Well you are sort of a bit stuffed there. Floats are stored in binary. What might appear to use only one decimal place is only because of the display rounding values. For example, the float representation of 0.1 is actually 0.100000000001 (give or take a decimal place or two).
Floats are approximate data types - if you are worried about decimal places then you have used the wrong data type and should have used something like DECIMAL (DECIMAL is a fixed precision data type).
You'll have to ask Rudy - I think it is perfectly valid. It is only a wrapper for DECIMAL(19, 4) really. The only objection I could imagine Rudy would have is I think it is proprietary to SQL Server, although the convention of four decimal places for money is not.