Multiple Lineair Regression - Another variable needed ?


I'm doing a research on 3 variables, exchange rates, macroeconomic news and net speculative positions.

So far I have a regression with exchange rates changes being the dependent variable, news and positions being the independent variables. The exchange rate changes are measures in a period of 15 minutes around the news announcements, so short time period. The net positions are taken on tuesday closing time, weekly.

I have figured out that news happening on a day closer to the positioning data should have a higher effect then announcements later on. So say that the net speculative positions are taken on tuesday, and there are news announcements on wednesday and on friday. Then the exchange rate should react more on wednesday then on friday.

I'm struggling at the moment how to implement this construction to test for the "time" effect. I hope someone can give me a push in the right direction!