NIST seems perfectly comfortable with having only one cryptographic primitive for a given category as long as there's high confidence in it. The reason we have two hash algorithms is that back when the SHA-3 competition was created, there was some uncertainty about the long-term security of SHA-2. That uncertainty has since subsided, I would say. But if SHA-2 does end up being considered insecure, as long as there's no reason to suspect that the same will happen to SHA-3, there's no reason to create a SHA-4 yet.