Geoffrey Wiseman

Are Markets Rational?

Planet Money #70: The Myth of Rational Markets (from mid-2009) discusses the idea that markets are rational, suggesting that you can, for instance, use the price of a house on the real estate market to determine the true value of a house.

I have a few bones to pick with this.

We Make Markets

Perhaps the platonic ideal of a market has some attributes one can generically describe, but for all intents and purposes, the formal markets we interact with are markets of our own creation, with rules that we made. Those rules may encourage the market to behave 'rationally', or not; they may also encourage different behaviours in different conditions. It doesn't seem terifically useful to have a discussion without dealing with the fact that whether markets behave 'rationally' or not is something we influence when we structure these markets and their rules.

Markets perhaps more like computer programs than they are logic; markets do what we tell them to do. In this way, I wouldn't really call them rational any more than I'd call a computer rational. They're just following a set of rules to process certain inputs and create certain outputs.

However, unlike computer programs, that operate in a very tightly controlled domain, markets exist in the real world. The real estate market includes both the formal market, with things like real estate agents and the Toronto Real Estate Board (TREB) as well as people simply giving one another some kind of compensation and transfering the title legally. This means that these rules only influence the behaviours of people, and if the rules run strongly counter to what the people want, they'll bypass them.

Garbage In, Garbage Out

Even if we do a fantastic job of creating a market that behaves in a way that we want to label 'rational', markets are still based strongly on their inputs. If we make terrible inputs, then we can expect the outputs to be bad.

In extreme scenarios, our inputs can seem like "what if?" scenarios. For instance, we recently ran a what-if scenario to find out what would happen if we encouraged people who have traditionally been unable to afford a home to buy one anyway by offering them mortgages at really low rates and not really checking their ability to pay. We ran that through the market and discovered that that results in significantly higher home prices, and a large number of people who couldn't afford the mortgages they had.

Simultaneously we were running another what-if scenario on what happens to seemingly low-risk mortgage backed securities when the mortgage market's inputs change significantly. Say, if lots of people owed significantly more on their homes than their homes were now worth.

We certainly can adjust the rules of these markets somewhat to help avoid these issues and cushion the blow when problems happen, but fundamentally, the output of markets will always going to be subject to its inputs, and if we provide garbage inputs, we can probably expect garbage results.