Hi,I start a new project and am currently looking into FOREX Trading Strategies.The strategy is intended to diversify a huge multi asset portfolio.As I work for a large institution I need to scan trading strategies from the major houses with at least 1yr track record.I know the Citi and JPM products, further Deutsche and UBS has some stuff. Most of the stuff is carry or momentum/macro strategies. Does any one have experience with institutional FOREX strategies or can give me some more references ?thanksP.

Lehman's have a good Foreign Exchange quant model called Marcus. The idea is to have 6 portfolios, each portfolio have their own determinants. For example, one porfolio is related to the commodity market, and has currencies like AUD, NZD, etc. Although it has only traded for about 2 years, it has done fairly well, but nothing special. YTD is down -0.7%.

Another interesting model is that of RBS's. They have a variety of models, including an Emerging Market Yield model. It calculates a risk-adjusted yield for each emerging market country by regressing the 1-month forward implied yield against measures of credit, market and liquidity risk across all countries at a particular point in time. A buy or sell signal is triggered when the market and risk-adjusted yield differ significantly. Allocations to each currency are then determined in a mean-variance framework using historic signal returns for the mean and all daily forward returns for the covariance.

In foreign exchange trading, firms using quantitative/technical models will offer a higher success rate as the ones using fundamental analyses usually are'nt that profitable. The reason is that fundemental models tend to work 3yrs out as there is a lot of noise in the fx market. I believe meese & rogoff wrote a paper on this. George soros also spoke of this as he was caught wrong footed when weak currencies were just appreciating because of their yield attractivenes whereas their trade deficits, gdp, inflation numbers, etc indicated that they should loose value. Also it seems like the "beauty contest " theory works with fundamental models since if everyone focuses on certain economic factors and disregard the other factors that were in fashion 2yrs ago then the market moves according to those economic factors, example, you might find that if everyone starts focusing on trade deficits, then currencies with high yields will loose value and some managers will be left scratching their heads why high yielding currencies are dropping when for the last eight yrs yield was the main variable. I think frankel & froot called this speculative bubbles in fx markets.Therefore you find that quant models make money during both destabilising and stabilising speculation in fx markets. So focus on this area and you won't go wrong...hopefully.

Nashequillibrium,I think you will be surprised to hear that Dr. Meese right now runs an FX strategy at BGI, and as far as i remember, his paper referred to inability of forecasting based on PPP differential, whereas right now there is a large nr of factors involved in forecasting equilibrium exchange rates. From what i have seen, returns on the fundamental/macro/discretionary strategies tend to be far above those based on quant models - think Soros's trade against BoE.I also would not lump quant and technical models into the same basket, there are a lot of quant managers who are using fundamental factors as well as technical (if you consider momentum as a technical factor) in ccy as well as in equities. I think it all depends on your investment horizon, if you are long term investor willing to weather short term technical pressures, then i think there is no other way then to invest based on fundamentals, if your investment horizon is short, concentrate on denoising short term/high freq data like RenTech or other superquant shops.

It seems there is a lot of quant models running, specially blackboxes using technical analysis and fondamentals factors... Another strategy is the G10, used by deutsche bank. The strategy simply involve shorting the low interest rates and buying the high interest rates currency pairs of the G10 countrys.

QuoteThe strategy simply involve shorting the low interest rates and buying the high interest rates currency pairs of the G10 countrys.Yes, that is actually called a Carry and it works untill events like July/August when it gets murdered, very similar to writing deep OTM calls...

QuoteLehman's have a good Foreign Exchange quant model called MarcusCan arsenalboi please refer me some documents on how that model works?

Lehman's model is actually spelled MarQCuSI have no idea how it works.