Realistically - how locked in are developers to their current engine?

I imagine >50% of video game development is art. And I imagine some decent percentage of code can be translated without nearly as much effort as the original implementation.

Still, I doubt many developers are going to switch engines mid or late development.

But a lot of times, developers re-use much from their old games in development of their new games.

Is this realistically going to hold developers back from moving future development to a new engine?