Mechanism design theory - Eric Maskin
ฝัง
- เผยแพร่เมื่อ 22 ธ.ค. 2013
- Nobel Prize winning economist Eric Maskin from Harvard University on privatization of the radio spectrum, history of the field, and mistakes in decision making.
Read the text version here: serious-science.org/mechanism...
Get a bigger picture:
1. Consequences of Decentralization - Daniel Treisman: • Consequences of Decent...
2. Explaining Modern Economic Growth - Deirdre McCloskey: • Explaining Modern Econ...
3. Models in Decision Theory and Economics - Itzhak Gilboa: • Models in Decision The...
Follow us:
LinkedIn: /
Twitter: / scienceserious
FB: / serious.scie. .
Instagram: / serious.sci. .
Support us on Patreon to see more videos: / seriousscience - วิทยาศาสตร์และเทคโนโลยี
Eric Maskin thanks Very much for explaining the complex theory in an easy and meaningful way, hope you will share us in detail in a feature
Hi. Thank you much for your videos. Is it possible to take a part of this talk to integrate it in a course on multiagent systems and mechanism design? Thank you much!
Brilliant Solution!
To be honest - the example is called an „auction“, explained in many words. You set yourself a highest price and see how far your competitors will go. Looking forward to more examples to make use of, e.g. to prevent people from spreading fake news in social media.
He was explaining second price auction. Second price auction has a subtle difference with auction yet achieves the optimal results.
Great channel :) greetings from germany
You are welcome!
I don't understand: is the mechanism design something useful in helping the market to do its job or it is about substituting it with a total planned economy? Is it a way to replace the free market or it's something to help the free market itself? Thanks in advance for your answer
Any market is a mechanism itself. As you say the market has a job to do. So let's say the job of the market is to do X (I actually don't know exactly what that job is but it doesn't matter here). Then in mechanism design you would start with the question: What is the best mechanism to do X? And then you would study this question until you find the answer. In this case the answer could be: A free market is the best mechanism to do X but it could also be: a market with this and that type of regulation is the best mechanism to do X.
Note that finding the answer can be a very hard question. It is possible that we cannot find the exact answer.
The marker is an unintended process, The beneficial results are largely Unintended results as individuals pursue their own goals in the presence of rule of law. If it could be modelled, it would be easy to replace it. Hayek's Competition as a discovery procedure hits on this theme
Fucking awesome!
Why would there be a risk of another company overbidding if you overbid? That would only happen if the other company knows your bid. Referring to the argument that this method prevents overbidding.
There is a risk to overbidding because another company might make a bid in between your actual value and your inflated bid. Thus, if you value the contract at $10m, but bid $12m (you overbid), another company might make an $11m bid. In this scenario, you are awarded the contract, but you must pay $11m, which is higher than the actual value you place on the contract.
Not everyone values the asset the same. Some have higher-value uses. So if a company could make $11m off the asset, that's what they'll bid. If you value it at 50 cents and you bid $100m, you'll have to pay $11m (if that's the second highest bid).
If you are referring to why a company wouldn’t underbid in this case, it’s because as an agent in the “game”, you have no way to know what other companies bid. Assume you value x at “y”. Then if your bid “b” is less than y, and another company bids a value between y and b, you lose the auction and earn no utility. If you bid exactly y, then one of two things will happen: a). Some company bids higher than you, in which case nothing happens to you, you pay nothing nor do you gain anything. Or b). You bid the highest, which means you win! But you don’t have to pay “y” you only pay whatever the 2nd highest bidder paid.
Where can I find material to learn more about this topic? Can someone recommend a good book for example? :)
Krishna's book on auction theory is probably a good starting point. It's "Auction Theory" by Vijay Krishna
An Introduction to the Theory of Mechanism Design Tilman Börgers
Borgers, Tilman. An Introduction to the Theory of Mechanism Design . Oxford University Press.
Mas-Colell Microeconomic theory has a good intro treatment I think.
Why would competitive bidders not bid the value in the first price auction?
I am not sure what you mean by "competitive bidders" but here is the general intuition:
The first price auction says that you pay what you bid, that is it takes the highest bid and asks the person to pay their bid amount and give them the item. Think of yourself as a bidder, say you value a painting at $1000 - if you bid your value and get the painting you are in essence getting nothing! (You loose $1000 to gain $1000) or you don't get the painting and still get nothing - so you get 'nothing for sure'. Which is not bad, but can you do better? You have two options: bid higher than $1000 or bid lower than $1000. Obviously bidding higher than $1000 is a bad idea because you value the painting at $1000, you might loose money if you win and get nothing if you loose! That is certainly worse than getting - 'nothing for sure'.
But what happens if you bid less than $1000? Say you bid $900 and nobody else has a value for the painting above $900 (by previous argument no one will bid more than $900) so you can actually win by bidding and by the rules paying $900! Why is that better? You get $1000 and loose $900 and make $100 profit. So you could get $100 if you win and nothing if you loose - that is certainly better than getting 'nothing for sure'. This should give you an intuition why bidding the value is not ideal if there is even a small chance that you could bid less and win!
If you mean "competitive bidders" to be "infinite bidders", then people will in fact bid their valuation. As number of competitors become large, the lower bound on the bid that could win converges to the value of the item for any individual player! The proof is a little more nuanced, but any auction theory book will have it.
@@mwazra6625 thanks for the reply. Yes I meant a continuum of bidders is what I was thinking.
This guy says fixing outcomes and then working backwards to get the inputs this is nothing but economic planning.how much it is effective is really a debatable issue.
It is more like designing the rules and tweaking the payoffs in order to achieve a certain objective defined by some entity - a government, a company, or any other running this mechanism.
I believe it was already proven that bidding and voting mechanisms are always less than ideal. A pity that this branch of economics actually needs a bit of mathematics
This first few mins is full of hot air, cotton wool and game theory.
People make mistakes, but economists don't.