A fundamental point that I make in this book is that individuals tend to be smart and are capable of making decisions that are in their own best interest.a But their decisions are impacted by their decision-making environment and their decision-making capabilities (which include bargaining power and human and gender rights). Gaps in this environment and their capabilities, which should be expected in the real world of complexity and differential power relationships, can result in decisions that are not best or ‘optimal’ from the point of view of the individual. But this has more to do with circumstances that are exogenous; largely beyond the control of the individual. Ultimately, people end up making smart decisions, doing the best they can (or satisficing) given their particular circumstances.b But these smart individuals can make poor or error-prone decisions given the constraints and opportunities (or lack thereof) which they face. And, these sub-optimal decisions can be improved upon by changing their decision-making environment and their decision-making capabilities (Altman, 2017a, 2017b).
This does not mean that smart decision-makers (concerned with their own wellbeing or even that of their families or close friends) will be intent on doing the best for their communities, firms, or even households. Their smart, intelligent, or rational behaviour can be harmful to others. As economists would put it, individual behaviour can have negative externalities. This problem can be fixed by repairing the decision-making environment so that individuals bear the cost of their socially pernicious behaviour (Thaler & Sunstein, 2008).
Overall, I argue, one should respect the preferences of individuals. Individuals, however, require the environment and capabilities to construct informed preferences and make choices that improve their wellbeing without causing harm to others. One should not be bent on forcing and manipulating individuals to make choices that experts prefer, unless free choice causes harm to others.
This book builds upon the research of Herbert Simon and, more generally, the Carnegie-Mellon school of behavioural economics. I refer to this approach to behavioural economics as the bounded rationality methodological approach to behavioural economics (Altman, 1999, 2005, 2015, 2017a, 2017b).c In this perspective, the prior assumption is that decision-makers are relatively rational, intelligent and smart (satisficing, boundedly rational and evolutionarily rational). As one of the intellectual leaders of the Carnegie-Mellon school, James March (1978, p. 589) stated, it is of primary importance to determine if we can explain human behaviour in terms of rationality, broadly defined, even if at first glance such behaviour does not appear rational and might even appear to be error-prone or ‘biased’. More generally, I refer to this methodological approach as smart decision-making (Altman, 2017a, 2017b), which encompasses bounded rationality, procedural rationality, fast and frugal heuristics, slow and fast thinking, the brain as a scarce resource (following the insights of Freidrich Hayek) and the institutional, sociological and psychological-neurological determinants of decision-making. This is counter-posed to the world view of conventional or neoclassical rationality as well as the heuristics and biases perspective on behavioural economics, pioneered by Kahneman and Tversky (Kahneman, 2003, 2011), that now dominates contemporary behavioural economics. This perspective is closely tied to the nudging approach to public policy wherein one serious point of focus is assuming that experts can drive welfare enhancing decisions on behalf of highly biased and, moreover, persistently biased individuals.
Smart decision-making encompasses intelligent or smart decision-makers or agents, who develop or adopt decision-making processes and make decisions given their cognitive limitations, the decision-making mechanism of the brain, individuals (or economic agents) decision-making capabilities, decision-making experience, environmental factors, which include institutional and legal parameters, culture and norms, relative power in the decision-making process and related sociological factors. It is also recognized that cognitive limitations are affected by technology (computers and calculators, for example), the capabilities to effectively use new or improved technologies and the learning processes that affect how the brain is hardwired (neuroplasticity). Smart decision-makers or agents do the best they can, given the pertinent circumstances that affect the decision-making process and related outcomes. Herbert Simon refers to the act of doing the best we can as satisficing behaviour. Satisficing, however, need not result in the best possible or optimal outcomes for the firm, household, society or individual; but it can, depending on circumstances.
Deviations from optimality do not imply that decision-makers are not smart and in this sense irrational. Nor does establishing that decision-makers are smart imply that decision-making outcomes are optimal. Here rationality, broadly defined, relates to the choices people make and the decision-making processes adopted by individuals given their various constraints and opportunities as well as their decision-making environment. Optimality in production and consumption at an individual, firm, household or social level need not necessarily flow from smart decision-making. Smart decision-making, however, would often be a necessary but not a sufficient condition for optimality to be obtained. What these sufficient conditions might be are critically important to research that stems from the smart agent or smart decision-making perspective.
