Input Specificity and the Propagation of Idiosyncratic

Input Specificity and the Propagation ofIdiosyncratic Shocks in Production Networks

The MIT Faculty has made this article openly available Please share how this access benefits you Your story matters

Citation Barrot Jean-Noeumll and Sauvagnat Julien ldquoInput Specificity and thePropagation of Idiosyncratic Shocks in Production Networksrdquo TheQuarterly Journal of Economics 131 3 (May 2016) 1543ndash1592 copy2016 The Author(s)

As Published httpdxdoiorg101093qjeqjw018

Publisher Oxford University Press

Version Authors final manuscript

Citable link httphdlhandlenet17211111134

Terms of Use Creative Commons Attribution-Noncommercial-Share Alike

Detailed Terms httpcreativecommonsorglicensesby-nc-sa40

INPUT SPECIFICITY AND THE PROPAGATION OFIDIOSYNCRATIC SHOCKS IN PRODUCTION NETWORKS

Jean-Noel Barrot and Julien Sauvagnat

This article examines whether firm-level idiosyncratic shocks propagate inproduction networks We identify idiosyncratic shocks with the occurrence ofnatural disasters We find that affected suppliers impose substantial outputlosses on their customers especially when they produce specific inputs Theseoutput losses translate into significant market value losses and they spill overto other suppliers Our point estimates are economically large suggesting thatinput specificity is an important determinant of the propagation of idiosyncraticshocks in the economy JEL Codes L14 E23 E32

I Introduction

The origin of business cycle fluctuations is a long-standingquestion in economics Starting with Long and Plosser (1983) anumber of studies have explored whether sectoral linkages mayhelp explain the aggregation of sector-specific shocks and havefound mixed empirical evidence of the importance of such link-ages Relative to the measurement of spillovers across sectorsspillovers within networks of firms have received little attentionin the empirical literature The main reason for this is the diffi-culty of identifying firm-specific shocks Whether firm-level idio-syncratic shocks propagate in production networks thereforeremains an open question

We thank Robert Barro the editor three anonymous referees Nittai BergmanBruno Biais Murillo Campello Thomas Chaney Philippe Chone Arnaud CostinotPierre Dubois Xavier Gabaix Nicola Gennaioli Christian Hellwig JohanHombert Augustin Landier Erik Loualiche Kambiz Mohkam SebastienPouget Ezra Oberfield Matthew Rognlie Ronan le Saout Fabiano SchivardiAntoinette Schoar and David Thesmar and participants at the MIT SloanFinance Retreat CREST internal seminar IDC Summer conference EARIE con-ference ASSAAFA meetings Stockholm School of Economics Oxford SaıdBocconi University EIEF CEMFI Copenhagen Business School WashingtonUniversity in St Louis UNC Kenan-Flagler York University and BostonUniversity for their valuable comments and suggestions Julien Sauvagnat grate-fully acknowledges financial support from the Agence Nationale de la RecherchemdashInvestissements dAvenir (ANR-11-IDEX-0003Labex EcodecANR-11-LABX-0047)

The Author(s) 2016 Published by Oxford University Press on behalf of Presidentand Fellows of Harvard College All rights reserved For Permissions please emailjournalspermissionsoupcomThe Quarterly Journal of Economics (2016) 1ndash50 doi101093qjeqjw018

1

The Quarterly Journal of Economics Advance Access published July 11 2016 at U

niversita Com

merciale L

uigi Bocconi on Septem

ber 20 2016httpqjeoxfordjournalsorg

Dow

nloaded from

On one hand firm-level idiosyncratic shocks should bequickly absorbed in production networks Firms plausibly orga-nize their operations to avoid being affected by temporary disrup-tion to their supplies Even when they face such disruptions theyshould be flexible enough to recompose their production mix orswitch to other suppliers The gradual decrease in trade tariffsand transportation costs and the development of online businessshould make it even easier for firms to adjust their sourcing Onthe other hand frictions might prevent firms from quicklymaking adjustments in the event of supply disruptions If firmsface switching costs whenever they need to replace a disruptedsupplier idiosyncratic shocks might propagate from firm to firmand gradually be amplified

This article studies whether firm-level shocks propagate orwhether they are absorbed in production networks To identifyfirm-level idiosyncratic shocks we consider major natural disas-ters in the past 30 years in the United States1 These events havelarge short-term effects on the sales growth of affected firms Wetrace the propagation of these shocks in production networksusing supplier-customer links reported by publicly listed USfirms If disrupted intermediate inputs can be easily substitutedwe should not expect input shocks to propagate significantly

Yet we find that suppliers hit by a natural disaster imposesignificant output losses on their customers When one of theirsuppliers is hit by a major natural disaster firms experience anaverage drop by 2 to 3 percentage points in sales growth followingthe event Given that suppliers represent a small share of firmsrsquototal intermediate inputs in our sample these estimates arestrikingly large We show that these estimates are robust to con-trolling for the location of firmsrsquo establishments In addition wedo not find any evidence of propagation from suppliers to cus-tomers when they are not in an active relationship which sug-gests that these estimates are not driven by common demandshocks triggered by natural disasters In robustness tests weshow that the estimates are similar when we control for hetero-geneous trends across firms with many or few suppliers when we

1 Natural disasters have already been used in prior work to instrument forschool displacement (Imberman Kugler and Sacerdote 2012) positive localdemand shocks (Bernile Korniotis and Kumar 2013) temporary shocks to locallabor markets (Belasen and Polachek 2008) changes in uncertainty (Baker andBloom 2013) and changes in risk perception (Dessaint and Matray 2013)

QUARTERLY JOURNAL OF ECONOMICS2

at Universita C

omm

erciale Luigi B

occoni on September 20 2016

httpqjeoxfordjournalsorgD

ownloaded from

weight regression by size when we restrict the sample toeventually treated firms only and whether we include localsupplier-customer relationships in the sample Given that weare interested in the propagation of firm-specific shocks we alsocheck that we are not picking up sector-level or even macroeco-nomic shocks Instead we find that the effect is not driven byevents that affect many suppliers at the same time or a largeshare of the same industry

We investigate whether the drop in firmsrsquo sales caused bysupply disruptions translates into value losses If input disrup-tions simply cause a delay in sales they would have little effect onfirmsrsquo cash flows and ultimately on firm value We do not observeany sort of overshooting in sales on average following disasterssuggesting that these sales are lost indeed We also conduct eventstudies and estimate firmsrsquo cumulative abnormal returns arounddisaster events affecting one of their suppliers We find that inputdisruptions cause a 1 drop in firmsrsquo equity value

We show that input specificity is a key driver of the propa-gation of firm-level shocks To do so we construct three measuresof suppliersrsquo specificity The first one borrows from the Rauch(1999) classification of goods traded on international marketsSecond we use suppliersrsquo RampD expenses to capture the impor-tance of relationship-specific investments Finally we use thenumber of patents issued by suppliers to capture restrictions onalternative sources of substitutable inputs We also check thatthe intensity of shocks affecting suppliers or the supplierrsquos rela-tive size do not systematically vary with our measures of inputspecificity in a manner that could drive the results We find thatthe propagation of input shocks varies strongly with our mea-sures of specificity Firmsrsquo sales growth and stock prices signifi-cantly drop only when a major disaster hits one of their specificsuppliers

We also ask whether the shock originating from one supplierpropagates horizontally to other suppliers of the same firm thatwere not directly affected by the natural disaster Even thoughfirms reduce output when one of their suppliers is hit they couldkeep buying from other suppliers and even start buying moreEven if the customer reduces purchases from all its suppliersfollowing the disruption of one of its inputs other suppliersmight be able to find alternative buyers for their productionInstead we find large negative spillovers of the initial shock toother suppliers The effect is only observed when the disaster hits

INPUT SPECIFICITY AND IDIOSYNCRATIC SHOCKS 3

at Universita C

omm

erciale Luigi B



ownloaded from

a specific supplier We show that our estimates are robust to con-trolling for the location of suppliersrsquo establishments Moreoverwe do not find evidence of horizontal propagation when the eco-nomic link between firms is inactive which confirms that ourestimates are not driven by common demand shocks

A potential concern with our analysis is the selected nature ofour network structure We obtain firmsrsquo network relationshipsfrom the obligation that publicly listed US firms have under reg-ulation Statement of Financial Accounting Standards (SFAS) No131 to report selected information about operating segments ininterim financial reports issued to shareholders including theidentity of any customer representing more than 10 of total re-ported sales Hence our sample comprises only suppliers withmajor customers and only publicly listed firms which might biasour estimates To ensure that our results are not driven by thisselection issue we run similar analysis using an alternative net-work structure and confirm that our results are not sensitive to therestriction of the sample to publicly listed firms

To check whether our estimates fall within a reasonable rangewe present a general equilibrium network model based on Long andPlosser (1983) and Acemoglu et al (2012) and find that our reduced-form estimates are consistent with the model predictions for highlevels of complementarity across intermediate input suppliers Wefinally assess the economic importance of the propagation channel bycomputing the aggregate dollar value of sales lost for suppliers andcustomers in our sample after suppliers are hit by natural disastersWe find that $1 of lost sales at the supplier level leads to $240 of lostsales at the customer level which indicates that relationships inproduction networks substantially amplify idiosyncratic shocks

Overall our findings highlight that the specificity of inter-mediate inputs allows idiosyncratic shocks to propagate in pro-duction networks They echo numerous press reports indicatingthat natural disasters have important disruptive effects thatpropagate along the supply chain2 They also highlight the pres-ence of strong interdependencies in production networks which

2 See for instance lsquolsquoHurricane Isaac Lessons For The Global Supply Chainrsquorsquo(Forbes August 31 2012) (Available at httpwwwforbescomsitesciocentral20120831hurricane-isaac-lessons-for-the-global-supply-chain431fd68b2515)and lsquolsquoA Storm-Battered Supply Chain Threatens Holiday Shoppingrsquorsquo (New YorkTimes April 11 2012) (Available at httpwwwnytimescom20121105businessa-storm-battered-supply-chain-threatens-the-holiday-shopping-seasonhtml_rfrac140)


at Universita C

omm

erciale Luigi B



ownloaded from

are highly relevant to assess the implications of corporatebailouts3

This article contributes to several strands of the literature Itrelates to a growing body of work assessing whether significantaggregate fluctuations may originate from microeconomic shocksThis view has long been discarded on the basis that these shockswould average out and thus would have negligible aggregate ef-fects (Lucas 1977) Two streams of papers challenge this intui-tion the first is based on the idea that large firms contributedisproportionately to total output (Gabaix 2011 Carvalho andGabaix 2013) the second stream posits that shocks are transmit-ted in the economy through industry linkages (Long and Plosser1987 Jovanovic 1987 Durlauf 1993 Bak et al 1993 Horvath1998 2000 Conley and Dupor 2003 Di Giovanni andLevchenko 2010 Carvalho 2010 Caselli et al 2011 Acemogluet al 2012 Bigio and LarsquoO 2013 Caliendo et al 2014 Baqaee2015) However the empirical evidence on the importance ofsector linkages for the aggregation of sector-specific shocks ismixed and depends on the level of aggregation (Horvath 2000)the way linkages are modeled (Foerster Sarte and Watson 2011)and the specification of the production function (Jones 2011Atalay 2013) Whereas earlier work has focused on the linkagesacross sectors4 we carefully estimate linkages within networks offirms5 In contemporaneous work Todo Nakajima and Matous(2014) Carvalho Nirei and Saito (2014) and Boehm Flaaenand Pandalai-Nayar (2015) study the supply-chain effects of theJapanese earthquake of 2011 Our setting which encompassesmultiple natural disasters over a period of 30 years allows usto disentangle input disruptions from common demand shocksand cleanly identify the importance of input specificity for thepropagation and amplification of idiosyncratic shocks We addto this literature by documenting that in addition to propagating

3 In testimony to the Senate Committee on Banking Housing and UrbanAffairs on December 4 2008 Ford CEO Alan Mulally said lsquolsquoThe collapse of one orboth of our domestic competitors would threaten Ford because we have 80 percentoverlap in supplier networks and nearly 25 percent of Fordrsquos top dealers also ownGM and Chrysler franchisesrsquorsquo

4 Di Giovanni Levchenko and Mejean (2014) is a recent exception5 While this article takes the network structure as given Chaney (2014)

Oberfield (2013) and Carvalho and Voigtlander (2014) among others explicitlymodel the formation of business networks


at Universita C

omm

erciale Luigi B



ownloaded from

to downstream firms idiosyncratic shocks also propagate hori-zontally into supplier networks6

Furthermore we build on earlier work that considers theimportance of switching costs for the propagation of firm-levelshocks A number of studies have analyzed the role of switchingcosts in banking relationships for the diffusion of financial shocks(Slovin Sushka and Polonchek 1993 Hubbard Kuttner andPalia 2002 Khwaja and Mian 2008 Fernando May andMegginson 2012) Amiti and Weinstein (2013) and Chodorow-Reich (2014) find that such frictions can explain a large share ofthe aggregate drop in investment and employment in the recentfinancial crisis We show that switching costs between trade part-ners are substantial and can explain the propagation of shocks innetworks of nonfinancial firms The existence of costs of searchingfor suppliers is a key parameter in recent studies of firmsrsquo sourc-ing decisions (Antras Fort and Tintelnot 2014 BernardMoxnes and Saito 2014) Our findings suggest that these costscan be large in the short run

We add to a growing body of work in financial economics thatstudies how firms are affected by their environment in particularby their customers and suppliers Recent studies have found ev-idence of comovement in stock returns within production net-works (Cohen and Frazzini 2008 Hertzel et al 2008 Menzlyand Ozbas 2010 Ahern 2012 Boone and Ivanov 2012 KellyLustig and Van Nieuwerburgh 2013) Our results which empha-size the importance of input complementarity and switchingcosts provide a foundation for this comovement In additionour results relate to prior studies of the implications of productmarket relationships for firmsrsquo corporate policies (Titman 1984Titman and Wessels 1988 MacKay and Phillips 2005 Kale andShahrur 2007 Campello and Fluck 2007 Banerjee Dasguptaand Kim 2008 Chu 2012 Moon and Phillips 2014 Ahern andHarford 2014) A key result of this literature is that firmswhose suppliers need to make relationship-specific investmentshold less leverage to avoid imposing high liquidation costs onthem Our results suggest that an alternative reason firmslinked to specific suppliers hold less leverage is to avoid the riskof financial distress brought about by input disruptions

