Message-ID: <15958225.1075858473107.JavaMail.evans@thyme> Date: Mon, 7 May 2001 13:00:00 -0700 (PDT) From: vince.kaminski@enron.com To: stinson.gibner@enron.com Subject: Newsletter Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-From: Vince J Kaminski X-To: Stinson Gibner X-cc: X-bcc: X-Folder: \Vince_Kaminski_Jun2001_10\Sent Items X-Origin: Kaminski-V X-FileName: vkamins.pst Stinson, Can you please take a quick look and fwd it to Sam Smith (wsmith2@enron.com) Vince ---------------------- Forwarded by Vince J Kaminski/HOU/ECT on 05/07/2001 09:59 AM --------------------------- VKaminski@aol.com on 05/06/2001 07:59:33 PM To: vkamins@enron.com cc: Subject: Newsletter A reminder: this column serves two objectives. The firstis introduction of new members of the Research Group, and of the associates andsummer interns rotating through the group. The second objective is to reviewinteresting publications related to energy markets, quantitative finance andother topics of interest to the Enron audience. I have run out of the newmembers to introduce and it's time to some real work. Today's column is devoted to the issue of technicalanalysis. Financial economists tend to dismiss technical analysis (chartism) asvoodoo science. Over the last few years, however, a significant amount ofevidence that technical analysis works at least in some markets. A paperpublished recently by Carol Osler of Federal Reserve Bank of New York (CurrencyOrders and Exchange-Rate Dynamics: Explaining the Success of Technical Analysis)reviews two cases in which technical analysis proved quite successful inoffering useful predictions in the foreign currency markets. The firstprediction is that trends tend to be reversed at identifiable support orresistance levels. The second is that if a support or resistance level isbreached, trends tend to accelerate. The explanation of this success can be found, according toCarol Osler, in the microstructure of the F/X market: take-profit and stop-lossorders tend to cluster around certain typical levels (like exchange rate endingin a specific market at 00 or 50 level). For example, the author found thatthat roughly 2.8% of all stop-loss buy orders are triggered when the rate hitsa round number, while 10.5% of take profit sales orders are triggered. Thisasymmetry explains the success of the first prediction of technical analysis. Asecond asymmetry is related to clustering of orders just above and just belowround numbers ending in 50 and 00. Stop-loss buy orders are infrequent at ratesjust below these levels, and cluster above these levels. Stop-loss sell orders,however, cluster just below 00 and 50 levels a relatively scarce at rates justabove these levels. The question remains, why conditional orders tend tocluster at certain levels. One possibility is cognitive advantage tradersderive from managing large volume of information using simplified rules.Another possibility is that they follow recommendations of technical analysisin managing their books. In this case, technical analysis becomes aself-fulfilling prophecy. Please, contact us for thecopy of the paper if this caught your attention. - wjk0506.doc