1、1外文题目SIXWAYSCOMPANIESMISMANAGERISK出处2009(3)HARVARDBUSINESSREVIEW作者RSTULZ原文SIXWAYSCOMPANIESMISMANAGERISKASINVESTORSTOTUPTHEIRLOSSESFROMTHENANCIALCRISIS,MANYWILLBEASKINGTHEMSELVES,HOWDIDWALLSTREETMESSUPSOBADLYWHATWENTWRONGWITHALLTHOSECOMPLICATEDMODELSEVENBACKINNOVEMBER2007,BEFORETHECRISISHADREALLYHITT
2、HESTOCKMARKETS,ONECOMMENTATORINTHEFINANCIALTIMESWROTE,“ITISOBVIOUSTHEREHASBEENAMASSIVEFAILUREOFRISKMANAGEMENTACROSSMOSTOFWALLSTREET”OFCOURSE,NANCIALINSTITUTIONSCANSUFFERSPECTACULARLOSSESEVENWHENTHEIRRISKMANAGEMENTISRSTRATETHEYARE,AFTERALL,INTHEBUSINESSOFTAKINGRISKSWHENRISKMANAGEMENTDOESFAIL,HOWEVER,
3、ITISINONEOFSIXBASICWAYS,NEARLYALLOFTHEMEXEMPLIEDINTHECURRENTCRISISSOMETIMESTHEPROBLEMLIESWITHTHEDATAORMEASURESTHATRISKMANAGERSRELYONSOMETIMESITRELATESTOHOWTHEYIDENTIFYANDCOMMUNICATETHERISKSACOMPANYISEXPOSEDTOFINANCIALRISKMANAGEMENTISHARDTOGETRIGHTINTHEBESTOFTIMESINTHEFOLLOWINGPAGESILLEXPLORETHESIXPA
4、THSTOFAILUREINDETAILDISTRIBUTIONOFHOUSEPRICECHANGESWOULDGRAPHICALLYMAPABELLCURVEAROUNDTHEMEANCHANGETHEHORIZONTALAXISOFTHEGRAPHWOULDMEASUREPRICECHANGES,ANDTHEVERTICALAXISWOULDMEASURETHEPROBABILITYTHATEACHCHANGEWOULDOCCURTHESHAPEOFTHEBELLWOULDBEDETERMINEDBYVOLATILITYTHEGREATERTHEVOLATILITY,THEATTERAND
5、WIDERTHECURVEYOURMODELINGEXERCISECOULDHAVEGONEWRONGINTWOWAYSFIRST,IFHOUSEPRICEVOLATILITYWASHIGHERINTHEFUTURETHANINTHEPASTWHICHPROVEDTOBETHECASE,YOUWOULDHAVESUBSTANTIALLYUNDERESTIMATEDTHEPROBABILITYOFAPLUNGEINPRICESYOURBELLCURVESHOULDHAVEBEENATTERANDWIDER,BECAUSETHEPROBABILITIESOFLARGERPRICEMOVEMENTS
6、UPORDOWNWEREGREATERTHANYOUTHOUGHTTHEYWOULDBEANDIFYOUHADOVERESTIMATEDTHEMEANCHANGE,YOURCURVEWOULDBETOOFARTOTHERIGHT,GENERATINGFAULTYESTIMATESOFTHEPROBABILITIESOFALLPOSSIBLECHANGES2SECOND,YOUMIGHTHAVEWRONGLYASSUMEDTHATTHEDISTRIBUTIONOFFUTUREHOUSEPRICECHANGESWASDESCRIBEDBYTHEBELLCURVEBUTIFCHANGESINPRIC
