im just curious on what is out there in terms quantitative trading algorithms for long term strategies (i.e not execution-speed-dependent & non-arbitrage strategies) What accepted methods using *quantitative* methods (not that technical analysis / moving average / out-of the box indicators bullshit ) for capturing long term movements (weeks, months) ??i guess a very simple example of what i mean would be an algorithm that trades EURUSD based on historical correlations between EURUSD and a basket of US and Euro-zone price indices ( historical month on month CPI / PPI / GDP in the last 10 years) .. obviously very simple, but just to clarify what i meant by long term 'numerical methods' and not TA . do any of the big houses have publicly available documentation about their long term quantitative models?

look at some fundamental econ models like FEER or ERER - these are FX models. I think GSAM was trading FEER at one point.Ev

Thanks Ev, that's the kind of models i was looking for.. found a paper on FEER (http://www.ecomod.net/conferences/ecomo ... itelli.pdf) will definitely review it.

thread still open for more references

I have worked on such strategies for a fundamental economic based hedge fund and the results were quite poor as one would expect. Well let me clarify, PPP worked extremely well at forecasting currecnies 8-10yrs out, but the variations in the mean time meant that you had to put up with being put massively on/offside. We then tried to incorporate some type of reversion variable with only limited success. So basically we said the EUR/USD is 20% away from its PPP valuation and therefore we would expect it to begin to revert. As a starting point we used FEER models. GS had extensive research on such topics and had a model called GSDEER (I am pretty certain that they never traded it, but was used as a research/sales tool). Most of the research can be found on the GS Portal in the yearly currency report. We were never able to find anyone steady relationship though, currency mkts (like most others) tend to fixate on certain variables at certain times (e.g. trade balances in the late 80's, interest rate differentials recently, etc). I remember there was an analyst from ABN (I can't recall his name, though in the early 2000's his currency research was quite well known), that put together a portfolio of very simple trading strategies that performed quite well. I think they were mainly based on earning positive carry and also some MA's