First of all, thanks for the suggestion. I will post this in the “Voting Theory Forum” (for papers I am also a participant in the “Decision Theory Forum”), and probably in electo wiki (the reddit looks too entropic).
I will read your paper and contact (by mail) you and Dr. Heitzig, including the gated version of my paper and some additional material and probably I will consult you on my next steps.
In any case, as commented before, in my view there is a massive difference between static voting and dynamic voting. With a single vote, Arrow is inevitable. When you vote many times in the i.i.d framework, you can communicate preference intensities, and the Arrow problem can be addressed. Unfortunately, when decisions interact, policy coordination by simple (sequential) voting looks intractable to me.
First of all, thanks for the suggestion. I will post this in the “Voting Theory Forum” (for papers I am also a participant in the “Decision Theory Forum”), and probably in electo wiki (the reddit looks too entropic).
I will read your paper and contact (by mail) you and Dr. Heitzig, including the gated version of my paper and some additional material and probably I will consult you on my next steps.
In any case, as commented before, in my view there is a massive difference between static voting and dynamic voting. With a single vote, Arrow is inevitable. When you vote many times in the i.i.d framework, you can communicate preference intensities, and the Arrow problem can be addressed. Unfortunately, when decisions interact, policy coordination by simple (sequential) voting looks intractable to me.
Thank you very much for your comment,
Arturo