7、EWERENOTNORMALLYDISTRIBUTED,THENAGRAPHOFTHEDISTRIBUTIONMIGHTBESKEWEDJUSTASTHEPROBABILITYDISTRIBUTIONOFACOINSFALLINGHEADSUPWOULDBESKEWEDIFTHECOINWERESLIGHTLYBENTBOTHKINDSOFERRORILLUSTRATEHOWDIFCULTITISTOUSEPASTDATATOPREDICTTHEFUTURESUPPOSEYOUHADLOOKEDAT30YEARSWORTHOFHOUSEPRICECHANGESTOESTIMATEYOURMEA
8、NANDSTANDARDDEVIATIONPERHAPSTHATTIMEFRAMEWASNTLONGENOUGHESPECIALLYIFTHETWODECADESPRECEDINGITWEREMOREVOLATILEABELLCURVEPROJECTEDFROM50YEARSWORTHOFDATAWOULDHAVEBEENATTERTHANTHEONEYOUPLOTTED,ORPERHAPSLESSFARTOTHERIGHTORSUPPOSEA50YEARANALYSISSHOWEDTHATLARGECHANGESWEREMORELIKELYTHANWITHTHENORMALDISTRIBUT
9、ION,BUTA30YEARANALYSISDIDNOTINTHATCASETHE30YEARANALYSISWOULDUNDERSTATETHERISKOFAPLUNGEINPRICESCOMPAREDWITHTHE50YEARANALYSIS,BECAUSEWITHTHELATTERTHEDISTRIBUTIONWOULDHAVEFATTAILSRELATIVETOTHENORMALDISTRIBUTION1RELYINGONHISTORICALDATARISKMANAGEMENTMODELINGUSUALLYINVOLVESEXTRAPOLATINGFROMTHEPASTTOFORECA
10、STTHEPROBABILITYTHATAGIVENRISKWILLMATERIALIZELETSASSUMEYOUWEREABANKSRISKMANAGERIN2006,ANDYOUWEREWORRIEDABOUTTHECHANCESTHATREALESTATEPRICESWOULDPLUNGEOVERTHECOMINGYEARYOURBANKSTOPEXECUTIVESNEEDEDTOKNOWHOWLIKELYSUCHAPLUNGEWASANDWHATLOSSESITWOULDCAUSEINORDERTODECIDEHOWMUCHEXPOSURETHEBANKSHOULDHAVETOREA
11、LESTATEPRICESYOUWOULDHAVEBEGUNBYEXAMININGTHEHISTORICALVOLATILITYOFHOUSEPRICESANDTHENCALCULATINGTHEONEYEARMEANPRICECHANGEANDITSSTANDARDDEVIATION,ONTHEASSUMPTIONTHATTHEYWOULDPROVIDEAGOODAPPROXIMATIONGOINGFORWARDTHEDATAMIGHTHAVELEDYOUTOCONCLUDETHATHOUSEPRICEMOVEMENTSARERANDOM,LIKETHEOUTCOMEOFACOINTOSST
12、HATINCREASESANDDECREASESOFTHESAMESIZEHAVETHESAMELIKELIHOODOFOCCURRINGANDTHATSMALLCHANGESAREMUCHMORELIKELYTHANLARGERONESIFALLTHISWERETRUE,THENTHEPROBABILITYYOURMODELINGEXERCISECOULDHAVEGONEWRONGINTWOWAYSFIRST,IFHOUSEPRICEVOLATILITYWASHIGHERINTHEFUTURETHANINTHEPASTWHICHPROVEDTOBETHECASE,YOUWOULDHAVESU
13、BSTANTIALLYUNDERESTIMATEDTHEPROBABILITYOFAPLUNGEINPRICESYOUR3BELLCURVESHOULDHAVEBEENATTERANDWIDER,BECAUSETHEPROBABILITIESOFLARGERPRICEMOVEMENTSUPORDOWNWEREGREATERTHANYOUTHOUGHTTHEYWOULDBEANDIFYOUHADOVERESTIMATEDTHEMEANCHANGE,YOURCURVEWOULDBETOOFARTOTHERIGHT,GENERATINGFAULTYESTIMATESOFTHEPROBABILITIE
