journal of nvestment anagement joim - andrew loinvestigations of transportation accidents and...

JOIMwww.joim.com

Journal Of Investment Management, Vol. 9, No. 1, (2011), pp. 17–49

© JOIM 2011

THE NATIONAL TRANSPORTATION SAFETY BOARD: A MODELFOR SYSTEMIC RISK MANAGEMENT

Eric Fielding a, Andrew W. Lo b and Jian Helen Yang c

We propose the National Transportation Safety Board (NTSB) as a model organizationfor addressing systemic risk in industries and contexts other than transportation. Whenadopted by regulatory agencies and the transportation industry, the safety recommenda-tions of the NTSB have been remarkably effective in reducing the number of fatalities invarious modes of transportation since the NTSB’s inception in 1967 as an independentagency. The NTSB has no regulatory authority and is solely focused on conducting forensicinvestigations of transportation accidents and proposing safety recommendations. Withonly 400 full-time employees, the NTSB has a much larger network of experts drawn fromother government agencies and the private sector who are on call to assist in accidentinvestigations on an as-needed basis. By allowing the participation in its investigations ofall interested parties who can provide technical assistance to the investigations, the NTSBproduces definitive analyses of even the most complex accidents and provides actionablemeasures for reducing the chances of future accidents. It is possible to create more effi-cient and effective systemic-risk management processes in many other industries, includingfinancial services, by studying the organizational structure and functions of the NTSB.

aAssociate Managing Director for Strategic Management,National Transportation Safety Board, 490 L’Enfant Plaza,S. W., Washington, DC 20594, [email protected](email).bHarris & Harris Group Professor, MIT Sloan School ofManagement, 100 Main Street, E62-618, Cambridge, MA02142, USA, [email protected] (email); and Chief InvestmentStrategist, AlphaSimplex Group, LLC.cSloan Fellow, MIT Sloan School of Management, 50Memorial Drive, E52-126, Cambridge, MA 02142, USA,[email protected] (email).

1 Introduction

The National Transportation Safety Board(NTSB) is an independent federal agency witha mandate to investigate accidents and pro-mote safety in the transportation industry. Withits reputation for independence and objectivity,the NTSB is widely regarded as an authori-tative voice in transportation safety, and oneof the most admired agencies in the federal

First Quarter 2011 17

18 Eric Fielding et al.

government. Through its detailed accident inves-tigations, direct and unequivocal recommenda-tions for safety improvements, and plainspokenreal-time communication with the public at thestart of a major accident’s media barrage, theNTSB has earned the public’s trust and confidencein ways that few other government agencies canmatch. The objective of this study is to under-stand the key factors underlying the effectivenessof the NTSB so as to derive lessons that may beprofitably applied to other industries.

Of course, we acknowledge at the outset that cer-tain unique features of the transportation industrycontribute to the NTSB’s success. Transportationaccidents are almost always limited in scope andtime, which makes it possible to conduct in-depthforensic1 investigations that result in concreteconclusions and actionable remedies. Such acci-dents are also typically well-defined, with causesthat are usually identifiable upon detailed exami-nation, allowing the NTSB to be largely reactiveyet still highly effective. Finally, and perhapsmost significantly, no one benefits immediatelyfrom a transportation accident. Therefore, allstakeholders are united in their desire to improvesafety. As Jeff Marcus, an NTSB safety special-ist, put it, “You can trust people to be honestand moral about not killing themselves.” This lastpoint may seem obvious, but in other industries,certain parties may profit handsomely from crisesthat inflict enormous pain on others.

Despite these features of transportation, theNTSB’s enviable record of success deserves fur-ther study to determine which of its methods areapplicable to other technology-based industriessuch as financial services, healthcare, and energy,all industries in which “accidents” arise from thefailure of complex systems, i.e., systemic risk.This is the challenge we undertake in this article.

By examining the structure and functions ofthe NTSB, and studying a specific accident

investigation in detail, we observe five majorfactors that seem to characterize the agency’s suc-cess: (1) the governance structures that give riseto the agency’s impartiality and singular focus;(2) the investigative “Go Team” as a cohesiveunit; (3) the collective intelligence of the NTSB’s“party system”; (4) effective media relations;and (5) employee satisfaction. While some ofthe NTSB’s practices are indeed specific to thetransportation industry, we believe that the mostimportant drivers of its success can be adapted toother industries and contexts.

We begin in Section 2 by providing a brief organi-zational overview of the NTSB. In Section 3, wefocus on the most important aspect of the NTSB:its accident-investigation process. To illustratehow this process works, we present a case study ofthe NTSB’s investigation of the Minnesota I-35Whighway bridge collapse in Section 4. Based onthis example and other observations, we summa-rize the organizational factors that contribute themost to the success of the NTSB in Section 5.In Section 6, we consider some of the currentchallenges facing the NTSB, and we conclude inSection 7.

2 The NTSB organization

The NTSB can be traced back to the Air Com-merce Act of 1926, which eventually led to theestablishment of the Civil Aeronautics Board’sBureau of Safety in 1940. Beginning in 1967,the NTSB emerged as an independent agencywithin the Department of Transportation (DOT),and was later reestablished as a completely inde-pendent entity outside of DOT by Congressthrough the Independent Safety BoardAct of 1974(see Appendix A.1 for key facts and figures ofthe NTSB). A lean organization of about 400employees, the NTSB is charged with investi-gating every civil aviation accident and all sig-nificant highway, marine, railroad, pipeline, and

Journal Of Investment Management First Quarter 2011

The National Transportation Safety Board: A Model for Systemic Risk Management 19

hazardous-materials accidents. From these inves-tigations, the NTSB first identifies the cause ofthe accidents and, thereafter, develops safety rec-ommendations for preventing similar accidents inthe future.

Significantly, the NTSB has no regulatory author-ity; the Federal Aviation Administration regu-lates the airline industry, the National HighwayTraffic Safety Administration and the FederalMotor Carrier Safety Administration regulatemotor vehicle transportation, the United StatesCoast Guard regulates civil waterborne trans-portation, and the Pipeline and Hazardous Mate-rials Safety Administration regulates pipelinesand hazardous materials. The NTSB is primarilyresponsible for conducting investigations to deter-mine the causes of accidents and making safetyrecommendations.2 This lack of regulatory pow-ers may seem to be a disadvantage; after all, howcan an agency have impact without the authorityto carry out its recommended courses of action? Infact, the absence of rule-making responsibilitiespreserves the NTSB’s objectivity in its investi-gations, allowing it to identify regulatory gapsand failings as well as engineering flaws, and cantherefore issue the most objective recommenda-tions to improve safety. This important feature isdescribed explicitly in the NTSB’s strategic plan(NTSB Strategic Plan 2010–2015, pages 6, 7):

In 1974, Congress reestablished the NTSB as a completelyseparate entity, outside the DOT, reasoning that ‘… Nofederal agency can properly perform such (investigatory)functions unless it is totally separate and independent fromany other … agency of the United States.’ Because theDOT has broad operational and regulatory responsibilitiesthat affect the safety, adequacy, and efficiency of the trans-portation system, and transportation accidents may suggestdeficiencies in that system, the NTSB’s independence wasdeemed necessary for proper oversight.The NTSB, which has no authority to regulate, fund, or bedirectly involved in the operation of any mode of transporta-tion, conducts investigations and makes recommendationsfrom an objective viewpoint.

This paradox of less regulatory authority yield-ing greater influence is one of the most strikingcharacteristics of the NTSB, and in this sectionwe shall attempt to deconstruct and explicate themechanisms by which this small agency has beenable to achieve so much.

2.1 The NTSB board

The NTSB is governed by a five-member board,with each member nominated by the Presidentand confirmed by the Senate to serve a five-yearterm. One of the board members is designated bythe President as the Chairman and another as ViceChairman, each for a two-year term, with separateSenate confirmation required of the Chairman.The Chairman oversees the entire organizationas its “Chief Executive Officer.” The ManagingDirector, reporting directly to the Chairman, is, ineffect, the “Chief Operating Officer” of the orga-nization, running the day-to-day operations of theNTSB and providing ongoing management to thecivilian workforce (see Figure 1 for the NTSB’scurrent organizational chart).

The NTSB’s board members do not report tothe Chairman of the Board and, except for theChairman, they are not involved in the day-to-daymanagement of agency operations. Therefore,once nominated by the President and confirmedby the Senate, each board member is free to focussolely on the mission of transportation safety. Thisstructure maximizes the independence and objec-tivity of each board member, one of the mostimportant factors in producing impartial investi-gations and safety recommendations. Moreover,the NTSB staff does not report to the boardmembers (with the exception of the reporting rela-tionship to the Chairman through the ManagingDirector), which imbues the investigative staff—whose technical expertise and judgment form thevery foundation of the NTSB—with the samesense of impartiality and purpose.

First Quarter 2011 Journal Of Investment Management


Figure 1 Organizational chart of the NTSB. Source: NTSB website http://ntsb.gov/Abt_NTSB/orgchart/org.htm.

The primary focus of the board members isto promote transportation safety by determin-ing the probable causes of accidents and issu-ing safety recommendations, which take centerstage at board meetings where each board mem-ber holds equal voting power and the major-ity prevails. In addition, one board memberis always on the scene of each major acci-dent investigation and serves as the NTSB’sspokesperson for that investigation. For exam-ple, Chairman Hersman has been the “Memberon-scene” for 19 major investigations during hertenure at the NTSB (as of October 27, 2010; seehttp://ntsb.gov/Abt_NTSB/member.htm).

Given the responsibilities of the board, mem-bers must have the necessary expertise to renderinformed judgments in accident investigations;

hence, it is mandated that at least three out of thefive board members be technical experts. By thesestandards, all five current board members qualify:Chairman Hersman had extensive transportationlegislative experience prior to the NTSB; ViceChairman Hart is a pilot with a Master’s Degree inaerospace engineering; Dr. Rosekind is an inter-nationally recognized expert in human fatigue;Dr. Weener is a long-time chief engineer inaerospace engineering; and Mr. Sumwalt servedas a commercial airline pilot for 32 years (seeAppendix A.2 for biographies of all the currentboard members). Each of the current board mem-bers came to the NTSB with extensive experiencewith or in the transportation industry, and duringtheir tenure at the NTSB, they develop a compre-hensive understanding of all aspects and modes ofthe industry, which is evident from board-meetingdiscussions.



To further ensure the objectivity of the board,it is mandated that no more than three mem-bers belong to the same political party. Last,but not least, the members are appointed forfixed terms, as opposed to serving at the plea-sure of the President, so they cannot be removedfor political reasons or unpopular views anddecisions.

The purpose of these safeguards is to help assurethat NTSB decisions are made as much as possi-ble on the basis of the evidence, and as little aspossible on the basis of the politics.

The board meetings of the NTSB are whereaccident investigations are discussed and con-clusions and recommendations are finalized (seeAppendix A.3 for a summary of a typical boardmeeting). Staff members present their findingsand proposed recommendations in the form of anaccident report, and board members vote on theprobable cause of the accident as well as safetyrecommendations stemming from that accident.The independence of each board member and thestaff, and the culture of openness at the NTSB,can make these meetings intense events with agreat deal of discussion and debate, often accom-panied by arcane technical details gleaned fromthe investigation.

Under the Government in the Sunshine Act, allNTSB board meetings are open to the publicand, under current practice, are available viawebcast. Moreover, while board members areallowed to consult with each other individu-ally in private, three members legally constitutea quorum for a meeting, and a quorum canmeet to discuss official Board business onlyif the public notice and other process require-ments of the Sunshine Act have been satisfied.3

The opening and closing comments, as well aspresentations, are also available at an NTSB web-site (http://www.ntsb.gov/events/Boardmeeting.htm).

2.2 The party system and public hearings

Given the resource constraints of the agency, it isimpossible for NTSB investigators to know allthe details of every aircraft, vehicle, or trans-portation mode involved in an accident, all of theoperational policies and procedures of the carri-ers involved, all the nuances of air traffic control,and all of the applicable regulatory guidance. Toleverage its limited resources, the NTSB desig-nates other organizations and external parties toparticipate in its investigations,4 creating a muchmore nimble, intelligent, and dynamic organiza-tion. This “party system” has become integral toNTSB accident investigations as transportationsystems have become more complex, knowledgemore specialized, and organizations more inter-connected. The party system allows the NTSBto focus on its core competencies of managingcomplex investigations and developing theoriesfrom myriad facts and data sources, and is one ofthe primary sources of the NTSB’s reputation andpride.

