The distinction between decision-making by rational individual deliberation and decision-making by following socially constructed canals or by copying others is a main basis of alternative methodologies in the social and behavioral sciences. Following conventional wisdom and the standard individualistic rhetoric of many economists, Bentley et al. locate economics in the northwest quadrant of their two-dimensional map.
The value of Bentley et al.’s schema does not depend on associating disciplinary paradigms with quadrants. However, it is worth taking trouble to jam such associations, especially as the authors offer remarks that encourage them, at least with respect to economics. The idea that economists in general model decision-making or choice as individual and transparent reprises the caricature of the discipline promoted by behavioral economists. In fact, top economics journals publish modeling approaches inhabiting all four of Bentley et al.’s quadrants; and models from the eastern side of their map are steadily gaining in importance.
Before elaborating on this main point, one semantic caveat is needed. Bentley et al. refer to their subject matter as “decision making.” Most economists would understand this to refer to processes modeled by formal decision theory, including decision theory that accommodates bounded rationality. On that usage, Bentley et al.’s contention that economists locate decision-making in the northwest quadrant becomes tautological. But Bentley et al. diverge from the economist's convention by saying that any decision-making occurs in other quadrants. In economists’ language, Bentley et al.’s topic is choice, that is, the family of information-processing patterns that link changes in environmental contingencies to changes in agents’ behavior at least partially by way of cognitive comparison of actual states, hypothetical states, and goal states. I follow this second convention here.
Behavioral economists who see themselves as leading a “paradigm shift” typically take methodological individualism for granted. That is, they suppose that models of economic phenomena on social or macroeconomic scales of representation and analysis are built from – aggregated from, in one formal sense or another – models of individual deliberation and/or neural processing. They then argue that traditional economists have exaggerated the logical transparency and consistency of individual choice, leading them to conflate real, empirical human choice with decision-making by a fictional ideal Homo economicus. In Bentley et al.’s terms, such critical behavioral economists then emphasize the importance of psychological framing influences, tolerances for inconsistencies, and conflicting motivations that tend to drag real choice in the southerly direction. (See Angner [Reference Angner2012] for a survey.) The rhetoric of some such theorists implies that most individual choice should be modeled as a southwest phenomenon (Ariely Reference Ariely2008). Most are more cautious, maintaining only that psychological influences on economic choices tend to generate substantial and thick tails in the southwest (Angner & Loewenstein Reference Angner, Loewenstein and Mäki2012).
Bentley et al. seem to endorse this picture. Echoing standard rhetoric of behavioral economists, they say that “[b]ecause Homo economicus is located at the extreme northwest, he does not appear to be the primary model for human culture” (sect. 4, para. 8). However, the behavioral economics literature tends to reflect unreliable history of economics. Homo economicus is much less important in the historical development of the technical apparatus of standard economic analysis than is alleged. Samuelson (Reference Samuelson1947), after analyzing H. economicus in one chapter of his classic book on theoretical foundations – arguably the most influential text in postwar economic theory – dismissed the construction as being of little working importance (p. 117), and gave it no role in later chapters. Samuelson synthesized a neoclassical tradition that emphasized consistency in choice with a Keynesian relaxation of informational transparency and an openness to acknowledgment of structural, market-scale constraints and influences on choice.
We are commonly told that the Samuelsonian synthesis unraveled in the 1970s and was displaced by a new modeling style that insisted on microfoundations for aggregate-scale (macroeconomic) models and based market-scale analysis on the behavior of a single “representative” agent who maximizes expected utility à la Savage (Reference Savage1954) using all available information, with the only concession sometimes made to Keynesian social influences being specific dynamic rigidities such as sticky prices. If this approach had in fact taken over economics, as its promoters often announced that it would or already had, then Bentley et al. would be correct to locate economics as a science of the northwest. But rational expectations models, with or without “New Keynesian” elements, never eliminated alternative approaches (Colander Reference Colander1996; Frydman & Goldberg Reference Frydman and Goldberg2007), and in the wake of the 2008–2009 financial disaster, they are in headlong retreat (Timbeau Reference Timbeau, Solow and Touffut2012).
The historically brief overemphasis on economic choice as paradigmatically individual and transparent has not merely passed in macroeconomics; it is also being pushed back in non-behavioral but firmly empirical, frequently experimental, microeconomics (Smith Reference Smith2008). The liveliest source of modeling innovation in microeconomics concerns effects of agent heterogeneity, which gives rise to emergent structural parameters. A particularly promising such approach estimates maximum-likelihood mixture models that yield distributions of types at the population level and uses this as a basis for predicting effects of incentive changes on markets and other social structures. (See, for example, papers in Cox & Harrison Reference Cox and Harrison2008.) Essentially, these are representative agent models with multiple representatives. The conditioning variables on which the representatives are clustered are demographic. Thus, the representatives should not be identified with individual people, and we find ourselves in Bentley et al.’s northeast quadrant. Models that accommodate heterogeneity can also be embedded in global games (Carlsson & van Damme Reference Carlsson and van Damme1993; Morris & Shin Reference Morris, Shin, Dewatripont, Hansen and Turnovsky2003), where agents often disregard and so fail to reveal their private information. Though players of global games are modeled as fully rational, exogenous noise can push them into Bentley et al.’s southeast quadrant, leading to such phenomena as market bubbles and bank runs. Finally, microeconomic modeling of labor, product, and technology diffusion markets increasingly borrows network theory (Goyal Reference Goyal2007), the dominant mathematics of the southeast.
There is irony here. Behavioral economists criticize standard economics as, in Bentley et al.’s terms, stuck in the northwest. They then creep cautiously south, while the economists they criticize range all over the map.
