0,FX Market Dynamics Methods for Trading, Forecasting & Risk Assessment,C1,Technical Analysts European Conference February 2007,1,Disclaimer,The information contained herein, including any expression of opinion, and any information which accompanies this presentation or which is supplied subsequently, has been obtained from or is based upon sources believed to be reliable but has not been independently verified and is not guaranteed as to accuracy or completeness, although Socit Gnrale and its affiliated companies (“SG”) believe it to be fair and not misleading SG is not, and will not be, responsible for ascertaining whether all risks associated with the transactions contemplated herein have been identified or disclosed, nor for providing advice to you as to whether you should enter into the transaction or on the documentation to be used for the transaction or on the merits of purchasing any investment or otherwise. You must make your own assessment of the transaction and the risks and benefits associated with it and of all the matters referred to in the preceding sentence. Neither SG nor any of its officers or employees makes any representation as to, or assumes any responsibility or liability for, the merits, suitability, expected or projected success, profitability, performance or benefit of any such transaction. SG recommends you enter into transactions only after having considered, with the assistance of external advisors, without reliance upon SG, the specific risks of any transaction, including but not exclusively, the legal, tax and accounting implications so as to enable you to appraise and understand the financial and legal terms of such transaction and to enter into such transaction in reliance on your own judgment and that of your advisers and not on any views expressed by SG This presentation was prepared exclusively for your benefit and your internal use. Neither the presentation nor any of its contents may be disclosed to, reproduced or used or relied upon by, any other person or used for any other purpose without the prior written consent of SG,2,Forex Markets Deep, Liquid and Accessible,$1.9trn average daily turnover Almost continuous trading 24/7Growth in currency overlay mandatesBroad “understanding” of market,3,Analysts Comes in All Shapes and Sizes,CommentatorsEconomistsTechnical Analysts Quant analysts,4,Styles of Analysis - All Are Important,CommentaryEconometrics / Statistical methodsTechnical Analysis Quantitative trading Signal generation using all available datasets Beg, borrow and steal from other areas of expertise,5,A Combined Approach to FX - A Scorecard,Build a committee / agenda / processSet up infrastructure for FX viewsAsset allocation made from aggregating viewsDeploy trades and risk management,6,Pitfalls of the Scorecard,Market timing and forecast objectivesArbitrary weighting schemeBalancing the number of tradesMore of a management tool than trading tool,7,The Commentators Approach,News for market consumptionSensitivity to new information Broader knowledge of anecdotal informationTypically hard to quantify Selling a view,8,Economic Analysis for Markets,Data released each week, month or quarterMonetary policy is decided by people Inflation targeting regime versus growth models Huge debate about market relevance Markets can deviate from the economic backdrop Markets are volatile Economics may not be the largest driving force,9,Econometrics and Statistical Methods,Linear regressionMany variables feed into economyOnly so many are relevantForecast target: average or volatility,10,Examples Include,ARMA, ARIMA family of models Seasonality Multiple VAR models Multifactor models ARCH GARCH type models Volatility models Parametric and non-parametric approaches Distribution assumptions,11,Problems Associated With Statistical Methods,Over fitting and parameterisationSample biasUnstable correlation structuresExplanation factor,12,Investigating Anecdotal Evidence,“adage value” Commodity price impact Political fall out for markets Relationships change over time Adages are hard to change Not always going to happen Markets wise up to inefficiencies There is always volatility in results Arguments backed by flows are best,13,Case Study Gold & AUD/USD,14,Case Study Conclusions,Relationships are volatile Correlation not the best What is the point of a test Underlying trends at work Framing the analysisre-basing a problem can yield stronger results Look forward for forecasting Instantaneous versus leading relationships But returns also have positive autocorrelation,15,Extending the Case Study Framework,What about using more relationships Filter market drivers for each currencyIncreasing explanatory variables Increasing risk of over fitting outliers Garbage in garbage out What really matters Is there an economic reasoning to inputs? Are inputs framed correctly?,16,Statistical Models Multiple VAR,Make the assumption of normal relationshipsAssume stable relationship over timeMost used explain use instantaneous relationsWeighting scheme problematic,