There are several ways to commute to Evanston from our house, and none of them is particularly simple or superior to the other. Usually I take the #22 bus 10 blocks down Clark to Belmont, where I catch the Evanston Express train. The KP Spouse, on the other hand, takes the #22 bus 6 blocks south to Addison, gets on the Howard line there, and proceeds south to work. Catching the el at Addison has the advantage of a shorter walk from the bus and a generally less frantic station. But for me, getting off the bus at Addison and switching to the el means taking it one stop south to Belmont and switching platforms to catch the Evanston Express northbound. Itís not clear that I gain anything from doing that.
But I gave it a shot this morning, because the traffic on Clark was intense. I didnít wait long at Addison for the el, switched platforms at Belmont, and caught an Evanston train immediately.
Pretty good, huh? But hereís the interesting intellectual question that arises: how do I evaluate the performance of this commuting route versus the alternative, on this particular day? I cannot. Why? Because I didnít take the alternative today, I took the one I took.
Think of it this way. If I wanted to get a precise read on which route is better in these weather and traffic conditions, Iíd like to take them both and compare them directly. But I canít.
This little parable illustrates an important point for economics in general and for policy analyses in particular: it is often really important to be able to evaluate something that didnít happen but could have, and to compare it with what did happen. Or, put another way, truthful evaluation of policy alternatives requires us to be able to evaluate what is not.
What does that mean? In Law, Legislation and Liberty, Hayek said that ďfruitful social science must be very largely a study of what is not.Ē Understanding what is not enables us to understand and evaluate what is. But understanding what is not is challenging because capturing information and data on what is not is difficult, as illustrated in my story above. Do I know how long my commute would have been if I had stayed on the bus? No. It is a datum on what is not, and I did not have a means to capture it this morning and use it to evaluate the commute I actually experienced.
That is precisely one area where performing experiments becomes valuable. In a laboratory setting, we can create environments that have the salient features of what is and what is not, and we can evaluate them against each other. My colleague Vernon Smith picked up on this insight in his Nobel address and in other writings since then. Our ongoing work continues to nibble at testing and analyzing and investigating what is not, specifically with reference to the creation of new markets and new institutions where they have not existed before.
Thatís what my morning commute got me thinking today; I wonder whatís in store for the ride home Ö