宏观经济成果和COVID19进度报告

M acr oeconomic O utcomes and C O VID-19: A P r ogr ess R epor t J es u ´ s F ern a ´ nde z-V illaver de Charles I.J ones ⇤ UP enn and NBER S tanfor d GSB and NBER S eptember 13, 2020—V ersion 0.5 A bstract This paper combines data on GD P , unemplo yment, and G oogle ’ s CO VID-19 Com-munity M obility R epor ts with data on deaths fr om C O VID-19 to study the macr oe-conomic outcomes of the pandemic.W e pr esent results fr om an international per-spectiv e using data at the country level as w ell as results for individual U.S.states and key cities thr oughout the world.The data fr om these differ ent lev els of geo-gr aphic aggr egation offer a r emar kably similar view of the pandemic despite the substantial heter ogeneity in outcomes.Countries like K or ea, J apan, G er many , and N or way and cities such as T oky o and S eoul hav e compar ativ ely few deaths and lo w macr oeconomic losses.A t the other extr eme , N ew Y or k City , Lombardy , the U nited K ingdom, and M adr id hav e many deaths and lar ge macr oeconomic losses.Ther e ar e few er locations that seem to succeed on one dimension but suffer on the other , but these include C alifor nia and S w eden and potentially offer useful policy lessons.⇤ W e ar e grateful to Andy Atkeson and J im S tock for many helpful comments and discussions...1.I ntr oduction This paper combines data on GD P , unemplo yment, and G oogle ’ s CO VID-19 Commu-nity M obility R epor ts with data on deaths fr o m C O VID-19 to study the macr oeconomic outcomes of the pandemic.W e pr esent r esults fr om an international perspectiv e us-ing data at the country lev el as w ell as r esults for individual U.S.states and ke y cities thr oughout the world.The evidence to date can be summar iz ed in a styliz ed way b y F igur e 1.O n the hor iz ontal axis is the number of deaths(per million population)fr om C O VID-19.The v er tical axis sho ws a cumulativ e measur e of the macr oeconomic losses apar t fr om the v alue of the loss in life;for simplicity her e w e call this the “ GDP loss.” Throughout this paper , w e will sho w data for v ar ious countr ies , U.S.states , and global cities to fill in this gr aph quantitativ el y.W e will also sho w the dynamics of ho w countr ies tr av erse thr ough this space o v er time.F or now , though, w e simply summar ize in a styliz ed way our main findings.F igur e 1: S ummar y of the T r ade-off E vidence GDP L OS S CO V I D D E A TH S O ne can divi de the gr aph into four quadr ants , based on many v ersus few deaths fr om C O VID-19 and on lar ge v ersus small losses in GD P.O ur first inter esting finding is that ther e ar e communities in all four quandr ants.I n the lo w er left cor ner of the di agr am—the quadr ant with the best outcomes—C alifor nia [lucky? too tight?] N ew Y or k City Lombardy U nited K ingdom M adr id [unlucky? bad policy?] G er many , N or way J apan, S.K or ea China, T aiwan Kentucky , M ontana [lucky? good policy?] S w eden [unlucky? too loose?]

