The problem with macro-economic composites is that at that level, everything is unfolding on different timescales. So around transitional moments, your end result is going to go wonky.

GDP plummets -> interest rates are dropped -> GDP recovers -> interest rates are raised

You're balancing between responsiveness, accuracy, and reliability: pick two.

I'd have thought the last few crises would have taught us that we should be thinking about the system more in terms of stability, or lack thereof.

The housing crisis could have not exploded, if key institutions and/or investors had acted differently. But it was objectively true that the entire system was in a very unstable state.