seems to have p

48、layed a major role in the minds of the architects of the recovery fund (see for example Buti, 2020). We use the 2019 debt-to-GDP ratio to measure levels of public indebtedness. We prefer this to using the fiscal stimulus implemented by governments to counter the crisis because the 2019 debt level is

49、 clearly unaffected by the COVID-19 crisis, while the fiscal stimulus is would not have happened without it7. Also, the size of the fiscal stimulus is not observed but must be estimated, and such estimates are not available for all the 26 EU countries considered here. In 2019, the debt-to-GDP ratio

50、in the 26 EU countries averaged 64 percent (using a simple rather than a weighted average) and ranged from 8 percent in Estonia to 180 percent in Greece. It averaged 132 percent for the S4 countries and only 47 percent for the F4 countries.The fourth and last factor that we consider is the quality o

51、f governance. We expect that, for a given degree of severity of lockdown measures, a given structure of the economy and a givenIn an earlier version of this paper, we used the number of deaths per million inhabitants to measure the severity of the COVID-19 crisis. Several commentators pointed out th

52、at this was problematic for two reasons. First, different countries account differently for their COVID-19 deaths. Second, and more importantly, what ultimately matters for the size of the economic shock is not so much the number of deaths as the severity of the measures taken by public authorities

53、in reaction to this number.Malta is missing from the Oxford database.In the language of econometricians, this implies that the 2019 debt-to-GDP is exogenous, while the fiscal stimulus is endogenous to the economic shock.level of public debt, the degree of resilience of the economy to the COVID-19 sh

54、ock, resulting from the quality of behaviour of both private and public economic agents, will be greater in countries with a higher quality of private and public governance and institutions8. Following Demertzis and Raposo (2018), we construct our indicator of governance by summingup for each countr

55、y its scores for the six parts of the World Bank Worldwide Governance indicator: voice and accountability; political stability and absence of violence; government effectiveness; regulatory quality; rule of law; and control of corruption. Scores for each of the six indicators can vary from -2.5 to 2.

56、5, so the overall indicator can vary from -15 to +15. In 2018, the latest year available, the quality of governance in the 26 EU countries averaged 6 (using a simple rather than a weighted average) and ranged from 0.9 in Romania to 10.6 in Finland. It averaged 4 for the S4 and 9.8 for the F4 countri

57、es.The appendices detail our model and results. In summary, we found that:The level of public indebtedness does not seem to contribute to explaining differences in the economic shocks suffered by EU countriesCountries with stricter lockdown measures witnessed a greater economic collapse, with the ca

58、usality running from the strictness of the lockdown measures to the economic collapse and not the other way round so if the countries where the economic collapse was greatest then reduced the strictness of the measures more than others, it was not sufficient to counterbalance the economic hit.The sh

59、are of tourism in the economy and the quality of governance were also significant factors. However, public indebtedness was not, implying that the level of public indebtedness does not contribute to explaining differences in the economic shocks suffered by EU countries. This suggests that the Europe

60、an Central Banks pandemic emergency purchase programme (PEPP) has been successful in countering the risk that high public debt euro-area countries would be cut off from the market if they attempted to expand their debt issuance to respond to the COVID-19 crisis.Together, the factors of lockdown stri