6 The finding that shocks propagate horizontally is related to Kee (2015) whodocuments that domestic firms can benefit from the entry of foreign rivals throughthe enhanced productivity of their shared domestic suppliers


at Universita C

omm

erciale Luigi B



ownloaded from

The remainder of the article is organized as follows SectionII presents our empirical strategy Section III presents the dataSection IV describes the results and Section V concludes

II Identification Strategy

The main source of identification in this article is the occur-rence of major natural disasters We identify disruptions to sup-pliersrsquo output in a given quarter with the event that a naturaldisaster hits the county where their headquarters is located Ofcourse firmsrsquo plants and establishments are not always located inthe same county as their headquarters This measurement erroris likely to bias the estimates against finding any effect of naturaldisasters on firmsrsquo output In addition using establishment-leveldata from Infogroup7 we find that in our sample of suppliers theaverage (median) firm has 60 (67) of its employees located atits headquarters (see Table II later)

There are many different but unobservable reasons disastersmight affect firmsrsquo output It might be that they trigger poweroutages disrupting production8 Perhaps assets including build-ings machines or inventories are damaged Finally firmsrsquo work-force or management might be prevented from reaching theworkplace Although we have no way to pin down the exact chan-nel through which disasters disrupt production we confirm inSection IV that such disasters have a temporary and significantnegative effect on these suppliersrsquo sales growth9

The main focus of the article is not the disruption to the sup-plying firm itself but the impact on the firmrsquos customers and onthe customersrsquo other suppliers Our identification strategy closelyapproximates the following example Assume that firm S1 is asupplier to firm C who also purchases input from firm S2Suppose however that S1 and S2 do not have any economiclinks other than their relationship with C We first analyze theresponse of C when S1 is hit by a natural disaster We then focus

7 We describe the data in more detail in Section III8 Hines Apt and Talukdar (2008) find that 44 of major power outages in the

United States are weather-related (ie caused by tornado hurricanetropicalstorm ice storm lightning windrain or other cold weather)

9 Following standard event methodology we also find that firms experience asignificant stock price decline following the date of a major disaster hitting thecounty location of their headquarters


at Universita C

omm

erciale Luigi B



ownloaded from

on the response of S2 In each case we contrast these effects withcharacteristics that capture the cost of replacing S1 with anotherprovider of the same input

To capture supplier-customer links we rely on the obligationthat publicly listed firms have in the United States to report anycustomer accounting for more than 10 of their sales10 We con-sider that S1 is a supplier to C in all years ranging from the firstto the last year when S1 reports C as one of its customers Wethen estimate the effect of the shock to S1 on Crsquos sales growth in adifference-in-differences framework at the firm level where thetreatment amounts to having at least one supplier hit by a natu-ral disaster

We run the following OLS regression at the firm-quarterlevel in our sample of customers11

Salesit4t frac14 0 thorn 1HitsOneSupplierit4

thorn 2DisasterHitsFirmit4 thorn i thorn t thorn iteth1THORN

where Salesit4t is the sales growth between the currentquarter and the same quarter in the previous year HitsOneSupplierit4 is a dummy taking the value of 1 if at least oneof the firmrsquos suppliers is located in a county hit by a naturaldisaster in the same quarter in the previous year DisasterHitsFirmit4 is a dummy equal to 1 if the firm is directly hit by anatural disaster in the same quarter in the previous year i

and t are year-quarter and firm fixed effects All regressionscontrol for fiscal quarter fixed effects and for the number ofsuppliers with dummies indicating terciles of the number ofsuppliers three years prior to date t In some specificationswe include stateyear fixed effects and industryyear fixed ef-fects We introduce lagged controls for size age and profitabil-ity interacted with year-quarter fixed effects12 We build thesecontrols by interacting year-quarter dummies with terciles offirmsrsquo assets age and return on assets three years prior to

10 We describe the data in more detail in Section III11 The benefit of using sales is that it is available at the quarterly level for all

publicly listed US firms which is the ideal frequency to study the temporary dis-ruptions caused by natural disasters The drawback is that sales reflect prices andquantities However in Section IV we show that similar results are obtained at thesector level using a quarterly index of industrial output

12 Including these controls ensures that the estimates are not driven by het-erogeneous trends among large old or profitable firms


at Universita C

omm

erciale Luigi B



ownloaded from

date t In all regressions standard errors are clustered at thefirm level to account for serial correlation of the error termwithin firms The coefficient of interest is 1 which measuresthe effect on the firmrsquos sales growth of a disruption to at leastone of its suppliers

For our strategy to consistently estimate the effect of theshock to S1 on C we need to make several identifying assump-tions First Crsquos sales growth would have been flat in the absenceof treatment (parallel trends assumption) We check whether wefind any effect in the quarter prior to the natural disaster and weformally test whether eventually treated and never treated firmsexperience diverging trends over the sample period

Second the natural disaster should affect C only through itsdisruptive effect on S1 (exclusion restriction) However this as-sumption might be violated if Crsquos own production facilities areaffected by the disaster We handle this problem by excludingfrom the sample any supplier-customer relationships whereboth partiesrsquo headquarters are located within 300 miles of eachother13 In addition we add a dummy in the regression that cap-tures whether the headquarter county location of C is hit by anatural disaster Finally we use establishment-level data to con-trol for the fact that plants of C might be directly hit by disastersaffecting S1 The exclusion restriction might otherwise be vio-lated if Crsquos demand is affected by the disaster hitting one of itssuppliers for instance because its customer base is located closeto its supplier base If this were the case disasters hitting thesupplierrsquos location would presumably affect the customer irre-spective of whether their economic link was active To addressthis concern we use the unique feature of our data relative toother studies of production networks namely that we observethe time series of relationships By means of illustration we pre-sent in Figure I the evolution of the supplier-customer networkfrom 1995 to 2000 (please see the online edition of this article toview the figure in color) Relationships that were active in 1995but not in 2000 are depicted in red Relationships that are activein both years are depicted in green Relationships that were notactive in 1995 but were active in 2000 are depicted in blue It isclear from this figure that a substantial share of relationships

13 We show in Table A7 in the Online Appendix that the estimates are insen-sitive to this cutoff


at Universita C

omm

erciale Luigi B



ownloaded from

start or end within this five-year window This allows us to checkwhether we only observe an effect of disruptions to S1 on Crsquosoutput when the link between S1 and C is active

One might also worry that firms endogenously select theirlocationmdashand the location of their suppliersmdashby taking into ac-count the fact that natural disasters will disrupt their productionThis is not a threat to the identification strategy if anything thisshould bias the results against finding any propagation effectsHowever it might affect the external validity of these estimatesa point that we discuss in Section IVE

FIGURE I

Network Evolution from 1995 to 2000

This figure illustrates the evolution of the supplier-customer network from1995 to 2000 Relationships that were active in 1995 but not in 2000 are de-picted in light gray Relationships that are active in both years are depicted indark gray Relationships that were not active in 1995 but were active in 2000are depicted in black (Please see the online edition of this article to view thefigure in color)


at Universita C

omm

erciale Luigi B



ownloaded from

We would expect to find an effect only when the firm facesrelatively large costs of searching for and switching to alternativesuppliers of the same input Otherwise following the disruptionof the supplier of a given intermediate input the firm would turnto other providers of the same input and maintain its first-bestlevel of output We thus contrast the effects with the extent towhich the customer can switch to other suppliers of a given inputWe hypothesize that suppliers are more likely to produce specificinputs if they operate in industries producing differentiatedgoods if they have a high level of RampD or if they hold patentsUsing these three different proxies to measure the specificity ofany given supplier we split the main variable of interest in equa-tion (1) Disaster Hits One Supplier into two dummy variablesDisaster hits one specific supplier and Disaster hits onenonspecific supplier indicating respectively whether at leastone specific and nonspecific supplier of the firm is hit by a naturaldisaster

Finally we study the effect of the initial shock on S1 on anyother supplier S2 of C To do so we run an OLS regression in oursample of suppliers at the firm-quarter level of sales growth be-tween the current quarter and the same quarter in the previousyear on Disaster hits firm a dummy equal to 1 if the firm is di-rectly hit by a natural disaster Disaster hits one customer adummy equal to 1 if (at least) one customer of the firm is hit bya natural disaster and Disaster hits one customerrsquos supplier themain variable of interest a dummy taking the value of 1 if (atleast) one other supplier of the firmrsquos customer(s) is hit by a nat-ural disaster In all specifications we control for fiscal quarterfixed effects and for the number of customersrsquo suppliers withdummies indicating terciles of the number of customersrsquosuppliers14

14 This test rests on the same assumptions needed to identify the effect of thenatural disaster on C In particular it needs to be the case that the natural disastershould affect S2 only through its disruptive effect on S1 and its indirect effect on CThe exclusion restriction might be violated if S2rsquos production facilities are affectedby the disaster hitting S1 We drop from the sample any relationship where S2 islocated within 300 miles of either S1 or C In addition we use establishment-leveldata to control for the fact that plants of S2 might be directly affected by disastersThe exclusion restriction might alternatively be violated if S2rsquos demand is affectedby the disaster hitting S1 for instance because its customer base is located close toS1 If this were the case disasters hitting S1 would presumably affect S2 irrespec-tive of whether they were linked through their relationship with C We address this


at Universita C

omm

erciale Luigi B



ownloaded from

III Data

IIIA Firm-Level Information

Financial data and information about firmsrsquo headquarter lo-cation are retrieved from Compustat North AmericaFundamentals Quarterly database We restrict our sample tononfinancial firms whose headquarters are located in theUnited States over the 1978ndash2013 period15 We restrict thesample to firms reporting in calendar quarters All continuousvariables are winsorized at the 1st and 99th percentiles of theirdistributions We adjust our computation of the growth in salesand cost of goods sold for inflation using the GDP deflator of theBureau of Economic Analysis

As already mentioned we use the county location of head-quarters to identify whether a firm is hit by a natural disasterWe make an important adjustment to the (county and state) lo-cation of the headquarters of the firms in our sample Compustatonly records the last available location of the headquarters ofeach firm We update the county and state of each firm in oursample using information gathered by Infogroup which goesback as far as 199716 In addition we use employment and estab-lishment information from Infogroup to construct controls forwhether more than 10 of employees of a firm across all estab-lishments are hit by a natural disaster17 Finally we constructthe 48 Fama-French industry dummies from the conversion tablein the appendix of Fama and French (1997) using the firmrsquos four-digit SIC industry code

We also examine the effect of input disruptions on stockprices For this we obtain data on daily stock prices from theCenter for Research in Security Prices (CRSP daily file) Wefocus on ordinary shares of stocks traded on NYSE AMEX andNASDAQ

concern by checking that disasters hitting S1 only affect S2 when the economic linkbetween S1 and C is active and when the economic link between S2 and C is active

15 Customer-supplier links detailed below are available only from 1978 2013is the last year for which data on major natural disasters are available

16 This leads to a nonnegligible adjustment Between 1997 and 2013 firmsrsquoheadquarter county location is corrected for 13 (15) of observations in oursample of customers (suppliers) For years before 1997 we update the county andstate location of firms using the nearest available year in Infogroup

17 Infogroup makes phone calls to establishments to gather among other dataitems the number of full-time equivalent employees


at Universita C

omm

erciale Luigi B



ownloaded from

IIIB Supplier-Customer Links

Crucial to our analysis is the identification of relationshipsbetween suppliers and their customers Fortunately regulationSFAS No 131 requires firms to report selected information aboutoperating segments in interim financial reports issued to share-holders In particular firms are required to disclose certain finan-cial information for any industry segment that makes up morethan 10 of consolidated yearly sales assets or profits as well asthe identity of any customer representing more than 10 of thetotal reported sales18

We take advantage of this requirement to obtain informationon supplier-customer links For each firm filing withthe Securities and Exchange Commission (SEC) we obtain thename of its principal customers and associated sales fromthe Compustat Segment files from 1978 to 201319 Given thatwe are mainly interested in publicly listed customers for whichaccounting data are available we associate each name to aCompustat identifier by hand More specifically we use a pho-netic string-matching algorithm to match each customer namewith the five closest names from the set of firms filing with theSEC and all their subsidiaries We then select the best match byhand by inspecting the firm and customersrsquo names and indus-tries Customers with no match are excluded from the sample

Customers in our data set represent approximately 75 ofthe total sales in Compustat over the sample period which makesus confident that the sample is representative of the US econ-omy There are limitations associated with these data In partic-ular we generally do not observe suppliers whose sales to thecustomer are lower than 10 of their revenues20 We discussthis selection issue in Section IVE and show that our estimateshold when we consider alternative network structures that arenot subject to this selection issue in Section A3 of the OnlineAppendix

18 Although the data set also includes the variable that captures the annualsales of the reporting supplier to the reported customer this information is pro-vided on a voluntary basis and often imputed

19 Other papers have used the customer-supplier data including Fee andThomas (2004) and Fee Hadlock and Thomas (2006) who analyze respectivelythe effect of mergers and corporate equity ownership on the value of suppliers

20 Some firms voluntarily report the names of other major customers whensales are below this threshold


at Universita C

omm

erciale Luigi B



ownloaded from

IIIC Natural Disasters

We obtain information on each major natural disaster hittingthe US territory from the SHELDUS (Spatial Hazard and LossDatabase for the United States) database maintained by theUniversity of South Carolina For each event the database pro-vides information on the start date the end date and the FederalInformation Processing Standards (FIPS) code of all affectedcounties We restrict the list to events classified as major disas-ters that occurred after 1978 which is when supplier-customerdata become available We also restrict the sample to disasterslasting less than 30 days with total estimated damages above $1billion 2013 constant dollars As evidenced in Table I we are leftwith 41 major disasters of all kinds including blizzards earth-quakes floods and hurricanes These disasters affect a broadrange of US states and counties over the sample periodHowever they are generally very localized and affect at most22 of US employment21 Figure II shows the frequency of oc-currence of major natural disasters over the sample period foreach US county Some counties are more frequently hit thanothers especially those located along the southeast coast of theUS mainland In comparison as evidenced in Figure III thelocation of suppliers in the sample spans the entire US main-land including counties that are never and counties that areoften hit by natural disasters

IIID Input Specificity

We rely on three different proxies to measure the specificityof any given supplier We first borrow from Rauch (1999) whoclassifies inputs into differentiated or homogeneous depending onwhether they are sold on an organized exchange This classifica-tion groups inputs into 1189 industries classified according to thefour-digit SITC Rev 2 system Each industry is coded as beingeither sold on an exchange reference priced or homogeneous Weuse the bridge between the SITC and SIC classification used inFeenstra (1996) to compute the share of differentiated goods pro-duced in each industry A supplier is thus considered specific if itoperates in an industry that lies above the median along this

