It is impossible to store the number 16.15 precisely in binary. However you handle it, you're going to get truncation and/or rounding problems. If you don't know why that is, take a course in computer science.
If you're building a financial application, amounts of money should be stored as integers. Your value of $16.15 (or Euros, or Pounds, or whatever it is) should be stored as 1615 in column with INTEGER affinity, and your C code should handle all those values as integer values. This will force you to handle truncation and rounding in your C code, without any ambiguity ("losing pennies").