Error probability of identification via channels at rates above capacity

In the identification via channels (the ID-channel), under some condition for noisy channels, the sum of two types of error probabilities of identification goes to one at rates of codes above capacity when the length of codes tends to infinity. This is known as a strong converse property of the ID-channel. In this paper we derive an exponential lower bound for the sum of two error probabilities to tend to one. In particular, for the class of discrete memoryless channels, we explicitly compute this lower bound.