Looks promising, @saudiwin, thanks for sharing! I can’t look at it right now, but I agree with @matti that this would easily be worth a publication since IMO it’s a much more elegant solution to “extremes inflation” than other alternatives. Even if some other method has tighter fits (OLS, etc.), there mere fact that you can generate the inflations as part of the ordered process rather than as an “independent” process is really strong.
It would be great with a solution that handles N cutpoints, i.e., an arbitrary number of response categories, which would render it a solution to this thread.