FIN 010 Lecture Notes - Lecture 23: Santa Barbara City College, Autocorrelation, Risk Premium
Document Summary
Sim: markowitz framework requires large# inputs when dealing with large # securities can simplify by assuming risk of individual security can be decomposed into firm-specific component + market-wide component. Returns as expected" + unexpected" component eg. ri=e(ri)+ui where ui has mean/std of 0. Can expand ui = m + zi (m=market/macros and 0 std/mean, zi=firm specific, i= each stock) Individual returns generally correlated (common set of factors eg. response to market, but assuming stock returns related to single/market influence) Asset return as constant" + sensitivity" to a factor, factor often market index (sim for portfolio) Sim requires 3(n)+2 rather than (n2+n)/2 using the full covariance matrix. Expected return-beta relationship: also common to work in excess return" form = -rf. > given e(ei)=0, can rid of rate form both asset & market return last symbol. Diversification: assuming portfolios equally weighted = xi=1/n. Eg see fig for cba est 50 monthly returns -> Beta estimates can be supplied through commercial vendors/calculated direct.