caballero on bubbles

Upload: ecrcau

Post on 03-Apr-2018

213 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/28/2019 Caballero on Bubbles

    1/30

    BubblesMacroeconomicsIV

    RicardoJ.CaballeroMIT

    Spring2011

    R.J. Caballero (MIT) Bubbles Spring 2011 1 / 29

  • 7/28/2019 Caballero on Bubbles

    2/30

    1

    2

    3

    References

    Allen,F.andD.Gale,BubblesandCrises, EconomicJournal,110:236-255,2000.Tirole,J.,AssetBubblesandOverlappingGenerations, Econometrica,53,(6),1499-1528,November1985.AbreuD.andM.Brunnermeier,BubblesandCrashes, Econometrica,71:173-204,2003.

    R.J. Caballero (MIT) Bubbles Spring 2011 2 / 29

  • 7/28/2019 Caballero on Bubbles

    3/30

    Introduction

    Historical: DutchTulipmania,SouthSea... GreatCrashof1929SouthSeaBubble(1710-1720)

    IsaacNewton: 04/20/1720soldsharesat7,000,profiting3,500.Re-enteredthemarket later endedup losing20,000Icancalculatethemotionsoftheheavenlybodies,butnotthemadnessofpeople

    Japanboom-bust(a lostdecade);EMEs,Nasdaq,realestate(allaroundthedevelopedworld),commoditiesWheredotheycomefrom? Whattodoaboutthem?

    R.J. Caballero (MIT) Bubbles Spring 2011 3 / 29

  • 7/28/2019 Caballero on Bubbles

    4/30

    Mainpoints

    Twobroad(andpolar)views:There isashortageofstoreofvalue bubbleshelpfixingthisproblemAgentsmisbehave(eitheranagencyproblemorabehavioralproblem)

    Myview: Theseviewsaremore intertwinedthan itmayseemThe former isaboutmacroenvironmentswherethere isshortageofassetsThe latter isaboutthe locationofbubbles

    Other: irrationalexuberance andmoreformalbehavioralstoriesMyview: More likelytoarisewhentheaboveconditionsarepresent

    R.J. Caballero (MIT) Bubbles Spring 2011 4 / 29

  • 7/28/2019 Caballero on Bubbles

    5/30

    Riskshifting

    Allen-Gale(2000) BubblesandcrisesThere isapattern:

    Phase1: financial liberalizationorsomeexpansionarypolicy fuelsabubblehase

    2:

    the

    bubble

    bursts

    and

    asset

    prices

    collapse

    hase3: widespreaddefaultsby leveragedassetbuyers, leadingtoabankingnd/orexchange ratecrisis,andapersistentrecession

    PPa

    Mainingredient(thisisallwelldiscusshere): Uncertaintyaboutpayoffs(realorfinancialsector)can leadtobubbles inan intermediatedfinancialsystem(riskshifting/assetsubstitution)

    R.J. Caballero (MIT) Bubbles Spring 2011 5 / 29

  • 7/28/2019 Caballero on Bubbles

    6/30

    Model

    Twodates,t=1,2andasingleconsumptiongoodTwoassets:

    Safeand invariablesupplyata rateriskyand infixedsupply. Stochastic return isR perunit,withdensityh(R)ndsupport [0,R ]

    R

    a MAXThereturnonthesafeasset isdeterminedbymarginalproductofcapital:r =f(x)wherex areunitsoftheconsumptiongood(standardassumptionsonf)Non-pecuniaryconvexcostof investing inriskyassetsc(x)(torestrictportfoliosizesandtoensureequilibriumprofitforborrowers)