14、SOFALLPOSSIBLECHANGESSECOND,YOUMIGHTHAVEWRONGLYASSUMEDTHATTHEDISTRIBUTIONOFFUTUREHOUSEPRICECHANGESWASDESCRIBEDBYTHEBELLCURVEBUTIFCHANGESINPRICEWERENOTNORMALLYDISTRIBUTED,THENAGRAPHOFTHEDISTRIBUTIONMIGHTBESKEWEDJUSTASTHEPROBABILITYDISTRIBUTIONOFACOINSFALLINGHEADSUPWOULDBESKEWEDIFTHECOINWERESLIGHTLYBE
15、NTBOTHKINDSOFERRORILLUSTRATEHOWDIFCULTITISTOUSEPASTDATATOPREDICTTHEFUTURESUPPOSEYOUHADLOOKEDAT30YEARSWORTHOFHOUSEPRICECHANGESTOESTIMATEYOURMEANANDSTANDARDDEVIATIONPERHAPSTHATTIMEFRAMEWASNTLONGENOUGHESPECIALLYIFTHETWODECADESPRECEDINGITWEREMOREVOLATILEABELLCURVEPROJECTEDFROM50YEARSWORTHOFDATAWOULDHAVE
16、BEENATTERTHANTHEONEYOUPLOTTED,ORERHAPSLESSFARTOTHERIGHTORSUPPOSEA50YEARANALYSISSHOWEDTHATLARGECHANGESWEREMORELIKELYTHANWITHTHENORMALDISTRIBUTION,BUTA30YEARANALYSISDIDNOTINTHATCASETHE30YEARANALYSISWOULDUNDERSTATETHERISKOFAPLUNGEINPRICESCOMPAREDWITHTHE50YEARANALYSIS,BECAUSEWITHTHELATTERTHEDISTRIBUTION
17、WOULDHAVEFATTAILSRELATIVETOTHENORMALDISTRIBUTIONEVENIFYOUCOULDHAVECAPTUREDTHEFULLIMPACTONYOURBALANCESHEETOFASIGNICANTDOWNTURNINREALESTATE,YOUWOULDHAVEBEENVULNERABLETOANOTHERKINDOFERRORINESTIMATINGTHEINDIRECTEFFECTSYOURBANK,LIKEANYOTHERNANCIALINSTITUTION,WOULDHAVEHADPOSITIONSINMANYASSETCLASSESWHOSEPR
18、ICECHANGESCORRELATEDWITHCHANGESINREALESTATEPRICEMOVEMENTSINALLASSETCLASSESCORRELATEWITHONEANOTHERMOREORLESS,ANDHECORRELATIONSHAVETOBEESTIMATEDCORRECTLYIFYOUARETODETERMINEYOURBANKSFULLEXPOSURETOAGIVENRISKTHEHIGHERTHECORRELATIONS,THERISKIERYOURTOTALPORTFOLIOOFASSETSANDTHEREFORETHEMOREEQUITYCAPITALYOUW
19、ILLNEEDTOPROTECTYOURINSTITUTIONAGAINSTLOSSESOBVIOUSLY,INESTIMATINGCORRELATIONSFROMHISTORICALDATAYOUWOULDHAVEBEENVULNERABLETOALLTHEPROBLEMSWEHAVEJUSTDESCRIBEDTOCOMPLICATEMATTERSFURTHER,CORRELATIONSARENOTCONSTANTINFACT,THEYAREKNOWNTOINCREASEDURINGPERIODSOFCRISISTHUS,INCALCULATINGTHEOVERALLRISKINESSOFY
