Prices rising with inflation is only keeping them stable in constant dollars. You'd have to apply an inflation discount to your future revenues as well.
My model isn't trying to argue that you'd value the actual real-world value of a book in terms of a constant, real-value, revenue stream. Reality is far less remunerative than that.
It's that even if that were an accurate model, the distant-future revenues would provide a minimal incentive to a current opportunity-cost analysis of writing for publication and income vs. alternative activities.
Which is to say: copyright terms are far longer than can be justified on an economic basis as incentive to authors.
That they're an incentive to present copyrightholders of aged works with some present value isn't at debate, but that's the only justification for post-hoc copyright term extension.
Also, apologies for not replying sooner, I thought you'd lost interest and had stopped checking.
If the work does live on, you won't be charging a constant price for them year after year - the price of a book, for example, will go up to match inflation, so if you sell a constant number of books per year, your "dollar" revenue will also go up with inflation (I'm ignoring technological advances in printing, etc., which are independent of this argument). It is essentially in this way that the Doyle estate continue(s/d) to make money from Arthur's works, albeit through licensing - the point is those things were priced at current market value.
However, I personally don't believe that the creators of copyrightable works take any of this into account when they are producing those works, partly because of the uncertainty involved in future revenues - it is highly likely that any given work will be one of the ones that just dies, and so anyone who might bank on a perpetual revenue stream is probably deluded, and most people are smarter than that.
I'm not sure, but I think your analysis might provide the expected value of a work, but that is essentially meaningless for someone who is only going to produce a relatively small number of works, which are highly unlikely to somehow represent a random sample of all possible works.
Assuming I understood correctly, I agree with everything in the post to which I'm replying, regarding the role of incentives, just not the present-value analysis you refer to.