The Experiment We Did Today - Our Class versus Economic Rationality

Before getting to the commentary here are the results.  You can look at them for yourself.  Below is a summary.

In the complete information scenario Cases A, B, C, and E, produced 3 units of trade, the efficient level.  Cases D, F, G, and I produced 5 units of trade, above the efficient level.  Case H did not trade, but I need to be filled in as to why that was.

In cases A, B, and C, the situation was symmetric for buyer and seller.  Absent any other information about the traders, the prediction would be split-the-difference pricing, so the pie is divided in half.  That happened in case A.  In case B, the seller got a slightly greater share of the pie.  In case C, the seller got the whole pie. The prices equaled the buyer's values.  This is sometimes referred to as first degree or perfect price discrimination, which can happen what a monopolist with all the bargaining power sets the price, while knowing the buyer's demand curve exactly.

Case E, which also produced 3 units of trade, was not symmetric.  If there were split-the-difference pricing here, then prices would have declined, as the seller's cost was constant but the buyer's value was declining.  However, this solution opted for was a uniform price alternative, and that price was higher than the buyer's value for the third unit.  So the buyer made negative surplus on that trade.  Negative surplus on a trade is inconsistent with economic rationality.

For the other four cases, where all 5 units were traded, there were some trades that yield negative surplus.  So each of those also were inconsistent with economic rationality.  This is puzzling to analyze.  I wonder why it happened.  Also, it seemed some of you didn't try to maximize your own surplus and thus sandbagged on the experiment.  It would be good to know this was.

The particular trader who captured the most surplus under complete information was the case C

------

Let's turn to the incomplete information case.  Here we had 5 cases with the efficient volume of trade and 4 cases where the trade volume was larger than that.   Among the cases that produced the efficient volume of trade, case 9 had one trade that produced negative surplus for the buyer and that buyer earned negative surplus overall.  In the other cases that produced the efficient volume of trade, all the trades produced nonnegative surplus for each side of the transaction.  That is consistent with economic rationality.

The overall surplus maximizer in the incomplete information case was the seller in case 6, but that is because this seller received hefty subsidies from the buyer, who lost out on every trade.

--------

The question I'm struck with here is why about half the class performed in this experiment like the theory predicts, but the other half performed in a way that was irrational.  I doubt that any student was deliberately trying to sabotage the experiment. (If you were, congratulations.  You achieved your goal.)  A more likely explanation to me is that some of the students weren't getting it about how they should behave in the experiment.

I would like to discuss that some in class on Thursday.  There is very little point in teaching something if the students never get it.  In contrast, sometimes a failure of this sort can be a stepping stone to doing things in a better way where students do get it.

You can't find that better way unless you do a full post mortem on the experiment and why things went wrong.  If you have thoughts on that you can leave them as a comment to this post. We'll use those as a launch point for the discussion on Thursday.