John Rapley's claim that economics has become a religion has kicked off another debate about the state of economics.

For me, there seems to be a presumption here which both sides seem to share but which I don't - or at least that I don't care about. It's that economics is a canon of work which is taught to students and the outside world. However, I think of economics differently - as a box of theories, evidence and mechanisms which help me make sense of the world. What I care about is: what are the facts? And: how do we explain them? I don't much care what's orthodox or heterodox.

I'll take an example from financial economics. We have a theory here - the capital asset pricing model - which predicts that riskier shares will on average out-perform safer ones, where risk is defined by a share's covariance with the market (or beta). I used to believe this.

But there's a problem with it. It's false. We know this not because of methodological diversity but because of the facts. We have a body of evidence which shows (pdf) that less risky stocks out-perform riskier ones - in flat contradiction to the CAPM*. I stress here that it's the body of evidence that matters. Single studies are prone to not being replicated, but the defensive anomaly has been found in different data and with different measures**. And the CAPM's failure is now accepted even by mainstream economists: Eugene Fama and Ken French have written that it "has never been an empirical success".

Now, maybe the CAPM is still taught uncritically in some business schools and universities. To the extent that it is, Rapley is right. But I don't really care. The purpose of university is to acquire credentials and a grounding for later learning rather than to get a complete education.

What I do care about is why the CAPM is wrong. A big part of the story lies in market imperfections.

Researchers at AQR Capital Management point out (pdf) that many investors cannot borrow and lend freely as the CAPM assumes. They express their bullishness not by borrowing to buy the market portfolio but by buying high-beta stocks - those which are in effect a geared play on the market. This causes such stocks to be over-priced on average, because most investors are usually bullish; they know about the equity premium. The counterpart to this is that defensives are under-priced.

A second imperfection is that short-selling is a damned sight harder in practice than in theory. Even if you're right in the long-term (a few weeks) that a stock is over-priced, margin calls can wipe you out in the near-term. This is especially true for volatile stocks.

Such frictions mean that mispricings persist. Again, we know this not because of abstract theorizing but because economists have done the hard yards of seeking the facts. David McLean and Jeffrey Pontiff show that only around half (pdf) of mispricings in the US are eliminated after they've been discovered. The proportion is even smaller in Europe.

What does all this tell us? In part, it leads me to side with some "heterodox" economists***. Market frictions such as borrowing and short-sales constraints (pdf) are not minor quirks that can be ignored until later in one's studies. They are instead part of the essential nature of the beast which generate the most important thing in economics - facts.

But I also sympathize with the "orthodoxy". Flaws with "mainstream" theory are to be found - in the first instance - not by armchair theorizing but by gathering facts.

Now, you might object that financial economics is a minor tributary of economics that nobody much cares about. For me, this is its appeal. Macroeconomics is, for many people, so tied up with their political passions that progress is more difficult.

* This isn't the only evidence against the CAPM: there's also momentum, among other things.

** In the day job, I've ran my own real-time test. I've simply taken an equal-weighted basket of the 20 lowest-beta stocks in the FTSE 350 and measured their performance. My chart shows that they have hugely beaten the FTSE 350.

*** The scare quotes are because I don't care about the distinction between heterodox and orthodox. What matters is what explains the facts.