1、信贷衍生产品：建模和信贷衍生产品：建模和计算计算2022-6-22摘要摘要l引言引言l信贷产品信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品复杂信贷产品l定价，计算技术定价，计算技术和对冲(Hedge)l风险管理风险管理2022-6-22摘要摘要l引言引言l信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品l定价，计算技术和对冲(Hedge)l风险管理2022-6-22引言引言l风险市场（credit ris
2、k market)l公司债券l信贷衍生产品(credit derivatives)l一种能把信贷风险从基础资产中剥离出来，便于交易和管理的衍生工具l基本功能：风险有效转换，聚集，分散和重新包装2022-6-22信贷衍生产品的增长信贷衍生产品的增长单个信贷CDS信贷指数相关性波动性Source: ISDA2022-6-22市场的参与者市场的参与者Source:BBA 2003/2004 survey2022-6-22市场的参与者市场的参与者2022-6-22信贷衍生产品的创新Source:BBA 2003/2004 survey2022-6-22交易的产品？交易的产品？l信贷违约互换(Credi
3、t Default Swap, CDS)l完全收益互换(Total Return Swap, TRS)l信贷短期债券(Credit Linked Note, CLN)l信贷组合互换/债券First-to-default basket, nth-to-default basketl信贷组合批次债券l信贷产品创新Option, futures, indices, constant maturity etc.2022-6-22市场最新发展市场最新发展lCDS matching and confirmationlStandardisation of documentationlTradable Cre
4、dit fixingslMarket regulation2022-6-22摘要摘要l引言引言l信贷产品信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品l定价，计算技术和对冲(Hedge)l风险管理2022-6-22信贷违约互换信贷违约互换(CDS)Protection 买家Protection 卖家银行甲银行乙某公司债券 信贷溢价 credit spread违约事件发生时 100债券2022-6-22信贷短期债券信贷短期债券(CLN)Special Purpose Vehicle(SPV)息票+到期
5、本金（没有违约）6% + 100100（开始时）3.6% + 100100（开始时）利率互换2.4%CDS2022-6-22摘要摘要l引言引言l信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品复杂信贷产品l定价，计算技术和对冲(Hedge)l风险管理2022-6-22信贷组合第一个违约互换信贷组合第一个违约互换(FTD)Protection 买家Protection 卖家银行甲银行乙公司1债券公司5债券 信贷溢价 credit spread违约事件发生时 100违约债券2022-6-22Senior
6、Class ACreditTranchedSecurities SpecialPurposeVehicle 衍生债务抵押债券 (CDO)Diversified Pool of , typically, fixed income assets Credit Risk Transfer through: - Cash “True Sale” - Synthetic using “Credit Default Swaps” Assets may comprise: Investment Grade Bonds / Loans HY Bonds Leveraged Loans Emerging Mar
7、ket Debt ABS / MBSAssetsLiabilitiesTransfer Credit Risk Transfer for: - Balance Sheet Management - Credit Arbitrage Mezzanine Class B/C/DSubordinatedThe above is indicative capital structure onlyAAAAA toBBBNotRatedRatingsCDO : Collateralized “Debt” Obligations, more encompassingterm than other terms
8、 such as CBO (“Bonds”) and CLOs (“Loans”)2022-6-22衍生债务抵押债券的特点 Efficient Portfolio Diversification Tool Gain Access to Assets, Otherwise Difficult to Access Choose Tranche depending upon Risk Appetite Customized “Portfolio” meets Investors Requirements Higher Spread than similarly Rated Assets Invest
9、ors receive higher spread premium relative to single name investments for a similar level of risk2022-6-22摘要摘要l引言引言l信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品l定价，计算技术和对冲定价，计算技术和对冲(Hedge)l风险管理2022-6-22CDS公平溢价公平溢价l违约概率(p)l恢复率(recovery rate)lCDS 现金流Fixed legFloating legSSSS