20、OURBANKSPOSITIONS,YOUWOULDALSOHAVEHADTOMODELTHEPROBABILITYDISTRIBUTIONSOFALL4THECORRELATIONSACROSSYOURVARIOUSPORTFOLIOSASWELLASTHEPRICEDISTRIBUTIONSINEACHASSETCLASS,FURTHERINCREASINGTHELIKELIHOODOFERRORALLTHISILLUSTRATESTHEINCONVENIENTTRUTHTHATRAPIDNANCIALINNOVATIONOVERRECENTDECADESHASMADEHISTORYANI
21、MPERFECTGUIDEFINALLY,ADAILYMEASUREDOESNTCAPTURETHERISKOFAPORTFOLIOWHENTHERMISSTUCKWITHTHEPORTFOLIOFORAMUCHLONGERPERIODDAILYVARMEASURESASSUMETHATASSETSCANBESOLDQUICKLYORHEDGED,SOARMCANLIMITITSLOSSESWITHINADAYBUTASWEHAVESEENIN2008ANDASWESAWINOTHERCRISESSUCHASTHEONEIN1998,ADRAMATICWITHDRAWALOFLIQUIDITY
22、FROMTHEMARKETSLEAVESRMSEXPOSEDFORWEEKSORMONTHSONPOSITIONSTHEYCANNOTEASILYUNWINDTHEMARKETFORMANYCDOSBACKEDINPARTBYSECURITIZEDSUBPRIMEMORTGAGESESSENTIALLYVANISHED,SOBANKSWITHSUCHCDOSONTHEIRBALANCESHEETSCOULDNTUNLOADTHEMEXCEPTATRESALEPRICESTHERISKOFTHOSECDOSDURINGTHATPERIODWASUNRELATEDTOADAILYARRATHER,
23、ITLASTEDASLONGASTHEMARKETFORHEMWASMORIBUNDANDRISKINCREASESOVERTHETIMEHORIZONFORWHICHITISCOMPUTEDTHEREARESEVERALOTHERPROBLEMSWITHTHISAPPROACHFIRST,THESIMPLEFACTTHATUBSEXPERIENCEDALLTHOSEVAROVERRUNSSAYSLITTLEABOUTTHECOMPANYSACTUALNANCIALHEALTHTHEOVERRUNSCOULDHAVEBEENSMALL,ANDARAPIDINCREASEINVOLATILITY
24、COULDHAVECREATEDMANYLARGEGAINSASWELLALTERNATIVELY,UBSCOULDHAVEREALIZEDMANYLARGELOSSESANDONLYAFEWLARGEGAINSINTHEFORMERCASEITMIGHTACTUALLYBEAHEADATTHEENDOFTHEYEARINTHELATTERCASEITMIGHTBEINSERIOUSTROUBLE5外文题目SIXWAYSCOMPANIESMISMANAGERISK出处2009(3)HARVARDBUSINESSREVIEW作者RSTULZ译文六种公司管理风险由于投资者的损失高过经济风险带来的危
25、机,许多人都会问自己,华尔街是如何陷入困境而变得如此糟糕的呢出了什么差错,所有那些复杂的模式即使早在2007年11月之前也没有如此的严重,这场危机确实在股票市场里蔓延,一位评论家在金融时报上写道“很明显华尔街出现了一个巨大的风险管理的失败。“当然,金融机构可能遭受最惨痛的的损失,及时当他们的风险人是行业内最为一流的也不能幸免遇难。他们毕竟是在为企业承担风险。华尔街的风险管理确确实实的失败了,但是,它只是度过了六分之一的危险,金融危机并没有展现出他所有的恐怖之处。有时,问题在于数据或措施的不正确,风险管理的措施是否依靠得住。有时,它涉及到如何识别和沟通风险管理措施对于一个金融企业是最不安全的。财
