"Using data on foreign exchange (FX) order flow, we study how carry trading lifts" high interest rate currencies but also increases the probability of a dramatic reversal. Furthermore, using data on market participants' forecasts of future currency values, we can decompose the forward bias into one part associated with time-varying risk premia as a function of order flow, and one part associated with forecast errors.

We find that order flow can account for a substantial portion of the forward bias – particularly for currency pairs typically associated with carry trading. The basic idea, from the seminal paper by Evans and Lyons (2002) on the microstructure approach to FX, is that carry trading is a portfolio shift that requires a change in risk premia in order to be willingly accommodated by counterparties. Furthermore, we see that order flow predicts (in sample) shifts in the skewness of FX returns so that carry trading also leads to increased risk of sudden reversals.

Our empirical approach combines the Reuters survey of market participants' forecasts of future currency values and FX transactions data from Electronic Broking Services (EBS), for euro, yen and sterling against US dollars, over a period of ten years between January 1997 and April 2007. Although the main focus of this study is to combine these data sets, it is worth noting that individually they are arguably superior to most data sets previously used in the literature. For example, Burnside, Eichenbaum, and Rebelo (2009), which also applies a microstructure approach to the forward bias puzzle, use indicative bid-ask quotes released by a large FX dealer. In our paper we have access to data on actual transactions completed on the main electronic trading platform which currently dominates spot FX markets for the major crosses. Compared to other studies using data on FX order flow this is the longest data set to date. Compared to work using survey data, e.g. Bacchetta, Mertens, and van Wincoop (2008), our survey of exchange rate forecasts, while shorter in length, focuses almost entirely on nancial institutions and contains information on all individual forecasts rather than sample averages."

Are you looking for more strategies to read about? Check http://quantpedia.com/Screener

Do you want to see performance of trading systems we described? Check http://quantpedia.com/Chart/Performance

Do you want to know more about us? Check http://quantpedia.com/Home/About