This unique party system is “by invitation only”;as described by the NTSB (http://www.ntsb.gov/abt_ntsb/invest.htm):

Other than the FAA, which is by law automatically des-ignated a party, the NTSB has complete discretion overwhich organizations it designates as parties to the investi-gation. Only those organizations and corporations that canprovide expertise to the investigation are granted party sta-tus and only those persons who can provide the Board withneeded technical or specialized expertise are permitted toserve on the investigation; persons in legal or litigationpositions are not allowed to be assigned to the investiga-tion. All party members report to the NTSB during theinvestigation.

The parties are responsible for the accuracy ofthe facts summarized by the NTSB in its factualreports (called “group factual reports” in majoraccident investigations undertaken by multipleparties).



By using outside experts and interested partieswho can provide technical assistance to accidentinvestigations, the NTSB also creates a muchmore inclusive atmosphere for all stakeholdersinvolved. This cooperative approach to accidentinvestigation is largely responsible for transform-ing a naturally adversarial finger-pointing exer-cise into a collaborative effort to uncover the truth.Most parties are more than willing to participatesince it provides them with initially privileged andtimely access to information from the ongoinginvestigation, as well as an opportunity to providethe investigators with additional information thatmay affect the course of the investigation.5 How-ever, even parties who are reluctant to participatebecause of potential liabilities usually do so sincethe relevant information is likely to emerge even-tually (in particular, the NTSB has the authorityto issue subpoenas to obtain the information theyneed).

However, while parties are used in the fact-gathering phase of an investigation, they are notpermitted to participate or be involved in theformal analysis, which is performed by internalNTSB staff only and is not disclosed to the publicuntil the board meeting.6 This limitation allowsthe NTSB to maintain full control of the investi-gation despite the active involvement of externalparticipants. Parties are, however, allowed to sub-mit analyses—their “side of the story”—to theNTSB for consideration. Any analysis the partieschoose to offer is limited to a “submission” underNTSB rules: the submission must be served to theother parties involved in the investigation, and theresults are put in a public docket item that dis-plays the offerings made by the party under thislimited role in NTSB analysis. After this publicsubmission, the parties have no further role in theNTSB’s deliberations.

Another method used by the NTSB for gather-ing information, and for providing transparency

and accountability to the public and the mediawith respect to its investigation, is the publichearing. As described by the NTSB (http://www.ntsb.gov/abt_ntsb/invest.htm):

The Board may hold a public hearing as part of a majortransportation accident investigation. The purpose of thehearing is two-fold; first, to gather sworn testimony fromsubpoenaed witnesses on issues identified by the Board dur-ing the course of the investigation, and, second, to allow thepublic to observe the progress of the investigation. Hearingsare usually held within six months of an accident, but maybe delayed for complex investigations.

Hearings are focused on information-gatheringrather than fault-finding, hence they tend to behighly technical and less like legal or congres-sional hearings. During these proceedings, boardmembers, assisted by a Technical Panel consist-ing of the NTSB staff, will question the witnesses,who have been selected because of their ability toshed light on the accident under investigation.7

The NTSB has a preferred practice of releas-ing some factual information through open publicdockets before a board meeting is held to supportformal hearings or to provide the public greaterinformation about the ongoing investigation. Thispractice minimizes interference with or intrusionsinto the analytical process by external parties dueto discovery needs in litigation, public or con-gressional calls for information, or regulatoryoversight.

2.3 Conflicts of interest

While conflicts of interest inevitably arise andmust be addressed by the appropriate policiesand procedures, the integrity of the board mem-bers and NTSB employees is the ultimate gate-keeper of such conflicts. Each of the boardmembers and employees undergoes a rigorousethical review process and is subject to ongo-ing restrictions such as not having any financialinterests in transportation-related companies so



as to minimize actual and potential conflicts (seeAppendix A.4 for details on the ethical reviewprocess for board members and employees).

The NTSB pays special attention to the potentialconflict of interest when its employees and boardmembers come from or return to the private sector.For example, an NTSB employee who is newlyhired from Boeing will not be permitted to serveas a lead investigator on an accident involving aBoeing aircraft during the first year of her employ-ment (a “cooling off” period), and may never beallowed to participate in NTSB investigations ofaircraft components she was directly responsiblefor during her tenure at Boeing. Other restrictionsapply to departing NTSB staff. The general prin-ciples upon which the restrictions are based areoutlined in the “14 Principles of Ethical Conductfor Government Officers and Employees,” and theNTSB has a team of ethics officials to ensure com-pliance and address more complex conflicts on acase-by-case basis.

To further reduce the potential for conflictingagendas, stakeholders’ attorneys, and insurancecompanies are excluded from the NTSB partysystem so as to help keep the investigation pro-cess from becoming adversarial and to insulate theNTSB’s deliberations from the legal and financialconsiderations that might be associated with anaccident.

3 The investigative process

While the independence of the NTSB gives ita unique vantage point, the lack of regulatoryauthority over any transportation sector poses aninteresting challenge to the organization, and yetit has yielded some remarkable results. As notedin its strategic plan (NTSB Strategic Plan 2010–2015, page 7), “Because the NTSB has no formalauthority to regulate the transportation indus-try, the effectiveness of the NTSB depends onits reputation for conducting thorough, accurate,

and independent investigations and for produc-ing timely, well-considered recommendations.”In the absence of regulatory powers, the onlymeans for the NTSB to maintain its relevance isthe quality of its analysis, which is manifested inits accident investigations.

Although major transportation accidents are rareevents, they capture tremendous media attentionand public interest when they occur. If not man-aged properly, they can become fertile ground forspeculation, rumor, and panic, which explains theNTSB’s congressional mandate to investigate allaviation accidents as well as significant accidentsin other modes of transportation.8 Given the lim-ited resources of the NTSB, judgment must beexercised in deciding which accidents should beinvestigated, and which accidents are chosen isstill more of an art than a science. For all trans-portation modes except aviation, the agency usesformal risk criteria to ensure a number of issuesare considered before making a final determi-nation. Over the years, and partly in responseto recommendations from the U.S. GovernmentAccountability Office (GAO), the NTSB’s selec-tion process for its accident investigations havebecome more transparent, objective, and system-atic. The clarity and openness of the investigativeprocess are crucial not only to the mission of theNTSB, but also to its credibility.

3.1 The NTSB “Go Team”

Once an accident investigation is launched, anInvestigator-in-Charge (IIC)—typically a seniorinvestigator with considerable technical and man-agement experience—is appointed to lead andoversee the entire investigation. For larger acci-dents, the IIC will assemble a “Go Team” ofseveral forensic engineers who will accompanythe IIC to the accident scene as quickly as possi-ble. For example, the Go Team for the accidenton February 12, 2009 involving Colgan Air, Inc.



Figure 2 Colgan Air, Inc., operating as Continental Connection Flight 3407, crashed into a house near Buffalo,NY at 22:17 EST on February 12, 2009, causing a post-crash fire, and killing all 49 on board and 1 personin the house. The NTSB was notified immediately, and the NTSB Go Team, headed by IIC Lorenda Ward,arrived at the scene early the next morning. Former board member Steven Chealander was the spokespersonaccompanying the Go Team on the scene. See Appendix A.5 for details about the accident and the investigation.Source: NTSB Annual Report (2009, page 31).

operating as Continental Connection Flight 3407(Figure 2) was led by IIC Lorenda Ward, andstaffed with 14 investigators having expertise inthe areas of operations, human performance, airtraffic control, and meteorology among others.Each specialist in the Go Team heads a workinggroup or “subcommittee” in his or her domainof expertise, and each subcommittee is staffed bythe representatives of the parties involved, provid-ing an unusual level of employee empowerment(see Section 2.2 for details of the party system).In addition to these technical experts, other par-ties to the Colgan Air accident investigation werethe FAA, Colgan Air, the Air Line Pilots Associa-tion (ALPA), the National Air Traffic ControllersAssociation, and the United Steelworkers Union(flight attendants). Bombardier, the manufacturerof the aircraft, participated as the technical advi-sor to theAccredited Representative from Canada.

A public hearing for Colgan Air Flight 3407 wasalso held from May 12 to May 14, 2009.

IICs and investigators are assigned to Go Teamson a rotational basis, and during their duty rota-tion, they are on-call 24 hours a day with the toolsof their trade—flashlights, tape recorders, cam-eras, personal protective equipment, etc.—readyto be deployed at a moment’s notice. This rota-tional arrangement fosters cross-pollination andprevents information silos by giving investigatorsopportunities to work with different colleagues,through which they develop broad expertise anda wide network of valuable contacts.

3.2 Communicating with the media

Communicating with the media and the publicis a critical component of the NTSB’s responsi-bilities, especially at the early stages of a major



accident investigation. This is an often under-appreciated aspect of crisis management in whichrumors, fear, and panic are counteracted withhard facts and deep expertise. The NTSB activelyengages in informing the public through thefollowing mechanisms:

• Board Member Spokesperson. Upon launch-ing a major investigation, one of the membersof the board accompanies the Go Team andserves as the NTSB spokesperson, holdingmedia briefings at least once a day. The Mem-ber on-scene is supported by an NTSB publicaffairs officer who serves as full-time liaisonto the press throughout the duration of theon-scene accident investigation.

• Press Releases. The NTSB offers regularpress releases and email alerts to inform thosewho sign up of the latest happenings at theNTSB.

• Transparency. Thanks to the 1976 Govern-ment in the Sunshine Act, much of the Board’sbusiness is conducted in the public domain,with Board meetings and public hearings opento the public and available via webcast. A pub-lic docket is published during the investigationto support the factual findings and analyticalwork of the NTSB.

• Web-Based Archive. The NTSB maintains anextensive web-searchable public docket whichincludes all factual information from accidents,the analyses from party members, and the finalaccident reports. The NTSB also offers publicaccess to its extensive safety recommendationdatabase.

Only factual information is released during theon-scene investigation; conjectures, conclusions,and editorial remarks are strictly prohibited. Overtime, the NTSB has discovered—as any emer-gency room doctor can attest—that timely, clear,and honest communication of the facts has acalming effect on the general public, building

credibility and creating trust in the NTSB inves-tigation process.

3.3 The accident report and safetyrecommendations

The culmination of an NTSB Go Team’s acci-dent investigation is the accident report, a publicdocument that outlines the probable cause ofthe accident and lists specific safety recommen-dations based on the data gathered from theinvestigation and subsequent analysis conductedby internal NTSB staff investigators. Accidentreports make clear distinctions between facts andanalysis, and typically contain four major sec-tions starting with factual information, followedby analysis, then conclusions and, finally, recom-mendations. As described in Section 2.2, externalparties who may have been involved in the fact-gathering phase of the investigation are not partof analysis and report writing, but are allowed tosubmit their own analyses and proposals for prob-able causes and safety recommendations, whichbecome part of the public docket.

Significantly, while there are no restrictionsregarding the use of the facts that the NTSB placesin the public docket, the NTSB’s accident reportis inadmissible as evidence in lawsuits for civildamages, which prevents the NTSB from playingany judicial role and allows it to focus solely on itsprimary mandate of information-gathering, anal-ysis, and safety recommendations.9 Moreover,the NTSB’s regulations provide that its employ-ees cannot be subpoenaed to testify as witnesses insuits for civil damages, otherwise its staff wouldlikely be mired in endless legal entanglementsfrom criminal and civil proceedings associatedwith transportation accidents. However, it does, inaccordance with its rules, permit extremely lim-ited depositions on factual matters only in civillawsuits. The NTSB receives numerous requestsfor such depositions, which are typically held dur-ing the discovery phase of litigation and attended



by a court reporter and attorneys from both sides,during which factual information is reviewed, butno analysis is provided.10

Safety recommendations are, of course, the mostimportant product of the NTSB, and need not belimited to direct causes of the accident.11 Some-times the investigation reveals recurring issuesthat have been reflected in safety recommenda-tions from previous accident reports, and theBoard may reiterate certain recommendations tospeed up the adoption process. If, in the courseof an accident investigation, the NTSB discoverspotential safety concerns that warrant immediateattention, it will issue recommendations imme-diately regarding those potential concerns ratherthan awaiting the completion of the final accidentreport.12

A draft accident report is prepared by career(i.e., not politically appointed) NTSB investiga-tive staff, and before being submitted to theboard for approval, it must also be approvedby the NTSB’s Office of Research and Engi-neering (RE), a separate group that serves as ageneral resource for all the NTSB modal offices.RE provides technical support for investigations,conducts separate safety studies, and serves asa repository for broad-ranging expertise, espe-cially in the area of emerging safety issues. REstaff tend to be specialists, but with experience inall modes of transportation and with many foren-sic techniques, their expertise is applied generallyacross the NTSB. RE is considered the scientificcore of the NTSB, providing yet another layerof independence and analytic oversight for theorganization.

