Comment by p4wnc6 | Hacker News Reader

Comment by p4wnc6 | Hacker News Reader

Parent Full thread

Why do you exclude economics? I see economics as practically leading the way with regard to dismissals like you suggest.

One example I have first-hand experience with is the low-volatility anomaly. CAPM-like theories predict a roughly linear, positive-sloping relationship between risk and return.

Yet decades worth of various empirical measurements of the riskiness of assets and portfolios unequivocally demonstrates that it is a negative-sloping relationship (e.g. for your willingness to bear an additional unit of risk, you actually receive less return).

But everyone says basically the same thing. "Things are more complicated in real markets."

It would be reasonable to say that CAPM is a mathematical result, and the evidence suggests the mathematical model required for CAPM simply fails to correspond to reality. But many finance folks, especially in academia, won't admit it. For them, CAPM must also be a descriptive theory.

Two sources that illustrate what I'm talking about:

[0] Baker, Bradley, and Wurgler, "Benchmarks as Limits to Arbitrage: Understanding the Low-Volatility Anomaly" FAJ Volume 61, 2011. < http://www.cfapubs.org/doi/pdf/10.2469/faj.v67.n1.4 >

[1] Baker, Malcolm P. and Bradley, Brendan and Taliaferro, Ryan, The Low Risk Anomaly: A Decomposition into Micro and Macro Effects (September 13, 2013). Financial Analysts Journal. < http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2210003 >

*disclaimer: The authors of these two papers were my bosses in a previous role, and I wrote the code and generated the results for the second paper.

What you are describing in economics - while perhaps a failure mode - is the exact opposite of what this article discusses. The critique of this article is that valid abstractions and theories are rejected due to a lack of "nuance", rather than a failure to explain and predict reality:

Second is the ever more extensive expansion of some theoretical system in a way that effectively closes it o from rebuttal or disconrmation by anything in the world...It is an evasion of the demand that a theory be refutable.

Your critique of some economic theories is a totally different one - that they make specific predictions which don't come true.

Yes, but as far as the economists are concerned, asking for someone to rectify CAPM with the low-volatility anomaly is just nuance.

The author of this piece is saying, Let's not just throw unconstrained nuance at a theory: let's look at a holistic and sort of complexity-adjusted assessment of how well it's doing before we start asking it to behave like a Swiss Army knife, with a special mode to satisfy every empirical whim.

I'm saying that this -- the argument against nuance of this type -- is used to defend otherwise flatly wrong theories. (And I misread your comment to be saying that about other fields of social science too.)

In other words, when someone says, "But wait a minute, can you augment CAPM to handle the particular messy real-world case that I care about?" the response is: sorry, but we have to Occam's Razor away your request, we can't tailor it to everything.

The problem is that that messy real-world case is not some bespoke, uncommon corner, but is the most up-front qualitative property of the whole theory.

I guess what I'm saying is that the author says we should Occam's Razor-ify requests for nuance, yet in economics people masquerade as if they are doing legit Occam's Razoring (we can't tailor our analysis to every whim) when really they are just protecting pet theories that simply are bad.

It's actually a caution against the way the OP's advice about constraining nuance can be misused and politically abused to defend models that are simply wrong.

>t would be reasonable to say that CAPM is a mathematical result, and the evidence suggests the mathematical model required for CAPM simply fails to correspond to reality. But many finance folks, especially in academia, won't admit it.

In other words, "Things are more complicated in real markets". I don't see any problem with wanting descriptive theories. Economists are interested in building models whose parameters can be tweaked in ways that correspond in a clear meaningful way to a possible actions (e.g. change government expenditure on X by Y). If you try to just maximize predictive power you're missing the point. Economics can be used to predict the future, but I would say the main goal of the field is to produce theory that will provide valid methods of altering the future.

This doesn't make a lot of sense in this example with CAPM. The economic theory provides something that is abjectly wrong. It is literally claiming there is a positive relationship between two quantities when the evidence suggests the relationship is negative.

The theory is grossly qualitatively wrong, not merely just on the edge of being quantitatively wrong in a few cases or something.

So you're saying that we need wrong theories that can guide policymakers in the right direction? That sounds like a bad idea. It's basically "I want you to do this thing. Let me invent a theory that makes me sound more persuasive. Don't worry about the details!"

Yes, we need theories that are easy enough to reason about that they provide useful insight, but I don't think this is even close to being too complicated.

Reply on news.ycombinator.com