26、务风险管理,其实是最难进行风险管理控制的。在下面的页面,我将详细探讨六个可能导致失败的路径。1。建立模型模依托历史数据的风险管理通常涉及概率,推断从过去到现在预测一个给定的风险,那么这样的计算是一定会实现的。让我们假设你是一家银行任职的风险管理经理,你要担心的机会可能会流失,房地产价格将下跌在未来的一年。您的银行的最高执行将延迟而你需要知道有可能这样的暴跌将会带来怎么样的损失,它将导致决定多少银行破产这些你作为风险管理经理都应当了解。你会开始通过审查房价的历史波幅然后计算一年的平均价格变化及其标准差,假设,他们将提供一个很确定的近似向前发展的趋势的一组数据。该数据可能导致您认为房价变动的结果,
27、像掷硬币的结果,这和减少有相同大小的发生的可能性相同,并认为小的变化更容易比大的可能性大一点。如果这一切是真的,那么概率将分配房屋价格变动的平均变化以图形德方式映射出周围的钟形曲线。该图的横轴将做衡量价格变动,垂直轴会显示每个测量的概率会发生变化的可能性。钟形状的波动将取决于更大的波动,曲线平坦,更广泛。而这些被证明模拟演习的方式6可能错了两个。首先,如果房价在未来变化得比过去更都(这被证明是如此),你会大大低估了价格的概率,概率可能会大幅下挫。您的钟形曲线平坦,本来应该做关系较大的价格变动(上涨或下跌均大于你认为他们会出现的那种情况。如果你高估平均变化,你的曲线会表现出波动太大,可能发生的变
28、化产生错误估计了所有的可能出现的情况。其次,你可能错误地认为,变化的分布情况对未来房价的描述钟形曲线是绝对正确的。但是,如果价格变动呈现出不正常分布式,那么分布图可能扭曲,就像概率存款保险如同一个硬币的下降行动将偏斜,如果投硬币的人略微发生了弯曲那么这种情况将不再出现。种种误差都说明是多么困难的利用过去的数据来预测未来。假设你有30年的时间看价格的价值,估计你的房子均值和标准差。也许那段时间不够长,尤其是如果两个十年前它是更不稳定的。那么按照钟形曲线的数据推测,从50年的价值将是正确的数值比你所策划出来的数据,或可能不太遥远的。或者,假设一个50年的分析表明,变化太大,更多的分销可能比正常的要
29、小,但30年的分析没有这样的变化。在这种情况下,30年的分析将低估现在的价格,在50年的分析,因为后者分布在经历大跌的风险面前将有正常的尾数分布,相对于历史数据矫正的困难,当你开始想动摇想去影响次级抵押贷款时却发现大量尚未行使。当许多业主发现邻里少有家庭资产(几乎总是次级借款人的情况下)和相当难度使他们的抵押金降低水平,取消抵押品赎回权可能性的下降房产价格下降,导致更多的抵押品赎回权的增多,因为业主有负资产。你和他的数据不能预言。此外,如果您的机构已具有复杂的证券持有期限不存在七种)支持您的数据作为付款的抵押债务行动(债务抵押债券,如由次级抵押贷款的一部分,你不能预料到价格影响房子的价值就在一
30、个数控范围之内。就算你可以完全理解房地产低迷会影响您资产负债表上表现,你会收到脆弱的另外估机错误的间接影响。您的银行交易,和其他金融机构一样,比如房地产类似的许多资产类别的交易,其价格的变化而变化的相关性。资产类别的价格走势都彼此相关,或多或少,以及相关性,要估计正确是很困难的,如果你要确定你的银行的充分暴露给给定的风险,那么风险越高,对应风险较高的资产组合也一定越多,因此,相对于总股本你更需要保护你的机构不受损失。从很明显的数据中不难发现,从历史的相关性估计分析你会被相关的问题所困惑,我们刚刚描述过,面对更加复杂的问题,相关性不是恒定的,事实上,它们只会增加麻烦在危机时期。因此,在计算银行上
31、的整体的风险时,你也不得不服从模型,以及所有的相关关系,系统的使用你的各种投资组合的价格分布在每个资产类别的7概率分布上,进一步增加了错误的可能性。所有这一切都说明了难以忽视的真相,迅速发展的金融创新在最近几十年取得了并不是不完善的发展。一个房价下降对你的资产价值将会有很大打击,你的数据会影响被错误估计,在大多数情况下,因为它们没有包括一个在此期间房地产市场可能会遇到的问题。2集中于使用创造的模型的数据来解决问题是不靠谱的。实际措施使您维持您风险管理的安全措施,证券交易,也可能导致你忽略的风险,你应该考虑到,每日价值在风险(VAR)的措施,是金融机构最常见的方式,也是买卖证券的风险评估的一种方式。VAR的主要措施,使用最高金额的钱,你可能会失去在给定的概率水平上所得到的。例如,A级风险值1亿美元在1意味着你只有一个,第二天1这一数额将意味着你失去更多的机会。如果银行指定一个上限的风险,那么它愿意接受,比如说,125亿美元,在这种情况下,将用1万美元去交换。银行申报每季在季度前的次数的损失风险值高于日报。如果一家银行风险措施水平在1,它超过其风险值大约是1的时间。在其2006年年度报告中,瑞士银行瑞银集团表示,它从未有过的亏损可能超过其每日的风险值。