Behavioral Finance

Financial systems are composed of agents and the exchange of financial instruments within institutions and markets. Thus to understand this system we need to first understand something about the logic of the agents in the system; how agents make decisions about the allocation and exchange of financial assets is clearly important. Agents have motivations that drive them to value things. In order for agents to pursue their valued ends, act and affect their environment towards achieving those outcomes they need some logic under which to do this. That is to say, they need to take in and process information according to rules so as to generate a response that will lead to their ultimate desired end. If we want to understand how agents behave within a financial context we thus need to define, to some extent, how this is done.

There are two fundamentally different paradigms with respect to understanding how agents make choices and evaluate financial resources. Our actions may derive from individual deliberative reasoning, and this would be called a rational action, or they may derive from some other non-deliberative source, such as instinct and emotion, heuristic or social cues etc. The assumption of rationality – and market efficiency – is central to modern portfolio theory (the CAPM), and to the Black–Scholes theory for option valuation. Rational means designed or conducted according to reason. Reasoning is a process whereby data is amassed, processed according to some logic in order to produce a conclusion that is both logically consistent and in accordance with objective data.

A rational decision is one where an agent amasses all relevant information, processes it according to a consistent and objective logic and then acts in accordance with the outcome of that process. In so doing the agent acts independently, they act on their own internal logic in an autonomous fashion. Thus for a decision to be rational, there are a number of requirements; firstly that the agent has all the relevant information and that any information that is not fully known can have a probability distribution assigned to it. Secondly, the agent must act according to a consistent and objective logic set, which means that the choices made will not change unless there is some alteration to the objective factors determining the decision. Agents have to have a fixed set of preferences and these preferences have to be complete – the person can always say which of two alternatives they consider preferable or that neither is preferred to the other. An actor is acting rationally when they take account of available information, probabilities of events, and potential costs and benefits in determining preferences, and act consistently in choosing the self-determined best choice of action.

Although the term rationality simply means according to reason, the requirements for achieving this are only met in some circumstances or some of the time. Rationality requires that we have intelligent calculating agents operating in relatively simple environments. In such circumstances, it is reasonable to say that people often act rationally in pursuing the things they value. However, just as often we will be dealing with contexts wherein agents with limited intelligence and limited propensity for reasoning will find themselves in relatively complex environments. In such circumstances, agents do not use reason to determine their actions but use a variety of alternative means, that are contingent upon the social, physical or cultural context within which the decisions are being made. That is to say, that the rationality of an agent is bounded, when it reaches a limit it switches to alternative means for making decisions. This limit is both contingent on the particular subject, i.e. their propensity to use reason, and the environment, i.e. how complex the environment is.

Bounded Rationality

While traditional economic and financial theory posits an objectivist rational understanding of agents decision making, complexity theory sees the environment within which agents make choices as often being fundamentally too complex for them to make rational choices and is thus based upon more of a subjectivist model to human behavior which is the idea of bounded rationality. Bounded rationality is the idea that agents do not have complete information and/or cannot rationally process all the information available to them. Agents find themselves in a system that they do not fully understand and have incomplete knowledge. Because they a part of complex systems that they can not fully know they use all sorts of shortcuts to try and find some basis for action; they create narratives, they copy others, they use simple rules that have worked for them in the past etc.

Much of the time people operate in environments where there is incomplete information, radical uncertainty may exist, where they do not wish to expend the time and energy to reason through their actions, we don’t want to take the responsibility for our actions, there are time limitations, social power dynamics or a series of other limiting factors involved. In such circumstances we defer our decisions to heuristics – which are shortcuts – we use social and cultural norms, we copy what others do, we allow random events to determine our decisions.

Behavioral finance deals with people as they are, i.e. with many psychological biases that means their actions and behaviors deviate from what would be expected if they were basing their decisions on rational thought processes alone. Some of these include anchoring, meaning the person places an arbitrary value as an anchor point and then bases future expectations around that. Another is mental accounting, referring to the process whereby people divide up their money mentally into different accounts based on subjective criteria, such as the source of the money or the proposed usage of it, this can prompt biases and systematic departures from rational, value-maximizing behavior.