Inadequate decision-making environments, for example, would preclude smart agents from achieving optimal results from their own and from society’s perspective (where externalities exist). For example, you might wish to increase your savings for retirement, but you invest in high-risk high-return financial paper because of the false or misleading financial information provided to you, resulting in you losing much of your savings. You might employ a low wage strategy for your firm because you are adopting (with good intent) a false mental model (theoretical framework) on the impact of higher wages of firm competitiveness. Women might want to have one child, but they end up giving birth to four or five, because they are not empowered to realize their preferences. A firm’s productivity might not be maximized because decision-makers are maximizing a complex utility function that includes managerial slack and short-term returns. None of the above is a product of irrationality. They are a product of preferences, misleading information, false mental models, decision-making capabilities, experience and the overarching decision-making environment.
Conventional theory’s point of focus is on very generalized concepts related to how humans should behave and are expected to behave to generate optimal outcomes. As long as the analytical prediction is correct, all is well. This is effectively the correlation-based analysis promoted by Friedman (1953). If you get the prediction correct, you can assume for reasons of simplicity that humans behave as if they are maximizing profits, minimizing costs and maximizing utility (which is often assumed to be identical wealth maximization, controlling for risk). The realism of the simplifying assumptions we make about decision-makers, the decision-making processes and the decision-making environment are not of importance from this perspective. We can simply assume that individuals behave as if they are maximizing profits or utility, as long as the analytical prediction is the correct prediction (Berg & Gigerenzer, 2010). The assumption here is that individuals ideally should behave ‘neoclassically’, if they are rational, which they are assumed to be. Rationality is defined in terms of neoclassical rationality. Apart from this, what transpires in the decision-making process is not of substantive interest. We simply abide methodologically with neoclassical simplifying assumptions of how individuals behave within the firm and in the household. Moreover, it is further assumed that the decision-making environment allows for the realization of optimal outcomes, given neoclassical rationality, for the individual, the household and the firm.
The analytical focus, therefore, is on correlation as opposed to true causation, where the latter relates to determining what particular behaviours and decision-making environments generate particular outcomes. Modelling true causation would address issues of spurious correlation, omitted variables and the possibility of alternative behaviours, yielding similar sustainable outcomes. What is key is the determination of what specific behaviours, decision-making processes and institutional and sociological variables yield specific outcomes. This deeper modelling agenda is part of the bounded rationality approach to behavioural economics.
The bounded rationality tradition in behavioural economics plays particular attention to identifying the actual decision-making process that generates particular outcomes. It ventures into the black box of the firm, the household and the individual. Only by understanding how individuals actually behave, how they make decisions, can we determine if these decisions are smart and in this broad sense rational. Hence, rationality here is contextualized. Benchmarks for what is rational are, therefore, not constructed by some imagined ideal unrelated to the decision-making capabilities and environments of the individual, household or firm.
For this reason, a core attribute of the approach taken in this book is, following from Simon, the overall importance of reasonable, reality-based, simplifying modelling assumptions for robust economic analysis. Related to this is the significance of situating our definition of rationality and smart decision-making in context. Simon writes (1986, p. s209):
The judgment that certain behavior is ‘rational’ or ‘reasonable’ can be reached only by viewing the behavior in the context of a set of premises or ‘givens.’ These givens include the situation in which the behavior takes place, the goals it is aimed at realizing, and the computational means available for determining how the goals can be attained. In the course of this conference, many participants referred to the context of behavior as its ‘frame,’ a label that I will also use from time to time. Notice that the frame must be comprehensive enough to encompass goals, the definition of the situation, and computational resources.
The smart agent, smart decision-making approach to decision-making and behavioural economics not only stands in contrast to what we find in much of conventional economics, it also stands in contrast, as mentioned above, to a theme running through much of contemporary behavioural economics where much of the typical individual’s behaviour is deemed irrational and error-prone (Altman, 2017a, 2017b). This is the heuristics and biases approach pioneered by Kahneman and Tversky (Kahneman, 2003, 2011). A common thread running through this approach and conventional economics is adopting neoclassical benchmarks for rationality and, flowing from this, benchmarks for optimal outcomes in the domain of consumption and product...