COGNITIVE�FAILURE?

THE�FINANCIAL�CRISIS:�MORAL�FAILURE�OR� COGNITIVE�FAILURE?�

ARNOLD�KLING*�

This�may� be� our� first� epistemologically�driven� depression.� (Epistemology� is� the� branch� of� philosophy� that� deals�with� the� nature� and� limits� of� knowledge,� with� how� we� know� what�we�think�we�know.)�That�is,�a�large�role�was�played�by� the� failure�of� the�private�and�corporate�actors� to�understand� what�they�were�doing.�Most�heads�of�ailing�or�deceased�finan� cial� institutions�did�not�comprehend�the�degree�of�risk�and� exposure�entailed�by� the�dealings�of� their�underlings—and� many�investors,�including�municipalities�and�pension�funds,� bought� financial� instruments� without� understanding� the� risks�involved.1�

There� are� two�major� competing� narratives� for� the� financial� crisis.� One� narrative� focuses� on� moral� failure,� in� which� the� compensation� structure� for� executives� at� financial� institutions� encouraged�them�to�place�their�own�and�other�firms�at�risk�to� reap�short�term�gains.2�The�other�narrative�focuses�on�cognitive� failure,� in�which� executives� and� regulators� overestimated� the� risk�mitigating� effects� of� quantitative� modeling� and� financial� engineering.� It� is� important� to� sort� out�which� of� these� narra� tives�deserves�more�credence.� Those� who� emphasize� moral� failure� have� highlighted� a�

number�of�distortions�between�private�and� social�benefits,� in� cluding:�that�executive�pay�at�financial�institutions�is�not�tied�to� long� term� viability,3� the� “originate� to� distribute”� model� of� mortgage� financing� gives� the� originator� an� incentive� to�make� bad�loans�that�are�passed�down�the�line�in�the�system�of�struc� ������������������������������������������������������������������������������������������������������������������ *�Adjunct�scholar,�Cato�Institute.�Mr.�Kling�has�worked�as�an�economist�at�the�

Federal�Reserve�and�at�Freddie�Mac.� 1.�Jerry�Z.�Muller,�Our� Epistemological�Depression,�AMERICAN,� Jan.� 29,� 2009,� http://�

www.american.com/archive/2009/february�2009/our�epistemological�depression.� 2.�See,�e.g.,�Lucian�A.�Bebchuk�&�Holger�Spamann,�Regulating�Bankers’�Pay,�98�

GEO.�L.J.�247,�249�(2010).� 3.�Lucian�Bebchuk�has�emphasized�this�disconnect.�See�id.��

508� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

tured�financing�of�mortgage�securities,4�and�rating�agencies�are� overly�generous�in�granting�AAA�and�AA�ratings�because�they� were�paid�by�the�issuers�of�mortgage�related�securities.5� � Under� the�moral� failure� theory,� the�essential�problem�is� the� misalignment�between�the�incentives�of�executives�to�maximize� their�own�salaries�and�the�long�term�best�interest�of�the�finan� cial�firms�they�led.6�In�this�narrative,�regulators�were�either�sti� fled�by� ideological� faith� in�markets�or�hampered�by�organiza� tional� flaws—most� notably,� the� alleged� absence� of� anyone� charged�with�monitoring�systemic�risk.� The� other� narrative� is� one� of� cognitive� failure.� Under� this�

view,�key�individuals�believed�propositions�that�turned�out�to� be� untrue.� Propositions� that� were� falsely� believed� included:� that� a� nationwide� decline� in� housing� prices,� having� not� oc� curred� since� the�Great�Depression,�was� impossible;� increased� home�ownership�rates�were�a�sign�of�economic�health;�the�use� of�structured�finance�and�credit�derivatives�had�reduced�risk�to� key�financial�institutions;�monetary�policy�only�needed�to�focus� on�overall�economic�performance,�not�on�asset�bubbles;�banks� were�well� capitalized;� and� quantitative� risk�models� provided� reliable� information�on� the�soundness�of�mortgage�backed�se� curities�and�of�the�institutions�holding�such�securities.7�In�hind� sight,� these� propositions� were� wrong.� Policymakers� were� caught� up� in� the� same� cognitive� environment� as� financial� ex� ecutives.�Market�mistakes�went�unchecked�not�because�regula� tors�lacked�the�will�or�the�institutional�structure�with�which�to� regulate,�but�because�they�shared�with�the�financial�executives� the�same�illusions�and�false�assumptions.� Under�the�narrative�of�moral�failure,�the�financial�crisis�was�

like� a� fire� started� by� delinquent� teenagers,�with� the� adults� in� charge� not� sufficiently� inclined� or� positioned� to� exercise� ade�

������������������������������������������������������������������������������������������������������������������ 4.�See,�e.g.,�Antje�Berndt�&�Anurag�Gupta,�Moral�Hazard�and�Adverse�Selection�

in� the�Originate�to�Distribute�Model�of�Bank�Credit� 5� (Nov.� 2008)� (unpublished� manuscript),�available�at�http://ssrn.com/abstract=1290312.� 5.�See,� e.g.,�Credit� Rating� Agencies� and� the� Financial� Crisis:� Hearing� Before� the� H.�

Comm.�on�Oversight�&�Gov’t�Reform,�110th�Cong.�31–32�(2008)�[hereinafter�Hearing]� (statement� of� Frank�L.�Raiter,� former�Managing�Director,�Residential�Mortgage� Backed�Securities�Ratings,�Standard�&�Poor’s).� 6.�Bebchuk�&�Spamann,�supra�note�2,�at�249.� 7.�For�what� is,� in�my� view,� the� best�work� on� the� crisis� thus� far,� see�GILLIAN�

TETT,� FOOL’S�GOLD:�HOW� THE� BOLD�DREAM�OF� A� SMALL� TRIBE� AT� J.P.�MORGAN� WAS�CORRUPTED�BY�WALL�STREET�GREED�AND�UNLEASHED�A�CATASTROPHE�(2009).�

No.�2]� Moral�Failure�or�Cognitive�Failure?� 509�

quate� supervision.� The� solution� is� thus� to� reorganize� and� re� energize�the�regulatory�apparatus.� Under�the�narrative�of�cognitive�failure,�it�is�as�if�the�authori�

ties� supplied� the� lighter� fluid,�matches,� and�newspapers�used� to� start� the� fire.� In� particular,� housing� policy� encouraged� too� many�households� to�obtain�homes�with� too� little�equity.�Bank� capital�regulations�steered�banks�away�from�traditional�lending� toward�securitization.�Moreover,�these�regulations�encouraged� the�banks’�use�of�ratings�agencies�and�off�balance�sheet�entities� to�minimize� the� capital�held� to�back� risky� investments.� If� this� narrative� holds,� then� financial� regulation� itself� is� inherently� problematic.� Regulators,� sharing� the� same� cognitive� environ� ment�as�financial�industry�executives,�are�unlikely�to�be�able�to� distinguish� evolutionary� changes� that� are� dangerous� from� those� that�are�benign.� It�may�not�be�possible� to�design�a� fool� proof�regulatory�system.�

I. FREDDIE�MAC�

Perhaps� the� best� illustration� of� the� tension� between� moral� and� cognitive� failure� narratives� is� the� response� to� Freddie� Mac’s�rapid�decline.�Freddie�Mac,�a�company�chartered�by�the� government� in�1970�but�sold� to�private� investors� in�1989,�was� one�of�the�institutions�that�suffered�catastrophic�losses,�in�part� because� it� relaxed� credit� standards� from� 2002� through� 2007.8� Was�this�relaxation�a�moral�or�cognitive�failure?� In�August�2008,�the�New�York�Times�reported�that�in�deciding�to�

become�more� active� in� the� subprime�mortgage�market,� Freddie� Mac�s�CEO,�Richard�Syron,�had�ignored�the�warnings�of�the�com� pany’s� Chief� Risk� Officer,� David� Andrukonis.9� Early� in� 2004,� Andrukonis�had�sent�Syron�memoranda�that�argued�against�pur� chasing�mortgages�that�were�originated�with�reduced�documenta� tion.10� Shortly� afterward,�Andrukonis� left,� and� Freddie�Mac� ex� panded�its�purchases�of�various�high�risk�mortgage�products.11�

������������������������������������������������������������������������������������������������������������������ 8.�STAFF�OF�H.�COMM.�ON�OVERSIGHT�&�GOV’T�REFORM,�111TH�CONG.,�THE�ROLE�

OF�GOVERNMENT�AFFORDABLE�HOUSING�POLICY�IN�CREATING�THE�GLOBAL�FINAN� CIAL�CRISIS�OF�2008,�at�24�(2009).� 9.�Charles� Duhigg,�At� Freddie�Mac,� Chief� Discarded�Warning� Signs,� N.Y.� TIMES,�

Aug.�5,�2008,�at�A1.� 10.�Id.� 11.�Id.�

510� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

The�narrative�of�moral� failure�would�suggest� that�Syron�was� motivated� by� the� desire� for� short�term� profits� and� bonus� pay� ments� to� the� detriment� of� his� obligations� to� shareholders� and� other� long�term� constituencies.� Certain� reports,� however,� such� as�one�that�appeared�in�the�Boston�Globe,12�paint�a�different�pic� ture.�According�to�this�alternative�account,�Syron�focused�on�his� responsibility�to�keep�Freddie�Mac�active�in�a�mortgage�market� that�was�shifting�away� from�traditional�safe�mortgages�and� to� ward�riskier�products.13�Moreover,�he�believed�that�Freddie�Mac� had�a�mission�to�serve� the�needs�of�minorities�and�low�income� home� buyers.14� One� could� therefore� argue� that� his� decisions� were�driven�by�moral�considerations,�not�by�personal�greed.� The�ultimate�difference�between�David�Andrukonis�and�Rich�

ard�Syron,�however,�was�not�that�one�had�a�moral�backbone�that� the� other� lacked.� The� difference� was� cognitive.� Andrukonis,� a� twenty�year� employee� of� the�mortgage� company,� knew� of� the� bad�experience�Freddie�Mac�once�had�with�low�documentation� loans�in�the�late�1980s—an�experience�that�resulted�in�agreement� between�Freddie�Mac�and�Fannie�Mae�not�to�purchase�reduced� documentation� loans.� He� was� also� skeptical� of� the� ability� of� Freddie�Mac�to�safely�expand�its�share�of�loans�to�so�called�“un� der�served”�borrowers.�By�contrast,�Syron,�who�became�CEO�in� 2003,�thought�that�Freddie�Mac�had�been�too�conservative�in�the� past�and�needed�to�demonstrate�greater�commitment�to�the�mis� sion�of�making�home�ownership�more�affordable.15�

II. INSIDE�THE�CREDIT�RATING�AGENCIES�

The�history�of�credit�rating�agencies�also�highlights�the�moral� and�cognitive�failure�dichotomy.�These�agencies�played�a�cen� tral�role�in�the�buildup�to�the�crisis.16�Financial�engineers�struc� tured�mortgage�backed� securities� to� try� to�maximize� the� pro�

������������������������������������������������������������������������������������������������������������������ 12.�Robert�Gavin,�Syron�s�side�of�the�story,�BOSTON�GLOBE,�Aug.�6,�2008,�at�C1.� 13.�Id.� 14.�Id.� 15.�Andrukonis�was�a�colleague�of�mine�when�I�worked�at�Freddie�Mac�in�the�

late�1980s�and�early�1990s,�and�we�have� remained� friends�since.�My�reconstruc� tion�of� the� controversy� is�based� in�part�on�conversations�with�Andrukonis�after� the�story�broke�in�the�New�York�Times.� 16.�See�Hearing,�supra�note�5,�at�1–2� (statement�of�Rep.�Henry�Waxman,�Chair�

man,�H.�Comm.�on�Oversight�and�Gov’t�Reform).�

No.�2]� Moral�Failure�or�Cognitive�Failure?� 511�

portion�of�securities�that�could�obtain�a�rating�of�AA�or�AAA.17� In� this� endeavor,� they� received� close� cooperation� from� rating� agency� staff.� The� high� ratings� allowed� these� securities� to� be� sold�to�a�broad�spectrum�of�institutional�investors�at�relatively� low� interest� rates.� As� it� turned� out,� many� of� these� securities� subsequently�suffered�substantial�losses.� Frank�Raiter,�Standard�and�Poor’s�former�Managing�Director�

and� Head� of� Residential� Mortgage�Backed� Securities� (RMBS),� suggested�in�congressional�testimony�that,�with�the�best�model� ing� techniques,� his� rating� agency�might� have� begun� to� take� a� more�conservative�approach�to�rating�structured�mortgage�secu� rities� in�2003�or�2004.18�He�also�pointed�out� that�upgrading�his� agency’s�modeling� capability�would�have� added� costs�without� increasing� market� share.19� This� position� is� consistent� with� the� moral�failure�narrative.�Raiter�further�pointed�out,�however,�that� “[t]he�Managing�Director�of�the�surveillance�area�for�RMBS�did� not�believe�loan�level�data�was�necessary�and�that�had�the�effect� of� quashing� all� requests� for� funds� to� build� in�house� data� bases.”20�This�position�is�consistent�with�the�cognitive�narrative.� More�generally,�there�seems�to�be�evidence�of�both�moral�fail�

ure�and�cognitive�failure�at�credit�rating�agencies.�Morally,�cer� tain�internal�documents�from�various�credit�rating�agencies�indi� cate�that�at�least�some�employees�knew�of�problems�with�rating� methodology.21� Cognitively,� there� were� indications� of� a� belief� that�a�nationwide�housing�price�decline�would�never�occur.22� Most�notably,�regulators�appear�to�have�supported�the�use�of�

credit� rating� agencies.� Capital� regulations� explicitly� encour� aged�banks�to�hold�securities�rated�AA�or�AAA.�In�a�comment� letter�to�regulators,�Fannie�Mae�warned�that�the�use�of�ratings� on� untraded� securities� solely� for� regulatory� purposes� would� create�an� incentive�to�distort�ratings�because�the�ratings�agen� cies� would� be� accountable� only� to� the� creators� of� the� securi�

������������������������������������������������������������������������������������������������������������������ 17.�See�id.�at�2.� 18.�Id.�at�37–39� (statement�of�Frank�L.�Raiter,� former�Managing�Director,�Stan�

dard�&�Poor’s).� 19.�Id.�at�38.� 20.�Id.� 21.�See� id.� at� 2–4� (statement�of�Rep.�Henry�Waxman,�Chairman,�H.�Comm.�on�

Oversight�and�Gov’t�Reform).� 22.�See�id.�at�68�(testimony�of�Sean�J.�Egan,�Managing�Director,�Egan�Jones�Ratings).�

512� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

ties—not�to�any�buyers�in�the�market.23�Along�the�same�lines,�a� group� of� economists� that� regularly� provided� commentary� on� bank�regulatory�matters�wrote:�

[T]he�use�of�private�credit�ratings�to�measure�loan�risk�may� adversely�affect�the�quality�of�ratings.�If�regulators�shift�the� burden� of� assessing� the� quality� of� bank� loans� to� ratings� agencies,� those� regulators� risk� undermining� the� quality� of� credit�ratings� to� investors.�Ratings�agencies�would�have� in� centives�to�engage�in�the�financial�equivalent�of�“grade�infla� tion”� by� supplying� favorable� ratings� to� banks� seeking� to� lower� their� capital� requirements.� If� the�ratings�agencies�de� base�the�level�of�ratings,�while�maintaining�ordinal�rankings� of�issuers’�risks,�the�agencies�may�be�able�to�[avoid]�a�loss�in� revenue� because� investors� still� find� their� ratings� use� ful.�.�.�.�In�short,� if� the�primary�constituency�for�new�ratings� is�banks�for�regulatory�purposes�rather�than�investors,�stan� dards�are�likely�to�deteriorate.24�

Notwithstanding�this�commentary,�a�white�paper�recently�is� sued� by� the� regulatory� community� states:� “Market� discipline� broke� down� as� investors� relied� excessively� on� credit� rating� agencies.”25�This�statement�seems�to� imply�that� the�use�of� rat� ing�agencies�reflected�a�moral�failure�within�the�private�sector.� As�the�historical�record�demonstrates,�however,�cognitive�fail� ures�may�have�played�just�as�significant�a�role.�

III. COGNITIVE�FAILURES�IN�THE�REGULATORY�COMMUNITY�

Today,�we�know�that�certain�financial�practices�were�unsafe� and�unsound.�Mortgage�securities�were�created�without� suffi� cient� due� diligence� concerning� the� quality� of� the� underlying� loans.� Banks� were� able� to� use� structured� finance� and� off� balance�sheet�entities� to�reduce�regulatory�capital� for�risky�in� vestments.�Credit�default�swaps�created�excess�risk�concentra� tion.�At� the� time,� however,� regulators� viewed� all� of� these�de� velopments� positively.� The� regulatory� community� accepted,� and�even�encouraged,�mortgage�securities,� structured� finance,� off�balance�sheet�entities,�and�credit�default�swaps.� ������������������������������������������������������������������������������������������������������������������ 23.�Corine�Hegland,�Why�it�Collapsed,�NAT’L�J.,�Apr.�11,�2009,�at�12,�16.� 24.�SHADOW�FIN.�REGULATORY�COMM.,�REFORMING�BANK�CAPITAL�REGULATION�

(2000),�http://www.aei.org/article/16542.� 25.�U.S.� DEP’T� OF� THE� TREASURY,� FINANCIAL� REGULATORY� REFORM:� A�NEW�

FOUNDATION�2�(2009).�

No.�2]� Moral�Failure�or�Cognitive�Failure?� 513�

Regulators�considered�mortgage�securities�a�safer,�more�effi� cient�form�of�mortgage�finance�than�traditional�mortgage�lend� ing.�They�viewed�the�decline�of�the�savings�and�loan�industry� in� the� 1970s� and� 1980s� as� a� result� of� the� mismatch� between� short�term�deposits�and� long�term�mortgages.�Mortgage�secu� rities,� in� contrast,� seemed� to� avoid� this� shortcoming� because� they�could�be�placed�with�pension�funds�and�other�institutions� with�long�term�investment�horizons.� In�reality,� the�growth�of�mortgage�securitization�was�not�so�

benign.�Distortions� in� bank� capital� requirements� fueled�much� of�that�growth.�For�high�quality�mortgages�issued�and�held�by� banks,�capital�requirements�were�too�high.26�As�a�result,�banks� were�inhibited�from�undertaking�traditional�mortgage�lending.� To� compensate� for� the� disincentive� to� invest� in� mortgages� caused�by�high�capital�requirements,�regulators�permitted�banks� to� reduce� their� capital� requirements—but� only� for� mortgages� held�as�securities.�This�approach�had�a�perverse�effect.�In�addi� tion�to�lowering�the�capital�requirements�for�holding�safe�mort� gages� in� the� form� of� mortgage�backed� securities,� the� reduced� capital� requirements� for� securities� enabled� banks� to� hold� less� capital�for�risky�mortgages�as�well,�including�subprime�loans.� A� given� pool� of�mortgages,� for�which� a� bank�might� other�

wise�be�required�to�hold�four�percent�capital�(that�is,�$4�in�capi� tal� for� each� $100� in�mortgage�principal),� could�be� carved� into� tranches,�each�with�a�separate�capital�requirement,�based�on�its� rating�by�a�credit�rating�agency.�When�added�together,�the�sum� of�these�capital�requirements�would�be�less�than�three�percent.� Banks� were� also� able� to� dodge� capital� requirements� alto�

gether� by� putting� mortgage� securities� into� off�balance� sheet� entities.�Known�as�Structured� Investment�Vehicles,� these�enti� ties�issued�short�term�commercial�paper�to�fund�their�holdings� of�mortgage�securities.�A�line�of�credit�from�the�bank�backed�the� commercial�paper,�but�because�the�line�of�credit�was�in�force�for� less�than�a�year,�no�capital�was�required�for�regulatory�purposes.� Regulators�clearly�were�aware�of�this�regulatory�capital�arbi�

trage.27� Fannie� Mae� and� Freddie� Mac� complained� in� January� 2002� about� the� potential� for� regulatory� capital� arbitrage� in�

������������������������������������������������������������������������������������������������������������������ 26.�See�David� Jones,�Emerging� problems�with� the� Basel�Capital�Accord:�Regulatory�

capital�arbitrage�and�related�issues,�24�J.�BAN.�&�FIN.�35,�36–37�(2000).� 27.�See�id.�at�48–49.�

514� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

comments�about� rules� that�gave�official� sanction� to� the�use�of� ratings�to�reduce�capital�requirements�on�mortgage�securities.28� Regulators� also� were� aware� of� the� banks’� growing� use� of�

credit� derivatives,� such� as� credit� default� swaps,� to� transfer� away� risk.� Today,� the� regulatory� community� refers� to� the� in� vestment�banks�and�insurance�companies�that�absorbed�credit� risk� as� the� “shadow� banking� system,”� suggesting� a� financial� network�that�was�stealthy,�if�not�downright�illicit.�At�the�time,� however,� lending� regulatory� authorities� acknowledged� and� even�applauded�the�use�of�these�techniques.�In�fact,�regulators� were�proud�of�the�role�they�played�in�stimulating�and�spread� ing�these�innovations.� For� example,� in� June� 2006,� Federal� Reserve� Chairman� Ben�

Bernanke�said:� � The� evolution� of� risk�management� as� a� discipline� has� thus� been� driven� by�market� forces� on� the� one� hand� and� develop� ments�in�banking�supervision�on�the�other,�each�side�operating� with� the� other� in� complementary� and� mutually� reinforcing� ways.�Banks�and�other�market�participants�have�made�many�of� the�key�innovations�in�risk�measurement�and�risk�management,� but�supervisors�have�often�helped�to�adapt�and�disseminate�best� practices�to�a�broader�array�of�financial�institutions.�.�.�.�

� The�interaction�between�the�private�and�public�sectors�in� the� development� of� risk�management� techniques� has� been� particularly�extensive�in�the�field�of�bank�capital�regulation,� especially�for�the�banking�organizations�that�are�the�largest,� most�complex,�and�most�internationally�active.�.�.�.�

.�.�.�.�

.�.�.�Moreover,�the�development�of�new�technologies�for�buy� ing�and�selling�risks�has�allowed�many�banks�to�move�away� from�the�traditional�book�and�hold�lending�practice�in�favor� of�a�more�active�strategy�that�seeks�the�best�mix�of�assets�in� light�of�the�prevailing�credit�environment,�market�conditions,� and�business� opportunities.�Much�more� so� than� in� the�past,� banks�today�are�able�to�manage�and�control�obligor�and�port� folio�concentrations,�maturities,�and�loan�sizes,�and�to�address� and�even�eliminate�problem�assets�before� they� create� losses.� Many�banks�also�stress�test�their�portfolios�on�a�business�line� basis�to�help�inform�their�overall�risk�management.�

������������������������������������������������������������������������������������������������������������������ 28.�Hegland,�supra�note�23,�at�16.�

No.�2]� Moral�Failure�or�Cognitive�Failure?� 515�

� To�an�important�degree,�banks�can�be�more�active�in�their� management�of�credit�risks�and�other�portfolio�risks�because� of�the�increased�availability�of�financial�instruments�and�ac� tivities�such�as�loan�syndications,�loan�trading,�credit�deriva� tives,� and� securitization.�For� example,� trading� in� credit�de� rivatives� has� grown� rapidly� over� the� last� decade,� reaching� $18�trillion�(in�notional�terms)�in�2005.�The�notional�value�of� trading�in�credit�default�swaps�on�many�well�known�corpo� rate�names�now�exceeds�the�value�of�trading�in�the�primary� debt�securities�of�the�same�obligors.29�

At� about� the� same� time,� the� International� Monetary� Fund� wrote� that� “[t]here� is� growing� recognition� that� the�dispersion� of�credit�risk�by�banks�to�a�broader�and�more�diverse�group�of� investors,� rather� than�warehousing� such� risk� on� their� balance� sheets,� has� helped� to�make� the� banking� and� overall� financial� system�more�resilient.”30� Regulators�were�aware�of�the�ways�that�banks�were�using�se�

curitization,� agency� ratings,� off�balance�sheet� financing,� and� credit�default� swaps� to� expand�mortgage� lending�while�mini� mizing�the�capital�necessary�to�back�such�risks.�Like�the�bank� ers� themselves,� the� regulators� believed� that� these� innovations� were�making�financial�intermediation�safer�and�more�efficient.�

IV. CAPITAL�REGULATION�AS�A�FUNDAMENTAL�� CAUSE�OF�THE�CRISIS�

Capital� regulations� played� a� fundamental� role� in� fostering� the�behavior�that�created�the�financial�crisis.�They�discouraged� traditional�mortgage�lending�and�instead�encouraged�securitiza� tion.� They� created� a� role� for� credit� rating� agencies� to� enable� banks� to� take� credit� risk� on� mortgages,� including� subprime� mortgages,� without� having� to� hold� the� requisite� capital.� And� they�allowed�banks�to�further�reduce�capital�by�undertaking�the� transactions�that�we�now�think�of�as�“shadow�banking,”�includ� ing�structured�investment�vehicles�and�credit�default�swaps.� Bank� capital� regulation�made� traditional�mortgage� origina�

tion�of�low�risk�loans�uneconomical�in�comparison�with�securi� ������������������������������������������������������������������������������������������������������������������ 29.�Ben� S.� Bernanke,� Chairman,� Fed.� Reserve,� Speech� at� the� Stonier� Graduate�

School:�Modern�Risk�Management�and�Bank�Supervision�(June�12,�2006),�available� at�http://www.federalreserve.gov/newsevents/speech/Bernanke20060612a.htm.� 30.�INT’L�CAPITAL�MKTS.�DEP’T,�INT’L�MONETARY�FUND,�GLOBAL�FINANCIAL�STA�

BILITY�REPORT�51�(2006).�

516� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

tization.�Banks�were�thus�discouraged�from�simply�originating� and�holding� low�risk�mortgages.� Instead,� they�were�rewarded� for� holding�mortgage� loans� in� the� form� of� securities,�without� regard�to�how�or�by�whom�those�loans�were�originated.� Capital�regulations�also�shifted�focus�away�from�the�risk�on�

the�underlying�mortgages�and� instead�put� emphasis�on�grad� ing�by�credit�rating�agencies�of�slices�of�mortgage�backed�secu� rities.�The�quality�of� the�underlying� loans�grew�progressively� worse,� and� originators� relaxed� the� requirements� for� down� payments,� extended� eligibility� to� borrowers� with� more� trou� bled�credit�histories,�and�abolished�requirements�for�borrowers� to�provide�documentary�proof�of� their� income,�assets,�and�em� ployment�status.�None�of�this�deterioration�in�loan�quality,�how� ever,�kept�financial�engineers�from�carving�AA�rated�and�AAA� rated� mortgage� security� tranches� out� of� loan� pools.� In� turn,� banks�were�eager�to�supply�funds�to�fuel�the�housing�boom.� Moreover,�capital�regulations�created�a�situation�in�which�the�

banking� system� became� highly� fragile.� Because� of� regulatory� capital� arbitrage,� banks� were� not� required� to� hold� sufficient� capital�relative�to�the�risks�that�they�were�taking.�When�the�cri� sis� hit,� there� were� consequently� justifiable� doubts� about� the� solvency�of�many�large�banks,�which�in�turn�caused�a�freeze�in� inter�bank�lending.�If�banks�instead�had�been�required�to�hold� sufficient�capital�reserves,�an�adverse�shock�would�have�raised� fewer�questions�about�bank�solvency.� Additionally,� capital� regulations� stimulated� the� use� of� struc�

tured� investment� vehicles� and� credit� default� swaps,� enabling� banks�to�present�a�lower�risk�profile.�At�the�time,�regulators�were� pleased�with�the�way�these�instruments�were�reconfiguring�credit� risk.� When� the� crisis� hit,� however,� regulators� were� just� as� tor� mented�by�risks�embedded�in�the�large�position�in�credit�default� swaps�at�AIG�or�the�off�balance�sheet�entities�of�the�leading�inter� national�banks�as�they�would�have�been�had�those�risks�been�on� the�books�of� the�banks.�Officials�at� the�Fed�and�at� the�Treasury� found�themselves�confronted�by�the�sorts�of�domino�effects�and� bank�runs�that�they�thought�had�long�since�been�made�impossible� by�deposit�insurance�and�other�market�developments.� Lastly,�capital�regulations�encouraged�cyclicality.�Assets�main�

tained�high�ratings�during�the�boom,�but�were�downgraded�when� the�housing�market�turned.�This�reversal�forced�banks�to�sell�as� sets�to�restore�regulatory�capital.�Those�asset�sales,�however,�fur� ther�depressed�asset�values,�which�meant�that�banks�had�to�mark�

No.�2]� Moral�Failure�or�Cognitive�Failure?� 517�

down�their�equity�even�further.� In�other�words,�during�a�boom,� the�value�of�bank�capital�may�have�seemed�higher�than�it�really� was,�and�during�the�crash�the�value�of�bank�capital�may�have�ap� peared�lower�than�it�really�was.�In�view�of�the�way�things�worked� out,�several�economists�have�proposed�countercyclical�capital�re� quirements�designed�to�mitigate�these�effects.31�

V. HOUSING�POLICY�

Capital� regulations�were� the� primary� locus� of� cognitive� er� rors�leading�to�the�financial�crisis,�but� it� is�worth�commenting� on�the�role�that�housing�policy�played.�The�irrational�efforts�to� promote� home� ownership� certainly� contributed� to� the� boom� and�crash�in�the�housing�market.�The�proportion�of�households� in� the� United� States� owning� their� dwellings� rose� from� sixty� four� percent� in� 1994� to� sixty�nine� percent� in� 2006.32� Among� politicians,�there�was�bipartisan�pride�in�this�development.�The� policies� that� pushed� up� the� home� ownership� rate,� however,� were�rather�questionable�in�retrospect.� In�particular,� the�Com� munity�Reinvestment�Act33� and� regulatory� oversight� of� Fannie� Mae�and�Freddie�Mac�were�used�to�impose�quotas�on�lenders�in� segments� of� the� housing� market� where� households� had� diffi� culty�affording�the�homes�that�they�were�buying.�Moreover,�the� policies� did� not� distinguish� owners� from� speculators,� and� the� proportion�of� loans� for�non�owner�occupied�housing�rose� from� five�percent�in�the�1990s�to�fifteen�percent�in�2005�and�2006.34� Increasing�home�ownership�also�encouraged�costly�mortgage�

