Risk and Return: Behavior's Impact on Value

 |  Includes: DIA, QQQ, SPY
by: Nick Gogerty

Here is an interesting and pretty lengthy paper by Eric Falkenstein. It consists of 2 major parts: a synopsis of failures of the typical risk reward relationship assumed in finance and an explanatory thesis involving individual agents who have envy in the pursuit of positional goods.

The refutation of the risk / reward relationship is well trodden ground and Falkenstein provides a nice summary by rounding up quite a few example papers. These include everything from bond yields, call options and even the risk reward of investing in G, PG and R rated film productions. If you are interested in getting to grips with where modern finance fails, this is a good paper to review and then pursue the cited sources.

Refuting existing dogma by rounding up the relative papers is well and good, but sadly efficient market theory in regards to risk and reward will remain taught even if with caveats. The meme survives until replaced by another.

Admitting ignorance, even when correct doesn't sell. Admitting ignorance doesn't fill $40k/year MBA classroom seats. Admitting ignorance doesn't sell books, consulting hours or videos. Admitting to shared ignorance doesn't sell shares, close the deal, fill the conference center or win the day in the job interview.

No. For most it is better to play the game, admit the dogma may be farce and replicate the farce that is the return risk relationship held up by efficient market theory.

The more interesting part of the paper is Falkenstein's attempt to bind the micro and the macro of finance using individual behavior and macro outcomes. Behaviouralist economics typically focuses on the counter intuitive nature of the individual. This novelty then gets distilled into entertainment a la Freakonics type books.

Falkenstein rightly acknowledges and attempts to model systems using agent behavior, including not the "homogenous" popular agent of the rational man, homo economus, found in traditional economics but rather an agent who seeks positional goods and perhaps purchases ascribed or achieved status among his or her peers. This complex agent approach is probably a better reflection of the real drivers of economy. It is also devilishly complicated model.

The paper's conclusion is that the model put forward better approximates real outcomes. I must profess to having neither the skill nor the resource to credit or discredit the thesis in the paper or the underlying model.

I like where Falkenstein is going and wish others would follow. It is the realm of complexity and, therefore, potentially intractable, but I would rather have a narrative or weak theory lacking in mathematical rigour, than a rigourous model that is wrong.

Better to have no instruments when flying a plane than overly believing in a flawed instrument and acting on it.

A good empirical argument for behavior driving value is the significant value created by brands. Brands have no rational economic value, but are merely social signals used to convey some value to others or ourselves. The brand message or function varies based on the positional objective of the good or service. Branding often allow firms who own a position to extract a price premia and thus build value at the bottom line.

My own thesis on the function of brand varies depending on the good or service, but breaks down crudely into the following:

  • Scarcity and desired social status signally via a visible considered purchase: Rolex, Mercedes, Abercrombie & Fitch, Trump etc.
  • Consistency and familiarity of a repeat / impulse purchase: Tide, Marlboro, Kellogg's corn flakes
  • Shared social ritual and repeatable quality signal: Dunkin Donuts, Starbucks, etc.
  • Personal and social attribute signalling via a halo effect: Nike, Addidas, Abercrombie & Fitch, Grey Goose vodka, etc.

Efficient market theory is the output of a human process and therefore a social area of research. This puts market theory beyond most of the simpler models thrown at it.

Most finance and economics participants will admit that the volatility, risk reward model is flawed. Unfortunately ignorance doesn't sell and bad dogma is better than no dogma for most.

The agent based models approach while probably correct is currently too complex for most to be consumed and regurgitated in the education and research milieu. The problem of a general economic belief system itself may thus be intractable until the proper language, framework and tools makes it easier to disseminate and digest. In economics we are still trying to explain the behaviour of individual particles when what we want is a way to take the temperature of a system. Reward vs. risk (volatility) is no thermometer.