21 Most of the events affect less than 10 of US employment which providesus with an ideal setting to cleanly identify input disruptions from general equilib-rium effects We further check that the estimates are similar for relatively smalland relatively large natural disasters (see Online Appendix Table A5)


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from

4 or more321Never Hit

FIGURE II

Major Natural Disaster Frequency by US Counties

This map presents the number of major natural disaster strikes for eachcounty in the US mainland over the sample period The list of counties af-fected by each major natural disaster is obtained from the SHELDUS databaseat the University of South Carolina Table I describes the major natural disas-ters included in the sample

101 to 35011 to 1002 to 1010

FIGURE III

Location of Sample Suppliersrsquo Headquarters

This map presents for our sample the number of suppliersrsquo headquarterslocated in each US county Data on the location of headquarters are obtainedfrom Compustat and Infogroup databases


at Universita C

omm

erciale Luigi B



ownloaded from

dimension We also proxy for the level of specificity with the ratioof RampD to sales and we classify suppliers as specific if this ratiolies above the sample median in the two years prior to any givenquarter Finally suppliers holding patents are more likely to pro-duce inputs that cannot be easily replaced by other suppliersHence in each quarter we also sort firms based on the numberof patents they issued in the three previous years and consider asspecific those lying above the sample median To do so we re-trieve patent information from Google patents assembled byKogan et al (2012)22

IIIE Summary Statistics

Table II presents summary statistics for our sample Panel Apresents the customer sample which consists of 80574 firm-quarters between 1978 and 2013 There are 2051 firms in thissample A firm is included in the sample in each quarter betweenthree years before and three years after it appears as a customerin the Compustat Segment files On average a firm is reported by138 suppliers in a given year The main variables of interest arethe growth in sales and cost of goods sold over the previous fourquarters The sample averages for these variables are 102 and106 and their medians are 40 and 38 The probability that(at least) one of the suppliers of a given firm is hit by a naturaldisaster in any quarter is 14 This compares with the probabil-ity of 16 that the customer is directly hit by a natural disaster

There are on average seven years between the first and thelast year a supplier reports a firm as a customer The averagesales of suppliers to their customers (identified with variableSALECS in the Compustat Segment files) represents around25 of firmsrsquo cost of goods sold Given that wages and associatedcosts represent a large share of cost of goods sold this is probablyan underestimate of the importance of these suppliers in cus-tomers inputs However this suggests that suppliers are smallwith respect to customers There is no significant difference in theshare that specific and nonspecific suppliers represent in firmsrsquocost of goods sold across our three measures of input specificityFinally suppliers are located on average a little over 1250 milesaway from their customers irrespective of whether they arespecific

22 We thank the authors for making the data available to us


at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II

DESCRIPTIVE STATISTICS

Obs Mean Std Dev p1 p50 p99

Panel A Customer sampleSales growth (t 4t) 80574 0102 0375 0606 0040 1927Cogs growth (t 4t) 79358 0106 0411 0651 0038 2193Disaster hits firm (t) 80574 0016 0126 0000 0000 1000Disaster hits one

supplier (t)80574 0014 0118 0000 0000 1000

Number of suppliers 80574 1383 4162 0000 0000 19000

Diff RampD Patent

S NS S NS S NS

Av duration ofrelationships

7125 6692 6373 8335 7821 6618

Av supplier-customerHQs distance

1332 1210 1502 1214 1388 1219

Av suppliersrsquo input share 0022 0025 0017 0023 0025 0022

Eventually treated Never treated

Obs Mean Std dev Obs Mean Std dev

Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128

Panel B Supplier sampleObs Mean Std dev p1 p50 p99

Sales growth (t 4t) 139976 0188 0814 -0876 0045 4568Disaster hits firm (t) 139976 0017 0127 0000 0000 1000Disaster hits a customer (t) 139976 0008 0088 0000 0000 0000Disaster hits a customerrsquos

supplier (t)139976 0042 0200 0000 0000 1000

Number of customers 139976 0711 0964 0000 0000 4000 Employees at HQs

county102279 0597 0365 0000 0667 1000

Notes This table presents the summary statistics for our sample Panel A presents the customer samplewhich consists of 80574 firm-quarters between 1978 and 2013 There are 2051 firms in this sample A firmis included in the customer sample for each quarter between three years before the first year and three yearsafter the last year it appears as a customer in the Compustat Segment files The main variables of interestare the growth in sales and cost of goods sold relative to the same quarter in the previous year Panel A alsoreports for customer firms the average duration of relationships with their suppliers (computed as thenumber of years between the first and last year the supplier reports the firm as a customer in theCompustat Segment files) the average distance in miles (computed using the Haversine formula) betweenthe headquarters (HQs) county of the firm and the headquarters county of its suppliers and the averagesuppliersrsquo input share (measured as the ratio of the suppliersrsquo sales of the supplier to the firm over the firmrsquoscost of goods sold) separately for relationships with specific (S) and nonspecific (NS) suppliers In columns (1)and (2) a supplier is considered as specific if its industry lies above the median of the share of differentiatedgoods according to the classification provided by Rauch (1999) In columns (3) and (4) a firm is consideredspecific if its ratio of RampD expenses over sales is above the median in the two years prior to any givenquarter In columns (5) and (6) a firm is considered as specific if the number of patents it issued in the pastthree years is above the median The last part of Panel A compares the size age and return on assets (ROA)of eventually treated firms namely those with suppliers that are hit by a major natural disaster at leastonce over the sample period and never treated firms Panel B presents the supplier sample which consists of139976 firm-quarters between 1978 and 2013 There are 4686 firms in this sample A firm is included in thesupplier sample for each quarter between three years before the first year and three years after the last yearit reports another firm as a customer in the Compustat Segment files The main variable of interest is thegrowth in sales relative to the same quarter in the previous year


at Universita C

omm

erciale Luigi B



ownloaded from

The last part of Panel A compares the size age and return onassets of eventually treated and never treated firms23 Eventuallytreated firmsmdashthose having one supplier hit by a major disasterat least once during the sample periodmdashare larger older andslightly more profitable than never treated firms This makes itall the more important to ensure in the empirical analysis thatfirm-level characteristics are not driving the results

Panel B presents the supplier sample which consists of139976 firm-quarters between 1978 and 2013 There are 4686firms in this sample A firm is included in the sample in eachquarter between three years before and three years after it reportsanother firm as a customer in the Compustat Segment files Thesefirms report an average of 07 customers The main variable ofinterest is the growth in sales over the previous four quartersThe sample average for this variable is 188 and the median is45 The probability that a firm in this sample is hit by a naturaldisaster in any quarter is 17 The probability that one of a firmrsquoscustomers is hit in any given quarter is 08 Finally the proba-bility that one of its customersrsquo suppliers is hit is 42

We investigate the distribution of suppliers and customers rel-ative to the entire Compustat universe in Table A12 of the OnlineAppendix In Panel A we present the number and share of quarter-firm per 48 Fama-French industries for suppliers customers andthe Compustat universe We do not find very large deviations acrossthe three samples This makes us confident that our sample is fairlyrepresentative of the Compustat universe In Panel B we furthersplit the supplier and customer samples depending respectively onwhether suppliers are hit and whether customers are treated in agiven quarter Again we do not find any patterns indicating thatour estimates might be driven by any specific industry

IV Results

IVA Effect on Affected Suppliers

We first explore the extent to which suppliersrsquo production isaffected when the county where their headquarters are located is

23 Size is defined as total assets (Compustat item AT) Age is defined as thenumber of years since incorporation when the date of incorporation is missing ageis defined as the number of years since the firm has been in the Compustat databaseReturn on assets (ROA) is operating income before depreciation and amortization(item OIBDP) divided by total assets


at Universita C

omm

erciale Luigi B



ownloaded from

hit by a natural disaster24 As already discussed we have no wayto formally pin down the channel through which natural disas-ters translate into disruptions to suppliersrsquo production functionsInstead we consider their effect on firmsrsquo sales

In our sample of suppliers we regress firmsrsquo sales growth (rel-ative to the same quarter in the previous year) on a series of dum-mies indicating whether a major natural disaster hits the firm ineach of the current and previous five quarters as well as fiscal quar-ter year-quarter and firm fixed effects The results are presented inTable III In the first column the coefficient on the dummies indi-cating that a disaster hits the firm in the previous three quarters arenegative and significant ranging from 33 to 45 percentage pointswhich indicates that suppliersrsquo sales growth drops significantly forthree consecutive quarters following a disaster We introduce con-trols for size age and profitability interacted with year-quarterfixed effects in the second column The coefficient range does notchange which suggests that differences in the types of firms thatare hit do not drive the patterns in sales growth In the third andfourth columns we introduce stateyear fixed effects and indus-tryyear fixed effects The effect goes down slightly in magnitudebut remains significant in quarter (t 1) Taken together the re-sults suggest that relative to firms in the same state or the sameindustry firms with headquarters located in a county directly af-fected by the natural disaster seem to do worse

One purpose of the following section is to assess whether sup-pliersrsquo specificity is a driver of the propagation of firm-level shocksHowever if shocks to specific suppliers were on average largerthan shocks to nonspecific suppliers this would lead us to mechan-ically overestimate the effect of input specificity on the propaga-tion of shocks We check in Table IV that the disruption caused bynatural disasters is not larger for specific than for nonspecific sup-pliers To do so we consider the sample of suppliers and regressfirmsrsquo sales growth on a dummy indicating whether the firm is hitby a disaster (in the previous four quarters) a dummy taking thevalue of 1 if the firm is specific and the interaction between the

24 It is important to note that the effect of a natural disaster on productioncould a priori go either way since the destruction triggered by disasters sometimesgenerates a local increase in demand (Bernile Korniotis and Kumar 2013)Anecdotal evidence indeed suggests that providers of basic supplies experienceboosts in sales in the period around the disaster (see for instance BloombergAugust 26 2011 lsquolsquoHome Depot Lowersquos stocks get hurricane boostrsquorsquo) (Availableat httpmoneycnncom20110826marketstweets_stocktwits)


at Universita C

omm

erciale Luigi B



ownloaded from

two We run the same regression for our three measures of inputspecificity The coefficient on the interaction term is always posi-tive although not statistically significant which suggests thatshocks to specific suppliers are if anything of smaller magnitudethan shocks to nonspecific suppliers25

IVB Downstream Propagation Effect on Customersrsquo Sales

In this section we estimate the effect on firmsrsquo sales of shocksaffecting their suppliers We first illustrate the results in Figure IV

TABLE III

NATURAL DISASTER DISRUPTIONSmdashSUPPLIER SALES GROWTH

Sales Growth (t ndash 4t)

Disaster hits firm (t) 0006 0004 0001 0011(0018) (0018) (0018) (0018)

Disaster hits firm (t 1) 0045 0045 0032 0039(0016) (0016) (0017) (0018)





Firm FE Yes Yes Yes YesYear-quarter FE Yes Yes Yes YesSize age ROA

year-quarter FENo Yes Yes Yes

State-year FE No No Yes YesIndustry-year FE No No No YesObservations 139976 139976 139976 139976R2 0177 0192 0212 0233

Notes This table presents estimates from panel regressions of firmsrsquo sales growth relative to the samequarter in the previous year on a dummy indicated whether the firm is hit by a major disaster in thecurrent and each of the previous five quarters All regressions include fiscal quarter year-quarter andfirm fixed effects In the second and fourth columns we also control for firm-level characteristics (dummiesindicating terciles of size age and ROA respectively) interacted with year-quarter dummies In the thirdand fourthcolumns we include state dummies interacted with year dummies In the fourth column weinclude 48 Fama-French industry dummies interacted with year dummies Standard errors presented inparentheses are clustered at the firm level Regressions contain all firm-quarters of our supplier sample(described in Table II Panel B) between 1978 and 2013 and denote significance at the 10 5and 1 levels respectively

25 The coefficient on Specific firm is omitted in the first and second columnsbecause firmsrsquo industry classification is fixed over time and therefore absorbed byfirm fixed effects


at Universita C

omm

erciale Luigi B



ownloaded from

which compares the growth in sales (relative to the same quarterin the previous year) at different quarters surrounding a majornatural disaster for both directly affected suppliers and their cus-tomers The graph highlights that input disruptions translate intolost sales for the firm a few quarters after the supplier is hit

1 Baseline Results We then run the OLS panel regressiondetailed in equation (1) and present the results in Table V InPanel A we consider the effect of input disruption on salesgrowth The variable of interest is the dummy Disaster hits onesupplier (t 4) which takes the value of 1 if (at least) one of thefirmrsquos suppliers is hit by a natural disaster in quarter t 4 and 0otherwise The estimates in the first column indicate that salesgrowth drops by 31 percentage points Given the sample mean of10 the estimate is economically large In the second column weintroduce controls for lagged size age and profitability inter-acted with year-quarter fixed effects The estimate decreases

TABLE IV

NATURAL DISASTERS DISRUPTIONSmdashSPECIFIC VERSUS NONSPECIFIC SUPPLIERS


Supplier specificity Diff RampD Patent

Disaster hits firm(t 4t 1)

0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)

Disaster hits specificfirm (t 4t 1)

0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)

Firm FE Yes Yes Yes Yes Yes YesYear-quarter FE Yes Yes Yes Yes Yes YesSize age ROA

year-quarter FENo Yes No Yes No Yes

Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192

Notes This table presents estimates from panel regressions of firmsrsquo sales growth relative to the samequarter in the previous year on a dummy indicated whether the firm is hit by a major disaster in one ofthe previous four quarters In the first and second columns a firms is considered as specific if its industrylies above the median of the share of differentiated goods according to the classification provided by Rauch(1999) In the third and fourth columns a firm is considered specific if the ratio of its RampD expenses oversales is above the median in the two years prior to any given quarter In the fifth and sixth columns afirm is considered as specific if the number of patents it issued in the previous three years is above themedian All regressions include fiscal-quarter year-quarter and firm fixed effects In the second fourthand sixth columns we also control for firm-level characteristics (dummies indicating terciles of size ageand ROA respectively) interacted with year-quarter dummies Standard errors presented in parenthesesare clustered at the firm level Regressions contain all firm-quarters of our supplier sample (described inTable II Panel B) between 1978 and 2013 and denote significance at the 10 5 and 1levels respectively


at Universita C

omm

erciale Luigi B



ownloaded from

slightly to 27 percentage points and remains significant In thethird column we control for stateyear fixed effects and obtainsimilar results This confirms that the effect of input disruptionon sales is not related to temporary shocks at the state level or to

-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters

Direct effect (suppliers) Propagation (customers)

