As you increase the number of samples, you reduce the variance. So a small government of, say, 1 person, can either be 100% corrupt or 100% pure, but a larger g

As you increase the number of samples, you reduce the variance. So a small government of, say, 1 person, can either be 100% corrupt or 100% pure, but a larger government would approach the population level of corruption in expectation. Why is a larger governance with each individual holding less power less desirable than fewer individuals holding more power?