Ah, I see what you’re saying Pav, I think this was a comprehension failure on my part. To simplify further, just so I’m sure: If there only 5 bonds being sold, and the top 5 wealthiest people have $10m, $9m, $8m, $7m, and $6m, respectively, no one will buy for more than $6,000,000.01 because they can still guarantee they’ll get a bond at that price. Is that the idea?
I can see a few problems with this. First, where the bonds are transferrable, the there might be speculative value in buying at a higher price. So if the wealthiest person can buy multiple bonds at a higher price, they could hold them in anticipation that people will be able to buy at a higher price later. For example, if the bonds are expected to maintain their value and other assets are expected to have negative return, the bonds could act as a hedge.
Second, as I mentioned in my reply to Tab, people may pool their resources to buy bonds, so several people whose net resources are blow the current price could buy one bond together (e.g. parents and granparents on behalf of their child).
Third, in practice there isn’t full transparency between buyers, so buyers can estimate how much others have to spend, but they can’t know and will sometimes rationally overpay when they’re uncertain about what other people can/will pay.
Fourth, I don’t see any cost to starting the bidding too high. If you start it too low, you leave money on the table, because people that would have purchased at a higher price instead purchase at a lower price.
I agree with your final point, that adding in all the realistic considerations complicates the question a lot. But dutch auctions are used in the real world to do things like IPOs, we could probably find a rich literature on how those sales behave in practice – do you think that the reproduction bond auction would function the same way? My gut reaction is yes, but I could also see it being sufficiently different. I’ll try to find some case studies.