FIGURE IV

Natural Disaster Strikes and Sales Growth

This figure presents difference-in-differences estimates of quarterly salesgrowth in the year before and the two years after a major natural disasterfor both directly affected suppliers and their customers Sales growth is thegrowth in sales relative to the same quarter in the previous year Thedashed line connects estimated coefficients of the following regression per-formed in the supplier sample

Salesit4t frac14 thornX9

frac144

HitsFirmit thorn i thorn t thorn it

The solid line connects estimated coefficients g of the following regressionperformed in the customer sample


frac144

HitsFirmit thornX9

frac144

gHitsSupplierit thorn i thorn t thorn it

where t and i are year-quarter and firm fixed effects respectively HitsFirmit is a dummy equal to 1 if a natural disaster hits firm i in year-quartert and HitsSupplierit is a dummy equal to 1 if a natural disaster hits atleast one supplier of firm i in year-quarter t Standard errors are clusteredat the firm level in both regressions and denote significance at the10 5 and 1 levels respectively The sample period spans 1978 to 2013


at Universita C

omm

erciale Luigi B



ownloaded from

the fact that treated firms might be closer to the disaster zonethan other firms In the fourth column we add industryyearfixed effects The point estimate is 19 percentage points whichsuggests that the effect is not driven by an industry-wide shock

TABLE V

DOWNSTREAM PROPAGATIONmdashBASELINE

Panel A Sales Growth (t ndash 4t)Disaster hits one

supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)


Number of suppliers Yes Yes Yes YesFirm FE Yes Yes Yes YesYear-quarter FE Yes Yes Yes YesSize age ROA



Panel B Cost of Goods Sold Growth (t ndash 4t)Disaster hits one

supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)





Notes This table presents estimates from panel regressions of firmsrsquo sales growth (Panel A) or cost ofgoods sold growth (Panel B) relative to the same quarter in the previous year on a dummy indicatingwhether (at least) one of their suppliers is hit by a major disaster in the same quarter of the previous yearAll regressions include a dummy indicating whether the firm itself is hit by a major disaster in the samequarter of the previous year as well as fiscal quarter year-quarter and firm fixed effects All regressionsalso control for the number of suppliers (dummies indicating terciles of the number of suppliers) In thesecond third and fourth columns we control for firm-level characteristics (dummies indicating terciles ofsize age and ROA respectively) interacted with year-quarter dummies In the third and fourth columnswe include state dummies interacted with year dummies In the fourth column we include 48 Fama-French industry dummies interacted with year dummies Regressions contain all firm-quarters of ourcustomer sample (described in Table II Panel A) between 1978 and 2013 Standard errors presented inparentheses are clustered at the firm level and denote significance at the 10 5 and 1levels respectively


at Universita C

omm

erciale Luigi B



ownloaded from

Across specifications the coefficient on the dummy Disaster hitsfirm (t 4) is negative which reflects the finding presented inTable III Similar results are obtained in Panel B when wereplace the dependent variable with the growth in the cost ofgoods sold Altogether the results indicate that disruptions totheir suppliersrsquo production strongly affect firmsrsquo sales growthwhich drops by a little over 25 with respect to the sampleaverage Since suppliers in the sample represent approximately25 of firmsrsquo cost of goods sold these estimates are strikinglylarge

The drop in sales growth should show no prior trends andshould be temporary for the parallel trends assumption to be sat-isfied As their suppliers restore their productive capacity firmsrsquosales growth should recover To test whether this is indeed thecase we analyze the dynamics of the effects We regress the firmrsquossales growth on dummies indicating whether a major disasterhits (at least) one of their suppliers in each of the current andthe previous five quarters The results presented in Table VI in-dicate that the coefficient in the same quarter of the previous year(Disaster hits one firm (t 4)) is the largest in absolute value Noeffect on firmsrsquo sales growth is found contemporaneously or priorto the quarter when the effect of natural disasters is found onsuppliers (which occurs in (t 1) see Table III) This confirmsthat the drop in firmsrsquo sales growth is not driven by prior trendsbut is indeed caused by the natural disaster affecting one of itssuppliers

We go a step further to test the validity of the parallel trendassumption We check whether eventually treated firms andnever treated firms experience diverging time trends in the ab-sence of major natural disasters To do so we regress firmsrsquo salesgrowth on a treatment dummy that equals 1 for firms eventuallytreated in our sample interacted with the full set of year-quarterfixed effects Ti dt and estimate the regression only over pe-riods for which no major natural disaster has hit the US terri-tory in the current or previous four quarters The regression alsoincludes terciles of the number of suppliers fiscal quarter fixedeffects and firm fixed effects We are mainly interested in theF-statistics of the joint significance test of all the Ti dt (seecolumn (6) of Table A2) If we fail to reject the null hypothesisthat they all equal 0 this would provide strong support for theparallel trend assumption Results are reported in Table A2 ofthe Online Appendix In all cases F-tests are small and we


at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI

DOWNSTREAM PROPAGATIONmdashSALES GROWTH DYNAMICS


Disaster hits one supplier (t) 0012 0010 0007 0003(0008) (0008) (0008) (0008)

Disaster hits one supplier(t 1)

0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)










Notes This table presents estimated coefficients from panel regressions of firmsrsquo sales growth relativeto the same quarter in the previous year on dummies indicating whether (at least) one of their suppliers ishit by a major disaster in the current and each of the previous five quarters All regressions includedummies indicating whether the firm itself is hit by a major disaster in the current and each of theprevious five quarters as well as fiscal-quarter year-quarter and firm fixed effects All regressions alsocontrol for the number of suppliers (dummies indicating terciles of the number of suppliers) In the secondthird and fourth columns we control for firm-level characteristics (dummies indicating terciles of sizeage and ROA respectively) interacted with year-quarter dummies In the third and fourth columns weinclude state dummies interacted with year dummies In the fourth column we include 48 Fama-Frenchindustry dummies interacted with year dummies Regressions contain all firm-quarters of our customersample (described in Table II Panel A) between 1978 and 2013 Standard errors presented in parenthesesare clustered at the firm level and denote significance at the 10 5 and 1 levelsrespectively


at Universita C

omm

erciale Luigi B



ownloaded from

always fail to reject at conventional levels the null hypothesisthat all Ti dt are 0 in the absence of major natural disastersThis makes us confident that never treated firms provide a goodcounterfactual for eventually treated firms in periods of majornatural disasters

One might be concerned that the results are driven by thelocation of customersrsquo plants close to the headquarters of theirsuppliers In Panel A Table VII we introduce a dummy takingthe value of 1 if more than 10 of the customerrsquos workforce acrossall establishments is hit by a natural disaster If headquartersrsquolocations are poor proxies for the true location of customersrsquo es-tablishments and if the economic link with the supplier proxiesfor the true location of the customer this variable should absorbthe effect The results indicate that this is not the case as thecoefficient on Disaster hits a supplier remains remarkably stable(compared with Table V) and statistically significant in allspecifications26

Another concern is that the estimates from Table V mightreflect common demand shocks affecting the firm and its suppli-ers for instance because their customer base is located in thesame area To handle this issue we augment our OLS regressionswith a dummy called Disaster hits any eventually linked suppli-ersrsquo location which takes the value of 1 if any headquartersrsquocounty locations of all suppliers once in a relationship with thefirm is hit by a natural disaster If the effects that we are pickingup in Table V reflect common demand shocks this variableshould subsume the main variable of interest Disaster hits onesupplier This is arguably a very conservative test of our hypoth-esis since it is likely that some of the supplier-customer relation-ships that we observed from the SFAS No 131 were initiatedearlier (at a time where the customer represented less than10 of the suppliersrsquo sales) or maintained later We present theresults of this specification in Table VII Panel B The coefficienton the additional variable is insignificant whereas the coefficienton Disaster hits one supplier remains stable and significant in allspecifications Hence input disruptions caused by natural disas-ters propagate only when there is an active business relationshipbetween the disrupted supplier and the firm

26 The results are similar when instead of a 10 threshold we use a 1 2 or5 threshold They are also similar when we restrict the sample to firm-years forwhich establishment data are available from 1997 to 2013


at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII

DOWNSTREAM PROPAGATIONmdashROBUSTNESS

Sales Growth (t 4t)

Panel A Controlling for share of the workforce hitDisaster hits more than 10

of firmrsquos workforce (t 4)0006 0005 0002 0009(0010) (0010) (0010) (0010)


0031 0027 0030 0020(0009) (0009) (0009) (0008)





Panel B Controlling for whether any lsquolsquoeventually linkedrsquorsquo supplier is hitDisaster hits any eventually

linked suppliersrsquo location(t 4)

0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)





Notes This table presents estimates from panel regressions of firmsrsquo sales growth relative to the samequarter in the previous year on a dummy indicating whether (at least) one supplier is hit by a majordisaster in the same quarter of the previous year All regressions include a dummy indicating whether thefirm itself is hit by a major disaster in the same quarter in the previous year as well as fiscal quarteryear-quarter and firm fixed effects Panel A also includes a dummy indicating whether 10 or more of thefirmrsquos workforce is hit Panel B includes a dummy indicating whether (at least) one location of any sup-plier once in a relationship with the firm is hit In the second through fourth columns we control for firm-level characteristics (dummies indicating terciles of size age and ROA respectively) interacted with year-quarter dummies In the third and fourth columns we include state dummies interacted with yeardummies In the fourth column we include 48 Fama-French industry dummies interacted with yeardummies Regressions contain all firm-quarters of our customer sample (described in Table II Panel A)between 1978 and 2013 Standard errors presented in parentheses are clustered at the firm-level and denote significance at the 10 5 and 1 respectively


at Universita C

omm

erciale Luigi B



ownloaded from

2 Input Specificity The propagation of input shocks should bestronger when the supplier is specific and thus harder for thefirm to replace We use our three measures of specificity to testwhether this is the case We expect the coefficient on the dummyDisaster hits one specific supplier to be positive significant andlarger than the dummy on the coefficient Disaster hits onenonspecific supplier The results are presented in Table VIIIOverall the effect is indeed much stronger when a disaster hitsa specific supplier rather than a nonspecific one The effect ofnonspecific suppliers is generally insignificant whereas theeffect of specific suppliers is greater than the baseline estimatesHence the results suggest that input specificity is a key driver ofthe propagation of shocks from suppliers to their customers

TABLE VIII

DOWNSTREAM PROPAGATIONmdashINPUT SPECIFICITY


Supplier Specificity Diff RampD Patent

Disaster hits onenonspecificsupplier (t 4)

0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)

Disaster hits onespecific supplier(t 4)

0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)

Disaster hits firm(t 4)

0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers

Yes Yes Yes Yes Yes Yes



Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262

Notes This table presents estimates from panel regressions of firmsrsquo sales growth relative to the samequarter in the previous year on two dummies indicating whether (at least) one specific supplier andwhether (at least) one nonspecific supplier is hit by a major disaster in the same quarter of the previousyear In the first and second columns a supplier is considered as specific if its industry lies above themedian of the share of differentiated goods according to the classification provided by Rauch (1999) In thethird and fourth columns a supplier is considered specific if its ratio of RampD expenses over sales is abovethe median in the two years prior to any given quarter In the fifth and sixth columns a supplier isconsidered as specific if the number of patents it issued in the previous three years is above the medianAll regressions include a dummy indicating whether the firm itself is hit by a major disaster in the samequarter in the previous year as well as fiscal quarter year-quarter and firm fixed effects All regressionsalso control for the number of suppliers (dummies indicating terciles of the number of suppliers) In thesecond fourth and sixth columns we control for firm-level characteristics (dummies indicating terciles ofsize age and ROA respectively) interacted with year-quarter dummies Regressions contain all firm-quarters of our customer sample (described in Table II Panel A) between 1978 and 2013 Standarderrors presented in parentheses are clustered at the firm level and denote significance at the10 5 and 1 levels respectively


at Universita C

omm

erciale Luigi B



ownloaded from

3 Robustness We perform a number of robustness tests andpresent the results in the Online Appendix We start with anadditional test of the parallel trend assumption which consistsof estimating the difference-in-differences specification usingonly observations of eventually treated firms We show in TableA3 that our results are similar to those in Table V when we re-strict the sample to eventually treated customer firms One mightalso be concerned that firms with many suppliers and firms withfew suppliers could be subject to differential trends We augmentthe baseline regression with dummies indicating terciles of thenumber of firmsrsquo suppliers interacted with year-quarter fixed ef-fects and present the results in Table A4 The estimates remainstable which indicates that the results are driven by the treat-ment rather than the number of firmsrsquo links

We next check that our results are not driven by large natu-ral disasters which would affect customers in our samplethrough their aggregate effect on the US economy To do sowe first interact the dummy Disaster hits one supplier with thevariable Large nb of affected firms which takes the value of 1 fordisasters that lie in the top half of the distribution of the totalnumber of directly affected Compustat firms The results are re-ported in columns (1) and (2) of Table A5 The coefficient on theinteraction term is positive and insignificant indicating that theeffect of input disruption does not vary with the importance ofdisastersmdashif anything it is smaller for more important ones Wealso look at whether the results differ for exporters and nonex-porters To do so we interact the dummy Disaster hits one sup-plier with the variable gt50 sales abroad which takes the valueof 1 if the customer firm reports sales abroad that represent morethan 50 of its total sales in the two years prior to any givenquarter As shown in columns (3) and (4) we find that the effectof the treatment is virtually the same for exporters and nonex-porters indicating again that the results are driven by input dis-ruptions rather than demand effects due to natural disasters onthe US economy

We also check whether our results are not driven by largenatural disasters through their effect at the sector level If sectorstend to be clustered geographically so that all firms tend to beaffected by natural disasters then the effects we are picking upshould be interpreted as sector-specific rather than firm-specificshocks In our sample the average and median share of affectedsales at the four-digit SIC level are 19 and 8 respectively


at Universita C

omm

erciale Luigi B



ownloaded from

We interact the dummy Disaster hits one supplier with variousdummies for whether more than 10 30 or 50 of the sup-plierrsquos four-digit SIC sector is affected by the disaster If what weare picking up is the effect of disruptions to geographically clus-tered sectors the coefficient on the interaction term should in-crease with the share of the sector being affected We find in TableA6 that this is not the case