Once approved, the accident report is submittedto the five NTSB board members, and each boardmember can meet with staff to discuss the report.The draft report conclusions and recommenda-tions are finalized during a public board meetingwhere board members meet with NTSB staff to

resolve any outstanding issues, and this publicmeeting is the first time that the board reviewsand considers the report as a full board.

3.4 What the investigations reveal

The direct causes of transportation accidents areoften technical at first glance, e.g., birds strik-ing an airplane’s engines, unusual winds, or icingconditions. However, the invention and adop-tion of technology are uniquely human endeavors,and more often than not, human factors are theroot cause. Some accidents are due to limitationsin individual skills or organizational capacities,which is understandable given the complexity ofmodern transportation systems and the level ofcoordination required. One example of such lim-itations is the case of Continental Airlines Flight1404, which veered off the runway during take-offfrom Denver International Airport on December20, 2008, resulting in a post-crash fire that seri-ously injured the captain and five passengers. OnJuly 13, 2010, the NTSB issued a press releasefollowing its board meeting in which it reportedthe following conclusions:

The National Transportation Safety Board today determinedthat the probable cause of the … accident was the cap-tain’s cessation of rudder input, which was needed … whenthe plane encountered a strong and gusty crosswind thatexceeded the captain’s training and experience.Contributing to the accident was the air traffic control sys-tem that didn’t require or facilitate the dissemination ofkey available wind information to air traffic controllersand pilots, and inadequate crosswind training in the airlineindustry…

This example illustrates surprisingly basic defi-ciencies in individual skills and organizationalcapacity that pilots and air traffic controllerswere unaware of as recently as in 2008. Clearly,transportation safety is an ongoing challenge.

Accident investigations sometimes reveal chronicshortcomings in operations, either as a result



of “unintentional” mismanagement or intentionalshortcuts. For example, the Colgan Air accidentis the latest in a string of accidents involv-ing “code-sharing”, a marketing arrangement inwhich regional airlines operate flights for majorairlines.13 A growing trend in the industry, code-sharing is used by the major airlines as a way tocut costs. However, the safety implications arenot well understood and regulations have not yetcaught up with this practice. The NTSB hosteda two-day symposium, “Airline Code-SharingArrangements and Their Role in Aviation Safety,”in October 2010, taking a leadership role in bring-ing parties together to publicly discuss the issuessurrounding code-sharing and potentially shapethe industry and regulatory views of this practice.

4 Minnesota I-35W highway bridgecollapse: A case study

OnAugust 1, 2007, the eight-lane I-35W highwaybridge over the Mississippi River in Minneapoliscollapsed, falling 108 feet into the 15-foot-deepriver. This tragedy instantly captured nationalattention. Since the safety of bridges and otherpublic infrastructure is taken for granted, catas-trophic failures like this directly threaten thecredibility of the government, and can easily turninto widespread panic and public outrage. In thissection, we consider in some detail the NTSB’sresponse to this extraordinary and politicallycharged accident.

4.1 Taking the initiative

While the NTSB has primacy in the investiga-tion of all aviation accidents, it does not enjoythe same level of authority for other modes oftransportation such as highways, where state andlocal authorities are not even obligated to informthe NTSB that a major accident has occurred.In such cases, the NTSB must take the initia-tive in establishing itself as a participant in the

Figure 3 The Minnesota I-35W highway bridge col-lapsed onAugust 1, 2007 (photo credit unknown). TheNTSB Go Team was headed by Investigator-in-Charge(IIC) Mark Bagnard. Mark Rosenker, Chairman atthe time, was the spokesperson accompanying the GoTeam on the scene. Source: NTSB Annual Report(2008, page 54). See Appendix A.6 for details on theaccident and investigation.

accident investigation process. This was the casefor the Minneapolis I-35W highway bridge col-lapse, in which the NTSB defined its own roleamong all the other federal and local authoritiesinvolved, including the Minnesota Departmentof Transportation (MN DOT), the state police,and the FBI. In such circumstances, the NTSB isusually a welcome participant because its focuscomplements that of the police, who are pri-marily concerned with which parties are at faultand whether any laws were broken, while theNTSB’s interest is in promoting long-term safetyby determining the underlying causes of theaccident.

The NTSB’s Office of Highway Safety—a rel-atively small division with only 24 people—quickly assembled a Go Team across its regionaloffices with experts in structural engineering,bridge design, construction oversight, and sur-vival factors. Because of the inevitable public



outcry for bringing the responsible parties to jus-tice, Tim Pawlenty, then-governor of Minnesota,and the MN DOT hired the engineering firm,Wiss, Janney, Elstner Associates, within hours ofthe collapse to provide analysis that would paral-lel the NTSB investigation.14 As the investigationprogressed, the NTSB quickly established itselfas the authority, with the engineering firm col-laborating in its area of expertise on behalf of theMN DOT, a party to the investigation. Through theFederal HighwayAdministration, another party tothe investigation, the NTSB was able to locate andcollaborate with an engineer who wrote his Ph.D.thesis on possible failure scenarios for this partic-ular bridge, an example of the kind of expertisethe NTSB can leverage. Once again, the NTSB’sreputation for impartiality and technical expertisewas instrumental in establishing its legitimacy inthis accident investigation.

4.2 Time constraints

However, the NTSB’s lack of primacy did hamperits efforts significantly. For example, unlike avia-tion accident investigations, the NTSB was unableto dictate the pace of the investigation, hence itsnormal investigative process had to be adjustedto take into account the constraints imposed bythe other stakeholders, including law enforce-ment agencies and politicians. Participating inthis particular investigation were representativesfrom the Federal Highway Administration, theMN DOT, the state and local police, and thebridge maintenance contractor, Progressive Con-tractors, Inc. (PCI). The bridge design company,Jacobs Engineering, which had acquired the orig-inal designer of the bridge (Sverdrup & Parceland Associates), became a party to the investi-gation five months into the investigation as thefocus shifted to the bridge design. While theNTSB is well-known for its party system (seeSection 2.2), one consequence of the constraintsimposed by other parties in this case was the

NTSB’s decision to forgo a public hearing soas to make faster progress on the investigationitself.15 This decision drew heavy criticism fromCongressman James Oberstar, the representa-tive of Minnesota’s eighth Congressional district,and Chairman of the House Transportation andInfrastructure Committee (Veracifier, 2008).

Cooperation and competition among the par-ties can be both challenging and productive.For example, an interesting exchange occurredbetween two parties to the investigation: MNDOT and PCI, the contractor repaving the bridgeat the time of the collapse. Because construc-tion equipment and materials were stockpiledon the bridge at the time of the accident, thisexcess weight was a source of contention andone of the primary focuses of the investigationearly on. MN DOT claimed that PCI did not getapproval for stockpiling materials on the bridge,while PCI claimed that its requests for an addi-tional lane closure (which would have spread theload) and permission to stockpile its materialsat another location were denied (NTSB HAR-08/03, page 13). While each party might havechosen to reveal facts selectively at the start ofthe investigation, competition between these twoparties eventually led to a number of importantrevelations regarding the accident.

4.3 Faulty gusset plate

With technical support from Dr. Carl Schultheisz,a materials research engineer, and others in theNTSB’s Office of Research and Engineering, theinvestigation ultimately concluded that the prob-able cause of the accident was the concentratedconstruction weight placed directly on the weak-est link of the bridge, a faulty gusset plate (a 4′×8′metal plate used to connect and secure multiplebeams, see Figure 4). This plate was half as thickas it should have been, the result of an error in theoriginal design by Sverdrup & Parcel and Asso-ciates in 1961 that somehow made its way through



Figure 4 Gusset plate for typical five-member nodeon the Minnesota I-35W highway bridge. Source:NTSB/HAR-08/03 Report (2008, Figure 6).

the quality control process within the firm, andthe many rounds of design reviews by federal andstate transportation officials.

To make this conclusion irrefutable from all stake-holders’ perspectives, the NTSB team had to ruleout several other possible explanations, which isone of the main reasons for the length of NTSBinvestigations. According to the accident report(NTSB/HAR-08/03):

Before determining that the collapse of the I-35W bridgeinitiated with failure of the gusset plates at the U10 nodes,the Safety Board considered a number of potential explana-tions. The following factors were considered, but excluded,as being causal to the collapse: corrosion damage in gussetplates at the L11 nodes, fracture of a floor truss, preexistingcracking, temperature effects, and pier movement.

Some parties welcomed this conclusion, e.g.,URS Corporation, the bridge inspection companythat was sued by the families of the victims,16

while others were disappointed because a bridge-design flaw did not provide as much support fortheir efforts to raise taxes to fund bridge repairsand related infrastructure projects.17 However,the NTSB managed to rise above political andeconomic considerations and withstood pressures

from all sides. NTSB officials recognize that itsmost valuable asset is its reputation for objectivityand thoroughness, which can easily be damagedor destroyed if there were even an appearanceof bias, favoritism, or acquiescence to politicalpressure.

4.4 Unanswered questions

On September 18, 2008, only one year afterthe collapse, the replacement I-35W St. AnthonyFalls Bridge was opened to the public. While theNTSB report was conclusive, it also raised newquestions that have yet to be answered. Are thereother problems lurking in the gusset plates of thethousands of similar bridges across the UnitedStates? A key fact about the failure of the I-35Whighway bridge was the lack of load redundancyin its design; are the many other bridges withthe same single-point-of-failure designs also injeopardy of collapsing, and how do we deter-mine which are safe and which should be closedimmediately? Finally, the load on the I-35Whighway bridge increased by 20% during thetwo prior renovations, a significant change forwhich the implications for the collapse is stillunclear; how do we assess the impact of futureload changes on structural integrity and how oftenshould re-assessments be done?

As with all good scientific research, the investi-gatory process often raises more questions than itanswers, but all stakeholders agree that these areimportant questions that must be asked, even if theanswers may not be immediately forthcoming.

5 Organizational effectiveness

Having described and illustrated the basic orga-nizational structure and functions of the NTSBin Sections 2–4, in this section we attempt todistill the most important factors for the organiza-tion’s effectiveness. Our observations suggest at



least five distinct factors: impartiality and singu-lar focus, the Go Team, the collective intelligenceof the party system, effective media relations, andemployee satisfaction. We describe each of thesefactors in more detail in Sections 5.1–5.5.

5.1 Impartiality and singular focus

Given the overriding importance of impartialityand trust to the NTSB’s mission, maintainingindependence and managing real and perceivedconflicts of interest are critical to the organi-zation’s success. Perhaps the most significantfactor contributing to this impartiality is the inde-pendent and non-regulatory status conferred tothe NTSB by Congress through the Indepen-dent Safety Board Act of 1974, which renderedthe NTSB a separate entity from DOT. With theprimary focus on accident investigation, deter-mination of probable cause, and formulation ofsafety recommendations, the NTSB is relativelyfree of ongoing economic and political influ-ences, which, in turn, reduces its profile amongindustry lobbyists and special interest groups.Also, without regulatory authority, the NTSBhas no stake in past or current regulations andis free to identify, analyze, and critique regula-tory deficiencies and failures, and has done soon numerous occasions.18 Because the NTSB isnot expected to promote the transportation indus-try, its focus is narrower and less conflicted. Incontrast, regulators do have a vested interest inthe economic health of the industries they regu-late (by design), which can lead to conflicts andcompromises as they attempt to satisfy multipleconstituents.19

A closely related source of the NTSB’s inde-pendence is its subpoena power, without whichaccident investigations would be considerablyless informative and conclusive. The NTSB hasthe ability to subpoena practically any entity—whether it is a party participant in the investigation

or not—to obtain information vital to an accidentinvestigation.

