Opposite-signed simple effects average to a zero main effect

When the conditional effect (also known as sample effect) for one of the factors in a $2^2$ design differ only in sign, then the corresponding main effect estimate will be: