ABSTRACT
This chapter uses the omnibus tests developed in Hong, Li and Zhao; Egorov, Hong and Li. It examines whether commonly used Autoregressive Conditional Duration (ACD) models can well forecast the probability density of foreign exchange price durations. The foreign exchange market is one of the most important financial markets in the world, with trading taking place 24 hours a day around the globe and trillions of dollars of different currencies transacted each day. Transactions in the foreign exchange market determine the rates at which currencies are exchanged, which in turn determine the costs of purchasing foreign goods and assets. Density forecasts for the price durations of exchange rates are particularly useful for many outstanding issues in international economics and finance. Price durations measure how long it takes for the price of an asset to move beyond a certain threshold. Density forecasts have become a standard practice in many economic and financial applications.
