This seems to misrepresent what incentives are: latent unobservable things. Incentives are not what you or anyone else says they are. It makes sense to simply state the incentives in academic games which are of course useful at times.

So the author is making some point about wrong incentive models and meta models.