    R.J. Caballero (MIT) Bubbles Spring 2011 6 / 29

  • 7/28/2019 Caballero on Bubbles

    7/30

    Model

    There isacontinuumofsmall,riskneutral investors; idemforbanks.InvestorshavenowealthwhilebankshaveafixedamountB (whichtheysupply inelastically). Only investorsknowhowto invest,sobanks onlychoice isto lendto investorsBanksand investorsarerestrictedtousedsimpledebtcontracts(inparticular,theydontdependonsize)Since investorscanborrowasmuchastheywantatthegoingrate, inequilibriumthecontractedrateon loansmustbeequaltotheriskless interestrateSymmetriceq. All investorsare identicalex-post. Xs andXR aretherepresentative investorsholdingsofthesafeandriskyassets

    R.J. Caballero (MIT) Bubbles Spring 2011 7 / 29

  • 7/28/2019 Caballero on Bubbles

    8/30

    RiskshiftingBecausebanksusedebtcontractsandcannotobserve investmentdecisionsbyborrowers,the latterdoesnotbearthefullcostof investment iftheoutcome isbad,while itgetsthebenefit iftheoutcome isgoodIfrepresentative investorbuysXs andXR, itborrowsXs +PXR (whereP isrel. priceofriskyasset)andtherepayment(ifnotbankrupt) isr(Xs +PXR)The

    liquidation

    value

    of

    the

    portfolio

    is

    rXs +RXR,sothepayofffortheinvestor is:

    max{(RrP)XR,0}

    andthe

    decision

    problem

    is:

    RMAX

    max XR

    (RrP)h(R)dRc(XR)XR0 R=rP

    R.J. Caballero (MIT) Bubbles Spring 2011 8 / 29

  • 7/28/2019 Caballero on Bubbles

    9/30

    EquilibriumMarketclearingconditions:

    XR = 1XS +P = B

    r = f(XS),

    thefocsevaluatedattheequilibriumare:R

    MAX

    (RrP)h(R)dR =c(1)R=rP

    r =f(BP)fromwhichwecansolvefor(r,P)

    R.J. Caballero (MIT) Bubbles Spring 2011 9 / 29

  • 7/28/2019 Caballero on Bubbles

    10/30

    Wecanre-writethefocwrttoP toget:RMAX Rh(R)dR c(1)

    1 R=rP

    P =r

    Pr[RR]

    1 c(1)= E[R|RR]

    r

    Pr[R

    R

    ]

    Definethefundamentalasthepriceanagentwouldbewillingtopay intheabsenceofriskshifting,then:

    1 Pf = E[R]c(1)r

    It iseasytoshowthat,as longasPr[RR]>0,P>Pf

    P

    R.J. Caballero (MIT) Bubbles Spring 2011 10 / 29

  • 7/28/2019 Caballero on Bubbles

    11/30

    Proof

    rPPr[R R]=RMAX

    R=rPRh(R)dRc(1)

    =>

    rPf R

    0Rh(R)dR

    rPf (rP)(1Pr[RR])

    rP>rPf

    R.J. Caballero (MIT) Bubbles Spring 2011 11 / 29

  • 7/28/2019 Caballero on Bubbles

    12/30

    Finalremarks

    Hence,duetoriskshifting,P ishigherthanfundamental(bubble)Thecounterpartofthebubble isthebank losses,andhencetherestofthestory...Inasense it isnotaGEbubble,asthepriceofbanksshouldgodown... butitmaywellbethathouseholdsarestuck... thistakesustothestandardmodelofREbubbles inmacro,whichhighlightstheshortageofassets..

    R.J. Caballero (MIT) Bubbles Spring 2011 12 / 29

  • 7/28/2019 Caballero on Bubbles

    13/30

    Assetshortages

    Letsremoveuncertaintytohighlightthefactthatthenatureofthesebubbles isverydifferentfromtherisk-shiftingargumentReadTiroles1982"OnthepossibilityofspeculationunderRE"(EMA).... soyourealizethatrationalbubblesarenoteasytoget...ButweknowfromSamuelsons(1958)consumption-loanmodelthatbubbles (i.e. assetswithpositivepricebutno intrinsicvalue)canexist inOLGstructures(infinitenewtraders inthehorizon)andthattheycanbegood

    Money inSamuelsonsmodel,butnot for itstransactionservicebuttostorevalue. Paretogain from solvingdynamic inefficiency(nocapitalwastedtostorevalue).

