I do an informal experiment testing whether LessWrong karma scores are susceptible to a form of anchoring based on the first comment posted; a medium-large effect size is found although the data does not fit the assumed normal distribution & the more sophisticated analysis is equivocal, so there may or may not be an anchoring effect.

Full writeup on gwern.net at http://www.gwern.net/Anchoring

Facebook IPO'd at a price of 38 dollars a share, which apparently gave it a price-to-earnings ratio in the range of 100 - extremely, fantastically high. The price dropped pretty rapidly and is currently somewhere around 20 dollars; which still, presumably, gives it a very high P/E ratio somewhere in the forties. Now, suppose it had IPO'd at a more historically-reasonable P/E of, say, 20 - still high, but not stratospheric. That would put the initial share price somewhere around 10 or 12 dollars. Is there any strong reason to believe that the price would then have *risen* to where it is now? It is not obvious to me that the current price is supported by anything but the historical price - in other words, it's trading around 20 because it has recently traded around 25. 

My point: I can't help but wonder if someone connected to the IPO had read Kahneman on anchoring. Somebody, clearly, was buying the stock at 33, just as someone is still buying at 20; I wonder if the chain of thought had that apparently-arbitrary number "38" in it somewhere, making 33 look cheap - fundamentals be damned! And if this happened, who benefited, and what ought we to conclude about the efficiency of markets? 

Reading Schelling's the Strategy of Conflict, a useful social purpose for anchoring occurred to me.

First some background.

You and your husband/wife lose each other in the mall. The two of you have not before this agreed on a place to meet in case you lose each other. Still, there is a good chance both of you would decide to meet up at some salient/prominent place, say the main information desk. This is coordination without communication and with aligned interests.

You and your partner/rival are independently given the choice of "Heads" or "Tails". Neither knows the choice of the other. If both of you choose "Heads" you'll get $1 and your opponent $3. If both of you choose "Tails", you'll get $3 and your opponent $1. Otherwise neither of you gets anything. Most pairs would coordinate on "Heads" (again because of convention/salience), with the Tails player not insisting on Tails because she has no way to coordinate (arrange for compensation, say) with the Heads player on this. This is coordination without communication and with some conflict of interest. But not total conflict, as they would rather coordinate than get no payment at all.

Schelling then goes on to consider explicit bargaining situations where there is of course actual communication between the parties deciding on a mutually acceptable outcome. But he notes that even here, "focal points" seem to exert a huge influence. Furthermore, these focal points are often quite partial towards the interests of a certain party. Yet often, the 'losing' party still accepts this less than stellar bargaining outcome. For example, a nation conceding some territory because the only prominent landmark permitting a stable division was some river partial to the interests of the other nation.

He then proceeds to show how explicit bargaining is not so different from the Heads/Tails coordination game. In a bargaining situation, both sides would rather reach agreement than none at all. And there is a range of possible points of agreement, where the 'losing' party would rather concede than forfeit any agreement. But how to decide among these points of potential agreement? A stable point of agreement for an outcome of the bargaining would be one in which neither expects the other to make further concessions. But a party decides if they would concede based on their expectations of the other party's likelihood of conceding. And so it goes back and forth. Hence, Schelling argues, even in explicit bargaining, focal points play an important role in coordinating expectations, and in ensuring that an agreement is reached, to the mutual (if lopsided) benefit of both parties.

This is where anchoring comes in. The proposal is that, sure, by letting yourself be influenced by a subpar anchor, you are forgoing a much better bargaining outcome for yourself. But this is better than no agreement at all! If you accepted a price from a merchant who proposed high price to set an anchor, it was only because this was a price you were willing to accept before you started bargaining. And if instead the lowest price the merchant would allow was too high, you would simply have rejected the transaction, and perhaps found a better offer from his competitors.

Anchoring is of course, not limited to such explicit bargaining situations. But then so is the principle of 'focal points'! In many situations throughout life, there are situations where participants share some interests and diverge on others, and where bargaining is not entirely explicit. To coordinate on these at all we require the ability to respond to anchors. Of course, this would only create an incentive to manipulate anchors, and subsequently an incentive to be resistant to such manipulation. But resistance is not total non-susceptibility! If one does not respond to anchors at all, one would be unnecessarily forgoing many mutually beneficial bargaining outcomes, to one's own detriment.