We also check that our estimates are not sensitive to the300-mile cutoff we use to exclude supplier-customers that aregeographically close In Table A7 we vary this cutoff from 0 to500 miles and find that the results remain unchanged We alsofind results similar to the baseline in Table A8 where we onlyconsider treatments when the customer and supplier are neverjointly hit by a disaster in the same quarter throughout thesample period In Table A9 we also control for linkages acrossfirms via product- and input-market competition and find virtu-ally identical estimates of the coefficient on our main variable ofinterest Disaster hits a supplier We show in Table A10 that thecoefficients go down slightly but remain significant when weweight regressions by customersrsquo size (inflation-adjusted sales)This ensures that the effects we are picking up are not concen-trated among the smallest of the customers in our sampleFinally we confirm in Table A11 that the estimates are robustto an alternative definition of our main dependent variablenamely the difference in the logarithm of firm sales

IVC Downstream Propagation Effect on Customersrsquo Value

The drop in sales growth could simply reflect the fact thatsales are delayed which would have few consequences for firmsrsquocash flows and value However the estimates in Table VI indicatethat firmsrsquo sales growth does not overshoot in the quarters follow-ing disasters suggesting that these sales are lost indeed We goone step further and ask whether the disruption to specific sup-pliers is reflected in firmsrsquo stock returns We follow standardevent study methodology and consider the first day when agiven major disaster hits a county in which a linked supplierrsquosheadquarters is located Under the efficient market hypothesisthe news of input disruption should be quickly reflected in thefirm share price allowing us to compute the associated drop infirm value


at Universita C

omm

erciale Luigi B



ownloaded from

1 Returns Analysis We select all firm-disaster pairs in oursample satisfying the following requirements (i) (at least) one sup-plier of the firm is hit by the disaster (ii) the firm is not hit by thedisaster (iii) the firm and its suppliers are not hit by another majordisaster in the previous or following 30 trading days around theevent date and (iv) the firm has no missing daily returns in theestimation or event window The event date is the day consideredas the beginning of the disaster in the SHELDUS database27 Wefind 1082 events satisfying the above requirements For each firm-disaster pair we then estimate daily abnormal stock returns usingthe Fama-French three-factor model

Rit frac14 i thorn iRMt thorn si SMBt thorn hi HMLt thorn iteth2THORN

where Rit is the daily return of firm i RMt is the daily returnof the market portfolio minus the risk-free rate SMBt is thedaily return of a small-minus-big portfolio and HMLt is thedaily return of a high-minus-low portfolio28 The three-factormodel is estimated over the interval from 260 to 11 tradingdays before the event date We use the estimates of the modeli i si hi ui to construct abnormal returns in the eventwindow as

ARit frac14 Rit ethi thorn iRMt thorn siSMBt thorn hiHMLtTHORNeth3THORN

We then aggregate daily abnormal stock returns by averagingthem over all firm-disaster pairs (N) and summing them overthe trading days of different event windowsmdashfrac12101 frac12010frac121120 frac122130 frac123140 and frac121040 where frac12t0frac1410Tfrac1440 isa 51-trading-days window starting 10 trading days before theevent datemdashto obtain cumulative average abnormal returns(CAAR) Formally

CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN

27 If the day reported as the beginning of the disaster in SHELDUS is a non-trading day we use the next trading day as the event date If more than one supplieris hit by the same disaster the earliest beginning date in SHELDUS is consideredas the event date

28 RM SMB and HML returns are obtained from Kenneth Frenchrsquos websiteSMB and HML returns are meant to capture size and book-to-market effectsrespectively


at Universita C

omm

erciale Luigi B



ownloaded from

We also examine whether the effect on firmsrsquo stock returns differswith the specificity of affected suppliers To do so we computefirmsrsquo cumulative abnormal returns separately for natural disas-ters affecting or not (at least) one specific supplier

Because natural disasters hit several firms at the same timethis is likely to generate cross-sectional correlation in abnormalreturns across (indirectly affected) customer firms To address thisissue we test for statistical significance using the ADJ-BMP t-statistic proposed by Kolari and Pynnonen (2010) which is a mod-ified version of the standardized test developed in BoehmerMasumeci and Poulsen (1991) Kolari and Pynnonen (2010)show that the ADJ-BMP test accounts for cross-sectional correla-tion in abnormal returns and is robust to serial correlation29

2 Results Table IX illustrated in Figure V reports cumula-tive average abnormal returns over different event windowsmdashaswell as their respective ADJ-BMP t-statisticsmdashseparately fortreated firms their (directly affected) suppliers and untreatedfirms that is for which all linked suppliers are not affected bya given major disaster CAAR for treated customer firms on the 51trading days event window frac12t0 frac14 10T frac14 40 are negative andstatistically significant indicating a drop of around 1 in the firmstock price when one of its supplier(s) is hit by a major naturaldisaster A large fraction of this drop occurs in the 21 trading daysfrac12t0 frac14 10T frac14 10 around the event for which CAAR are highlystatistically significant which is consistent with investorsquickly reacting to the news30 These findings indicate thatfirmsrsquo sales are not simply postponed in reaction to input disrup-tions but materialize into sizable value losses

We find that directly affected suppliers experience an abnor-mal drop in returns of around 25 over the same event windowIn the third column of Table IX we consider the average stockprice reaction of untreated customers Reassuringly the size ofthe effect is small in all event windows for these firms

29 Note also that simulations presented in Kolari and Pynnonen (2010) sug-gest that the ADJ-BMP test is superior in terms of power to the commonly usedportfolio approach to account for serial correlation

30 Earthquakesrsquo striking dates might be considered truly unexpected eventsHowever in the case of hurricanes for instance stock price valuation might incor-porate forecasts about the passage and severity of the hurricane in the few daysprior to the striking date


at Universita C

omm

erciale Luigi B



ownloaded from

Finally Table X presents the results separately for eventsaffecting specific and nonspecific suppliers For our three mea-sures of input specificity we find that firms experience a largerdrop in returns when disasters hit their specific suppliers thantheir nonspecific ones

Overall these findings indicate that stock prices react to sup-plier risk especially when linked suppliers are specific Thesefindings provide to the best of our knowledge the first cleanlyidentified evidence that input disruptions have an effect on firmvalue and that input specificity is a key determinant thereof

IVD Horizontal Propagation Effect on Related Suppliers

Here we explore whether the effects documented above spillover to other related suppliers that are not directly affected by thenatural disaster but only indirectly through their common

TABLE IX

DOWNSTREAM PROPAGATIONmdashEFFECT ON FIRM VALUE

CAAR

Customers Suppliers Customers(Direct Effect) (Control Group)

(N frac14 1082) (N frac14 2004) (N frac14 6379)

[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)

Notes This table presents CAAR of customer firms around the first day of a natural disaster affecting(at least) one of its suppliers When more than one supplier is affected by the same natural disaster theevent day is the earliest date across affected suppliers reported in SHELDUS database Abnormal returnsare computed after estimating for each firm-disaster pair a three-factor Fama-French model over theinterval from 260 to 11 trading days before the event date We exclude firm-disaster observations withmissing returns in the estimation or event windows when the firm itself is hit by the disaster or whenthe firm or one of its suppliers are hit by another major disaster in the 30 trading days around the eventADJ-BMP t-statistics presented in parentheses are computed with the standardized cross-sectionalmethod of Boehmer Masumeci and Poulsen (1991) and adjusted for cross-sectional correlation as inKolari and Pynnonen (2010) The second column reports CAAR of directly hit supplier firms The thirdcolumn reports CAAR of unaffected customer firms namely firm-disaster pairs for which no suppliersreporting the firm as a customer have been hit by the disaster Computations of abnormal returns followthe same procedure as above The sample period is from 1978 to 2013 and denote significance atthe 10 5 and 1 levels respectively


at Universita C

omm

erciale Luigi B



ownloaded from

relationship with the same customer31 Going back to the settingdescribed in Section II we are interested in the response of S2 tothe drop in Crsquos sales triggered by a disruption to S1rsquos production

FIGURE V

Cumulative Average Abnormal Returns

This figure presents cumulative average abnormal returns (CAAR) of customerfirms around the first day of a natural disaster affecting (at least) one of its sup-pliers When more than one supplier is affected by the same natural disaster theevent day is the earliest date across affected suppliers reported in SHELDUSdatabase Abnormal returns are computed after estimating for each firm-disasterpair a three-factor Fama-French model over the interval from 260 to 11 tradingdays before the event date Firm-disaster observations with missing returns in theestimation or event windows for which the firm itself is hit by the disaster or forwhich the firm or one of its suppliers are hit by another major disaster in theprevious or following 30 trading days on either side of the event date are excludedWe find 1082 customer firm-disaster pairs satisfying these requirements

31 The nature of our data limits our ability to precisely estimate the effect ofdisruptions on customersrsquo customers there are only 012 of the observations inour sample for which the dummy Disasters hits a supplierrsquos supplier takes the value1 The alternative network structure we obtain from Capital IQ (see Section A3 inthe Online Appendix) is not subject to this limitation In that case there are 99 ofthe observations for which the dummy Disasters hits a supplierrsquos supplier takes thevalue 1 In Table A19 we runthis analysis and find that the effect on salesgrowth ofdisruptions affecting a firmrsquos suppliersrsquo supplier is negative but insignificant


at Universita C

omm

erciale Luigi B



ownloaded from

Note that the direction of the effect is unclear It could bepositive or negative depending on the degree of complementarityacross intermediate input suppliers If intermediate inputs arestrong substitutes the response might be positive that is thedisruption of supplier leads to an increase in sales growth ofother suppliers servicing the same customer Conversely whenthey are complement related suppliers could experience a de-crease in sales in particular if they cannot easily shift their pro-duction to other customers To estimate the direction of the effectwe run the OLS specification presented in Section II in thesample of suppliers

The results are presented in Table XI In the first column thecoefficient on Disaster hits one customerrsquos supplier is a negativeand significant 38 percentage point decrease in sales growth

TABLE X

DOWNSTREAM PROPAGATIONmdashINPUT SPECIFICITY AND EFFECT ON FIRM VALUE

Customersrsquo CAAR When DisasterHits at Least One Supplier

Supplier Specificity Diff RampD PatentN frac14 628 N frac14 454 N frac14 318 N frac14 764 N frac14 375 N frac14 707

At least onespecific supplier Yes No Yes No Yes No

[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)

Notes This table presents CAAR of customer firms separately for events affecting (at least one)specific supplier or only nonspecific suppliers In the first column a supplier is considered as specific ifits industry lies above the median of the share of differentiated goods according to the classificationprovided by Rauch (1999) In the third column a supplier is considered specific if the ratio of its RampDexpenses over sales is above the median in the two years prior to any given quarter In the fifth column asupplier is considered as specific if the number of patents it issued in the previous three years is above themedian Abnormal returns are computed after estimating for each firm-disaster pair a three-factor Fama-French model over the interval from 260 to 11 trading days before the event date Firm-disaster obser-vations with missing returns in the estimation or event windows for which the firm itself is hit by thedisaster or for which the firm or one of its suppliers are hit by another major disaster in the previous orfollowing 30 trading days on either side of the event date are excluded ADJ-BMP t-statistics presented inparentheses are computed with the standardized cross-sectional method of Boehmer Masumeci andPoulsen (1991) and adjusted for cross-sectional correlation as in Kolari and Pynnonen (2010) Thesample period is from 1978 to 2013 and denote significance at the 10 5 and 1 levelsrespectively


at Universita C

omm

erciale Luigi B



ownloaded from

This is consistent with substantial negative spillovers to relatedsuppliers In line with Table III the coefficient on Disaster hitsfirm (t 4t 1) is also negative and significant Results pre-sented in the second through fourth columns are obtained byaugmenting the model with a dummy Disaster hits one cus-tomerrsquos specific supplier which isolates the effect of disruptionsto specific suppliers of the customer The estimates indicate thatmost of the negative effect feeding back from the customer comesfrom initial shocks to specific suppliers (either differentiatedRampD-intensive or patent-intensive) These results uncover animportant channel through which firm-specific shocks propagatehorizontally across suppliers of a given firm

TABLE XI

HORIZONTAL PROPAGATIONmdashRELATED SUPPLIERSrsquo SALES GROWTH

Sales Growth (t 4t)


Disaster hits firm (t 4t 1) 0040 0040 0041 0040(0013) (0013) (0013) (0013)

Disaster hits one customer(t 4t 1)

0002 0001 0001 0002(0021) (0021) (0021) (0021)

Disaster hits one customerrsquossupplier (t 4t 1)

0038(0010)

Disaster hits one customerrsquosspecific supplier (t 4t 1)

0047 0048 0040(0013) (0014) (0013)

Disaster hits one customerrsquosnon-specific supplier(t 4t 1)

0011 0013 0015(0013) (0013) (0013)

Number of customersrsquo Suppliers Yes Yes Yes YesFirm FE Yes Yes Yes YesYear-quarter FE Yes Yes Yes YesSize age ROA

year-quarter FEYes Yes Yes Yes

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192

Notes This table presents estimated coefficients from panel regressions of firmsrsquo sales growth relativeto the same quarter in the previous year on one dummy indicating whether one of the firmrsquos customersrsquoother suppliers is hit by a major disaster in the previous four quarters The second and fourth columnssplit customersrsquo other suppliers into specific and nonspecific suppliers All regressions include two dum-mies indicating whether the firm itself is hit in the previous four quarters and whether one of the firmrsquoscustomer is hit in the previous four quarters All regressions also control for the number of customersrsquosuppliers (dummies indicating terciles of the number of customersrsquo suppliers) All regressions include fiscalquarter year-quarter and firm fixed effects as well as firm-level characteristics (dummies indicatingterciles of size age and ROA respectively) interacted with year-quarter dummies Standard errors pre-sented in parentheses are clustered at the firm level Regressions contain all firm-quarters of our suppliersample (described in Table II Panel B) between 1978 and 2013 and denote significance at the10 5 and 1 levels respectively


at Universita C

omm

erciale Luigi B



ownloaded from

These effects may be driven by the fact that some establish-ments of S2 are located close to S1 To address this concern weintroduce a dummy equal to 1 if at least 10 of S2rsquos workforce ishit by the disaster If the effect that we are measuring in Table XIis due to the fact that the link between S1 and S2 proxies for thelocation of S2rsquos plants then this variable should absorb the effectsof our main treatment variable In Table XII Panel A the intro-duction of this dummy does not affect the coefficient on Disasterhits one customerrsquos supplier

Another concern might be that these results are driven byunobserved economic links between S1 and S2 not related totheir common relationship with C The fact that S2rsquos salesgrowth is affected when S1 is hit by a natural disaster could bethe consequence of the fact that S2rsquos demand is located close tothe headquarters of S1 and is therefore affected by the disasterIn Table XII Panel B we augment the model with a dummycalled Disaster hits any customersrsquo eventually linked suppliersrsquolocation which takes the value of 1 if for any customer of S1 atleast one of all the locations of all suppliers once in a relationshipwith this customer is hit by a natural disaster If the effects thatwe are picking up in Table XI reflect the geographical clusteringof the demand to suppliers of C this variable should subsume themain variable of interest However we find that the results arerobust to the introduction of this variable