The NTSB’s impartiality is, of course, a productof its independence and singular focus on acci-dent prevention and safety. However, one of themost important ways it preserves its objectivityis by separating the fact-gathering function of anaccident investigation from the analysis of theaccident. While there may be genuine differencesof opinion in analyzing an accident, all partiesshould be basing their analyses on the same set offacts, otherwise it may be impossible to reach anydefinitive conclusions. This separation is achievedin many ways (see Sections 2 and 3):

• Member On-Scene. During the daily mediabriefings at on-scene accident investigations,the Member on-scene provides only factualinformation, and will not speculate on possi-ble explanations as to why or how the accidentoccurred.

• Party System. The party system is used solelyduring the fact-gathering phase, not for theanalysis phase, which is done by internal NTSBstaff only, and is not disclosed to the public untilthe board meeting.

• Separation of Facts and Analysis. The acci-dent reports make clear distinctions betweenfacts and analysis.

Another contributing factor to the NTSB’s impar-tiality is the fact that five board members—oftenof different perspectives, political affiliations, andbackgrounds—must vote on the reports, includ-ing probable causes and safety recommendations,hence their integrity and judgment are at stakeeach time they file a report.

5.2 The Go Team as a cohesive unit

Making teams, not individuals, the cornerstoneof organizational design and performance is



an important feature of the NTSB’s relativelyflat organization (see the organizational chart inFigure 1). Small to medium size teams as cohesiveunits yield many advantages:

• Accountability. A clearly identifiable group ofindividuals is accountable for the investigation,while within the team, the responsibilities ofeach individual are clearly delineated.

• Collaboration. Because the team is account-able for the investigation, the members are “init together” and cooperation and camaraderiearise naturally.

• Empowerment. The Go Team is in some partmade up of leaders, with each team memberleading his or her own workgroup in a specificarea of expertise, and staffed by party mem-bers. Given that party members are often someof the brightest experts in their organizations,handpicked to participate in the investigation,each Go Team investigator heads an impressivegroup of high-caliber professionals.

• Sense of Accomplishment and Pride. Theobjective of an investigation is clearly defined,and so is “mission complete,” often a lengthyand challenging process of putting the piecestogether. As a result, the sense of accomplish-ment and pride is tremendous.

Moreover, having team assignments on a rota-tional basis fosters an atmosphere of fairness andequality, promoting knowledge sharing and col-laboration while de-emphasizing individual statusin the organization.

5.3 The collective intelligence of the partysystem

The NTSB’s party system is a unique approach toanalyzing complex systems that no single individ-ual or group has the expertise to do. At first blush,allowing external parties to participate in an acci-dent investigation—especially parties that may

ultimately be found to be responsible—seemsrife with bias and conflicted interests. However,the clean separation between an investigation’sfact-gathering and analysis phases is an inge-nious organizational structure to address theseconflicts. The party system yields several notableadvantages over other approaches to accidentinvestigation:

• Leverage. With about 400 employees in anorganization that investigated 13 major acci-dents, 10 international accidents, and 195other accidents in fiscal year 2009 (NTSB 2009Annual Report, page 4), the NTSB is able toaccomplish this feat by temporarily secondingexperts from other organizations for their GoTeams.

• Expertise. The party system allows the NTSBto draw on a much larger pool of expertise,which is essential to gathering the most rele-vant and accurate facts about a complex systemover a short-time frame.

• Influence. One of the best ways to expand theinfluence of an organization is to invite othersin. By working alongside NTSB investigators,parties become temporary insiders to an acci-dent investigation and are exposed to the NTSBculture and philosophy, and this perspective isbrought back to the party’s home organizationafter the investigation is completed.

• Competition. Each party is eager to clear itselffrom blame, and will compete with other par-ties to provide as much information and partyanalysis as possible to that end. This healthycompetition greatly benefits the NTSB’s effortsto gather facts quickly and thoroughly, but doesnot compromise the NTSB’s independence inconducting its own analysis of those facts.

• Adoption. It is a lengthy and often frustratingprocess to get safety recommendations adoptedby regulators and the industry. Party members,while not directly involved in developing safetyrecommendations, are intimately familiar with



the accident and the investigation (at least froma factual perspective). These individuals oftenserve as informal advocates for the NTSB’s rec-ommendations, and may be among the first toembrace the recommendations and their impli-cations for improved safety. An interestingdevelopment in recent years, particularly inlight of new communications technologies, isthe growing role of victims and victims’ fami-lies in organizing to advocate for certain safetyimprovements, oftentimes those recommendedby the NTSB.

The team of experts involved in any given NTSBaccident investigation, and the social value theycreate, are among the most compelling examplesof “collective intelligence” available.

5.4 Effective media relations

At the very start of each accident investigation,the NTSB establishes itself as the clearing housefor all information related to the accident, com-municating regularly with the press and televisionmedia. By taking such an active role in providingthe media with factual information, the NTSBreduces the likelihood of panic, the perceptionof self-serving statements by various parties, andthe development among the public of a mistrust inthe validity of the investigation and the reliabilityof the results. The importance of this single fea-ture of the NTSB’s accident investigation protocolcannot be overemphasized.

For industries that rely on public confidence, e.g.,banking and finance, insurance, public health,and transportation, widespread fear and hysteriacan have disastrous economic and social con-sequences that can take years to recover from.Misinformation, or simply the lack of informa-tion, is one of the primary causes of such herdbehavior. This is particularly true in recent yearsas media coverage has become ubiquitous andrelentless thanks to new technologies such as

Internet-based communication, text-messaging,and other forms of technology-leveraged socialnetworking. The NTSB has learned over timethat news agencies will file reports whether or notinformation is available, hence forthright and fre-quent communication with the media is the mosteffective way of reducing panic and ensuring anaccurate portrayal of an accident investigation.

5.5 Employee satisfaction

In a knowledge-based organization like theNTSB, the level of dedication and motivationof the employees directly affects the quality andquantity of its work product. While financialcompensation is an important element to anyworkforce, highly knowledgeable employees arelikely to put more emphasis on autonomy andbeing able to do meaningful and/or challengingwork. In fact, because of the flat and lean orga-nizational structure of the NTSB, managers oftenengage in hands-on work in addition to managingtheir team, and Board members provide clear rolemodels of top management being deeply engagedin accident investigations.

One of the predominant characteristics of theNTSB staff is the “mission driven” ethos—theytake great pride in the NTSB and appreciate theagency’s reputation in the transportation indus-try. For example, Erin Gormley, an aerospaceengineer in the NTSB’s Office of Research andEngineering, said, “When I travel in an airplane,I know what I have done to make it safer.” JuliePerrot, a safety recommendation specialist, said,“I know the woman whose hand you see onthe pedestrian signals. We recommended usinga hand to make the signal more intuitive.” Manyindividuals in the industry make it their careergoal to work for the NTSB; the average tenure ofemployees at the agency is 17 years.

In fact, according to the 2009 rankings of best-placestowork.org, the most comprehensive rating



of employee satisfaction in the federal govern-ment, the NTSB ranked 7th out of 32 underthe category “Best Places to Work in the Fed-eral Government” for small agencies. One of thecomponent criteria in this ranking is “employeeskills/mission match,” in which the NTSB isranked 3rd in the small agencies category. For pro-fessionals who have studied and worked hard formany years to build their skills, nothing appearsmore fulfilling than being able to use those skillsin a meaningful way.20

In short, the typical NTSB employee’s position—particularly for technical experts—is the capstoneto a successful career, not a stepping stone to abetter job elsewhere.

6 Current challenges facing the NTSB

Although the NTSB has enjoyed an enviablerecord of success in improving safety in thetransportation industry, its leadership has longrecognized a number of challenges that need to beaddressed. The Rand Corporation, a well-knownthink tank, conducted a comprehensive review ofthe NTSB in 2000 with the support of then NTSBChairman Jim Hall and Managing Director PeterGoltz, and highlighted a number of issues thatthe organization could improve upon, which theNTSB has sought to address since the report’spublication (see Sarsfield et al., 2000). In thissection, we highlight four current challenges tothe NTSB’s mission and continuing relevance.The sometimes lengthy delay between an acci-dent and the NTSB report, and the even longerdelays before certain safety recommendations areadopted, often frustrate politicians and the pub-lic who seek immediate and decisive action. Wediscuss these challenges in Sections 6.1 and 6.2,respectively. In Section 6.3, we consider thetension between the NTSB’s mandated role ofproducing safety recommendations reactively anda more proactive role of attempting to reduce the

likelihood of accidents yet to occur. Finally, aswith any government agency, the NTSB faces stiffcompetition from the private sector in recruitingand retaining new talent, which we describe inmore detail in Section 6.4.

6.1 Duration of investigations

Due to the complex nature of major accidentinvestigations and resource constraints, produc-ing timely accident reports has been an ongoingchallenge for the NTSB. The 2006 GovernmentAccountability Office (GAO) report indicated thatit routinely took longer than two years to com-plete major investigations (GAO-06-801T, coverpage). The NTSB has made significant progressin improving the process; in 2009, the averagetime for completing aviation investigations wasreduced to 13 months (NTSB FY2010 OperatingPlan, page 13). Table 1 summarizes key factorscontributing to these delays and how the NTSBhas addressed them.

The NTSB recognizes the importance of timelysafety recommendations. As a partial remedy,it may issue safety recommendations duringthe investigation, before the final report ispublished.21

6.2 Adoption of safety recommendations

Issuing safety recommendations is only the firststep of the NTSB’s accident investigation process;its mission is not complete until its recommenda-tions have been accepted and become standardpractice in the industry either voluntarily orthrough new regulation. For example, flight datarecorders or “black boxes”—now standard equip-ment for all commercial aircraft—started out assafety recommendations issued in the mid-1950sand early 1960s by the Civil Aeronautics Board(CAB), the predecessor to the NTSB. This kindof success does not come easily; the adoption



Table 1 Key factors, challenges, and NTSB responses in producing timely accident-investigation reports.

Key factor Challenges Actions

Investigation Scope The NTSB does not limit itself to thedirect cause of an accident tomaximize lessons learned. But thetemptation of making the mostcomprehensive and in-depth accidentreports often lead to “scope creep”.

The NTSB will evaluate and refine itscapability to establish appropriatescope of its investigative activities andaccident reports (NTSB FY2010Operating Plan, page 12)∗.

EliminatingNegatives

It is often not enough to prove theprobable causes. In order to convinceall parties, the NTSB has to take onadditional burden to rule out otherpossibilities†, which is resourceintensive.

This feature of accident reports cannotbe eliminated. While it may seem to bea waste of time, such thoroughness iswhat makes NTSB accidentinvestigations so irrefutably conclusivefor all stakeholders.

Human Resources The NTSB is a small agency;investigators are often pulled awayfrom ongoing investigations when newaccidents occur.

Process improvements allow the NTSBto partially compensate for the impactof staff reductions in recent years(NTSB FY2010 Operating Plan,page 13).

Report ApprovalProcess

Paper-based report production process islabor intensive.

The NTSB is in the process of deployingan electronic information system tostreamline this process.

∗One example is the practice of code-sharing, which was listed as a factor in the Colgan Air accident on February 12, 2009 (NTSBAAR-10/01) (see Section 3.4). The NTSB decided to refrain from extensive study of code-sharing during this investigation and, instead,hosted a symposium on this topic in October 2010 (see http://www.ntsb.gov/events/symp_code-sharing/agenda.htm).†For example, as part of the investigation of the Minnesota I-35W highway bridge collapse in 2007, the following alternate explanationshad to be ruled out before the NTSB concluded that an incorrectly designed gusset plate was responsible: corrosion damage in gussetplates at the L11 nodes, fracture of a floor truss, preexisting cracking, temperature effects, and pier movement. See Section 4.3 forfurther details.

process is usually a lengthy and often frustratingprocess.

When the NTSB issues a safety recommendationto, for example, the FAA, the FAA has a mandatefrom Congress to acknowledge it within 90 days.Afterward, each safety recommendation takes ona life of its own, and it is not unusual for the regu-lator to take a year or substantially longer to reportback with an expected adoption status.22 Unlikethe NTSB, whose sole focus is on safety, the reg-ulators have other mandates including cost andbenefit analysis as well as addressing feedbackfrom the industry.23

In addition, the introduction of any new regula-tions has to go through a lengthy formal process. Ifthe FAA decides to proceed with a new regulationafter its own due diligence, it will then send it tothe Office of the Secretary in the DOT for review,where it is subject to a cost-benefit analysis byits own economists as well as industry lobbying.Once it is approved by the DOT’s Office of theSecretary, the regulation is sent to the Office ofManagement and Budget, a White House agency,for review, where it is once again subjectto a rigorous cost-benefit analysis, and poten-tially, to the influence of industry and industrygroups.