ar e G ermany , N or way , China, J apan, S outh K or ea, and T aiwan as w ell as U.S.states such as Kentucky , M ontana, and I daho.S ome combination of good luck and good policy means that these locations hav e exper ienced compar ativ ely few C OVID deaths as a fr action of their populations while simultaneously keeping the losses in economic activity r elativ ely lo w.I n the opposite quadr ant—the one with the worst outcomes—N ew Y or k City , Lombardy , the U nited K ingdom, and M adr id ar e emblematic of places that hav e had compar ativ ely high death r ates and lar ge macr oeconomic losses.S ome combination of bad luck and policy mistakes is likely r esponsible for the poor per for mance on both dimensions.These locations w er e unlucky to be hit r elativ ely early in the pandemic, per haps b y a str ain of the vir us that was mor e contagious.B eing hit early also meant that communities often did no t take appr opr iate measures in nursing homes and car e facilities to ensur e that the most susceptible w er e adequately pr otected and that the medical pr otocols at hospitals w er e less w ell-dev elop.The other two quadr ants of the char t stand out in inter esting ways , having good per for mance on one dimension and poor per for mance on the other.C ompar ed to N ew Y or k, Lombardy , M adr id, and the U.K., S w eden and S tockholm had compar able death r ates with much smaller losses in economic activity.B ut of course , that is not the only compar ison: r elativ e to N or way and G ermany , S w eden had many mor e deaths and compar able losses in economic activity.R elativ e to the worst outcomes in the nor theast quadr ant, S w eden is a success.B ut r elativ e to what was possib le—as illustr ated b y G ermany and N or way—S w eden could hav e done better.C alifor nia, in the quadr ant opposite of S w eden, also makes for an inter esting com-parison.R elativ e to N ew Y or k, C alifor nia had similarly lar ge losses in economic activ-ity but far few er deaths.I n r ecent months , both states had unemplo yment r ates on the or der of 15 per cent.B ut N ew Y or k had 1700 deaths per million r esidents while C alifor nia had just 300.F r om N ew Y or k ’ s perspectiv e , C alifor nia looks enviable.O n the other hand, C alifor nia looks less successful when comp ar ed to G ermany , N or way , J apan, and S outh K or ea.These places had similarly lo w deaths but much smaller losses in economic activity.O nce again, r elativ e to what was possible—as illustr ated b y the best-per for ming places in the world—C alifor nia could hav e done better.O ne of the most impor tant cav eats in this analysis is that the pandemic continues.This char t and the graphs below that it is based on may v er y w ell look quite differ ent six months fr om now.O ne of the most impor tant di mensions of luck is r elated to whether a location was hit early b y the pandemic or has not—y et?—been sev er ely affected.W ill a v accine or cheap , widespread testing end the pandemic befor e these places ar e impacted? S till, with this cav eat in mind, pr obably the most impor tant lesson of the paper is that ther e ar e a good number of places in the lo w er-left quadr ant of the graph: with the r ight policies , good outcomes on both the GDP and C OVID mor tality outcomes ar e possible.P laces like China, G ermany , J apan, N or way , S outh K or ea and T aiwan ar e heter ogeneous on v ar ious dimensions.The set includes lar ge , dense cities such as S eoul and T oky o.The set includes nations that w er e for ewar ned b y experiences with SARS and MERS, but also countr ies like G ermany and N or way that did not hav e this dir ect exper ience.Ther e ar e places that w er e hit early , like China and S outh K or ea, and places that w er e hit later , like G ermany and N or wa y.O ur paper does n ot highlight precisely what they did to get the se good outcomes , but it suggests wher e to look for these deeper lessons.I n the r emainder of the paper , w e pr esent the detailed evidence that underlies this styliz ed summar y.S ection 2 lays out a basi c fr amewor k for thinki ng about this diagram.S ection 3 pr esents evidence for countr ies using data on GDP fr om the first and second quar ters of 2020 to measur e the macr oeconomic outcomes.I t also sho ws evidence for U.S.states using monthly unemplo yment r ates.S ection 4 then turns to a comple-mentar y sour ce of data on economic activity , the G oogle C ommunity M obility R epor ts.W e sho w that these economic activity measures ar e highly corr elated with GDP and unemplo yment r ates.The G oogle measures hav e additional adv antages , ho wev er.I n particular , they ar e av ailable for a lar ge number of locations at v ar ying geogr aphic lev els of aggr egation, ar e r epor ted at t he daily fr equency , and ar e r epor ted with a lag of only just a f ew days , a particularly impor tant featur e given the natur al lags in NIP A r epor ting.W e r epr oduce our earlier finding s using the G oogle data and also pr oduce new char ts for key cities ar ound the world.The city-lev el data is impor tant because of concer ns about aggr egating to , say , the national lev el acr oss r egions of v ar ying densities.S ec-tion 5 sho ws the dynamic v ersion of our graphs at the monthly fr equency using the G oogle data, so w e can see ho w differ ent locations ar e ev olving o v er time.F inally ，S ection 6 offers some closing thoughts.Literatur e R eview.Ov er the last few months , a gigantic liter atur e on C O VID-19 and economics has appeared.I t is bey ond our scope to r eview such liter atur e , which tou ches on multiple questions , fr om the design of optimal mitigation policies(A cemoglu, Cher-no zhuko v , W er ning and Whinston , 2020)to C O VID-19 ’ s impact on gender equality(Alon , Doepke , O lmstead-Rumsey and T er tilt , 2020).I nstead, w e highlight thr ee sets of papers that hav e explor ed the inter action betw een C O VID-19, the policy r esponses to it, and economic outcomes.The first set o f papers has extended standar d economic models to incorpor ate an epidemiological block.Among those , early effor ts include A´ lv ar e z, Ar gente and Lippi(2020), E ichenbaum, R ebelo and T r abandt(2020), G lo v er , H eathcote , Kr u eger and R ´ ıos-R ull(2020), and F arboodi, J ar osch and S himer(2020).I n this tr adition, the contribu-tions of models with many differ ent sectors(B aqaee and F ar hi , 2020a , b;B aqaee , F ar hi, M ina and S tock , 2020)ar e particularly inter esting for the goal of mer ging micr odata with aggr egate outcomes and the design of optimal r eopening policies.These models will also ser v e , in the futur e , as potential labor ator ies to measur e the r ole of luck vs.policy that w e discuss abo v e.A second set of pape rs has attempted to measur e the effects of lockdown policies.This measur ement is vital to distinguish betw een the r eduction in economic activity tr igger ed b y economic agents ’ endogenous r eactions(e.g., the v oluntar y cancellation of tr av el)v ersus go v er nment-imposed mandates(e.g., an international tr av el pr ohibi-tion).A gr o wing consensus suggests that v oluntar y changes in behavior ar e the primar y dr iv er of outcomes.F or example , G oolsbee and S yv erson(2020)compar e consumer behavior within the same commuting z ones but acr oss boundar ies with differ ent policy r egimes to conclude that legal r estr ictions account only for 7 percentage points(p.p.)of the o v er all r eduction of o v er 60 p.p.in consumer tr affic.H o wev er , the authors docu-ment that legal mandates shift consumer activity acr oss differ ent industries(e.g., fr om r estaurants into gr ocer ies).E quiv alent r esults ar e r epor ted using smar tphone data b y Gupta, Nguy en, R ojas , R aman, Lee , B ento , S imon and W ing(2020b)and v acancy post-ing and unemplo yment insur ance claims in the U.S.b y F o rsythe , K ahn, Lange and W icz er(2020), although Gupta, M ontenov o , Nguy en, R ojas , Schmutte , S imon, W ein-