    R.J. Caballero (MIT) Bubbles Spring 2011 13 / 29

  • 7/28/2019 Caballero on Bubbles

    14/30

    AbarebonesversionofSamuelsonsmodel

    OLG.Individuals livefortwoperiods,theyarebornwithanendowmentwtWhichtheysave in itsentiretyandonlyconsumewhenold(hencewecanindexthegenerationswelfarebyct,t+1) .There isnopopulationgrowth,buttheendowmentgrowsatarate.

    wt+1 = (1+)wtct,t+1 = (1+rt)wt

    What isthe interestrate inthiseconomy?

    R.J. Caballero (MIT) Bubbles Spring 2011 14 / 29

  • 7/28/2019 Caballero on Bubbles

    15/30

    Mothernature...Theanswerdependsonwhichassetsareavailabletostorevalue.Samuelsonfirstobservedthattheyoungcouldnotsaveby lendingtotheoldsincethe latterwillnotbearoundtorepaythem later(financialmarketincompleteness). Theonlyoptionoftheyoung istotradewithmothernature, i.e. to invest inphysicalcapital.Letssimplifythetechnologysideandassumethat ithasconstantreturns:. That is,oneunitofsavingsatt produce1+ att+1(wecouldhaveamorestandardf(k)... butmain insightswouldbeunchanged). Itfollowsthatthe interestrate inthiseconomymustbe:

    rt =

    andutility is:UMN = (1+)wtt

    R.J. Caballero (MIT) Bubbles Spring 2011 15 / 29

  • 7/28/2019 Caballero on Bubbles

    16/30

    B bbl

  • 7/28/2019 Caballero on Bubbles

    17/30

    Bubbles

    Morebroadly: themarketeconomy isoveraccumulatingcapitaltofacilitatestoreofvalueDoesthismeanthatthemarketeconomy issuboptimal? Notnecessarily.Naturally, if,themarketcanreachthesameallocationasthesocialcontract,providedweenlargethesavingoptionsoftheyoungto includeoneirreproducibleanduselessobjectwithpriceBt suchthat:

    Bt+1 = (1+rt)Bt

    R.J. Caballero (MIT) Bubbles Spring 2011 17 / 29

    W lf

  • 7/28/2019 Caballero on Bubbles

    18/30

    Welfare

    LetxB/w. ThenBt+1 (1+rt)Bt 1+rt

    xt+1 = = = xtwt+1 (1+)wt 1+

    Ifx

  • 7/28/2019 Caballero on Bubbles

    19/30

    Finalremarks

    TherearetwoPareto-rankablestationaryequilibria(bubblebetterthanfundamental);andacontinuumofnon-stationaryequilibriathatconvergetothefundamentalequilibriumthatprovide intermediatewelfare(note: alltheseequilibriacontainbubbles,butthesebecomesmallrelativetotheeconomy)Bubblesariseasaresultofcoordinationacrossdifferentgenerations. Butthisisjustoneofthepossibleequilibria,andhencethepossibilityofacrash islatent

    R.J. Caballero (MIT) Bubbles Spring 2011 19 / 29

    A di Ab d B i

  • 7/28/2019 Caballero on Bubbles

    20/30

    Appendix: AbreuandBrunnermeier

    Behavioralbiases leadtobubbles(theytakethisasgiven)Assumingthatrationalarbitrageursunderstandthatthemarketwilleventuallycollapse,willtheystillridethebubble?Delayedarbitragemodel(ridingthebubbleforawhilemaybeoptimal)[connectionwithearlierdiscussionon limitedarbitrage]Amodelofmarkettiming

