A Statistical Arbitrage FX Trading System Based on Short Term FX Volatility Swings Forecasting with Institutional Data on JPY Based Investment Flows Into US Markets

Abstract:     
In this paper I explore the informational content in cross-currency flow maintained by large custodian banks with an objective to design a statistical arbitrage trading system that could exploit such information. After an initial simple test involving one-step ahead forecasts for JPYUSD FX pair with lagged I-Flow data series and concluding that such forecasts don’t measure up to a simple AR(1) model’s forecasts involving the FX pair time series itself, I introduce a 15-day moving standard deviation variables based off the I-Flow time series with a 5-day lag to the one-period I-Flow forecasting model to discover a considerable improvement in forecasts to the original model though still not bettering the simple AR(1) regression model of the FX time series. With the information from the initial tests at hand, we move on to explore the possibility of designing a system to forecast the swings observed in the 15-day moving standard deviation series of the JPYUSD FX pair. A partial dynamic equilibrium regression system involving a transformation of the individual I-Flow series similar to the FX pair series (15-Day moving standard deviation) is then experimented with to capture long-run stable 5-period ahead forecast for the 15-day moving standard deviation swings in the JPYUSD FX pair. Finally, I propose two models to exploit the volatility swing forecasting system; The first model involving volatility trades for the FX pair that could exploit an expected upswing in the short term volatility of the FX pair, and the second model involving an overlay of the swing forecasting system over traditional trend-forecasts involving technical rules to capture profitable long/short trades of the FX pair. Data between May 2007 and May 2009 was employed for the exercise.

Number of Pages in PDF File: 17

Keywords: Statistical Arbitrage, Partial Dynamic Equilibrium

JEL Classification: C22, C51, C52, C53, F31, G21

working papers series