Finally we check that the effect on S2 is not driven by acommon industry shock affecting both S1 and S2 by introducinga dummy called Disaster hits any eventually linked customerrsquos sup-pliers taking the value of 1 whenever C is affected by a shock to S1irrespective of whether there is an active business relationshipbetween C and S2 If the effects found on S2 are related tocommon shocks to S1 and S2 the inclusion of this variableshould absorb the effect of the variable Disaster hits one customerrsquossupplier However in Table XII Panel C the coefficients are robustto the introduction of this variable In addition the coefficient onthis variable is not different from 0 which indicates that the initialshock would not spill over to related suppliers in the absence of anactive economic link through their common customer

IVE Discussion

1 A General Equilibrium Network Model To check whetherour estimates fall within a reasonable range we present a general


at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII

HORIZONTAL PROPAGATIONmdashROBUSTNESS

Supplierrsquos Sales Growth (t ndash 4t)

Diff RampD Patent

Panel A Controlling for share of the workforce hitDisaster hits at least 10 of firmrsquos

workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)


Disaster hits one customer (t 4t 1) 0002 0001 0001 0002(0021) (0021) (0021) (0021)

Disaster hits one customerrsquos supplier(t 4t 1)

0038(0010)

Disaster hits one customerrsquos specificsupplier (t 4t 1)

0047 0048 0040(0013) (0014) (0013)

Disaster hits one customerrsquosnonspecific supplier (t 4t 1)

0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192

Panel B Controlling for whether any customersrsquo lsquolsquoeventually linked supplierrsquorsquo is hitDisaster hits any customersrsquo eventually

linked suppliersrsquo location (t 4t 1)0022 0022 0022 0022(0013) (0013) (0013) (0013)




0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192

Panel C Controlling for whether any lsquolsquoeventually linked customersrsquo rsquorsquo supplier is hitDisaster hits any eventually linked

customerrsquos supplier (t 4t 1)0008 0005 0019 0012(0016) (0014) (0012) (0014)




0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192Number of customersrsquo suppliers Yes Yes Yes YesFirm FE Yes Yes Yes Yes


at Universita C

omm

erciale Luigi B



ownloaded from

equilibrium network model based on Long and Plosser (1983) andAcemoglu et al (2012) in Section A1 of the Online Appendix thatdelivers predictions of the magnitude of the pass-through of sup-pliersrsquo disruptions to their customers (vertical propagation) andthe other suppliers of their customers (horizontal propagation)32

Firms have constant-returns-to-scale production functions andchoose the quantities of labor and intermediate inputs to maxi-mize profits while households provide labor and consume Wemodel the effect of natural disasters as the destruction of asmall fraction of the output of impacted firms We express thepass-through of supply disruptions to any given firm as theratio of its sales drop to the disrupted supplierrsquos sales drop as afunction of model parameters In Online Appendix A1B we pre-sent and discuss the predicted value of the downstream and hor-izontal pass-throughs and the ratio of both pass-throughs as afunction of the elasticity of substitution across intermediate

TABLE XII

CONTINUED


Diff RampD Patent

Year-quarter FE Yes Yes Yes YesSize age ROA year-quarter FE Yes Yes Yes YesObservations 139976 139976 139976 139976R2 0192 0192 0192 0192

Notes This table presents estimated coefficients from panel regressions of firmsrsquo sales growth relativeto the same quarter in the previous year on one dummy indicating whether one of the firmrsquos customersrsquoother suppliers is hit by a major disaster in the four previous quarters as well as in Panel A a dummyindicating whether 10 or more of the firmrsquos workforce is hit by a major disaster in the four previousquarters in Panel B a dummy indicating whether (at least) one location of any other suppliers once in arelationship with a customer of the firm is hit by a major disaster in the four previous quarters in Panel Ca dummy indicating whether (at least) one other supplier of any customer once in a relationship with thefirm is hit by a major disaster in the four previous quarters The second through fourth columns splitcustomersrsquo other suppliers into specific and nonspecific suppliers All regressions include two dummiesindicating whether the firm itself is hit in the previous four quarters and whether one of the firmrsquoscustomer is hit in the previous four quarters All regressions also control for the number of customersrsquosuppliers (dummies indicating terciles of the number of customersrsquo suppliers) All regressions include fiscalquarter year-quarter and firm fixed effects as well as firm-level characteristics (dummies indicatingterciles of size age and ROA respectively) interacted with year-quarter dummies Standard errors pre-sented in parentheses are clustered at the firm level Regressions contain all firm-quarters of our suppliersample (described in Table II Panel B) between 1978 and 2013 and denote significance at the10 5 and 1 levels respectively

32 Within our framework horizontal propagation combines the effect of thedemand feedback effect from the common customer and the effect of complemen-tarity across suppliers of intermediate inputs See Grossman and Rossi-Hansberg(2008) for a framework that allows for a clear decomposition of both channels


at Universita C

omm

erciale Luigi B



ownloaded from

input suppliers33 We compare these pass-throughs to the ratio ofthe estimates we obtain from Table VI and XI namely a down-stream pass-through close to 2

4frac14 05 a horizontal pass-through

close to 384frac14 095 and a ratio of the horizontal over downstream

pass-through of 09505 frac14 19 Our empirical estimates are compara-

ble yet slightly higher to the predictions of the model for valuesof nearing 0 the Leontief limit which are 03 05 and 15respectively Our reduced-form coefficients are therefore consis-tent with the predictions of a network model with high levels ofcomplementarity across intermediate input suppliers

2 Sample Representativeness An important concern with thenetwork production data used in this article is that it includesrelationships wherein the customer typically represents morethan 10 of the sales of the supplier and both firms are publiclylisted Even though the firms we consider are the largest in theeconomy this double selection issue might introduce some bias inour estimates A priori the fact that we are missing some suppli-ers introduces noise which is likely to bias the results againstfinding any sort of propagation34 Nonetheless in Section A3 ofthe Online Appendix we go one step further to ensure that thisselection issue is not driving the results We replicate our resultsusing an alternative network structure that is not prone to theselection issues highlighted above We consider an alternativefirm-level data set obtained from Capital IQ which providesfirm-to-firm relationships based on regulatory filings as well aspress reports and is therefore not subject to the 10 reportingthreshold Reassuringly we find similar estimates when we runour baseline tests using this network data (see columns (1) and (2)of Online Appendix Table A17) We also show that downstreampropagation does not depend on whether the supplier is publiclylisted and that horizontal propagation does not depend on

33 We set the share of intermediate inputs to 055 the elasticity of substitutionbetween labor and intermediate inputs to 1 and the cross-firm elasticity of demandto 2 See Online Appendix A1B for discussions of parameter values and of thesensitivity of model predictions to these values

34 Moreover Atalay et al (2011) use these data and show that the truncationissue does not affect the shape of the in-degree distribution the fraction of suppliersof each customer that we miss because of the 10 threshold is similar for customerswith many or few suppliers


at Universita C

omm

erciale Luigi B



ownloaded from

whether the customer is publicly listed (see columns (3) and (4) ofTable A17) A limitation of our study is that we cannot observethe output growth of privately held firms Reassuringly we findsimilar estimates when we consider the effect of supply disrup-tions at the industry level (Online Appendix Table A20) or at theindustry state level (Online Appendix Table A21)

3 Measurement of Output The benefit of using Compustatdata is that they allow us to measure sales growth at the quar-terly frequency Although we cannot disentangle quantity fromprices from Compustat data we also find significant effects ofintermediate input disruptions on real output growth when weperform our analyses at the industry level (Online AppendixTable A20) A related concern is that we cannot measure valueadded from Compustat Hence we cannot disentangle from thedrop in sales what comes from the drop in value added and whatcomes from the drop in intermediate input use The finding thatstate industry GDP growth reacts to intermediate input dis-ruptions (Online Appendix Table A21) suggests that supplyshocks ultimately reduce downstream value added35

4 Role of Inventories Inventories typically serve as a bufferfor production One might expect differential patterns of propa-gation depending on whether firms hold high or low inventoriesWe first ask whether suppliers holding high levels of inventoriestend to experience the drop in sales growth later than those hold-ing little inventories In Table A13 in the Online Appendix wefind that high inventory suppliers experience their largest drop insales growth in quarter (t 2) one quarter later than low inven-tory suppliers who experience it in quarter (t 1) We then turnto regressions at the customer level Given what we found at thesupplier level we would expect the effect to kick in later for cus-tomers sourcing from high inventory suppliers In OnlineAppendix Table A14 we split the Disaster hits supplier dummy

35 In addition we find in Table A16 in the Online Appendix that the ratio ofsales to capital and labor goes down following intermediate input disruptionsHence in contrast to Costinot Vogel and Wang (2013) who assume that down-stream errors (or disasters) are more costly because they destroy a longer chain ofvalue added our results suggest that upstream errors are more costly because theyprohibit downstream tasks from being performed We thank an anonymous refereefor highlighting this point


at Universita C

omm

erciale Luigi B



ownloaded from

into two dummies indicating whether a disaster hits a high or lowinventory supplier We find that the drop in sales growth occursin quarters (t 3) and (t 4) when a low-inventory supplier is hitwhereas it occurs in quarters (t 4) and (t 5) when a high-inventory supplier is hit This illustrates that inventories delaythe propagation of supply shocks in production networks

5 Economic Significance We first note that treated firms inour sample make up a large share of the US economy In anygiven quarter eventually treated firms represent 36 and 43 oftotal Compustat sales and total stock market value respectivelyIn quarters when natural disasters hit the US territory treatedfirms represent on average 9 of total Compustat sales and totalstock market value which is economically significant By con-trast eventually hit suppliers represent 12 and 15 of totalCompustat sales and stock market value respectively in an av-erage quarter and affected suppliers represent on average 1 oftotal Compustat sales and total stock market value in quarterswhen a disaster hits Another way to assess the economic impor-tance of propagation is to compare the aggregate output losses forsuppliers and customers in our sample To compute this multi-plier we first estimate the lost sales for each firm in the sampledue to direct or indirect exposure to natural disasters The drop insales growth is obtained for each firm by taking the residual of aregression of sales growth on fiscal quarter year-quarter andfirm fixed effects as well as controls for size age and return onassets interacted with year-quarter dummies in the four quartersfollowing any disaster We then apply these sales growth resid-uals to the 2013 constant dollar value of firmsrsquo sales to obtain thedollar value of lost sales We aggregate these lost sales acrosssuppliers and customers in our sample We find that lost salesamount to approximately $246 billion for suppliers and $580 bil-lion for customers Hence $1 of lost sales at the supplier levelleads to $24 of lost sales at the customer level in our sample Thissuggests that relationships in production networks substantiallyamplify idiosyncratic shocks Whether or not this amplificationmechanism is powerful enough to generate fluctuations in aggre-gate output is a question that we leave to future research

6 Trends in Input Specificity Figure VI draws from Nunn(2007) to quantify the importance of input specificity The


at Universita C

omm

erciale Luigi B



ownloaded from

author uses the US input-output table to identify which inter-mediate inputs are used and in what proportions in the produc-tion of each final good Then using data from Rauch (1999)inputs are sorted into those sold on an organized exchangethose reference priced in a trade publication and those that aredifferentiated As evidenced from the graph the share of differ-entiated inputs is large and increasing Hence the propagationchannel examined in this article is likely to play an important andgrowing role for the aggregation of idiosyncratic shocks in pro-duction networks

7 External Validity Our results are informative for thesekinds of idiosyncratic shocks and their propagation in the econ-omy Nonetheless these results can plausibly be extended toother forms of firm-specific idiosyncratic shocks such as strikes

0

20

40

60

1982 1987 1992 1997

Share of total inputs

Sold on organized exchange Reference priced Differentiated

FIGURE VI

Aggregate Input Specificity in the United States

This figure is based on the computation of Nunn (2007) The author usesthe US Input-Output Use Table to identify which intermediate inputs are usedand in what proportions in the production of each final good Then using datafrom Rauch (1999) inputs are sorted into those sold on an organized exchangethose that are reference priced in a trade publication and those that aredifferentiated


at Universita C

omm

erciale Luigi B



ownloaded from

or management turnover36 In addition the results presented inthis article also extend to the specificity of inputs within theboundaries of the firm While the customer-supplier links allowus to pin down the nature of the input we would expect similarresults to be obtained within a firm when the division producinga specific part of the final good is hit by a shock Yet the extrap-olation of the results should take into account that firms endog-enously select their location and the location of their suppliersThis does not threaten our identification strategy and should biasthe results against finding any propagation effects In fact weshow in Table A15 in the Online Appendix that propagationtends to be weaker when disasters hit areas that are frequentlyhit in our sample Although this is a nice confirmation that pro-duction networks react more strongly to shocks that are lesslikely to be anticipated it also suggests that one should be cau-tious in extrapolating our findings to estimate the impact oflarger shocks if firms devote more resources to shelter them-selves against those than against natural disasters

V Conclusion

This article explores whether firm-level shocks propagate inproduction networks Using supplier-customer links reported byUS publicly listed firms we find that customers of suppliers hitby a natural disaster experience a drop of 2ndash3 percentage pointsin sales growth following the event which amounts to a 25 dropwith respect to the sample average Given the relative size ofsuppliers and customers in our sample this estimate is strikinglylarge The effect is temporary shows no prior trends and is onlyobserved when the relationship between customers and suppliersis active It is significantly stronger when the affected supplierproduces differentiated goods has a high level of RampD or ownspatents and is thus plausibly more difficult to replace Saleslosses translate into significant value losses to the order 1 ofmarket equity value Finally the effect spills over to other sup-pliers who also experience a drop in sales growth following thedisaster

36 For narrative examples of the role of strikes at the largest US firms inexplaining GDP fluctuations see Gabaix (2011)


at Universita C

omm

erciale Luigi B



ownloaded from

We provide evidence that on average specific input disrup-tions do not seem to be compensated and translate into sector-wide output losses Given that a large share of firmsrsquo inputs in theUnited States are specific the amplification mechanism that wedescribe is likely to be pervasive Taken together these findingssuggest that input specificity is a key determinant of the propa-gation of idiosyncratic shocks in the economy

Massachusetts Institute of Technology and CEPR

Bocconi University and IGIER

Supplementary Material

An Online Appendix for this article can be found at QJEonline (qjeoxfordjournalorg)