ber g and W ing(2020a)find lar ger effects of go v er nment-mandated lockdo wns on em-plo yment.1 S imilar findings r egar ding the pr eponder ance of v oluntar y changes in behavior ar e r epor ted for E ur ope b y Chen, I gan, Pierr i and P r esbiter o(2020), S outh K or ea b y A um, Lee and S hin(2020), and J apan b y W atanabe and Yabu(2020).2 A tkeson, K opecky and Zha(2020)highlight, using a r ange of epidemiological models , that a r elativ ely lo w impact of go v er nment mandates is the only way to r econcile the obser ved data on the pr ogr ession of C OVID acr oss a wide cr oss-section of countr ies with theor y.O n the other hand, the r esults using Chinese data in F ang, W ang and Y ang(2020)in-dicate that early and aggr essiv e lockdo wns can hav e lar ge effects in contr olling the epi-demic and findings using G erman data b y M itz e , K osfeld, R ode and W a ¨ lde(2020)point out to the effectiv eness of face masks in slo wing do wn contagi on gr o wth.Amuedo-Dor antes, K aushal and M uch o w(2020)study U.S.count y-lev el data to ar gue that non-phar maceutical inter v entions hav e a significant impact on mor tality and infections.A subset of these papers has dealt with S w eden ’ s case , a country that implemented a much mor e lenient lockdown policy than its N or ther n E ur opean neighbors.Among the papers that offer a mor e fav or able assessment of the S w edish exper ience , J ur anek, P aet-z old, W inner and Z outman(2020)hav e gather ed administrativ e data on w eekly new unemplo yment and furlough spells fr om all 56 r egions of S w eden, D enmar k, F inland, and N or wa y.U sing an ev ent-study difference-in-differ ences , J ur anek, P aetzold, W inner and Z outman(2020)conclude that S w eden ’ s lighter appr oach to lockdo wns tr anslated into betw een 9, 000 and 32,000 seasonally and r egionally adjusted cumulativ e unem-plo yment plus furloughs per million population b y w eek 21 of the pandemi c.I f w e compar e , for example , S w eden with N or way , these numbers suggest a cr ude tr ade-off(without contr olling for any other v ar iable)of ar ound 61 jobs lost per life sav ed.3 On 1 S ince many of these pape rs r ely heavily on smar tphone data, C outur e , D ingel, G r een, H andbury and W illiams(2020)sho w that this data is a r eliable snapshot of social activities.2 N otice that ev en if most of the r eduction in mobility comes fr om v oluntary decisions , we might still be far fr om a social optimum(as agents do not fully account for the contagion externalities they cr eate)or that the go v er nment inf or mation cannot play a r ole in shaping agents ’ belie fs about the state of the epidemic and, ther efor e , influence v oluntary behavior.F ur ther mor e , go v er nment-mandated policies may increase the risky behavior b y agents thr ough a v ersion of the P eltzman effect: if all non-essential businesses ar e closed, ther e is less r eason to be cautious when patr onizing an essential business , as the total contagion exposur e is lo w er.3 Among many othe r elements , this com putation does not contr ol for the possibility that S w eden, b y getting closer to her d immunity , might hav e sav ed futur e deaths or , conv ersely , that higher death r ates today might hav e long-r un scarr ing effects on S w edish GDP and labor mar ket.the negativ e side , B or n, D ietrich and M u ¨ ller(2020)and Cho(2020), using a synthetic contr ol appr oach, find that st r icter lockdown measures w ould hav e been associated with lo w er ex cess mor tality in S w eden b y betw een a quar ter and a third.The thir d set of papers has studied ho w to monitor the economy in r eal time(C a-jner , C r ane , D ecker , G r igsby , H amins-P uertolas , H urst, K ur z and Yildir maz , 2020;S tock , 2020), ho w the sector al composition of each country matters for the r epor ted output and emplo yment losses(G ottlieb , G r obo vsek, P oschke and S altiel , 2020), and the im-pact of concr ete policy measures.Among the latter , Chetty , F r iedman, H endr en, S tep-ner and T eam(2020)ar gue that stimulating aggr egate demand or pr o viding liquidity to businesses might hav e limited effects when the main constr ained in the unwillingness of households to consume due to health r isks and that social insur ance pr ogr ams can be a super ior mitigation tool.2.F ramew or k W e focus on two outcomes in this paper : the loss in economic activity , as captur ed b y r educed GDP or increased unemplo yment , and the number of deaths fr om C O VID-19.E v en with just these simple outcome measures , it is easy to illustr ate the subtle inter actions that occur in the pandemic.F igur e 2: E conomic P olicy T r ade Off, H olding H ealth P olicy and L uck Constant GDP LOSS(PERCENT)COVID DEATHS PER MILLION PEOPLE N ote: H olding health policy and “luck ” constant, economic policy implies a tr adeoff between economic activity and deaths fr om C O VID-19.T o begin, F igur e 2 illustr ates a simple tr adeoff betw een economic act ivity and deaths fr om the pandemic.I n the shor t t erm, economic pol icy can shut the economy do wn sharply , which incr eases the economic losses on the v er tical axis but sav es liv es on the hor iz ontal axis.Alternativ ely , policy could focus on keeping the economy activ e to minimiz e the loss in GDP at the expense of mor e deaths fr om the pandemic.F igur e 3 sho ws that the stor y is mor e complicated when health policy and luck ar e br ought under consider ation.Ther e can be a positiv e corr elation betw een economic losses and C OVID deaths.G ood health policy—for example , masks , pr otecting nursing homes , and tar geted r eductions in super-spr eader ev ents such as choirs , bars , night-clubs , and parties—can r educe the number of deaths.F ur ther mor e , b y r educing the death r ate , such polic ies encour age economic activity and allo w peopl e to r etur n safely Shut down economy Keep economy open

