Time Series and Term Structure Data

These terms are used on this site to refer to different types of data that are available in Horizon and its related products. In order to clarify DTS’s definition of these 2 different data types we have attempted to define them here and to use a couple of images to demonstrate what we have put in text form.

Time Series

A time series is a set of historical data that has been collected over time. Each value has a date and/or time stamp.

A simple example is a series of daily end of day data where there is a single value collected for each day of trading. Another example would be a time series of hourly F/X data where rates are sampled at the same time each hour and stored in a database.

Traditionally, historical data is time series data as it is not possible to sample market data in the future. Every trader would like to be able to sample market data in the future!

An example of a time series graph is shown below:

Term Structure

A term structure is a representation of interest rates in the future. It is a representation because it is the markets “view” of rates in the future. In some currencies the term structure is modeled out 50 or more years into the future

Sometimes also called the “yield curve”, the term structure is a modeling of the representation of future expectations. Different institutions use different models to create term structures and we believe that DTS’s model is a reliable and accurate methodology that has been proven over time. We believe that the model we use is an “arbitrage free” model.

The instruments used to build the term structure are the following:

Cash instruments

Interest rate futures

Interest rate swaps

Our term structure is a representation of the future values of LIBOR or its equivalent in other currencies

Some models also use FRA’s and government bonds