References

Acemoglu Daron Vasco M Carvalho Asuman Ozdaglar and Alireza Tahbaz-Salehi lsquolsquoThe Network Origins of Aggregate Fluctuationsrsquorsquo Econometrica 80(2012) 1977ndash2016

Ahern Kenneth R lsquolsquoNetwork Centrality and the Cross Section of Stock ReturnsrsquorsquoUSC Marshall Working Paper 2012

Ahern Kenneth R and Jarrad Harford lsquolsquoThe Importance of Industry Links inMerger Wavesrsquorsquo Journal of Finance 69 (2014) 527ndash576

Amiti Mary and David E Weinstein lsquolsquoHow Much Do Bank Shocks AffectInvestment Evidence from Matched Bank-Firm Loan Datarsquorsquo NBERWorking Paper 2013

Antras Pol Teresa C Fort and Felix Tintelnot lsquolsquoThe Margins of Global SourcingTheory and Evidence from US Firmsrsquorsquo Working Paper 2014

Atalay Enghin lsquolsquoHow Important Are Sectoral Shocksrsquorsquo Working Paper 2013Atalay Enghin Ali Hortacsu James Roberts and Chad Syverson lsquolsquoNetwork

Structure of Productionrsquorsquo Proceedings of the National Academy of Sciences108 (2011) 5199ndash5202

Bak Per Kan Chen Jose A Scheinkman and Michael Woodford lsquolsquoAggregateFluctuations from Independent Sectoral Shocks Self-Organized Criticalityin a Model of Production and Inventory Dynamicsrsquorsquo Ricerche Economiche 47(1993) 3ndash30

Baker Scott R and Nicholas Bloom lsquolsquoDoes Uncertainty Reduce Growth UsingDisasters as Natural Experimentsrsquorsquo NBER Working Paper 2013

Banerjee Shantanu Sudipto Dasgupta and Yungsan Kim lsquolsquoBuyer-SupplierRelationships and the Stakeholder Theory of Capital Structurersquorsquo Journal ofFinance 63 (2008) 2507ndash2552

Baqaee David R lsquolsquoCascading Failures in Production Networksrsquorsquo Working Paper2015

Belasen Ariel R and Solomon W Polachek lsquolsquoHow Hurricanes Affect Wages andEmployment in Local Labor Marketsrsquorsquo American Economic Review 98 (2008)49ndash53

Bernard Andrew B Andreas Moxnes and Yukiko U Saito lsquolsquoProductionNetworks Geography and Firm Performancersquorsquo CEPR Working Paper 2014

Bernile Gennaro George M Korniotis and Alok Kumar lsquolsquoGeography of Firmsand Propagation of Local Economic Shocksrsquorsquo Working Paper 2013

Bigio Saki and Jennifer LarsquoO lsquolsquoFinancial Frictions in Production NetworksrsquorsquoWorking Paper 2013


at Universita C

omm

erciale Luigi B



ownloaded from

Boehm Christoph E Aaron Flaaen and Nitya Pandalai-Nayar lsquolsquoInput Linkagesand the Transmission of Shocks Firm-Level Evidence from the 2011 TohokuEarthquakersquorsquo Working Paper 2015

Boehmer Ekkehart Jim Masumeci and Annette B Poulsen lsquolsquoEvent-StudyMethodology under Conditions of Event-Induced Variancersquorsquo Journal ofFinancial Economics 30 (1991) 253ndash272

Boone Audra L and Vladimir I Ivanov lsquolsquoBankruptcy Spillover Effects onStrategic Alliance Partnersrsquorsquo Journal of Financial Economics 103 (2012)551ndash569

Caliendo Lorenzo Fernando Parro Esteban Rossi-Hansberg and Pierre-Daniel Sarte lsquolsquoThe Impact of Regional and Sectoral Productivity Changeson the US Economyrsquorsquo NBER Working Paper 2014

Campello Murillo and Zsuzsanna Fluck lsquolsquoProduct Market PerformanceSwitching Costs and Liquidation Values The Real Effects of FinancialLeveragersquorsquo AFA 2007 Chicago Meetings Paper 2007

Carvalho Vasco lsquolsquoAggregate Fluctuations and the Network Structure ofIntersectoral Tradersquorsquo Working Paper 2010

Carvalho Vasco and Xavier Gabaix lsquolsquoThe Great Diversification and its UndoingrsquorsquoAmerican Economic Review 103 (2013) 1697ndash1727

Carvalho Vasco Makoto Nirei and Yukiko U Saito lsquolsquoSupply Chain DisruptionsEvidence from Great East Japan Earthquakersquorsquo Working Paper 2014

Carvalho Vasco and Nico Voigtlander lsquolsquoInput Diffusion and the Evolution ofProduction Networksrsquorsquo NBER Working Paper 2014

Caselli Francesco Miklos Koren Milan Lisicky and Silvana TenreyrolsquolsquoDiversification through Tradersquorsquo Working Paper 2011

Chaney Thomas lsquolsquoThe Network Structure of International Tradersquorsquo AmericanEconomic Review 104 (2014) 3600ndash3634

Chodorow-Reich Gabriel lsquolsquoThe Employment Effects of Credit MarketDisruptions Firm-Level Evidence from the 2008ndash9 Financial CrisisrsquorsquoQuarterly Journal of Economics 129 (2014) 1ndash59

Chu Yongqiang lsquolsquoOptimal Capital Structure Bargaining and the SupplierMarket Structurersquorsquo Journal of Financial Economics 106 (2012) 411ndash426

Cohen Lauren and Andrea Frazzini lsquolsquoEconomic Links and Predictable ReturnsrsquorsquoJournal of Finance 63 (2008) 1977ndash2011

Conley Timothy G and Bill Dupor lsquolsquoA Spatial Analysis of SectoralComplementarityrsquorsquo Journal of Political Economy 111 (2003) 311ndash352

Costinot Arnaud Jonathan Vogel and Su Wang lsquolsquoAn Elementary Theory ofGlobal Supply Chainsrsquorsquo Review of Economic Studies 80 (2013) 109ndash144

Dessaint Olivier and Adrien Matray lsquolsquoDo Managers Overreact to Salient RisksEvidence from Hurricane Strikesrsquorsquo Working Paper 2013

Di Giovanni Julian and Andrei A Levchenko lsquolsquoPutting the Parts TogetherTrade Vertical Linkages and Business Cycle Comovementrsquorsquo AmericanEconomic Journal Macroeconomics 2 (2010) 95ndash124

Di Giovanni Julian Andrei A Levchenko and Isabelle Mejean lsquolsquoFirmsDestinations and Aggregate Fluctuationsrsquorsquo Econometrica 82 (2014) 1303ndash1340

Durlauf Steven N lsquolsquoNonergodic Economic Growthrsquorsquo Review of Economic Studies60 (1993) 349ndash366

Fama Eugene F and Kenneth R French lsquolsquoIndustry Costs of Equityrsquorsquo Journal ofFinancial Economics 43 (1997) 153ndash193

Fee Edward C Charles J Hadlock and Shawn Thomas lsquolsquoCorporate EquityOwnership and the Governance of Product Market Relationshipsrsquorsquo Journalof Finance 61 (2006) 1217ndash1251

Fee Edward C and Shawn Thomas lsquolsquoSources of Gains in Horizontal MergersEvidence from Customer Supplier and Rival Firmsrsquorsquo Journal of FinancialEconomics 74 (2004) 423ndash460

Feenstra Robert C lsquolsquoUS Imports 1972ndash1994 Data and Concordancesrsquorsquo NBERWorking Paper 1996

Fernando Chitru S Anthony D May and William L Megginson lsquolsquoThe Value ofInvestment Banking Relationships Evidence from the Collapse of LehmanBrothersrsquorsquo Journal of Finance 67 (2012) 235ndash270


at Universita C

omm

erciale Luigi B



ownloaded from

Foerster Andrew T Pierre-Daniel G Sarte and Mark W Watson lsquolsquoSectoralversus Aggregate Shocks A Structural Factor Analysis of IndustrialProductionrsquorsquo Journal of Political Economy 119 (2011) 1ndash38

Gabaix Xavier lsquolsquoThe Granular Origins of Aggregate Fluctuationsrsquorsquo Econometrica79 (2011) 733ndash772

Grossman Gene M and Esteban Rossi-Hansberg lsquolsquoTrading Tasks A SimpleTheory of Offshoringrsquorsquo American Economic Review 98 (2008) 1978ndash1997

Hertzel Michael G Zhi Li Micah S Officer and Kimberly J Rodgers lsquolsquoInter-Firm Linkages and the Wealth Effects of Financial Distress along the SupplyChainrsquorsquo Journal of Financial Economics 87 (2008) 374ndash387

Hines Paul Jay Apt and Sarosh Talukdar lsquolsquoTrends in the History of Large Black-Outs in the United Statesrsquorsquo Power and Energy Society General Meeting-Conversion and Delivery of Electrical Energy in the 21st Century 2008IEEE 2008

Horvath Michael lsquolsquoCyclicality and Sectoral Linkages Aggregate Fluctuationsfrom Independent Sectoral Shocksrsquorsquo Review of Economic Dynamics 1(1998) 781ndash808

mdashmdashmdash lsquolsquoSectoral Shocks and Aggregate Fluctuationsrsquorsquo Journal of MonetaryEconomics 45 (2000) 69ndash106

Hubbard Glenn R Kenneth N Kuttner and Darius N Palia lsquolsquoAre There BankEffects in Borrowersrsquo Costs of Funds Evidence from a Matched Sample ofBorrowers and Banksrsquorsquo Journal of Business 75 (2002) 559ndash581

Imberman Scott A Adriana D Kugler and Bruce I Sacerdote lsquolsquoKatrinarsquosChildren Evidence on the Structure of Peer Effects from HurricaneEvacueesrsquorsquo American Economic Review 102 (2012) 2048ndash2082

Jones Charles I lsquolsquoIntermediate Goods and Weak Links in the Theory of EconomicDevelopmentrsquorsquo American Economic Journal Macroeconomics 3 (2011) 1ndash28

Jovanovic Boyan lsquolsquoMicro Shocks and Aggregate Riskrsquorsquo Quarterly Journal ofEconomics 102 (1987) 395ndash409

Kale Jayant R and Husayn Shahrur lsquolsquoCorporate Capital Structure and theCharacteristics of Suppliers and Customersrsquorsquo Journal of FinancialEconomics 83 (2007) 321ndash365

Kee Hiau Looi lsquolsquoLocal Intermediate Inputs and the Shared Supplier Spillovers ofForeign Direct Investmentrsquorsquo Journal of Development Economics 112 (2015) 56ndash71

Kelly Bryan Lustig Hanno and Stijn Van Nieuwerburgh lsquolsquoFirm Volatility inGranular Networksrsquorsquo Working Paper 2013

Khwaja Asim Ijaz and Atif Mian lsquolsquoTracing the Impact of Bank Liquidity ShocksEvidence from an Emerging Marketrsquorsquo American Economic Review 98 (2008)1413ndash1442

Kogan Leonid Dimitris Papanikolaou Amit Seru and Noah StoffmanlsquolsquoTechnological Innovation Resource Allocation and Growthrsquorsquo NBERWorking Paper 2012

Kolari James W and Seppo Pynnonen lsquolsquoEvent Study Testing with Cross-Sectional Correlation of Abnormal Returnsrsquorsquo Review of Financial Studies23 (2010) 3996ndash4025

Long John B and Charles I Plosser lsquolsquoReal Business Cyclesrsquorsquo Journal of PoliticalEconomy 91 (1983) 39ndash69

mdashmdashmdash lsquolsquoSectoral vs Aggregate Shocks in the Business Cyclersquorsquo AmericanEconomic Review 77 (1987) 333ndash336

Lucas Robert E lsquolsquoUnderstanding Business Cyclesrsquorsquo Carnegie-RochesterConference Series on Public Policy 5 (1977) 7ndash29

MacKay Peter and Gordon M Phillips lsquolsquoHow Does Industry Affect FirmFinancial Structurersquorsquo Review of Financial Studies 18 (2005) 1433ndash1466

Menzly Lior and Oguzhan Ozbas lsquolsquoMarket Segmentation and Cross-Predictability of Returnsrsquorsquo Journal of Finance 65 (2010) 1555ndash1580

Moon Katie S and Gordon Phillips lsquolsquoOutside Purchase Contracts and FirmCapital Structurersquorsquo Working Paper 2014

Nunn Nathan lsquolsquoRelationship-Specificity Incomplete Contracts and the Patternof Tradersquorsquo Quarterly Journal of Economics 122 (2007) 569ndash600

Oberfield Ezra lsquolsquoBusiness Networks Production Chains and Productivity ATheory of Input-Output Architecturersquorsquo Working Paper 2013


at Universita C

omm

erciale Luigi B



ownloaded from

Rauch James E lsquolsquoNetworks versus Markets in International Tradersquorsquo Journal ofInternational Economics 48 (1999) 7ndash35

Slovin Myron B Marie E Sushka and John A Polonchek lsquolsquoThe Value of BankDurability Borrowers as Bank Stakeholdersrsquorsquo Journal of Finance 48 (1993)247ndash266

Titman Sheridan lsquolsquoThe Effect of Capital Structure on a Firmrsquos LiquidationDecisionrsquorsquo Journal of Financial Economics 13 (1984) 137ndash151

Titman Sheridan and Roberto Wessels lsquolsquoThe Determinants of Capital StructureChoicersquorsquo Journal of Finance 43 (1988) 1ndash19

Todo Yasuyuki Kentaro Nakajima and Petr Matous lsquolsquoHow Do Supply ChainNetworks Affect the Resilience of Firms to Natural Disasters Evidencefrom the Great East Japan Earthquakersquorsquo Journal of Regional Science 55(2014) 209ndash229


at Universita C

omm

erciale Luigi B



ownloaded from

INPUT SPECIFICITY AND THE PROPAGATION OFIDIOSYNCRATIC SHOCKS IN PRODUCTION NETWORKS

Jean-Noel Barrot and Julien Sauvagnat

This article examines whether firm-level idiosyncratic shocks propagate inproduction networks We identify idiosyncratic shocks with the occurrence ofnatural disasters We find that affected suppliers impose substantial outputlosses on their customers especially when they produce specific inputs Theseoutput losses translate into significant market value losses and they spill overto other suppliers Our point estimates are economically large suggesting thatinput specificity is an important determinant of the propagation of idiosyncraticshocks in the economy JEL Codes L14 E23 E32