indebtedness.�Arguably,�there�are�positive�externalities�associ� ated�with�having�people�own�rather�than�rent�their�dwellings.� But�a�high�ratio�of�mortgage�debt�to�house�price�is,�if�anything,� a� negative� externality,� because� it� reduces� the� stability� of� the� housing�market.�Public�policy� is�nevertheless�heavily�commit� ted� to� subsidizing�mortgage� indebtedness� through� the� income� tax�deductibility�of�mortgage�interest,�direct�federal�subsidies�in� ������������������������������������������������������������������������������������������������������������������ 31.�See,�e.g.,�Charles�Wyplosz,�The�ICMB�CEPR�Geneva�Report:�“The�Future�of�Financial�

Regulation,”�VOXEU,�Jan.�27,�2009,�http://www.voxeu.eu/index.php?q=node/2872.� 32.�U.S.� Census� Bureau,� Housing� Vacancies� and� Home� Ownership,� tbl.14,�

http://www.census.gov/hhes/www/housing/hvs/historic/index.html� (last� visited� Feb.�11,�2010).� 33.�Pub�L.�No.�95�128,�91�Stat.�1147�(codified�at�12�U.S.C.�§§�2901–2908�(2006)).� 34.�Robert�B.�Avery,�Kenneth�P.�Brevoort�&�Glenn�B.�Canner,�The�2006�HMDA�

Data,�93�FED.�RES.�BULL.�A73,�A87�(2007).�

518� Harvard�Journal�of�Law�&�Public�Policy� [Vol.�33�

the�Federal�Housing�Administration� and�Veterans�Affairs,� and� indirect�federal�subsidies�through�Fannie�Mae�and�Freddie�Mac,� which� enjoyed� special� status� as� Government�Sponsored� Enter� prises.�Had�there�not�been�such�political�support�for�home�own� ership� and� mortgage� subsidies,� the� housing� cycle� probably� would� have� been�much� less� severe,� and� this� mitigation� could� have�interrupted�one�of�the�key�triggers�of�the�financial�crisis.�

VI. THE�ISSUE�OF�NARRATIVE�

The�ultimate�outcome�of� the� financial�crisis�will�be�visible� in� the� high� school� history� textbooks� of� the� future.� If� those� books� convey� the� causes� of� the� crisis� only� in� terms� of�moral� failure,� then�as�a�society�we�will�have�entrenched�a�historical�narrative� that� is� excessively� skeptical� of�markets� and� excessively� credu� lous�of�the�effectiveness�of�regulation.� The�narrative�of�moral�failure�is�attractive�for�many�reasons.�

First,� for� those�who� are� inclined� to�distrust�markets� and� sup� port�vigorous�government�intervention,�the�narrative�provides� reinforcement�of�those�prejudices.�Second,�it�is�a�narrative�with� clear�villains,� in� the� form�of�greedy� financial� executives.� Such� villains�always�make�a�story�more�emotionally�compelling.�Fi� nally,�the�narrative�provides�a�comforting�resolution:�Once�we� reorganize� and� reinvigorate� the� regulatory� apparatus,�we� can� rest�assured�that�the�crisis�will�not�recur.� The�narrative�of�cognitive�failure�is�not�so�comforting.�Rather�

than� identifying� villains,� this� narrative� sees� the� crisis� as� the� outcome� of� mistakes� by� well�intentioned� people,� including� both�financial�executives�and�regulators.�Moreover,�this�narra� tive� carries� with� it� the� implication� that� human� fallibility� will� persist,� and�so�we�cannot�be� confident� that� regulatory� reform� can�make�our�financial�system�crisis�proof.� The� narrative� of� cognitive� failure� suggests� a� need� for� greater�

humility� on� the� part� of� policymakers.� They� should� perhaps� re� think�the�push�for�greater�home�ownership,�particularly�to�the�ex� tent� that� the�push�encourages�people� to�borrow�nearly�all�of� the� money�necessary�to�finance�the�purchase�of�a�home.�They�might� even�want�to�reconsider�the�corporate�income�tax,�which�penalizes� equity�relative� to�debt,� creating�an� incentive� for�banks�and�other� firms�to�look�for�ways�to�maximize�their�use�of�debt�relative�to�eq� uity.�Above�all,� the�public�should�not�be�deceived� into�believing� that�regulatory�foresight�can�be�as�keen�as�regulatory�hindsight.