This, however, is where I bring up my fussy objection to the loose use of the term "health care costs." The authors of the paper give it the title "when the cost curve bent," but the idea of bending the "cost curve" itself seems to me to be a fatally ambiguous concept. What's frequently meant is simply the aggregate spending curve. This starts with the observation that unless the aggregate spending curve starts bending downward, eventually all spending will be health care spending. But since it almost certainly won't ever be the case that all spending is health care spending, the bending of the aggregate spending curve is inevitable. People will need to spend some money on some other stuff. This then becomes a pretty uninteresting question.
What we actually have to know about is something about costs. I don't want to rehash my whole previous post on this, but it's just worth restating that there's a lot of different stuff that can be happening here. Are we getting more efficient at treating illness? Are people getting healthier? Are we just doling out less treatment? Is it ineffective treatment we're curtailing, or just random swathes of it? If you cut spending on shoes by finding a great bargain, then good for you. If you cut spending on shoes by walking around barefoot, then that's not so good and it has really nothing to do with "footwear costs" declining.