On 9 May 2002 01:22:11 -0700, [EMAIL PROTECTED] (Indian) wrote: > Dear Sir, > > Presently I am working on a research paper which > involves application of Zellner's (1962) Seemingly > Unrelated Regression (SUR) technique. Let me elaborate > my problem so that some of may provide further inputs. > Many Indian companies in recent years have raised > capital by listing their stocks(Depositary Receipts, > DRs) on the foreign stock markets (USA, London etc.).
Any time you are talking about time-series, you are far outside the reach of OLS regression. As I grasp it, the simple SUR was devised for a set of OLS equations. The google-search that I reported on a few days ago showed be a number of various hits for SUR, and IIRC, some of them *did* involve other than OLS, so there might have been something there for time-series. But stocks and markets are a notorious mess to model. An enormous amount has been tried; most of it has been bad; even the decent stuff stops working; and who will point you to the little that has worked? (I point you to the counterexample, the Nobel winners who took their investment firm to billions in debt.) SO I have been bored with stock-market questions, for years, and I won't try to say any more. Good luck. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
