60 years of wasted effort

Bold as this claim is, it has mainstream support. Eugene Fama and Ken French, two leading conventional financial economists, have written that the CAPM “has never been an empirical success” and that most applications of it are “invalid”.

The model’s biggest claim is that equities’ returns should be positively related to their sensitivities to market returns – their betas. This, however, is simply false. Economists have known since the early 1970s – pretty much when the model was first tested – that the opposite is the case. It is low-beta stocks, not high-beta ones, that do best. This has prompted some economists to believe, in the words of Eric Falkenstein of Pine River Capital Management, "there is generally no empirical risk-reward relation".

This would have been no surprise to Ben Graham. He thought investors could have both low risk and high returns if they bought shares with a good “margin of safety” – ones whose prices were well below companies’ intrinsic value. To him, high returns were not a reward for risk, as the CAPM says, but rather a reward for the skill and diligence required to find undervalued companies.

And it is defensive stocks that often have Graham’s margin of safety. Many of these have what Warren Buffett calls “economic moats” – sources of monopoly power such as big brands that allow them to fend competition and so maintain profits thereby limiting their downside risk.

Defensive stocks beat the market

There’s a second failure of the CAPM – the fact that momentum stocks do well. Again, this wouldn’t have surprised Graham. He saw what the CAPM denies – that investors can be driven by wages of sentiment. Often, he wrote, “Mr Market lets his enthusiasm or his fears run away with him”.

This raises the question: if Graham’s approach is so much better than the CAPM, why did academia choose the latter (and still teach it) and ignore the former?

One reason is that Graham’s approach is not so much a scientific theory as a guide to action. Its core concept – intrinsic value – is one that different investors estimate differently: that’s why rational people trade shares. This means it is difficult to test. The CAPM, by contrast, makes simple falsifiable predictions – albeit ones that have actually been falsified.

But there’s something else. Otuteye and Siddiquee’s criticisms of the CAPM remind me of criticisms of modern macroeconomics. Here, too, critics allege that the discipline has become dominated by theories that are mathematically elegant but practically useless – dynamic stochastic general equilibrium models. Paul Romer alleges that macroeconomics has become dominated by “mathiness” – an erroneous conflation of tricky equations with real science. Perhaps the same is true of financial economics.

The problem here might not be confined to economics. The philosopher Alasdair MacIntyre said in 1988 that universities “can on occasion effectively exclude from academic debate and enquiry points of view insufficiently assimilable by the academic status quo”. New work by George Akerlof and Pascal Michaillat shows how this can happen. If new researchers gain preferment by pleasing incumbents – as can happen with decisions on academic tenure or about which papers to publish – then they’ll prefer to work within the existing, inadequate theories of those in power rather than challenge those ideas. “When tenured scientists show favoritism toward candidates for tenure with similar beliefs, science may not converge to the true paradigm,” they write.

We shouldn’t, though, overstate the differences between the CAPM and Graham’s approach. Both agree upon one very important fact – that fund managers don’t beat the market and so don’t justify their fees. Back in 1972, Mr Graham wrote: “Even the majority of the investment funds, with all their experienced personnel, have not performed so well as the general market.”