Not necessarily, but if that's your use case (and the hashing speed isn't extremely critical), why not use a better algorithm with a larger output?