The communications trail of a typical safetyrecommendation is described in Appendix A.7—nine-month turnaround times are not conducive toquick resolutions. From time to time, the NTSBmay request meetings with the FAA to discuss theprogress of high-priority items. However, manyof the non-priority items are often open for yearsat a time, waiting for status updates from the FAA.

Perhaps the clearest indication of the challengesin getting safety recommendations adopted isthe lack of satisfactory progress on the NTSB’s“Most Wanted” list of recommendations. Out of16 items listed in the NTSB 2009 Annual Report(pages 8 and 9), 10 are color-coded red, indicat-ing unacceptably slow responses. None of themis color-coded green, indicating timely progress.

6.3 Reactive accident investigations versusproactive safety studies

Most of the work at the NTSB is reactive in nature,responses to accidents as they occur. As a result,the vast majority of its safety recommendationsare lessons learned from actual accidents insteadof proactive measures for prevention. Moreover,since the NTSB cannot investigate every acci-dent due to resource constraints, it has developedaccident launch criteria to determine which acci-dents to investigate (largely related to the numberof fatalities or other risk factors). Technologi-cal innovations also pose great challenges for theNTSB because the staff has limited time to keeppace with these innovations, which is necessaryfor identifying emerging issues before they causeaccidents. To that end, the NTSB develops andpublishes an emerging issues list each year, andemerging issues are often the focus of hearings,forums, and other NTSB-sponsored symposia.

The NTSB has long realized the need for a lessreactive approach to identifying safety issues, andsafety studies have been used for this purpose, butonly on a limited scale. In fact, in 2006 the GAO

recommended increasing the use of safety studies(GAO 07-118, page 58), and the NTSB has madesignificant progress, but according to the 2009GAO report, there is more to be done on this front(GAO 10-183T, pages 12 and 13):

NTSB has also made significant progress in implementingour recommendation to increase its use of safety studies,which are multiyear efforts that result in recommenda-tions. … Although NTSB has not completed any safetystudies since we made our recommendation in 2006, it hasthree studies in progress, one of which is in final draft, andit has established a goal of developing two safety study pro-posals and submitting them to its board for approval eachyear. … NTSB officials told us they would like to broadenthe term ‘safety studies’ to include not only the currentstudies of multiple accidents, but the research done for theother smaller safety-related reports and data inquiries. Sucha term, they said, would better characterize the scope of theirefforts to report safety information to the public.

On March 9, 2010, the NTSB issued a safety studyon the glass cockpit (the electronic flight display)in light aircraft to determine how the transition tothis new technology has affected the safety of suchaircraft. Safety studies like these help to shape theindustry and the regulatory landscape in antici-pation of further technological innovation. Theexpanded use of safety studies has the potentialto further establish the leadership of the NTSB inthe transportation industry.

6.4 Challenges in human capital management

Like all other organizations, the NTSB faces chal-lenges in recruiting and retaining its work force.One of the biggest challenges—common to mostgovernment agencies—is competition from theprivate sector for the best talent. The compen-sation in the private sector tends to be higher forhighly trained professionals. The salaries for mid-career NTSB professionals in particular appearto lag behind typical aerospace industry salaries(Sarsfield et al., 2000, pages 142, 143).24 Figure 5shows the discrepancies for various levels ofexperience, which can be substantial in some



Source: “Despite Consolidation, Aerospace OffersAttractive Employment Opportunities and Salaries,”Aviation Week & Space Technology, February 8, 1999,page 83; Personal communications with NTSB admin-istrative officers, Spring 1999; U.S. Office of Person-nel Management, 1998.

Figure 5 Comparison of NTSB and AerospaceIndustry Salaries. Source: Sarsfield et al. (2000,Figure 5.4).

cases. However, this gap is not nearly as large asit may be in other industries such as financial ser-vices, in which mid-career and senior personnelcan earn multiples of a GS-15 government salary.

In addition, companies typically offer reloca-tion packages while the NTSB typically doesnot, which can be a major factor in recruit-ing new talent. Also, while a flat organizationalstructure may be efficient from a workflow per-spective, it provides limited opportunities forpromotion, making career advancement more dif-ficult. Finally, anticipated retirements may havea significant impact on the NTSB’s managementranks during the next few years ((NTSB StrategicPlan 2010–2015, page 37), making recruitmentand employee development even more importantin maintaining the continuity of leadership aswell as the level of technical and managementexpertise.

7 Conclusion

As inevitable as accidents seem to be, how oneresponds to them can lead to profound differencesin the likelihood of reoccurrence. Dismissingthem as rare and isolated events does little toprevent them from happening again, and point-ing fingers without any evidence or analysisonly creates more obstacles to genuine progress.Through the NTSB, the transportation industryhas developed a different approach to dealingwith accidents. The independence, objectivity,and credibility that this small but powerful agencyexemplifies are essential elements worth consid-ering in other industries that hope to improve theirsafety records.

While there may be certain aspects of the trans-portation industry that allow the NTSB to be soeffective, we believe that the practice of indepen-dently and systematically reviewing failures—sifting through the wreckage and analyzing everystep of an accident to determine its proximate andultimate causes—can benefit every technology-based industry. Accident investigations do occurin other industries already, e.g., mortality andmorbidity review committees in medicine, butfew industries have institutionalized this impor-tant aspect of process improvement as systemat-ically as the transportation industry, and in mostof the industries that have an investigative processfor mishaps, the investigation is undertaken by theregulator rather than by an independent entity.

A case in point is the financial services industry,which is as highly regulated as the transportationindustry but has no organization like the NTSB.One reason may be the trauma and public outrageassociated with transportation accidents involv-ing the deaths of innocent victims—the 50 liveslost in the Colgan Air accident of February 12,2009 create a sense of urgency that outweighsmost other priorities, including financial loss.A conclusive accident investigation yields safety



recommendations that may save future lives, butalso provides some form of closure to victims’families. While the recent financial crisis maynot be directly linked to loss of life, the conse-quences of systemic shocks to the financial systemmay be far broader, devastating in its own right,and on a much larger scale considering the sizeof the financial industry.25 Moreover, in somecases, the sudden loss of wealth and employmentmay force individuals into adverse lifestyle andhealthcare choices, which do have implicationsfor life expectancy and mortality rates. And tothe extent that a dollar value may be placed ona human life—which is an inescapable aspect ofmany public policy decisions—the most recentestimate of the cost of the government’s TroubledAsset Relief Program of $50 billion is equivalentto 7,761 lives lost.26 This figure does not includethe impact of unemployment, home foreclosures,lost retirement savings, and other consequencesof the financial crisis.

Of course, the transportation and financial indus-tries differ in a number of ways. Perhaps the mostsignificant difference is the definition of an “acci-dent,” which is clear in transportation but not infinance. For example, was the bursting of the“Internet Bubble” that began on March 13, 2000an “accident” deserving of investigation? Unlikean airplane crash—which no one would argue isa necessary and unavoidable consequence of airtravel—financial loss is indeed a necessary andunavoidable corollary of risk, financial innova-tion, and economic growth (“nothing ventured,nothing gained”). Therefore, the first order ofbusiness for a financial NTSB is to define thescope of its investigations. One possible start-ing point is to define a “systemic event” as aset of circumstances that has the potential to dis-rupt the stability or proper functioning of thefinancial system or any critical subcomponent.While this notion is still ambiguous when com-pared to a transportation accident, nevertheless,

it does serve the purpose of narrowing the inves-tigative purview of a financial NTSB to a moremanageable subset of incidents. Of course, somejudgment and discretion will need to be exercisedat the outset of such an organization’s activities,but over time and after a number of investiga-tions, a clear and practical definition of a financialaccident should emerge.

The second difference between transportation andfinancial services is the role of intellectual prop-erty and client confidentiality. In the transporta-tion industry, intellectual property is typicallyprotected by patents, and client confidential-ity is not often threatened by any single NTSBinvestigation. However, in the financial indus-try, intellectual property is typically protected bytrade secrecy, and client confidentiality can bea significant concern even for one investigation.These differences suggest that a financial NTSBinvestigation must be conducted in a differentmanner, with greater protections for proprietaryinformation, e.g., anonymizing client data andlimiting the disclosure of unrelated trade secrets.

Finally, the breadth of financial services impliesa broader investigative mandate for a financialNTSB, requiring greater use of external partieswhich, in turn, implies a more complex web ofpotential conflicts of interest. A closely relatedissue is the much larger gap between governmentpay scales and those of the financial industry; insome cases, the differences may be two or threeorders of magnitude. These challenges can beovercome with greater use of the party system,perhaps with more involved ethical guidelinesthat must be actively managed. Contrary to thestereotypical view of Wall Street denizens as self-interested profit-maximizers who have no interestin government service, the recent financial crisishas drawn many highly compensated individu-als into government service because they wishto have a positive impact during this critical



period. Moreover, as with the NTSB’s party sys-tem, there is an even larger pool of industryprofessionals who would gladly take a three- tosix-month leave of absence—with the support oftheir employers—to assist in an official govern-ment investigation requiring their unique financialexpertise.

The complexity of the financial system now rivalsthat of the most technologically sophisticatedindustries, and truly systemic “accidents” are anunfortunate consequence. This state of affairs ispart of a much broader trend in which technolog-ical innovation is a double-edged sword that isresponsible for great prosperity, but has also cre-ated unintended consequences and systemic vul-nerabilities. The most sophisticated technologiesoften require equally sophisticated coordinationamong individuals with highly specialized skills,but the institutional and organizational structuresneeded to support that level of coordination havenot always kept pace, especially during peri-ods of rapid growth and innovation. As systemsbecome more complex, the number of pointsof failure inevitably also increases, and eventhe most experienced and intelligent individualcannot comprehend all possible failure scenarioswith proactive risk analysis. Accordingly, in-depth investigation of failures and collaborationon systemic remedies are essential to reducing thechances of catastrophe. As the sociologist CharlesPerrow argued over two decades ago, all complexand tightly coupled systems are prone to “nor-mal accidents” (Perrow, 1984); systemic failureis not only possible, but should be expected tooccur under such conditions. The financial sys-tem may well be the largest and most importanttightly coupled complex system today.

Accidents happen in virtually every technolog-ically advanced endeavor, and while this mayindeed be normal, we need not compoundour mistakes by failing to learn from them.

The NTSB’s example provides a compellingalternative.

A Appendix

In this appendix, we provide more detailed infor-mation about the NTSB. In Appendix A.1, wepresent some basic facts and statistics for theorganization, and include the biographies ofcurrent board members in Appendix A.2. Weprovide an example of a typical NTSB boardmeeting in Appendix A.3, and in Appendix A.4we describe the NTSB’s ethical review process.Excepted summaries from the NTSB accidentreports of the ColganAir crash and the Minneapo-lis I-35W Highway Bridge collapse are includedin Appendixes A.5 and A.6, respectively. Andin Appendix A.7, we present the NTSB/FAAcommunication trail for a typical NTSB safetyrecommendation.

A.1 Facts and Statistics of the NTSB

Source: NTSB website (http://ntsb.gov/Abt_NTSB/history.htm).

The NTSB originated in the Air Commerce Actof 1926, in which the U.S. Congress charged theU.S. Department of Commerce with investigat-ing the causes of aircraft accidents. Later, thatresponsibility was given to the Civil AeronauticsBoard’s Bureau of Aviation Safety, when it wascreated in 1940.

In 1967, Congress consolidated all transportationagencies into a new U.S. Department of Trans-portation (DOT) and established the NTSB as anindependent agency placed within the DOT foradministrative purposes. In so doing, Congressexpanded the NTSB’s purview to include alltransportation modes, not just aviation.

In 1974, Congress reestablished the NTSB as acompletely separate entity, outside the DOT. In1996, Congress assigned the NTSB the additional



responsibility of coordinating Federal assistanceto the families of aviation accident victims.

In 2000, the agency embarked on a major ini-tiative to increase employee technical skills andmake their investigative expertise more widelyavailable to the transportation community byestablishing the NTSB Academy, which wasrenamed to NTSB Training Center in 2006.