I Introduction

The origin of business cycle fluctuations is a long-standingquestion in economics Starting with Long and Plosser (1983) anumber of studies have explored whether sectoral linkages mayhelp explain the aggregation of sector-specific shocks and havefound mixed empirical evidence of the importance of such link-ages Relative to the measurement of spillovers across sectorsspillovers within networks of firms have received little attentionin the empirical literature The main reason for this is the diffi-culty of identifying firm-specific shocks Whether firm-level idio-syncratic shocks propagate in production networks thereforeremains an open question

We thank Robert Barro the editor three anonymous referees Nittai BergmanBruno Biais Murillo Campello Thomas Chaney Philippe Chone Arnaud CostinotPierre Dubois Xavier Gabaix Nicola Gennaioli Christian Hellwig JohanHombert Augustin Landier Erik Loualiche Kambiz Mohkam SebastienPouget Ezra Oberfield Matthew Rognlie Ronan le Saout Fabiano SchivardiAntoinette Schoar and David Thesmar and participants at the MIT SloanFinance Retreat CREST internal seminar IDC Summer conference EARIE con-ference ASSAAFA meetings Stockholm School of Economics Oxford SaıdBocconi University EIEF CEMFI Copenhagen Business School WashingtonUniversity in St Louis UNC Kenan-Flagler York University and BostonUniversity for their valuable comments and suggestions Julien Sauvagnat grate-fully acknowledges financial support from the Agence Nationale de la RecherchemdashInvestissements dAvenir (ANR-11-IDEX-0003Labex EcodecANR-11-LABX-0047)

The Author(s) 2016 Published by Oxford University Press on behalf of Presidentand Fellows of Harvard College All rights reserved For Permissions please emailjournalspermissionsoupcomThe Quarterly Journal of Economics (2016) 1ndash50 doi101093qjeqjw018

1

The Quarterly Journal of Economics Advance Access published July 11 2016 at U

niversita Com

merciale L

uigi Bocconi on Septem

ber 20 2016httpqjeoxfordjournalsorg

Dow

nloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from



FIGURE I




at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

III Data










at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from









at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

LIS

TO

FM

AJO

RD

ISA

ST

ER

S

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Mou

nt

St

Hel

ens

eru

pti

onM

ay

1980

200

3W

A

Hu

rric

an

eA

lici

aA

ugu

st1983

139

47

2T

XH

urr

ican

eE

len

aA

ugu

st1985

32

05

4A

L

FL

L

A

MS

Hu

rric

an

eJu

an

Oct

ober

1985

66

35

8A

L

FL

L

A

MS

T

XH

urr

ican

eH

ugo

Sep

tem

ber

1989

71

14

3N

C

SC

V

AL

oma

eart

hqu

ak

eO

ctob

er1989

825

6C

AH

urr

ican

eB

obA

ugu

st1991

54

70

6M

A

ME

N

C

NH

N

Y

RI

Oak

flan

dH

ills

fire

stor

mO

ctob

er1991

105

4C

AH

urr

ican

eA

nd

rew

Au

gu

st1992

51

26

7A

L

FL

L

A

MS

Hu

rric

an

eIn

iki

Sep

tem

ber

1992

100

2H

IB

lizz

ard

Marc

h1993

221

111

5A

L

CT

F

L

GA

M

A

MD

N

J

OH

S

C

VA

V

TN

orth

rid

ge

eart

hqu

ak

eJan

uary

1994

136

9C

AH

urr

ican

eA

lber

toJu

ly1994

41

06

6A

L

FL

G

AH

urr

ican

eO

pal

Oct

ober

1995

186

64

3A

L

FL

G

A

LA

M

S

NC

S

CB

lizz

ard

Jan

uary

1996

319

145

7C

T

DE

IN

K

Y

MA

M

D

NC

N

J

NY

P

A

VA

W

VH

urr

ican

eF

ran

Sep

tem

ber

1996

100

20

2N

C

SC

V

A

WV

Ice

stor

mJan

uary

1998

43

10

9M

E

NH

N

Y

VT

Hu

rric

an

eB

onn

ieA

ugu

st1998

43

12

6N

C

VA

Hu

rric

an

eG

eorg

esS

epte

mber

1998

78

36

8A

L

FL

L

A

MS

Hu

rric

an

eF

loyd

Sep

tem

ber

1999

226

156

8C

T

DC

D

E

FL

M

D

ME

N

C

NH

N

J

NY

P

A

SC

V

A

VT

Hu

rric

an

eA

llis

onJu

ne

2001

77

45

6A

L

FL

G

A

LA

M

S

PA

T

XH

urr

ican

eIs

abel

Sep

tem

ber

2003

89

49

9D

E

MD

N

C

NJ

NY

P

A

RI

VA

V

T

WV


at Universita C

omm

erciale Luigi B



ownloaded from

TA

BL

EI

(CO

NT

INU

ED)

Dis

ast

erD

ate

C

oun

ties

US

E

mp

loym

ent

Aff

ecte

d(

)L

ocati

on

Sou

ther

nC

ali

forn

iaw

ild

fire

sO

ctob

er2003

317

8C

A

Hu

rric

an

eC

harl

eyA

ugu

st2004

67

39

4F

L

GA

N

C

SC

Hu

rric

an

eF

ran

ces

Sep

tem

ber

2004

311

124

7A

L

FL

G

A

KY

M

D

NC

N

Y

OH

P

A

SC

V

A

WV

Hu

rric

an

eIv

an

Sep

tem

ber

2004

284

73

1A

L

FL

G

A

KY

L

A

MA

M

D

MS

N

C

NH

N

J

NY

P

A

SC

T

N

WV

Hu

rric

an

eJea

nn

eS

epte

mber

2004

160

88

DE

F

L

GA

M

D

NC

N

J

PA

S

C

VA

Hu

rric

an

eD

enn

isJu

ly2005

200

53

8A

L

FL

G

A

MS

N

CH

urr

ican

eK

atr

ina

Au

gu

st2005

288

92

1A

L

AR

F

L

GA

IN

K

Y

LA

M

IM

S

OH

T

NH

urr

ican

eR

ita

Sep

tem

ber

2005

123

37

5A

L

AR

F

L

LA

M

SH

urr

ican

eW

ilm

aO

ctob

er2005

24

35

5F

LM

idw

est

floo

ds

Ju

ne

2008

216

52

5IA

IL

IN

M

N

MO

N

E

WI

Hu

rric

an

eG

ust

av

Sep

tem

ber

2008

98

17

9A

R

LA

M

SH

urr

ican

eIk

eS

epte

mber

2008

163

41

1A

R

LA

M

O

TN

T

XB

lizz

ard

G

rou

nd

hog

Day

Feb

ruary

2011

210

146

3C

T

IA

IL

IN

KS

M

A

MO

N

J

NM

N

Y

OH

O

K

PA

T

X

WI

Hu

rric

an

eIr

ene

Au

gu

st2011

40

31

9C

T

MA

M

D

NC

N

J

NY

V

A

VT

Tro

pic

al

Sto

rmL

eeS

epte

mber

2011

110

52

3A

L

CT

G

A

LA

M

D

MS

N

J

NY

P

A

TN

V

AIs

aac

Au

gu

st2012

77

33

6F

L

LA

M

SH

urr

ican

eS

an

dy

Oct

ober

2012

274

220

8C

T

DE

M

A

MD

N

C

NH

N

J

NY

O

H

PA

R

IV

A

WV

Flo

odin

gan

dse

ver

ew

eath

er

Illi

noi

sA

pri

l2013

29

31

1IL

IN

M

O

Flo

odin

g

Col

orad

oS

epte

mber

2013

809

2C

O

Not

es

Th

ista

ble

des

crib

esth

e41

natu

ral

dis

ast

ers

incl

ud

edin

the

sam

ple

N

am

es

date

sn

um

ber

ofaff

ecte

dco

un

ties

an

dth

elo

cati

onof

each

natu

ral

dis

ast

erare

obta

ined

from

the

SH

EL

DU

Sd

ata

base

at

the

Un

iver

sity

ofS

outh

Caro

lin

a

Th

eli

stis

rest

rict

edto

even

tscl

ass

ified

as

Majo

rD

isast

ers

inS

HE

LD

US

w

ith

tota

ld

irec

tes

tim

ate

dd

am

ages

abov

e$1

bil

lion

2013

con

stan

td

olla

rsan

dla

stin

gle

ssth

an

30

days

Th

esh

are

ofto

tal

US

em

plo

ym

ent

aff

ecte

dby

each

natu

ral

dis

ast

eris

com

pu

ted

from

Cou

nty

Bu

sin

ess

Patt

ern

data

pu

bli

cly

pro

vid

edby

the

US

C

ensu

sB

ure

au

T

he

sam

ple

per

iod

isfr

omJan

uary

1978

toD

ecem

ber

2013


at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE II



101 to 35011 to 1002 to 1010

FIGURE III




at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

TABLE II




supplier (t)80574 0014 0118 0000 0000 1000


Diff RampD Patent

S NS S NS S NS


7125 6692 6373 8335 7821 6618


1332 1210 1502 1214 1388 1219




Assets 32061 12656 20013 48513 3254 7099Age 32061 27822 16623 48513 19233 15680ROA 32061 0145 0091 48513 0118 0128



supplier (t)139976 0042 0200 0000 0000 1000


county102279 0597 0365 0000 0667 1000



at Universita C

omm

erciale Luigi B



ownloaded from




IV Results





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from




TABLE III















at Universita C

omm

erciale Luigi B



ownloaded from



TABLE IV





0050 0044 0048 0048 0046 0041(0017) (0016) (0012) (0012) (0016) (0015)


0023 0013 0038 0044 0020 0011(0026) (0026) (0040) (0039) (0028) (0028)

Specific firm 0099 0090 0060 0030(0021) (0021) (0014) (0013)



Observations 139976 139976 139976 139976 139976 139976R2 0177 0192 0177 0192 0177 0192



at Universita C

omm

erciale Luigi B



ownloaded from


-08

3027

-02-05

-51

-38

-55

-37

-08-06

09

19

-13

04

11

-05

08

-12 -12 -11

-27-30

-20

-15

-05-02

-04

-6-4

-20

24

-6 -4 -2 0 2 4 6 8 10Quarters


FIGURE IV




frac144




frac144

HitsFirmit thornX9

frac144




at Universita C

omm

erciale Luigi B



ownloaded from


TABLE V



supplier (t 4)0031 0027 0029 0019(0009) (0008) (0008) (0008)






supplier (t 4)0031 0028 0029 0020(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VI





0013 0013 0011 0004(0008) (0009) (0009) (0009)


0013 0009 0010 0002(0009) (0009) (0010) (0010)


0028 0025 0025 0013(0009) (0009) (0009) (0009)


0031 0027 0030 0020(0009) (0009) (0009) (0009)


0016 0013 0014 0007(0010) (0010) (0010) (0010)












at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from

TABLE VII


Sales Growth (t 4t)




0031 0027 0030 0020(0009) (0009) (0009) (0008)







0003 0004 0004 0005(0007) (0007) (0007) (0007)


0033 0029 0032 0023(0010) (0010) (0010) (0010)







at Universita C

omm

erciale Luigi B



ownloaded from


TABLE VIII





0002 0002 0018 0011 0020 0016(0012) (0011) (0011) (0011) (0011) (0010)


0050 0043 0039 0032 0039 0034(0010) (0010) (0014) (0014) (0011) (0012)


0031 0029 0031 0029 0031 0029(0011) (0011) (0011) (0011) (0011) (0011)

Number ofsuppliers




Observations 80574 80574 80574 80574 80574 80574R2 0234 0262 0234 0261 0234 0262



at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from






CAAR frac14XT

tfrac14t0

eth1

N

XN

ifrac141

ARitTHORN




at Universita C

omm

erciale Luigi B



ownloaded from








at Universita C

omm

erciale Luigi B



ownloaded from





TABLE IX


CAAR



[101] 0487 0195 0176(1283) (0819) (2310)

[010] 0361 0548 0124(1911) (2215) (0302)

[1120] 0177 1452 0029(0269) (3340) (0205)

[2130] 0121 0385 0006(0583) (1088) (0392)

[3140] 0014 0120 0042(0215) (1123) (1208)

[1040] 1132 2459 0379(1982) (3029) (1563)



at Universita C

omm

erciale Luigi B



ownloaded from


FIGURE V





at Universita C

omm

erciale Luigi B



ownloaded from



TABLE X





[101] 0885 0064 0321 0556 0096 0694(1567) (0288) (0568) (1243) (0164) (1731)

[010] 0636 0018 0621 0253 0651 0208(2757) (0022) (1395) (1542) (2585) (0742)

[1120] 0168 0189 0181 0326 0177 0365(0608) (0250) (1482) (1021) (0479) (0674)

[2130] 0460 0348 1351 0391 0560 0112(1017) (0192) (2349) (0564) (0502) (0421)

[3140] 0219 0337 0340 0162 0391 0229(0331) (0006) (0896) (0213) (0381) (0038)

[1040] 2368 0578 2452 0582 1519 0926(2939) (0080) (1769) (1412) (1285) (1690)



at Universita C

omm

erciale Luigi B



ownloaded from


TABLE XI


Sales Growth (t 4t)




0002 0001 0001 0002(0021) (0021) (0021) (0021)


0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)



Observations 139976 139976 139976 139976R2 0192 0192 0192 0192



at Universita C

omm

erciale Luigi B



ownloaded from




IVE Discussion



at Universita C

omm

erciale Luigi B



ownloaded from

TABLE XII



Diff RampD Patent


workforce (t 4t 1)0007 0007 0008 0007(0016) (0016) (0016) (0016)




0038(0010)


0047 0048 0040(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0038(0010)


0047 0048 0039(0013) (0014) (0013)


0011 0013 0015(0013) (0013) (0013)

Observations 139976 139976 139976 139976R2 0192 0192 0192 0192






0032(0016)


0044 0041 0036(0014) (0015) (0013)


0008 0003 0008(0015) (0014) (0015)



at Universita C

omm

erciale Luigi B



ownloaded from



TABLE XII

CONTINUED


Diff RampD Patent





at Universita C

omm

erciale Luigi B



ownloaded from










at Universita C

omm

erciale Luigi B



ownloaded from






at Universita C

omm

erciale Luigi B



ownloaded from





at Universita C

omm

erciale Luigi B



ownloaded from



0

20

40

60

1982 1987 1992 1997



FIGURE VI




at Universita C

omm

erciale Luigi B



ownloaded from


V Conclusion




at Universita C

omm

erciale Luigi B



ownloaded from






References

















at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from



























at Universita C

omm

erciale Luigi B



ownloaded from







at Universita C

omm

erciale Luigi B



ownloaded from

Documents

Input Specificity and the Propagation of Idiosyncratic