Request for help from a stats guru about multicollinearity

[Zobags]

If I have a multiple linear regression model where one independent variable is perfectly correlated with the relationship between two other independent variables, will this create the problem of multicollinearity in my model?

As a hypothetical example, suppose I am trying to predict the income of two-person partnerships and have a theory that it can be predicted in part by 3 variables: (1) the age of the oldest partner; (2) the age of the youngest partner; and (3) the difference in age between the two partners.

If I include all 3 of these variables in my model, will it suffer from multicollinearity or any other problems?

Any help is greatly appreciated.