Since its inception, the NTSB has investigatedmore than 132,000 aviation accidents and thou-sands of surface transportation accidents. Oncall 24 hours a day, 365 days a year, NTSBinvestigators travel throughout the country and toevery corner of the world to investigate significantaccidents and develop factual records and safetyrecommendations with one aim— to ensure thatsuch accidents never happen again.

To date, the NTSB has issued over 12,900 safetyrecommendations to more than 2,500 recipients.In 2008, the NTSB continued to push for safetyimprovements as 67 recommendations (for allrecipients) were officially closed with one of thefollowing classifications: “exceeds recommendedaction,” “acceptable action,” or “acceptable alter-nate action.” The average acceptance rate forsafety recommendations remained at just over 82percent for 2008.

The NTSB publishes the Most Wanted List ofTransportation Safety Improvements each year.The brochure highlights important safety actionsthat the Department of Transportation, the U.S.Coast Guard, and States need to take to preventaccidents and save lives.

In addition to its central role as accident inves-tigators, the NTSB provides a fair and impartialadjudicatory process for appeal of FAA certifi-cate actions and denials, and some civil penaltyactions through proceedings before its adminis-trative law judges and appellate review of the

judges’ decisions. The five-member Board alsoreviews Coast Guard certificate actions.

A.2 Biographies of current NTSB boardmembers

Source: http://www.ntsb.gov/abt_ntsb/member.htm.

DeborahA. P. Hersman, Chairman. DeborahA.P. Hersman was sworn in as the 12th Chairman ofthe National Transportation Safety Board on July28, 2009, following her nomination to the postby President Barack Obama and confirmation bythe United States Senate. Her two-year term asChairman runs until July 2011. She is also serv-ing a second five-year term as a Board Member,which expires on December 31, 2013.

Chairman Hersman has been a Member of theNTSB since June 21, 2004. Since then, shehas chaired a number of public events hostedby the Board including public forums and pub-lic hearings. During her tenure at the Board, shehas been the Member on the scene of 19 majortransportation accidents.

Before joining the Board, Chairman Hersman wasa Senior Professional Staff Member of the U.S.Senate Committee on Commerce, Science andTransportation from 1999 to 2004 where she wasresponsible for a number of transportation issues,and earlier served as Staff Director and SeniorLegislative Aide to Congressman Bob Wise ofWest Virginia.

Chairman Hersman earned Bachelor of Artsdegrees in Political Science and InternationalStudies from Virginia Tech University in Blacks-burg, Virginia, and a Master of Science degreein Conflict Analysis and Resolution from GeorgeMason University in Fairfax, Virginia.

Christopher A. Hart, Vice Chairman. Christo-pher A. Hart was sworn in as a Member of the



National Transportation Safety Board on August12, 2009 and designated by the President for atwo-year term as Vice Chairman of the Board onAugust 18.

Member Hart joined the Board after a long careerin transportation safety, including a previous termas a Member of the NTSB. Immediately beforereturning to the Board, Member Hart was DeputyDirector for Air Traffic Safety Oversight at theFederal Aviation Administration. He was previ-ously the FAA Assistant Administrator for theOffice of System Safety.

He served as a Member of the NTSB from 1990to 1993. After leaving the Board, he served asDeputy Administrator of the National HighwayTraffic Safety Administration, before moving tothe FAA in 1995.

From 1973 until joining the Board in 1990, Mem-ber Hart held a series of legal positions, mostlyin the private sector. He holds a law degreefrom Harvard University and Master’s and Bach-elor’s degrees in Aerospace Engineering fromPrinceton University. He is a member of the Dis-trict of Columbia Bar and the Lawyer-Pilots BarAssociation.

Member Hart is a licensed pilot with commercial,multi-engine and instrument ratings. His termexpires December 31, 2012.

Robert L. Sumwalt. Robert L. Sumwalt wassworn in as the 37th Member of the NationalTransportation Safety Board on August 21, 2006,whereupon President Bush designated him asVice Chairman of the Board for a two-year term,ending August 2008. His term of office as a BoardMember will run until December 31, 2011.

Prior to coming to the Board, Mr. Sumwalt wasManager ofAviation for the SCANA Corporation,a Fortune 500 energy-based company.

Mr. Sumwalt was a pilot for 32 years with exten-sive experience as an airline captain, airline checkairman, instructor pilot, and air safety representa-tive. Mr. Sumwalt worked on special assignmentto the U.S. Airways Flight Safety Departmentfrom 1997 to 2004. He also served as a memberof the Air Line Pilots Association’s (ALPA) Acci-dent Investigation Board, and he chaired ALPA’sHuman Factors and Training Group. A trainedaccident investigator, Mr. Sumwalt participatedin several NTSB investigations prior to joiningthe Safety Board.

Mr. Sumwalt co-authored a book on aircraft acci-dents and published over 85 articles and papers inaviation trade publications. In 2003, Mr. Sumwaltjoined the faculty of the University of SouthernCalifornia’s Aviation Safety and Security Pro-gram, where he was the primary human factorsinstructor.

In recognition of his contributions to the aviationindustry, Mr. Sumwalt received the Flight SafetyFoundation’s Laura Taber Barbour Award in 2003and ALPA’s Air Safety Award in 2005. He is a2009 inductee into the South Carolina AviationHall of Fame.

Since joining the Board, Member Sumwalt hasserved as the Chairman of the Board of Inquiryfor several NTSB public hearings. He has servedas the Member on-scene for a number of NTSBaccident investigations.

Mr. Sumwalt is a graduate of the University ofSouth Carolina.

Mark R. Rosekind. Mark R. Rosekind, Ph.D.was sworn in as a Member of the National Trans-portation Safety Board on June 30, 2010. He wasnominated by President Obama and confirmed bythe United States Senate for a term that expireson December 31, 2014.



Prior to joining the Board, Dr. Rosekind was thePresident and Chief Scientist of Alertness Solu-tions, a scientific consulting firm that specializesin fatigue management. Before founding Alert-ness Solutions, Dr. Rosekind directed the FatigueCountermeasures Program and was Chief of theAviation Operations Branch in the Flight Manage-ment and Human Factors Division at the NASAAmes Research Center. Prior to his work atNASA, Dr. Rosekind was the Director of the Cen-ter for Human Sleep Research at the StanfordUniversity Sleep Disorders and Research Center.

Member Rosekind is an internationally recog-nized fatigue expert who has conducted researchand implemented programs in diverse settings,including all modes of transportation. He haspublished 150 scientific, technical, and industrypapers and provided hundreds of presentations tooperational, general, and scientific audiences. Hiscontributions have been acknowledged throughnumerous honors and awards, including theNASA Exceptional Service Medal, six otherNASA Group/Team Awards, two Flight SafetyFoundation honors (Presidential Citation for Out-standing Safety Leadership, Business AviationMeritorious Award), and as a Fellow of the WorldEconomic Forum in Davos, Switzerland.

Member Rosekind earned his B.A. with Hon-ors at Stanford University, his M.S., M.Phil.,and Ph.D. at Yale University, and completed apostdoctoral fellowship at the Brown UniversityMedical School.

Earl F. Weener. Earl F. Weener, Ph.D., tookthe oath of office as a Member of the NationalTransportation Safety Board on June 30, 2010.

Dr. Weener is a licensed pilot who has dedicatedhis entire career to the field of aviation safety.Most recently, he has been a consultant and fel-low for the Flight Safety Foundation, where he

worked to reduce accidents through coordinatedindustry programs.

From 1975 to 1999, Dr. Weener held a series ofpositions with The Boeing Company, includingthree Chief Engineer positions, in Airworthi-ness, Reliability and Maintainability, and Safety;in System Engineering; and in Safety Technol-ogy Development. He also served four years asBoeing’s Manager of Government Affairs.

He has served as a general aviation flight instruc-tor and Part 135 pilot.

Dr. Weener earned all three of his academicdegrees in Aerospace Engineering at the Univer-sity of Michigan—his bachelor’s, master’s, anddoctorate. Among his awards are a 1994 LaurelAward fromAviationWeek and Space Technologymagazine and, in 2005, the Honeywell BendixTrophy for Aviation Safety.

Dr. Weener’s term as a Member of the NTSBexpires on December 31, 2015.

A.3 Example of a typical NTSBboard meeting

Source: http://ntsb.gov/events/2010/Washington-DC-Metro/presentations.htm.

On July 27, 2010, the National TransportationSafety Board held a board meeting to deliberatethe draft accident report for the Metrorail traincollision which occurred on June 22, 2009 inWashington D.C. The board meeting was chairedby Chairman Debbie Hersman and joined byVice Chairman Christopher Hart, Member RobertSumwalt, Member Mark Rosekind, and MemberEarl Weener.

Chairman Hersman’s opening statement gave thebackground and laid out the structure of themeeting:

This morning, the Board meets in open session as requiredby the Government in the Sunshine Act. While this is a



public meeting, only the Board members and NTSB staffmay participate in today’s discussions. Three weeks ago,the staff presented to the Board the draft report we are con-sidering today: the collision of two Metrorail trains on theRed Line near the Fort Totten station on June 22, 2009.While the individual Board members have had the oppor-tunity to read the accident report and to meet with the staffindividually, today is the first time that all five Board mem-bers are meeting together to discuss it. Staff has preparedfive presentations, which will be followed by a round ofquestions from the Board members. We will then considerthe report’s conclusions, probable cause determination, andproposed safety recommendations.

The five presentations from the NTSB staff wereas follows:

• James A. Southworth, Chief, Railroad Divi-sion: Overview with animation of the accident.

• Ruben Payan, Electrical Engineer, Signal andTrain Control Investigation: Two technical pre-sentations on automatic train control systemand false signals.

• Loren Groff, Safety Research Division: SafetyCulture of the WMATA.

• Stephen Klejst, Railroad Division: SafetyOversight of the WMATA.

After four hours of discussion which resulted inmodifications to the draft report, the board votedon the probable causes of the accident and safetyrecommendations. Chairman Hersman concludedthe meeting with closing remarks.

After the board meeting, that same day, the NTSBissued a press release and synopsis of the NTSBaccident report, including the probable cause,conclusions, and safety recommendations on itswebsite. The full report was made available on itswebsite a few weeks later.

Because the investigation uncovered potentialsafety concerns that warranted immediate atten-tion, the NTSB also issued several recommen-dations on July 13, 2009, and on September22, 2009, well before the July 27, 2010 Board

meeting. These recommendations were approvedby vote of the members individually, without ameeting.

A.4 Ethical review process

Source: Designated Agency Ethics Official, NTSBOffice of the General Counsel.

Ethical Review Process for Board Members

New Board Members (Nominees): For individ-uals that the President wishes to nominate fora Board Member position, the Office of WhiteHouse Counsel simultaneously provides to theDesignated Agency Ethics Official (DAEO) ofthe NTSB and the Office of Government Ethics(OGE), copies of the individual’s Public FinancialDisclosure Report (SF 278).

• The DAEO and OGE review the draft finan-cial disclosure report, as well as the potentialnominee’s resume concerning his or her pastemployment and professional relationships. Inconsultation with the potential nominee, theDAEO and OGE ensure that the financial dis-closure report, which includes informationabout prior employment and employers, isaccurate and complete and then identifies finan-cial holdings that would present a conflict ofinterest and other relationships that could cre-ate issues with regard to the potential nominee’simpartiality.

• The DAEO and OGE identify financial inter-ests that the potential nominee must divest (toavoid a conflict of interest) and assist him or herin creating an Ethics Agreement that specifieswhat issues exist and what the potential nomi-nee will do to manage and/or avoid the conflictsand questions of the nominee’s impartiality.

• Upon completing this review, and after the Pres-ident announces his intent to nominate the indi-vidual for a position as Board Member,the DAEO forwards the nominee’s financial



disclosure report (SF 278) and ethics agree-ment to the Director of OGE, along with a letterarticulating whether the individual may servewithout conflicts and questions as to his or herimpartiality.

New Board Members (who have been confirmedby the Senate and sworn in): Within 30 days ofa Board Member assuming his or her office, theDAEO provides a complete, face to face brief-ing in accordance with the requirements of theStandards of Conduct and the Ethics in Govern-ment Statute. The briefing is required to last forat least one hour and typically extends for one totwo hours.