You have
Access
The distinction between decision-making by rational individual deliberation and decision-making by following socially constructed canals or by copying others is a main basis of alternative methodologies in the social and behavioral sciences. Following conventional wisdom and the standard individualistic rhetoric of many economists, Bentley et al. locate economics in the northwest quadrant of their two-dimensional map.
The value of Bentley et al.’s schema does not depend on associating disciplinary paradigms with quadrants. However, it is worth taking trouble to jam such associations, especially as the authors offer remarks that encourage them, at least with respect to economics. The idea that economists in general model decision-making or choice as individual and transparent reprises the caricature of the discipline promoted by behavioral economists. In fact, top economics journals publish modeling approaches inhabiting all four of Bentley et al.’s quadrants; and models from the eastern side of their map are steadily gaining in importance.
Before elaborating on this main point, one semantic caveat is needed. Bentley et al. refer to their subject matter as “decision making.” Most economists would understand this to refer to processes modeled by formal decision theory, including decision theory that accommodates bounded rationality. On that usage, Bentley et al.’s contention that economists locate decision-making in the northwest quadrant becomes tautological. But Bentley et al. diverge from the economist's convention by saying that any decision-making occurs in other quadrants. In economists’ language, Bentley et al.’s topic is choice, that is, the family of information-processing patterns that link changes in environmental contingencies to changes in agents’ behavior at least partially by way of cognitive comparison of actual states, hypothetical states, and goal states. I follow this second convention here.
Behavioral economists who see themselves as leading a “paradigm shift” typically take methodological individualism for granted. That is, they suppose that models of economic phenomena on social or macroeconomic scales of representation and analysis are built from – aggregated from, in one formal sense or another – models of individual deliberation and/or neural processing. They then argue that traditional economists have exaggerated the logical transparency and consistency of individual choice, leading them to conflate real, empirical human choice with decision-making by a fictional ideal Homo economicus. In Bentley et al.’s terms, such critical behavioral economists then emphasize the importance of psychological framing influences, tolerances for inconsistencies, and conflicting motivations that tend to drag real choice in the southerly direction. (See Angner [Reference Angner2012] for a survey.) The rhetoric of some such theorists implies that most individual choice should be modeled as a southwest phenomenon (Ariely Reference Ariely2008). Most are more cautious, maintaining only that psychological influences on economic choices tend to generate substantial and thick tails in the southwest (Angner & Loewenstein Reference Angner, Loewenstein and Mäki2012).
Bentley et al. seem to endorse this picture. Echoing standard rhetoric of behavioral economists, they say that “[b]ecause Homo economicus is located at the extreme northwest, he does not appear to be the primary model for human culture” (sect. 4, para. 8). However, the behavioral economics literature tends to reflect unreliable history of economics. Homo economicus is much less important in the historical development of the technical apparatus of standard economic analysis than is alleged. Samuelson (Reference Samuelson1947), after analyzing H. economicus in one chapter of his classic book on theoretical foundations – arguably the most influential text in postwar economic theory – dismissed the construction as being of little working importance (p. 117), and gave it no role in later chapters. Samuelson synthesized a neoclassical tradition that emphasized consistency in choice with a Keynesian relaxation of informational transparency and an openness to acknowledgment of structural, market-scale constraints and influences on choice.
We are commonly told that the Samuelsonian synthesis unraveled in the 1970s and was displaced by a new modeling style that insisted on microfoundations for aggregate-scale (macroeconomic) models and based market-scale analysis on the behavior of a single “representative” agent who maximizes expected utility à la Savage (Reference Savage1954) using all available information, with the only concession sometimes made to Keynesian social influences being specific dynamic rigidities such as sticky prices. If this approach had in fact taken over economics, as its promoters often announced that it would or already had, then Bentley et al. would be correct to locate economics as a science of the northwest. But rational expectations models, with or without “New Keynesian” elements, never eliminated alternative approaches (Colander Reference Colander1996; Frydman & Goldberg Reference Frydman and Goldberg2007), and in the wake of the 2008–2009 financial disaster, they are in headlong retreat (Timbeau Reference Timbeau, Solow and Touffut2012).
The historically brief overemphasis on economic choice as paradigmatically individual and transparent has not merely passed in macroeconomics; it is also being pushed back in non-behavioral but firmly empirical, frequently experimental, microeconomics (Smith Reference Smith2008). The liveliest source of modeling innovation in microeconomics concerns effects of agent heterogeneity, which gives rise to emergent structural parameters. A particularly promising such approach estimates maximum-likelihood mixture models that yield distributions of types at the population level and uses this as a basis for predicting effects of incentive changes on markets and other social structures. (See, for example, papers in Cox & Harrison Reference Cox and Harrison2008.) Essentially, these are representative agent models with multiple representatives. The conditioning variables on which the representatives are clustered are demographic. Thus, the representatives should not be identified with individual people, and we find ourselves in Bentley et al.’s northeast quadrant. Models that accommodate heterogeneity can also be embedded in global games (Carlsson & van Damme Reference Carlsson and van Damme1993; Morris & Shin Reference Morris, Shin, Dewatripont, Hansen and Turnovsky2003), where agents often disregard and so fail to reveal their private information. Though players of global games are modeled as fully rational, exogenous noise can push them into Bentley et al.’s southeast quadrant, leading to such phenomena as market bubbles and bank runs. Finally, microeconomic modeling of labor, product, and technology diffusion markets increasingly borrows network theory (Goyal Reference Goyal2007), the dominant mathematics of the southeast.
There is irony here. Behavioral economists criticize standard economics as, in Bentley et al.’s terms, stuck in the northwest. They then creep cautiously south, while the economists they criticize range all over the map.