Not really. Most econometrics seeks to test a given economic theory, and almost always theories involve several variables.
What you seem to be describing is known as "data mining" or "data snooping," i.e. creating an ad hoc model with just as many variables as necessary to "explain" a given phenomenon. Such a practice is generally frowned upon within the profession, as it leads to all kinds of spurious inferences.
If one is looking at investment one would first look at interest rates then add expected income or some other variable which would explain changes not "explained" by the first two.
You're not going to explain investment with single equation model that has interest rates as an independent variable. That's becaue there is significant reverse causality going on, i.e. investment affects interest rates just as much as interest rates affect investment. This is known in the econometric literature as "simultaneity bias" or "endogenetity bias." The upshot is that you have to estimate a multi-equation model.
The first description I ever heard (from my econ professor)of a practical use for econometrics was of a student who used one variable to predict electricity demand. His finding was so robust that Commonwealth Edison hired him before he even graduated. It is likely that the model is more sophisticated now.