>Does it make more sense to use the currency type or the numeric (14,2) ? It is for North America. Does it mater ?
>

Yes it does, especially when discounts or fractional cents are a factor. The currency data type has a fixed number of decimals (4) to the right of the decimal point, regardless of the number's magnitude, and will force any computation containing a currency data type to that precision, which can be murderous for certain types of computations. If you're dealing with fractional computations, especialy multiplication/division/transcendentals, don't use the currency type.

Currency pays off big time with addition where the magnitude can vary greatly; very large numbers expressed as a floating point value sometimes have inexact values - IOW, you may run into situations where two large integers are indistinguishable when represented as a floating point data type. The currency type avoids many of the rounding problems encountered with floats as well.