The answer is it minimizes type I error but increase the chance of Type II error. The normal margin of error is 5% and when a statistical test shows that it is smaller than 5%, then it commits type I error. When there is an lesser type I error, it will increase the probability to commit a type II error.