Continuing Ethics Review/ Training: After the ini-tial financial disclosure report and training, eachMember must file a public financial disclosurereport on an annual basis, which is reviewed andcertified by the DAEO and forwarded to the Direc-tor of OGE for final certification. The financialdisclosure report, with its sections concerninggifts, payments from outside sources, and outsidepositions provides the opportunity for periodicethics review and counseling throughout a Mem-ber’s tenure. Additionally, each year each BoardMember must also have face to face ethics train-ing of at least one hour with the DAEO or anotherethics official.

As a final matter, when a Board Member leavesthe government and his or her appointed posi-tion, he or she must, within 30 days of separatingfrom the NTSB, file a termination financial dis-closure report, which is reviewed and certified bythe DAEO and forwarded to the Director of OGEfor final certification.

Ethical Review Process for NTSB Employees

Initial Ethics Review/Screening: A member ofthe NTSB ethics staff conducts an interviewand briefing with each potential employee (anindividual that has received a preliminary offer

of employment) as part of the final screen-ing/clearance process—before a final job offer ismade. The interview/briefing, which is conductedin-person or telephonically, typically lasts from30 minutes to an hour (or more, depending onthe complexity of the employee’s background)and follows the format in the ethics interviewchecklist focusing on financial interest and rela-tionship. The NTSB is required to provide eachnew employee with a copy of the “14 Principlesof Ethical Conduct for Government Officers andEmployees” as well as contact information of theNTSB ethics staff.

Upon completing the briefing/interview, theethics staff creates a report that goes to the hir-ing authority. The report identifies ethics issuesand specifies methods for handling issues, if theyexist.

Additional Review/ Training: With regard to con-tinuing review for compliance with ethics guide-lines, if the individual is identified, because ofposition or duties, as a being required to file afinancial disclosure report (Public—SF 278; orConfidential—OGE Form 450), he or she mustfile his initial report within 30 days of reportingfor duty. The report is reviewed by ethics offi-cials. The employee must, thereafter, file annualfinancial disclosure reports which are reviewed byagency ethics officials each year.

Additionally, the employee must complete on-line ethics training within 30 days of reportingto duty and must, thereafter, complete it, as do allemployees, on an annual basis.

A.5 Continental connection flight 3407(Colgan Air) crash

Source: NTSB accident report (NTSB AAR-10/01).

On February 12, 2009, about 22:17 EST, a Col-ganAir, Inc., Bombardier DHC-8-400, N200WQ,



operating as Continental Connection flight 3407,was on an instrument approach to Buffalo-Niagara InternationalAirport, Buffalo, NewYork,when it crashed into a residence in Clarence Cen-ter, New York, about 5 nautical miles northeastof the airport. The two pilots, two flight atten-dants, and 45 passengers aboard the airplane werekilled, one person on the ground was killed, andthe airplane was destroyed by impact forces anda post-crash fire.

The National Transportation Safety Board wasnotified of this accident about 22:30 on February12, 2009. A Go Team launched early the nextmorning. Accompanying the team to Buffalowas former Board Member Steven Chealan-der. The following investigative teams wereformed: Operations, Human Performance, Struc-tures, Systems, Power plants, Air Traffic Control,Meteorology, Aircraft Performance, MaintenanceRecords, and Pipeline. Also, specialists wereassigned to conduct the readout of the flight datarecorder and transcribe the cockpit voice recorderat the NTSB’s laboratory in Washington, D.C.

Parties to the investigation were the Federal Avia-tion Administration (FAA), Colgan Air, Air LinePilots Association (ALPA), National Air TrafficControllers Association, and United Steelwork-ers Union (Flight Attendants). In accordance withthe Convention on International Civil Aviation,the Transportation Safety Board of Canada (TSB)participated in the investigations among others.Bombardier, the manufacturer of the aircraft, wasthe technical advisors to the TSB.

A public hearing was held from May 12 to 14,2009, in Washington, D.C. Former Acting Chair-man Mark Rosenker presided over the hearing;Board Member and current Chairman Debo-rah Hersman, former Board Member KathrynHiggins, and Board Member Robert Sumwaltalso participated in the hearing. The issues pre-sented at the hearing were the effect of icing

on airplane performance, cold weather opera-tions, sterile cockpit rules, flight crew experience,fatigue management, and stall recovery training.The technical panel comprised investigators fromthe NTSB and the TSB. Parties to the publichearing were the FAA, Colgan Air, ALPA, andBombardier.

The National Transportation Safety Board deter-mines that the probable cause of this accidentwas the captain’s inappropriate response to theactivation of the stick shaker, which led to anaerodynamic stall from which the airplane did notrecover. Contributing to the accident were (1) theflight crew’s failure to monitor airspeed in relationto the rising position of the low speed cue, (2) theflight crew’s failure to adhere to sterile cockpitprocedures, (3) the captain’s failure to effectivelymanage the flight, and (4) ColganAir’s inadequateprocedures during approaches in icing conditions.

The safety issues discussed in this report focuson strategies to prevent flight crew monitoringfailures, pilot professionalism, fatigue, remedialtraining, pilot training records, airspeed selec-tion procedures, stall training, Federal AviationAdministration (FAA) oversight, flight opera-tional quality assurance programs, use of personalportable electronic devices on the flight deck, theFAA’s use of safety alerts for operators to transmitsafety-critical information, and weather informa-tion provided to pilots. Safety recommendationsconcerning these issues are addressed to the FAA.

A.6 Minnesota I-35W highwaybridge collapse

Source: NTSB accident report (NTSB HAR-08/03).

About 6:05 p.m. on August 1, 2007, the eight-lane, 1,907-foot long I-35W highway bridge overthe Mississippi River in Minneapolis collapsed,resulting in 13 fatalities and 145 injuries.



On the day of the collapse, roadway work wasunderway on the I-35W bridge. In the earlyafternoon, construction equipments and construc-tion materials (sand and gravel for making con-crete) were delivered and positioned in the twoclosed inside southbound lanes.

The National Transportation Safety Board wasnotified of the Minneapolis, Minnesota, bridgecollapse on August 1, 2007. Investigative teamswere dispatched from the Safety Board’s Wash-ington, D.C.; Atlanta, Georgia; Arlington, Texas;and Gardena, California, offices. Separate groupswere established to investigate structural engi-neering, bridge design, construction oversight,and survival factors issues. Other groups wereformed to facilitate evidence documentation,structural modeling, and witness identification.Chairman Mark Rosenker at the time was theBoard Member on scene.

Participating in the on-scene investigationwere representatives of the Federal HighwayAdministration, the Minnesota Department ofTransportation, the Minnesota State Police, theMinneapolis Police Department, the HennepinCounty Sheriff’s Office, and the maintenancecontractor, Progressive Contractors, Inc. JacobsEngineering (the company that had acquired thefirm responsible for original design of the bridge)initially provided design plans and other relateddocuments and later, on January 17, 2008, wasincluded as an official party to the investigation.

The on-scene investigation, including documen-tation and analysis of the recovered bridge struc-ture, required Safety Board investigators andother support staff to remain at the accident sitefrom August 2 to November 10, 2007.

The National Transportation Safety Board deter-mines that the probable cause of the collapse ofthe I-35W bridge in Minneapolis, Minnesota, wasthe inadequate load capacity, due to a design error

by Sverdrup & Parcel and Associates, Inc., ofthe gusset plates at the U10 nodes, which failedunder a combination of (1) substantial increasesin the weight of the bridge, which resulted fromprevious bridge modifications, and (2) the trafficand concentrated construction loads on the bridgeon the day of the collapse. Contributing to thedesign error was the failure of Sverdrup & Par-cel’s quality control procedures to ensure that theappropriate main truss gusset plate calculationswere performed for the I-35W bridge and theinadequate design review by Federal and Statetransportation officials.

Contributing to the accident was the generallyaccepted practice among Federal and State trans-portation officials of giving inadequate attentionto gusset plates during inspections for conditionsof distortion, such as bowing, and of excludinggusset plates in load rating analyses.

A.7 NTSB/FAA communication trail of atypical safety recommendation

Source: NTSB Safety Recommendation Informa-tion System (SRIS) http://www.ntsb.gov/safety-recs/private/QueryPage.aspx.

Accident Background:

According to NTSB accident report (NTSBAAR-04/04), American Airline Flight 587 crash onNovember 12, 2001 killing 260 on board and fiveon the ground, caused by excessive rudder inputfrom the first officer when the plane flew into alarger jet’s wake. The NTSB issued safety recom-mendation A-02-01 to FAA regarding the issue ofexcessive rudder input.

Timeline:

Nov 12, 2001—Accident occurred and investiga-tion was launched.Feb 08, 2002—Safety recommendation A-02-01issued.Status (as of 8/11/2010)—Open acceptable.



Response Date: 4/15/2002From: AddresseeResponse:

Letter Mail Controlled 05/07/2002 10:58:22 AMMC# 2020458. The Federal Aviation Adminis-tration (FAA) agrees with the intent of thesesafety recommendations and has taken the fol-lowing action based on the results of the accidentinvestigation to date.

Response Date: 7/22/2002From: NTSBResponse:

The Safety Board thanks the FAA for the actions ithas taken in response to these recommendations.The Board notes that the FAA’s review of trainingprograms was limited to the programs of opera-tors of Airbus airplanes. The Board believes thatthe training programs of operators of other manu-facturers’ airplanes should also be reviewed. TheFAA’s plan to use non-regulatory means to meetthe intent of Safety Recommendation A-02-01may represent an acceptable alternative; however,the FAA also indicates that it may ultimately makea regulatory change. The Board will assume, untilthe FAA indicates otherwise, that the FAA willdevelop some regulatory changes in pilot trainingprograms in response to Safety Recommenda-tion A-02-01. Pending completion of changesto pilot training programs in response to therecommendations, the determination of whetherthese revisions will be implemented through theprocedures in HBAT 99-07 or through regula-tory changes, and the FAA’s consideration ofreviewing the training programs of operators ofother manufacturers’ airplanes, Safety Recom-mendations A-02-01 and A-02-02 are classified“Open–Acceptable Response.”

This is the last communication in the databasebetween the NTSB and the FAA on this safetyrecommendation as of August 15, 2010.

Acknowledgement

The views and opinions expressed in this reportare those of the authors only, and do notnecessarily represent the views and opinions ofMIT, AlphaSimplex Group, the National Trans-portation Safety Board, or any of their affiliatesand employees. We thank Mark Bagnard, ArnieBarnett, Jayna Cummings, John DeLisi, DavidEsch, Dwight Foster, Erin Gormley, Joseph A.Gregor, Chris Hart, Joseph Kolly, Joe Langsam,William Love, Bruce Magladry, Jeffrey H. Mar-cus, Harkey Mayo, Julie Perrot, Carl R. Schulthe-isz, Dan Walsh, Lorenda Ward, and participantsat the 2010 JOIM Fall Conference and the BostonUniversity Law School Workshop on FinancialReform for their helpful comments and discus-sion. We also thank Vice Chairman Christo-pher Hart, Managing Director David L. Mayer,and Deputy Managing Director Barbara E. Zim-mermann for their cooperation and hospitalitythroughout this project.

Notes1 Our use of the term “forensic” here and throughout this

article refers to the scientific methods with which theNTSB investigates matters of public safety, and is notmeant in the legal sense; in fact, the NTSB’s investiga-tions are not admissible evidence in court (see Section5.1).

2 The NTSB also serves as the appellate authority for air-men, aviation mechanic, and mariner certificate actions(e.g., to reinstate a suspended certificate) taken, respec-tively, by the FAA Administrator, and the Commandantof the U.S. Coast Guard. It also provides transportationdisaster assistance functions, which involves guidingcarriers in the development of plans to assist accidentvictims and families of accident victims.

3 Of course, individual board members, in accordancewith the NTSB’s preference for openness and accessibil-ity, often hold separate meetings with others, includingadvocacy groups and family members of the victims ofaccidents under investigation.

4 Parties are named as participants in the investigatoryprocess based upon their ability to provide needed infor-mation, skills, and assistance. The parties typically will



include a DOT modal administration (i.e., the FAA,the FRA, FHWA, etc.), as well as non-Federal enti-ties, including carriers, manufacturers, and a variety ofexperts.

5 Many of the participants who can provide technicalassistance are also likely to be defendants in the ensu-ing litigation. For that reason, families of the victims,i.e., potential litigation plaintiffs, complain that thisprivileged access to information by those prospectivedefendants gives those defendants an unfair litigationadvantage.