    Dispersion inexitstrategiesmakesthebubblepossibleAtsome random timet0 pricesurpassesthe fundamentalvalue. Thereafter,rationalarbitrageursbecomesequentiallyawarethatthepricehasdepartedfrom fundamentals. Theydontknowwhethertheyareearlyor late relativetoothersBubbleburstswhenasufficientmassofarbitrageurshavesoldout(coordination)

    R.J. Caballero (MIT) Bubbles Spring 2011 20 / 29

    Th S t

  • 7/28/2019 Caballero on Bubbles

    21/30

    TheSetup

    InBubblesandCrashes, theydiscussanirrationalexuberance episodewhereaftersamerandomdatet0 thepricecontinuestoriseatsomerateg>r,whilethefundamentalonlyrisesatrThemaineconomicforces intheirEMApaperarealsofound intheirsimpler,JFE,paper:SynchronizationRiskandDelayedArbitrage (wewilldevelopthisone,althoughtheconnectionwithabubble is lessdirect)There isasingleriskyassetwithpricept andfundamentalvt. Priortothearrivalofashockatarandomtimet ,thefundamental value isert0 andafterthat(1+)ert,with takingvaluesand withequalprob.,andF(t0) =1e t0Priortotheshockatt0,pt =vt. Aftert0 thepricedeviatesfromfundamentalsuntilfullarbitragetakesplace(thecrash if,whichweassumehenceforth)

    R.J. Caballero (MIT) Bubbles Spring 2011 21 / 29

    The Setup

  • 7/28/2019 Caballero on Bubbles

    22/30

    TheSetup

    Therearetwotypesofagents: rationalarbitrageursandbehavioraltradersTheonlyroleofthe latteragents istosupportthemispricingandmaintainthepriceatpt =ert as longasthesellingpressurebyrationalarbitrageursliesbelowathreshold(.)Thefocusofthepaper isontheformeragents. Arbitrageursareex-anteidenticalbutreceive informationaboutthedeviationsequentially(uniformly)betweent0 andt0+An individualarbitrageurwho learnsaboutthechange infundamentalatti(denoted

    by

    ti)thinks

    that

    t0

    isdistributed

    between

    ti

    andt

    i

    R.J. Caballero (MIT) Bubbles Spring 2011 22 / 29

    The Setup

  • 7/28/2019 Caballero on Bubbles

    23/30

    TheSetup

    R.J. Caballero (MIT) Bubbles Spring 2011 23 / 29

    (1+b)e

    rt

    ert

    p,v

    p0= 1

    t0 t0 + hk0 t0 + h

    1/h

    t0 + tt

    Random

    starting

    point

    k0arbs areaware of

    the mispricing

    All arbitrageurs

    are aware of

    the mispricing

    Pricing correction

    for exogenous

    reasons

    +b

    -b (1_b)e

    rt

    Image by MIT OpenCourseWare.

    The Setup

  • 7/28/2019 Caballero on Bubbles

    24/30

    TheSetup

    Arbitrageursare

    risk

    neutral

    but

    the

    maximum

    short

    position

    is

    xi = 1.Thenormal/neutral position isx i =0. Departingfromthisbenchmark

    generates(large)holdingcostsofcpt|xi|Thepricecorrectionoccursassoonastheaggregateorder imbalanceofallarbitrageursexceeds(tt0),with(reducedformfrombehavioralagents)

    (tt0) =0[1(1/ )(tt0)][Ifthetradingorderexceeds,there isarandomizationofthepriceatwhichordersareexecuted]Motivation: The longerthemispricingpersists,thesmaller isthemassofbehavioraltradersthatremainconfidentthattheprice isright Sincetherearenopricechanges,arbitrageurscannot infert0 fromthemwhilepressure isbelow(.)

