best variable to use for money

float or double?

Neither. Fractional powers of 10 (eg 0.1, 0.01) cannot be represented in any floating point type. So, if you use a floating point type to represent a decimal monetary type (eg dollars made up of 100 cents, pound made up of 100 pence) there will be an error of representation. That error tends to propagate and grow with various operations, such as addition or multiplication, when repeated a significant number of times.

Better to use some representation that allows precise representation of the various components. I'll leave working out options as an exercise.