Thursday, September 15, 2016

An Analysis of the Bargaining Experiment We Did in Class Today

In part because I was perplexed by the outcome, I did an analysis of what happened.  This is for the first round, which I called incomplete information, where buyers had the information about their values but not about the sellers' costs, and likewise sellers had the information about their costs but not about the buyers' values.  One team did share that information, but they said they did that only after they were done bargaining, to check that they had done okay.

There were 12 students who participated in these experiments (6 pairs of buyers and sellers).  I had planned for 18.  So some of the scenarios that I had planned for were not used.  That helps to explain the numbering in what was reported.  The scenarios we did use were: Cases 1 and 2, which are essentially the same but where in Case 2 demand and supply have been shifted up by $5, Cases 4 and 5, where demand was downward sloping but  now supply was taken to be perfectly elastic and the same sort of shifting between Case 4 and Case 5 by shifting everything up by $5 occurred, and finally Cases 7 and 8, where it is now demand that was perfectly elastic.

You can see the analysis here.  The preview is sufficiently good that you don't have to download the spreadsheet.  I color coded things so that as long as you are not colorblind you should readily recognize everything in the table.  Focusing on the column called Social Surplus, all the scenarios had social surplus maximized with 3 units being traded.  The puzzle that the experiments provided is that only one group reached this outcome, and no group had fewer than 3 units traded.  All the other groups had excess trading, where trade produced negative surplus and one or both of the parties were harmed by the trade.

My prior going in is that many groups would have only 2 units traded.  There was ample surplus to get that far, but not much surplus for the third unit, so I thought in haggling over that students might reach an impasse.  This is especially true under incomplete information where it was not possible to tell whether the other side was bluffing or not, unless the two sides did share their information.

But this prior intuition didn't predict what happened at all, so in looking back at this I want to focus on those possible explanations of what happened.

The first possibility is that students didn't understand what they were asked to do.  I did say at the outset that consumer surplus was such and such and likewise for producer surplus and your goal was to maximize your own surplus.  But just because I gave the instructions doesn't mean they were understood.  This parallels something we talked about in the last class when boots on the ground communicates with command and control, though in this case it is command and control sending the message.  It may be the message gets garbled.

The other possibility is that students didn't care so they didn't act to maximize the surplus even though they understood the message.  If you look at the results, there are a couple instances where each trade netted zero surplus to one side of the bargain.  Doing this made it easy to conclude the negotiation, but it surely didn't show effort in trying to attain surplus produced by that negotiation.  Students sandbag in a variety of circumstances.  If that was happening here, the results don't tell us anything about what would happen in a real negotiation.

The one team who produced the efficient outcome are the students who negotiated Case 2.  Let's give a hand to them.  And if I were giving a prize for the top negotiator, that would have to go to the seller in Case 2, who earned the lion's share of the surplus.  Absent any other information about student's bargaining skills, we'd predict an even division of the surplus.  Even in those cases where there were more trades than were efficient, it doesn't look like splitting the surplus happened.  If people didn't care, that's an easy explanation for the outcome.  But if they did care, that would pose a puzzle.

No comments:

Post a Comment