SAS CODE

options linesize=78; * read in the data *; data ibm; infile 'IBMclose.dat'; input price; return = dif(log(price)); t=_n_; * squared returns lag 1 and lag 2 *; retsq= return**2; retsql1 = lag (retsq); retsql2 = lag2(retsq); * estimate the ARCH(1) model *; proc reg; model retsq = retsql1; run; * estimate the ARCH(2) model *; proc reg; model retsq = retsql1 retsql2; run; 

Consider the market closing daily prices on IBM in file IBMclose.dat. Estimate the ARCH(1) and ARCH(2) models using the code in file Assgn3.sas. Write the two models. Discuss the statistical significance of the estimated parameters. Are the squared returns on IBM serially correlated? Is therefore IBM volatility predictable? If yes, compute the out-of-sample value of volatility at time T+1. Does your finding contradict the former conclusion that returns on IBM are white noise? Consider the market closing daily prices on IBM in file IBMclose.dat. Estimate the ARCH(1) and ARCH(2) models using the code in file Assgn3.sas. Write the two models. Discuss the statistical significance of the estimated parameters. Are the squared returns on IBM serially correlated? Is therefore IBM volatility predictable? If yes, compute the out-of-sample value of volatility at time T+1. Does your finding contradict the former conclusion that returns on IBM are white noise