Hard forks require every single node to update, including exchanges. As such they require a good amount of lead time, testing, etc. They also can't be done too often or the repeated mandatory updates become a big burden.

The next planned hard fork is intended to include the new SMT functionality, which is a major piece of development and is still undergoing testing. When it is ready to move toward deployment, the final build approved by witnesses will either include the power down change or not.

In short, the power down change itself is very simple, but it still has to wait for the rest of the update to be ready.