F igur e 3: H ealth P olicy D ecisions and L uck Can S hift the T r ade-off GDP LOSS(PERCENT)COVID DEATHS PER MILLION PEOPLE N ote: H eal th policy and luck can shift the tr adeoff between economic activity and deaths fr om C O VID-19.to wor k and to the marketplace.S imilarly , luck plays an impor tant but not y et fully-understood r ole.Wher e does the cor onavir us str ike early v ersus late? P er haps a cou ntr y is in the lo w er left cor ner today with lo w deaths and little loss in GDP but only because it has been lucky to av oid a sev er e outbr eak.Two months fr om now , things may look differ ent.Alternativ ely , is a r egion hit b y a str ain that is less infectious and deadly , or mor e(see our next subsec-tion)? This is another dimension of luck.4 F inally , all of these for ces play out o v er time , which giv es r ise t o impor tant dynamic consider ations.F or example , a community may keep the economy open in the shor t ter m, which may lead to a wav e of deaths , and then be compelled to shut the economy do wn to pr ev ent ev en mor e deaths.Two communities can end up with lar ge economic losses , but v er y differ ent mor tality outcomes , because of these timing consider ations.This can be thought of as being embodied in F igur e 3.F igur e 4 puts these mechanisms together in a single char t.I t r ev eals that the corr e-lation betw een economic losses and C OVID deaths that w e see in the data is go v er ned 4 Also , simple demogr aphic differ ences , giv en the steep age pattern of C O VID-19 mor tality r ates , mo v e the tr ade-off between deaths and GDP losses in significant ways.Bad policy or bad luck Good policy or good luck

F igur e 4: E conomic A ctivity , C o vid D eaths , H ealth P olicy , and L uck GDP LOSS(PERCENT)COVID DEATHS PER MILLION PEOPLE N ote: P utting the two together explains why the data can be har d to interpr et.b y a sophisticated collection of for ces , both static and dynamic.When w e see a cloud of data points in the empir ical v ersions of this graph, w e can think about ho w these v ar ious for ces ar e playing out.E vidence on the R ole of M utation.W e hav e mentioned sev er al times that a simple mechanism behind luck is the str ain of vir us that attacked one location.F r om M ar ch to M ay of 2020, a SARS-C oV-2 v ar iant carr ying the S pike pr otein G614 that likely appear ed in some moment in F ebr uar y r eplaced D614 as the dominant vir us for m globally(K or-ber et al., 2020).While the global tr ansition to the G614 v ar iant is a w ell-established fact, its pr actical consequences ar e still debated.K orber et al.(2020)pr esent exper imental evidence that the G614 v ar iant is associated with gr eater infectivity and clinical evidence that the new v ar iant is linked with higher vir al loads , although not with gr eater disease sev er ity.H u et al.(2020), O z ono et al.(2020), and Zhang et al.(2020)r epor t similar findings.H o wev er , these latter r esults r egar ding gr eater infectivity and higher vir al load ar e not y et the consensus among scientists(G r ubaugh et al., 2020).I n other wor ds , ther e is so me evidence—although not conclusiv e—that indi-Shut down economy Bad policy or bad luck Good policy or good luck Keep economy open

cates that the pandemic ’ s timing may hav e play ed a r ole deter mining the quadr ant wher e e ach place fall s in F igur e 1.I f indeed the or iginal D614 v ar iant is less infectious , Asian countr ies(who w er e exposed mor e to this earlier for m of the vir us)faced a mor e str aightfor war d tr ade-off betw een containing the epidemic and sustaining economic activity.E v en within the U.S., C alifo r nia, likely due to its closer ties to Asia, exper ienced a higher pr ev alence of lineages of D614 at the star t of the health cr isis than N ew Y or k, closer to E ur ope , and thus it had better outcomes r egar dless of the policies adopted.3.C umulativ e D eaths and C umulativ e E conomic Loss This section sho ws the empir ical v ersions of the tr ade-off graphs for v ar ious countr ies and U.S.states using GDP and unemplo yment as measures of the economic outcomes.3.1 I nter national E vidence W e use GDP data fr om the OECD(2020)5 and death data fr om J ohns H opkins U niv ersity CSSE(2020)to study the international evidence on C O VID-19 deaths and GD P.F igur e 5 plots the C O VID-19 deaths per million population as of A ugust 24 against the loss in GD P.“ GDP Loss ” is the cumulativ e loss in GDP since the star t of 2020(w e curr ently hav e data fr om Q1 and Q2)and is annualiz ed.F or example , a v alue of 6 means that the loss since the star t of 2020 is equiv alent to a six percent loss in annual GD P.F igur e 5: I nter national C o vid D eaths and Lost GDP GDP LOSS(PERCENT YEARS)7 6 5 4 3 2 1 0-1 0 100 200 300 400 500 600 700 800 900 COVID DEATHS PER MILLION PEOPLE N ote: “ GDP Loss ” is the cumulativ e loss in GDP since the star t of 2020 and is annualized.F or example , a v alue of 6 means that the loss since the star t of 2020 is as if the economy lost six per cent of its annual GD P.B efor e discussing our findings , some war nings ar e appr opr iate.F irst, w e only hav e 5 W e also use data fr om v ar ious national statistical agencie s for sever al countries that hav e r eleased 2020Q2 data that has not been integrated in the OECD database y et;see A ppendix A.Spain Philippines India France United Kingdom Italy Portugal Mexico China Slovakia Estonia Greece Austria Singapore Germany Fin.Israel Denmark Belgium Netherlands United States Poland Norway Japan Sweden Chile Korea, South Taiwan