6 Party members are not permitted to reveal to theiremployers any aspect of the investigation that is notpublic information. However, party members may facepressure from their superiors to divulge information inan effort to limit the liabilities of their companies.

7 Whether or not a public hearing is held can some-times be a point of contention between the NTSB,Congress, and other stakeholders. For example, theNTSB’s decision not to hold a hearing for the Min-neapolis bridge collapse was quite controversial (seehttp://www.youtube.com/watch?v=ZgCqh1OLcG0).

8 In some cases, an accident investigation may involvecriminal activities, e.g., the four airplane crashes onSeptember 11, 2001, and in these situations, theNTSB no longer plays the lead investigative role but,instead, provides support for the law enforcement agen-cies involved. As described on the NTSB website(http://ntsb.gov): “In cases of suspected criminal activ-ity, other agencies may participate in the investigation.The Safety Board does not investigate criminal activity;in the past, once it has been established that a trans-portation tragedy is, in fact, a criminal act, the FBIbecomes the lead federal investigative body, with theNTSB providing any requested support. …As the resultof recent legislation, the NTSB will surrender lead statuson a transportation accident only if the Attorney Gen-eral, in consultation with the Chairman of the SafetyBoard, notifies the Board that circumstances reasonablyindicate that the accident may have been caused by anintentional criminal act.”

9 Because NTSB investigations are focused on determin-ing probable causes of accidents, not legal liability,the accident reports are not meant to serve as evi-dence in legal proceedings, and certain conclusions andrecommendations may be considered prejudicial—andtherefore, inadmissible—under certain local and federalrules of evidence. The inadmissibility of NTSB accidentreports also allows parties engaged in legal proceedings

to conduct their own investigations in ways that are mosteffective in supporting their cases.

10 While NTSB accident reports are inadmissible as evi-dence in legal proceedings, they provide importantunderlying information on which lawsuits are based, andnot surprisingly, make NTSB investigators attractive tolitigants as potential sources of valuable information thatcan influence litigation strategies. See 49 U.S.C 1154(b)and 49 CFR Part 835 for more details.

11 For example, on March 4, 2008, a Cessna 500 corpo-rate jet crashed in Oklahoma City, killing all five peopleon board. At first glance this accident was caused bypelicans hitting the airplane, but the investigation soonrevealed numerous alarming inadequacies in charteroperations. The NTSB proposed four safety recom-mendations regarding wildlife strikes and six safetyrecommendations on the operations of chartered flights[NTSB AAR-09/05].

12 For example, during the investigation of the crashof TWA flight 800 in 1996, the NTSB issued anurgent safety recommendation regarding the plane’scenter fuel tank, four years before the investigation wascompleted.

13 In addition to the crash of the Colgan Air flight, whichwas operating as Continental Connection, there was a2007 accident in Traverse City, Michigan, in which aPinnacle Airlines flight was operated as Northwest Air-link; a 2007 accident in Cleveland, Ohio, in which aShuttle America flight was operated as Delta Connec-tion; and a 2006 accident in Lexington, Kentucky, inwhich a Comair flight was operated as Delta Connection.(NTSB Advisory, August 16, 2010).

14 This action was later criticized by some as politicallymotivated (see Kaszuba, 2007).

15 The time commitment required to prepare for an NTSBpublic hearing is typically three months or more of full-time preparatory work for most of the investigative team,which effectively brings the investigation to a full stopduring this period.

16 According to a statement issued by URS Corpora-tion to its investors on August 23, 2010, it settledthe litigation for a sum of $52.4 million, denying anywrong-doing, and citing the NTSB investigation (seehttp://www.urscorp.com).

17 For example, this explanation was not particularly sup-portive of Congressman Oberstar’s initiative to raisestate and federal gasoline taxes to fund bridge repairsand road reconstruction. As reported in the Minneapo-lis Star Tribune on August 22, 2007 (see Kersten,



2007):

…Minnesota Rep. Jim Oberstar—the powerfulchairman of the House Transportation and Infrastruc-ture Committee—has called for a ‘temporary’ 5-centincrease in the federal gas tax to raise what he says isa critically needed $25 billion over three years fora national bridge-repair trust fund. “If you’re notprepared to invest another five cents in bridge recon-struction and road reconstruction, then God helpyou,” he said after the bridge collapse.

18 During the investigation of the Colgan Air accident(see Section 3.1), the NTSB pointed out deficienciesin FAA oversight and surveillance. According to theaccident report (NTSB AAR-10/01, page 154), “Thecurrent Federal Aviation Administration surveillancestandards for oversight at air carriers undergoing rapidgrowth and increased complexity of operations do notguarantee that any challenges encountered by the car-riers as a result of these changes will be appropriatelymitigated.”

19 For example, “To ensure aviation’s future viability,FAA is now working with its federal and industrypartners to develop a flexible aerospace system thatfully responds to the changing needs of businessesand customers in the 21st Century.” (http://www.faa.gov/about/history/brief_history/). A more directillustration is the FAA’s 2004 “Customer Service Ini-tiative” in which the airliners were considered its “cus-tomers” instead of the traveling public. This initiativehas since been appropriately modified, but it providesa sobering illustration of the sometimes unavoidableconflicts faced by regulators.

20 The NTSB also ranked 6th in “pay and benefits,” 10th in“team work,” 10th in “performance based rewards andadvancement,” and 6th in “family friendly culture.”

21 According to the NTSB accident report for the Ameri-can Airline Flight 587 crash on November 12, 2001 inBelle Harbor, New York, which killed 260 aboard and 5on the ground (NTSB AAR-04/04), the first safety rec-ommendation was issued on February 8, 2002, whilethe final accident report was adopted on Oct 26, 2004.

22 For example, the NTSB determined that the proba-ble cause of the explosion of TWA Flight 800 shortlyafter take-off on July 17, 1996 was the ignition offlammable fuel vapors in the plane’s center wing fueltank, and issued safety recommendation A-96-174 onDecember 13, 1996 requiring airplanes to be fitted withsome form of “fuel inerting system.” In July 1998,an FAA study concluded that the cost to the industry

would be $20 billion and very difficult to retrofit intoexisting airplanes. Subsequent FAA studies produceda cheaper and effective alternative, and on Novem-ber 23, 2005 the FAA published a Notice of ProposedRule Making in the Federal Register (Vol. 70, No.225, pages 70921–70962) regarding its proposal torequire the installation of flammability reduction sys-tems in all commercial aircraft. The new FAA rule wasissued on July 21, 2008 (Federal Register Vol. 73, No.140, pages 42444–42504), 12 years after the NTSBsafety recommendation. See the NTSB article: http://www.ntsb.gov/recs/mostwanted/explosive_tanks.htm forfurther details.

23 Some have argued that this dual perspective can some-times lead to “regulatory capture,” in which regulatorsbecome overly influenced by industry they are supposedto regulate. For example, The NewYork Times publisheda story on June 4, 2009, about four months after theColgan Air accident, of an FAA inspector who raisedissues with Colgan Air more than a year prior to thecrash, but was overruled by his supervisor in an effort tohelp Colgan Air maintain its flight schedule (see Wald,2009). However, it is clear that most regulators do havemultiple competing objectives by charter, hence suchconflicts are inevitable and must be managed carefullyby the regulatory body.

24 The salary data is from 1998 and 1999.25 For example, in 2008, the value-added contribution

of the financial services industry to gross domesticproduct was $1,882.4 billion (the sum of: finance andinsurance; Federal Reserve banks, credit intermedia-tion, and related activities; and securities, commoditycontracts, and investments), which is over seven timesthe value-added of $252.3 billion of the transportationindustry (the sum of: air, rail, water, truck, transitand ground-passenger transportation). See the Bureauof Economic Analysis’s “Gross-Domestic-Product-by-Industry Accounts,” May 25, 2010 Release for furtherdetails.

26 See Office of Financial Stability (2010, page 1) for the$50 billion estimate of the economic impact of TARP.A recent estimate of the economic value of a life isprovided by Viscusi (2004), who incorporates workerfatality risk, and arrives at an estimate of $5.0 millionin 2000 dollars. With the Consumer Price Index at169.30 in January 2000 and 218.15 in August 2010 (seehttp://www.bls.gov/cpi/), this implies a current value of$6,442,705.26 which, when divided into $50 billion,yields 7,761 lives.



References

Kaszuba, M. (2007). “Bridge Consultant Helping FederalProbe.” Minneapolis-St. Paul Star Tribune, November 15.

Kersten, K. (2007). “Until I-35W disaster, Oberstar’s fund-ing focus wasn’t on bridges.” Minneapolis-St. Paul StarTribune, August 22.

Lo, A. (2010). “The Financial Industry Needs Its Own CrashSafety Board.” Financial Times, March 2.

National Transportation Safety Board (2008a). “Collapse ofI-35W Highway Bridge Minneapolis, Minnesota, August1, 2007.” NTSB HAR-08/03. Washington, DC: NationalTransportation Safety Board.

National Transportation Safety Board (2008b). AnnualReport to Congress. Washington, DC: National Trans-portation Safety Board.

National Transportation Safety Board (2009a). Operat-ing Plan Fiscal Year 2010. Washington, DC: NationalTransportation Safety Board.

National Transportation Safety Board (2009b). StrategicPlan, Fiscal Year 2010 through 2015. Washington, DC:National Transportation Safety Board.

National Transportation Safety Board (2009c). AnnualReport to Congress. Washington, DC: National Trans-portation Safety Board.

National Transportation Safety Board (2010a). “Loss ofControl on Approach, Colgan Air, Inc. Operating asContinental Connection Flight 3407 Bombardier DHC-8-400, N200WQ, Clarence Center, NewYork, February 12,2009.” NTSB/AAR-10/01. Washington, DC: NationalTransportation Safety Board.

National Transportation Safety Board (2010b). “NTSBADVISORY: NTSB to Hold Symposium onAirline Code-Sharing Arrangements and their Role in Aviation Safety.”August 16. Washington, DC: National TransportationSafety Board.

Office of Financial Stability (2010). Troubled Asset ReliefProgram: Two-Year Retrospective. Washington, DC: U.S.Treasury.

Perrow, C. (1984). Normal Accidents: Living with HighRisk Technologies. New York: Basic Books.

Sarsfield, L., Stanley, W., Lebow, C., Ettedgui, E., andG. Henning (2000). Safety in the Skies: Personnel and

Parties in NTSB Aviation Accident Investigations (MasterVolume). Santa Monica, CA: Rand Corporation.

United States Government Accountability Office (2006a).“Testimony before the Subcommittee on Aviation, Com-mittee on Commerce, Science and Transportation, U.S.Senate: Report to Congressional Requesters: NationalTransportation Safety Board.” GAO-06-801T, May 24.Washington, DC: Government Accountability Office.

United States Government Accountability Office (2006b).“Report to Congressional Requesters: National Trans-portation Safety Board.” GAO-07-118, November. Wash-ington, DC: Government Accountability Office.

United States Government Accountability Office (2008).“Testimony before the Subcommittee on Aviation, Com-mittee on Transportation and Infrastructure, House ofRepresentatives: National Transportation Safety Board.”GAO-08-652T, April 23. Washington, DC: GovernmentAccountability Office.

United States Government Accountability Office (2009).“Testimony before the Subcommittee on Aviation Oper-ations, Safety, Security, Committee on Commerce,Science, and Transportation, U.S. Senate: NationalTransportation Safety Board.” GAO-10-183T, Octo-ber 29. Washington, DC: Government AccountabilityOffice.

United States Government Accountability Office (2010).“Testimony before the Subcommittee on Aviation, Com-mittee on Transportation and Infrastructure, House ofRepresentatives: National Transportation Safety Board.”GAO-10-366T, January 27. Washington, DC: Govern-ment Accountability Office.

Veracifier (2008). “Congressman Rips Into NTSBChair Over Collapse Hearing.” http://www.youtube.com/watch?v=ZgCqh1OLcG0, April 24.

Viscusi, K. (2004). “The Value of Life: Estimates withRisks by Occupation and Industry.” Economic Inquiry42, 29–48.

Wald, M. (2009). “Inspector Predicted Danger BeforeBuffalo Crash.” The New York Times, June 3.

Keywords: Risk management; systemic risk;forensic finance; NTSB


journal of nvestment anagement joim - andrew loinvestigations of transportation accidents and...

Documents