Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site proper.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!sdcsvax!sdcrdcf!hplabs!amdahl!proper!walt_sak
From: walt_sak@proper.UUCP (Walt Sakai)
Newsgroups: net.invest,net.math.stat
Subject: Forecasting Capital Markets
Message-ID: <1624@proper.UUCP>
Date: Sun, 18-Nov-84 19:39:25 EST
Article-I.D.: proper.1624
Posted: Sun Nov 18 19:39:25 1984
Date-Received: Wed, 21-Nov-84 01:32:31 EST
Reply-To: walt_sak@proper.UUCP (walt_sak~{)
Organization: Proper UNIX, San Leandro, CA
Lines: 22
Xref: sdcsvax net.invest:490 net.math.stat:68


I would like to call for comments regarding FORECASTING the
various areas of the CAPITAL MARKETS, especially with respect
to TRADING.   Comments on negative results are very welcome.
We can all learn from mistakes or deadends.

Particular areas include:
     BOND MARKET                        TIME-SERIES ANALYSIS
     MONEY MARKET                       ARMA / BOX-JENKINS
     FOREIGN EXCHANGE                   MULTIVARIATE METHODS
     FINANCIAL FUTURES                  KALMAN / SPACE-STATE

In working with the T-BOND futures, I have found the price
series (using a very limited data)  to basically in line with 
the RANDOM WALK model.  After removing the nonstationarity
of the series, the autocorrelations of the first-difference
series were not significantly different from zero.  This is
very preliminary, and subject to more extensive testing.
Your comments in this area would be most valuable.

Comments sent will be summarized and posted.   Thank you.   
------------------------------------------------------------
( :-) Among economists, the world is a special case.  :-)    
#! rnews 1551