obser v ations fr om a limited number of countr ies , as the 2020Q2 data is still being r e-leased.S econd, these early numbers ar e likely to be r evised substantiall y.E v en in normal times , the r evisions of GDP early r eleases ar e considerable(Ar uoba , 2008).The difficulties in data collection dur ing the last few months suggest that the r evisions for 2020 ar e bound to be ev en lar ger.6 Thir d, GDP is only an imperfect measur e of eco-nomic activity.Ther e ar e r easons to believ e that those imper fections ar e ev en mor e acute dur ing a pandemic.Think, consider go v er nment consumption.This item is measur ed b y the sum of emplo y ee compensation, consumption of fixed capital, and inter mediate goods and ser vices purchased.M any go v er nme nt ser vices , fr om the local DMV to public schools , w er e not offer ed(or o nly offer ed under a v er y limited schedule)dur ing the lockdo wns.H o wev er , most go v er nment emplo y ees w er e still paid(furloughs w er e r ar e in OECD countr ies), and the consumption of fixed capital is imputed accor ding to fixed depr eci-ation tables.Thus , ex cept for some r eduction of inter mediate goods and ser vices pur-chased, go v er nment consumption r emained unchanged fr om the perspectiv e of GD P.I ndeed, in the U.S., r eal go v er nment consumption increased 0.6% in 2020Q2 while GD P dr opped 9.5%.While par t of the increase can be attr ibuted to the fiscal stimulus and the fight against C O VID-19, a substantial par t of go v er nment consumption oper ated w ell below normal lev els dur ing that quar ter and such a change has had little impact on measur ed GD P.W ith these consider ations in mind, F igur e 5 suggests that ther e has not been a simple tr adeoff betw een deaths and GD P.R ather , countr ies can be seen to fall into sev er al gr oups.F irst, w e hav e countr ies with lo w deaths and moder ate GDP losses: T aiwan(with actual GDP gr o wth), K or ea, I ndonesia, N or way , J apan, China, P oland, and G erman y.S uch countr ies illustr ate an impor tant lesson fr om the cr isis: it was possible to emer ge with r elativ ely good per for mance on both dimensions.I mpor tantly , this gr oup is het-er ogeneous.I t includes countr ies in both Asia and E ur ope.I t includes countr ies with lar ge , densely populated cities.And it includes countr ies that ar e globally highly con-6 R ecall, for example , the note on the C or onavir us(C O VID-19)I mpact on J une 2020 Establishment and H ousehold S ur v ey D ata: “ The house hold sur v ey is gener ally collected thr ough in-person and telephone inter views , but personal inter views w er e not conducted for the safety of inter view ers and r espondents.The household sur v ey r esponse r ate , at 65 per cent, was about 18 per centage points lo w er than in months pr ior to the pandemic.” https://www.xiexiebang.combination of bad luck and impe r fect policy led these r egions to suffer on both dimensions dur ing the pandemic.The U nited K ingdom, as an example , has suffer ed fr om mor e than 600 deaths per million people and alr eady lost the equiv alent of 6 percent of a y ear ’ s GD P.Also , high C O VID-19 incidence might tr igger nonlinear effects on mor talit y.Ther e is evidence that the I talian and S panish health systems w er e o v er whelmed in M ar ch 2020, leading to many deaths that could hav e been av oide d.Ciminelli and G ar cia-M andic o ´(2020)sho w that mor tality in the I talian municipalities that w er e far fr om an ICU was up to 50% higher and ar gue that this was a pr oof the congestion of the emer gency car e system dur ing those crucial w eeks.A few countr ies in F igur e 5 ar e har der to classif y.I ndia and the P hilippines hav e exper ienced a consider abl e r eduction in GD P , but compar ativ ely few deaths per million people.As w e wi ll see later , ho wev er , the situation in I ndia is still v er y much ev olving.The U nited S tates and S w eden also stand out, w ith many C O VID-19 deaths but smaller r eductions in GDP than F r ance , I taly , or S pain.As with I ndia, ho wev er , the dynamic graphs w e sho w later suggests that the position of the U nited S tates is still ev olving.The case of S w eden is particularly inter esting because its go v er nment defied the consensus among other adv anced economies and imposed a much milder set of r e-str ictions and e...

机电专业成果进度汇报

成果报告

“宏观经济治理”：中国自主的经济学知识体系重要成果

调研成果报告

毕业设计进度