Hello, I am attempting to regress future currency returns against changes in economic variables (GPD, inflation, index values etc). I am having some trouble with the units and interpretation, and have exhausted the internet to no avail so hoping any of you can help.

My objective is to find how changes in monthly, quarterly, and yearly variables affect future percentage returns in currency pairs.

Currently, my dependent future return is a percentage change (t+1 - t_0)/t_0, and my independents are monthly, quarterly, and yearly absolute changes (t_0 - t-1).

For example, at a given observation I would have a future percentage change against a change from the last month (this month - last month), last quarter, and last year.

I am not sure this data structure is valid, and how I would interpret the coefficients.

Currently, I interpret it as a unit change in X is associated with a B percentage point change in Y. Most of my X variables are percentages, but they are not percent changes, so it is just the difference from a month/quarter/year ago.

I am fearing that because it is the change from last quarter for example, I am finding the rate of change.

Any help and guidance would be appreciated.