    R.J. Caballero (MIT) Bubbles Spring 2011 24 / 29

    Market Timing and Delayed Arbitrage

  • 7/28/2019 Caballero on Bubbles

    25/30

    MarketTimingandDelayedArbitrageArbitrageurti specifiesatradingstrategyasfunctionofi =tti. A-Bfocusontriggerstrategiessuchthatthearbitrageursetsxi =0untiladateti +i andxi =1afterthat(untilthecorrectiontakesplace)Anarbitrageurthattradesjustbeforethecorrectionachievesthehighestpayoff. Bypostponingthetradehereducesholdingcostsbutrisksmissingthearbitrageopportunity(Keynes: beatthegun terminology)Leth(t|ti)bearbitrageurtisperceivedhazardratethatthepricecorrectionoccurs inthenext instantt. Thus,hisestimateofacorrection inthenext(small)time interval ish(t|ti),whiletheholdingcost iscpt

    Thusthearbitrageurwillonlytrade iftheexpectedbenefitpth(t|ti)exceedstheexpectedcostofholdinganondiversifiedportfolio(1h(tti))cpt

    |Ofcoursethehazardratedependsonotherarbitrageurs tradingstrategies.A-Brestrictattentiontosymmetrictriggerstrategyequilibria(basedonEMAarticle)

    R.J. Caballero (MIT) Bubbles Spring 2011 25 / 29

    Abreu-Brunnermeier: Market Timing and Delayed

  • 7/28/2019 Caballero on Bubbles

    26/30

    Ifallarbitrageurstradewithadelay,thenthepricecorrectionoccursatt0+(),wherethe latter isdefined implicitlyfrom

    ( ) = +(())

    Usingthe linearexpressionfor(.),wehave+

    () = 0 +0

    Abreu-Brunnermeier: MarketTimingandDelayedArbitrage

    R.J. Caballero (MIT) Bubbles Spring 2011 26 / 29

    Market Timing and Delayed Arbitrage

  • 7/28/2019 Caballero on Bubbles

    27/30

    Arbitrageurti knowsthat, inequilibrium,thepricecorrectionwilloccurnolaterthanti +()butafterti +()Giventhepriordistributionont0,the latterobservationyieldsasimpleposterior:

    0 for t

  • 7/28/2019 Caballero on Bubbles

    28/30

    MarketTimingandDelayedArbitrage

    R.J. Caballero (MIT) Bubbles Spring 2011 28 / 29

    bh(ti + t|ti)

    c

    * j

    Arbitrageur ti's Hazard Rate

    Image by MIT OpenCourseWare.

    Final Remarks

  • 7/28/2019 Caballero on Bubbles

    29/30

    FinalRemarks

    Arbitrageis

    delayed.

    This

    is

    possible

    because

    mispricing

    is

    never

    common

    knowledge,whichpreservesthedisagreementaboutthetimingofpricecorrections

    Thearbitrageurwhobecomes immediatelyawareofthemispricingatt0 knowsthatatt0+ everybodyknowsaboutthemispricing. However,thetraderwhoonlybecomesawareatt0+ thinksthathemightbethefirsttohearofitandhedoesnotknow thatalltradersalreadyknow it. Hence,even ifeverybodyknowsofthemispricingatt0+,onlythefirsttraderknowsthateverybodyknowsAtt0+2,eventhe lasttraderknowsthateverybodyknows,buthedoesnotknowthateverybodyknowsthateverybodyknowsofthemispricing,andsoon

    Themaindistinctionwithnoise-tradersisthatmostoftheactioncomesfromtherationaltraders. It istheuncertaintyaboutthebehaviorofotherrationaltradersthat leadstodelayedarbitrage

    R.J. Caballero (MIT) Bubbles Spring 2011 29 / 29

  • 7/28/2019 Caballero on Bubbles

    30/30

    MIT OpenCourseWare

    http://ocw.mit.edu

    14.454 Economic CrisesSpring 2011

    For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

    http://ocw.mit.edu/http://ocw.mit.edu/termshttp://ocw.mit.edu/termshttp://ocw.mit.edu/