On quantitative finance, truth, and American pragmatism

The implicit assumption in the dominant school of thought in quantitative finance suggests that analytics can be studied separately from making decisions. In other words, “what it is” is separable from “what to do about it”.

For example, we use various numerical methods to estimate market volatility, correlations, default probabilities, we build predictors of volume, spread and momentum, we price derivatives, regress changes in asset prices against changes in economic data, etc. Hordes of quantitative analysts produce and code ever more granular and precise models whereas huge data processing farms spew gigabytes and terabytes of predictors and estimators. And then we pass the exhaust of these analytics engines to appropriate decision makers, most often algorithms. Continuously operating decision making machines make use of analytics and realtime data to interact with markets. Analytics are plugged into algorithms to allow them to act to produce expected results.

Production of quantitative analytics in this framework is akin to exact sciences: both claim to yield objective and absolute knowledge about the world. It is not surprising that a lot of quantitative analytics is done by people with at least some background in exact sciences.

If “what it is” is the objective basis for decision making, then achieving success and competitive advantage is possible by extracting better, faster and more complete quantitative analysis. In this framework no trade will ever occur if everyone has the same perfect information, the truth. To put it bluntly, in this framework every trade is a trade between a fool and a scoundrel.


Yet, such quantitative, scientific approach is not the only possible philosophical foundation for making disciplined decisions in finance.

Instead of viewing the world as a set of things to study and then to act upon, in an alternative approach we view the world as the arena of human agency, as the place where we act:

  • Every market participant is characterized by his or her objectives, preferences, tastes, and tolerances to various forms of risk;
  • Market activity in general and trades in particular occur as the exchange of ideas about the relative value of assets, as the exchange of information about preferences and objectives and tastes of individual actors — rather than as the process of discovery of an absolute truth, e.g. the price of an asset;
  • It is not self-evident that observables which we extract from markets and which we massage to produce quantitative analytics actually correspond to anything real and persistent. Yes, the averages are more stable than the raw data, but this fact does not prove that any average exists outside of our mind;
  • For example, it is not self-evident that volatility or probability of default or market volume or average duration of a bond exist in the same sense as the concentration of carbon dioxide or the atmospheric temperature, or electric and magnetic field in the solar corona exist;
  • Even if we adopt the view that quantitative analytics do reflect some reasonably persistent characteristics of markets, it is not self-evident that the extraction of these analytics from market data can be fully separated from the decision making process that serves a particular purpose;
  • Selection of observables is deeply rooted in our beliefs of what is relevant and what is not relevant for a particular purpose. These beliefs are imperfect and are affected by objectives, tastes, and preferences;
  • If this is indeed the case, rather than imposing our beliefs onto our decision process, we may as well ask algorithms to separate relevant factors from irrelevant as part of machine learning on a case by case basis (which is already done with different degrees of success in finance);
  • And while we are at it, we might as well ask algorithms to learn best decisions given the circumstances of markets and our goals and objectives.

In the alternative framework we don’t claim the existence of analytics independent of our objectives and preferences. We do away with the scientific stage of learning “what it is”. Rather, from the ground up we operate in the world of actions and try to achieve the best we can given the circumstances of our actions and following our beliefs and goals.


To distinguish this approach from quantitative finance I would call such framework pragmatic or instrumental finance.

In it I follow the tradition of American pragmatism:

Any idea upon which we can ride, so to speak; any idea that will carry us prosperously from any one part of our experience to any other part, linking things satisfactorily, working securely, simplifying, saving labor; is true for just so much, true in so far forth, true instrumentally. (William James, 1907, emphasis is mine, VSG)

Or, in a more modern interpretation:

Pragmatism is based partially on the idea that there is no one intellectual “ring to bind them all.” Pragmatism denies that there is “Truth,” with a capital “T.” It denies that science, philosophy, religion, or any other form of inquiry should seek truth as the holy grail of human study. Rather, pragmatism argues that while truth may exist some place in the universe, we mere evolved primates must be content with finding ideas and theories that work for us, that serve as tools for our different tasks. (Daniel Everett, 2012, emphasis is mine, VSG)