0 does double duty as meaning itself and to indicate underflow due to multiplying very small numbers.

If you have a function that can experience underflow it can be useful to preserve the sign of the underflowing value.

Otherwise you'd need more checks to get that information.