Market Design: worthwile reading

McMillan's proceedings paper from the 2003 meetings is still worth reading. I can't introduce the topic of market design better than McMillan does:

Herbert Stein, advisor to presidents, noted that "most of the economics that is usable for advising on public policy is about at the level of the introductory undergraduate course." My subject is the remainder of usable economics: the part that is not elementary.

McMillan reviews several instances of the economist acting as engineer (a turn of phrase that I think is properly attributed to Roth). I list them in my own order, descending roughly in the order of most-to-least designed, i.e. where economists had the most direct influence to where they had the least.

* Spectrum sales (FCC to start with; now UK and Continental Euro markets as well)
* Electricity sales (peak load pricing, etc.)
* FTC divestitures
* Treasury experiments (uniform vs. discriminatory when there is multi-unit demand)
Note here: The notion of demand-reduction is what makes the uniform format no longer truth-revealing. This was pointed out by Ausubel and Cramton in what now must be the most well cited unpublished manuscript ever. Anybody know another?
...
and the rest:
* Procurement at the Pentagon
* Privatization of British Telecoms and former Soviet economies (the whole shebang)