10、iiniiiniiidfprdfppiniiiiiniiSNrdfppNSdfp111)1()(111)1 ()()1 (2022-6-22Semi-analytic Model for STCDOln obligorslRandom vector of default time: 1, nlJoint Distribution and Survival functions:F(t1,tn)=Q(1t1, n tn)S(t1,tn)=Q(1t1, n tn)Q pricing measurelF1,Fn; S1,Sn; marginal distribution and survival fu
11、nctionslCopula function C: F(t1,tn)=C(F1(t1), , Fn(tn)lEi nominal; i recovery rate; Mi = Ei * i loss given default;lA latent factor V such that conditionally on V, default times are indep:pti|V = Q(i t | V) cond default prob; qti|V = Q(i t | V) cond survival prob; So cond joint survival prob:S(t1,tn
12、 | V)= qtii|V 2022-6-22STCDO (continued)lAggregated loss process:L(t) = Mi Ni(t)Ni(t) - default indicator processlPV of default leg = E (L(t)-K)+ , where K is the tranches attachmentlSemi-analytic techniques applied for the computation of loss expection FFT and recursivelHow to represent default tim
13、e?2022-6-22David Lis 1-factor modellGaussian vector, v1, , vniiVVv21)(1iiivF)(211)(|tFVVitipGaussian cdf2022-6-22Stochastic Correlation)1)(1 ()1(22iiiiiVVBVVBviiiiiiVBBVBBv2)1 (1)1 (iB10iiiiVVv21iIndependent stochastic correlation with distribution function F101)(|)()(21dFptFVViti2022-6-22Student t
14、Copula)1(2iiVVWv)(212/11)(,|tFtWVWVitivp)(1iviivtFW - independent from the 2nd part and an inverse Gamma distribution with parameters equal to v/2.V, - independent Gaussian random variables.iVvt- the distribution function of the standard univariate Student t2022-6-22Double t CopulalProposed recently
15、 by Hull & White (2004)lLatent variables:nvv ,1ivvvviVVv2/1222/12)(1)()(22/1211)()(2/12|VtFHvvvVitvviitpiVV ,vv,- independent t-distribution random variables.- degrees of freedom for iVV ,)(1iiiivHFDefault time:2022-6-22Clayton Copula0,1transformLaplacexxexfx0,)(/ )1()/1(1Consider a positive random
16、variable V (1-fatcor), std. Gamma distribution withShape parameter of )(1(exp(),()()1()()(|1ln0/1tFVpVFUvsdxexfsiVitiiiiVUisxiIndependent uniform random variables also independent from VDefault time.2022-6-22摘要摘要l引言引言l信贷产品l法规和文档(Legal and Documentation)l行业规范（Regulatory environment)l复杂信贷产品l定价，计算技术和对冲
17、(Hedge)l风险管理风险管理2022-6-22市场风险管理市场风险管理(market risk)lCredit VaRlInterest rate risk monitor2022-6-22信贷风险管理信贷风险管理(credit risk)l多维性信贷风险控制l信贷证券组合的风险分析2022-6-22国内金融资产结构国内金融资产结构 l国内金融市场融资结构(2003)l贷款 3万亿 (85%);l直接融资 5340亿(15%):股票、国债、企业债券 l直接金融 (2003 12515亿）l政府债券比重 ：6280亿元（占比50.2） l政策性银行金融债发行4520亿元（占比36.1） l股票发行1357亿元（占比10.8） l企业债券发行量358亿元（占比2.9）2022-6-222003年直接融资比重年直接融资比重 2022-6-22CD离我们有多远？离我们有多远？l建立风险产品交易市场 周小川：国有商行改制后如何抑制不良贷款增长(2004-07-26 中国经济周刊 )l吴晓灵:稳步发展企业债券市场 优化金融资产结构 (2005-01，第九届中国资本市场论坛 ）l苏宁：科学规划 统筹安排 积极推进中国债券市场发展 （2005年6月26日首届中国债券市场论坛 ）l周小川 ：吸取教训 以利再战 （2005年10月20日中国债券市场发展高峰会 ）2022-6-22Q & A
