Well here as are a few predictions made by game theory.
1. There wasn't a nuclear war today
2. Every telco in the world gives large chunks of money to governements due to auctions
3. Ebay sellers get higher prices than they should
Also, I can read association rules, decision trees and lisp programs ok... and yet my machine learning programs (ok ok nowadays the ones I get out of various apis and packages) have written them...
1 and 3 aren't predictions, but post-hoc explanations.
There is an argument that game theory helped lead to 1, I suppose. But, in general, I think that is giving them more credit than they would even ask for. Mutually assured destruction is far from their central theorem.
2 could just as easily be explained by standard supply/demand scenarios. There is a very limited supply of that is auctioned off to telcos and they have a very vested interest in buying as much as they can. Even if they are not going to utilize it.
I'm looking forward reading that more in depth, but I am curious what you think my claim was.
I gladly admit I am using a layperson view that the spectrum auctioned off to telcos is a) likely to just get bought up by entrenched actors, and b) is unlikely to lead to any innovations in the marketplace.
Evidence that I am wrong on either of those points would be fun to read. My second claim, in particular, is one that I suspect is ill-founded.
Yeah, that is what the linked paper was about. However, I'm not sure I care too much for the goal of raising more money. In some ways, I would view money raised as heat generated for a mechanism. That is, a necessary by product that is certain to correlate heavily with any outcome, but ultimately waste that should be minimized.
The cynic in me sees this as yet another way that economists have found to increase the wealth generation capabilities of valuable things, while at the same time losing sight of the utility that was there.
So the money either goes to government or the telcos' employees and shareholders. It's a valid goal for the government to attempt not to undersell the asset, but as you say it's primarily a transfer. If you want an example that changes more than market prices, have a look at Roth's work on matching for hospitals, schools and organs:
1. There wasn't a nuclear war today
2. Every telco in the world gives large chunks of money to governements due to auctions
3. Ebay sellers get higher prices than they should
Also, I can read association rules, decision trees and lisp programs ok... and yet my machine learning programs (ok ok nowadays the ones I get out of various apis and packages) have written them...