To use past data to predict future events. Researchers use back testing to find relationships between apparently unrelated events and determine if one causes the other. One may conduct back testing to inform one's investment decisions or strategy, though the practice is not always accurate because a great number of inputs cause economic events.
Using historical data to determine the relationship of specific variables. For example, a researcher might use historical data to determine if changes in the money supply have influenced changes in stock prices. This relationship, if positive, can be used to develop an investment policy.