高盛因子模型ppt课件

,Title,Body Text(shift return),line spacing check,line spacing check,Third Level,Fourth level,Fourth level,Third Level,line spacing check,*,Project presentation,Linear Factor Model and More,Hong Cai,October 14,2010,1,Project presentation1,OUTLINE,Introduction to Linear Factor Models,Introduction to BARRA-Type Linear Factor Models,BARRA-Type Chinese Equity Linear Factor Models,Beyond,BARRA-Type Linear Factor Models,Summary,2,OUTLINEIntroduction to Linear,Introduction to Linear Factor Models,A linear factor model(LFM)is an analytical device borrowed from the world of statistics and,employs single or multiple,mostly multiple,factors linearly in its computations to explain market phenomena and/or equilibrium asset prices,especially in equity market.LFMs can be used to explain either an individual security or a portfolio of securities by comparing the factors to analyze relationships between variables and the securitys resulting performance.LFMs are also used to construct portfolios with certain characteristics,such as risk,or to track indexes.Theycan be dividedintothree categories:macroeconomic,fundamental and statistical models.,Major Application of LFMs,Predict returns of,an individual security or a portfolio of securities.,Generate estimates of abnormal returns(so-called Alpha Model),Identify risk sensitivities.,Estimate the variability and convertibility of returns(Further,VaR computation by Delta-Normal method).,Used in various trading strategies,like Stat-Arb,Quant/Algo Trading(such as momentum,liquidity and etc).,3,Introduction to Linear Factor,Linear Factor Model Specification,The general form of the linear(multiple)factor model is,R,it,=,i,+,1i,f,1t,+,2i,f,2t,+,Ki,f,Kt,+,it,where:,R,it,=return or excess return on asset,i,(,i,=1,n),f,kt,=,k,th,common factor(,k,=1,K),ki,=factor loading on asset,i,for,k,th,factor,it,=asset specific factor for asset,i,Assumptions,The factor realizations,f,t,are,I(0),with unconditional moments,.,The asset specific error terms,it,are uncorrelated with each of the common factors,f,kt,.,The error terms,it,are serially uncorrelated and contemporaneously uncorrelated across assets.,4,Linear Factor Model Specificat,Three Types of Linear Factor Models,Macroeconomic factor models,Use observable economic time series like interest rates and inflation as measures of pervasive or common factors in asset returns.,Examples:Sharpes Single Index Model,CAPM,and APT.,Factor variables,f,t,are observable but factor loadings,i,must be estimates.,Fundamental factor models,Use observable firm or asset specific attributes such as firm size,dividend yield,and industry classification to determine common factors in asset returns.,Examples:Fama-French Factor Model and,BARRA,Factor Model.,In Fama-French-Type Factor Model,firm attributes are used to create factor mimicking portfolios that are used to proxy the factor variables,f,t,.Factor loadings,i,must be estimated.,In BARRA-Type Factor Model,factor loadings,i,are firm attributes and observable.The factor variables,f,t,must be estimated.,Statistical factor models,Treats the common factors as unobservable or latent variables.,Examples:Factor Analysis and Principle Component Analysis.,Both factor variables,f,t,and factor loadings,i,must be estimated.,5,Three Types of Linear Factor M,Introduction to BARRAs Models,BARRA,who had developed BARRAs Models,is an investment consulting firm specializing in quantitative risk management tools.,The,factors,are the models statistically derived building blocks to capture the sources of equity market returns.The BARRA USE3 model employs 67 explanatory factors.Most of them are intuitive although some not.,The model is considered,linear,because,in general,it assumes that the factors influencing the market work in a linear fashion.,BARRA does performance analysis by first estimating the returns to its factors and then by linking these factor returns with each managers active policies.Each month,BARRA runs a big statistical regression that simultaneously produces estimates of the prior months return for each factor.,These factor returns can be interpreted as“while adjusting for the other factors in the USE3 model,the return for factor X was Y%last month.”Taking the dividend yield factor as an example,one can say that net of the 64 other factors,the return attributable to dividend yield was 1.2%last month.,Now,BARRA Factor Models will be the focus of the presentation due to their importance and popularity in financial industry as a factual standard.,6,Introduction to BARRAs Models,Introduction to BARRAs Models,BARRAs model isused to predict returns and risk for equity,fixed income,cash and derivative instruments,at both the asset and portfolio level.,Major Components of BARRAs Models,Barra Global Equity Model(GEM2 S/L),-Captures the effects of global common factors,such as the world market,styles,countries,industries and currencies,on portfolio return.,Barra Europe Equity Model(EUE3),-Provides a unified perspective on risk