A fully rational agent has a fully comprehensive view of the system or environment within which they are operating. However, this is not the case for most investors whose patterns of activity follow a vision that is more narrow and a function of their psychology, e.g. what often dominates peoples’ behavior is immediate losses and gains when what really matters is overall wealth. This is the so-called disposition effect, which relates to the tendency of investors to sell shares whose price has increased while keeping assets that have dropped in value. Given rational expectations, the price at which you purchase a stock should not determine when or whether you sell it.


Complexity theory sees uncertainty as something inherent to nonlinear complex systems and environments, not something that can be modeled using probability and statistical methods. Much of mainstream finance assumes uncertainty can always be modeled mathematically, and that everyone should arrive at the same assessment of an uncertain event’s probability, at least given the same information. The Austrian School notes every individual always perceives a unique information set, and that each individual values every item of information in a unique manner. Even if participants had exactly the same information, their assessment of its importance would be subjective. One person’s behavior may be quite different from another’s, even when presented with identical choices.1


The efficient markets hypothesis is that people bring large amounts of information into the market and cannot be systematically incorrect – that prices reflect all available information. It views actors as rational and independent and it is possible to use statistical averages to abstract away the underlining diversity and particularities to arrive at a single average that can then be used to represent the whole sample. This is often done in financial models where a number of heterogeneous assets – like mortgages – are bundled together and given a single value based upon the average.  This only works well if the collection of variables is random, independent and identically distributed (IID) if each random variable has the same probability distribution as the others and all are mutually independent.

However, if you view agents as not necessarily rational and as being interconnected than the assumptions behind this no longer hold. An actor in a market is often not disconnected nor acting in an isolated fashion but instead communicating and interacting with others creating interdependencies and feedback loops. Self-fulfilling speculative attacks by investors expecting other investors to follow suit given doubts about a nation’s currency peg is an example of a feedback loop that creates macro-level disequilibrium in the market defying statistical averages.

Value Theory

Finance is an abstraction of the real economy. The foundations of finance is the abstraction of value, i.e. it is built on an accounting system that is designed as an information record of value within the real economy. Whereas real economic assets have value in use, finance is based upon an assessment of the value of something, thus there is always a subjective factor involved in finance. As soon as we come to exchange an asset it also comes to have a subjective value which is the value that the particular individual places upon it based upon their subjective evaluation of it, which is a product of their particular desire for it and many psychological and contextual factors that contribute to that assessment. This being noted many of the traditional models – efficient market hypothesis – in finance are based upon the idea that financial assets derive their value from market fundamentals. Fundamental analysis is a method of valuing securities that attempts to discover their true value by examining related economic and financial factors; looking for the intrinsic or “true” value of the asset. This is an objectivist view of value theory, that the value of something derives from objective factors.

The subjective theory of value is a theory which advances the idea that the value of a good is determined by the importance an acting individual places on a good for the achievement of his or her desired ends, what is called utility. The theory of extrinsic value posits that value cannot be measured or observed directly but the value is simply given to things based on people subjective perception of them.


As noted by George Soros, financial markets involve a feedback loop between the subjective models of agents and the objective structures of the market. The subjective models that actors create to understand the market creates the actions they take which then affect the state of the market which then feeds back to shape their understanding and acting. There is a constant feedback between agents and structure, agents acting, affecting the overall structure, with that then feeding back to change their behavior, with a continuous feedback loop between the subjective and objective.

Models and theories are not real. There is a real world and models do not exist there. They simply help us to interpret and give structure to it, sometimes even predict it. But this is not to say that models do not affect the world – quite the contrary, within the social sciences, they have a very significant effect. We create these models. People adopt them and go around seeing the world through them and acting on them. In so doing, the models change the world. Thus, in creating models we are responsible for creating the future state of the system.

This is one insight from Andrew Lo’s adaptive market hypothesis that integrates an aspect of the efficient market hypothesis and behavioral finance’s understanding – that both have aspects to contribute but much is dependent upon the kind of market environment that is being operated within. In simpler environments that are knowable and limited in dynamic change and unpredictability, an objective analysis may well reflect outcomes in the market. However, as the environment becomes more complex, more dynamic, uncertain and unknowable and the actions of agents more interconnected, then more subjective factors come into account for which a more complex model incorporating psychology and contextual factors must be used.

  1. (2018). [online] Available at: [Accessed 9 Aug. 2018].