Tuesday, August 25, 2015

Realist Theory and Complex Methods: The Path Toward Integration

In its purest form, economic theory describes its domain of study as it actually exists. Reality is not comprised of homogeneous agents who choose quantities of goods according to the mariginal unit of utility gained per dollar spent. If only life were so simple! In the real world, agents have no option but to make choices that confront scarcity of time and resources. These choices are often aided by some standard or algorithm. In some cases, choices appear to be random. Despite this, order in society is apparent enough to suggest to the observer with even a slight inclination toward introspection that the persons around her act with ends in mind; that there must be some process by which society organizes itself toward the ends of its agents.

James Buchanan identifies much of this order in the form of rules that govern human action and interaction. Douglas North took this a step further, arguing that the rules and order that appear in society first arise in a mind or minds. The rules of the mind come to impact the rules of the game and their structure. This may happen simply through duplication of the rule governed action by others, or may rely on a complex, iterative processes of instantiation. Herbert Simon showed us that humans are ecologically rational, using the information present in their surroundings to guide their actions. This is a process of necessity for humans to function in a world that is best described as an open system and where one's own computational power is limited both by physical and structural constraints of the mind. Hayek (1952; 1960; 2014) delved further into the nature of rules, perception, interpretation, and action, providing a general language that made many readers confuse his economics for something approaching philosophy.

The wisdom of those great economists who have come before us serves as foundation for a new framing of economics. This starts with a reconstruction of the agent along the lines suggested by the economists above mentioned, as well as researchers in the adjacent disciplines which include philosophy, sociology, cognitive science, computational and computer science. Carl Menger dreamed of building a robust economic system that integrated essential innovation in formal knowledge of mind, human action that flows from the mind, and the society that arises as a result of human action. In some ways, Mises (1949) fulfilled this dream in human action. However, unaddressed problems concerning epistemology made his contribution fodder for those who do not appreciate the implications of his argument. Hayek corrected much of this with his work from the Abuse of Reason project. It is no coincidence that he is highly cited by those who study the emergence of institutions.

Many recognize that Hayek's work in this domain is valuable, but I have yet to see a substantial integration of his ideas with the core of economic theory. Many researchers in institutional economics have picked up the torch that he passed. Elinor Ostrom credits Hayek (1937, 1945) for her understanding of the emergence of social institutions that help to govern the commons. A similar appreciation is true for North and Denzau, as well as more recent work from Boettke, Coyne, and Leeson. The latter of these note that Hayek "was among the first to emphasize these aspects of spontaneously emergent institutions (333)." Koppl (2002) advises us to consider the internal state of the economic agent with reference to his perception of the external world. He argues that agents form expectations according to types that they have constructed or borrowed from others. (When I approach a cashier, I do not necessarily expect to make a friend, but I can be nearly certain that he will facilitate my purchase.) Agents have mental maps - private ontologies - which are comprised of a mixture of anonymous and personal typifications and rules that are believed to govern the existence and interactions of objects represented by these typifications.

The work of all of these mentioned have helped move economics toward the cusp of a change that has been building throughout the last half-century. Their work is part of a line realist theories that trace their lineage back to early economists like David Hume and Adam Smith who were not afraid to recognize, and even awed at, the complexities inherent in social interaction. These contributions represent a more scientific rendition classical political economy. This is synonomous with Wagner's "entangled political economy."  Their work is integral to the resuscitation and development of a paradigm of realism to underlie the work of economists.

Not long after the Samuelsonian turn in economics, theorists on the margins of the discipline - and some like Arrow who was certainly not on the margin - began to employ a new set of tools. These tools are part of the complexity paradigm. Researchers within this practice investigate emergent processes and the objects and structures that the processes generate. Of primary concern for economic theorists are agent based models. Some popular models include Shelling's (1971) segregation model, Axelrod's (1997) conflict model, and Gode and Sunder's (1993) zero-intelligence trader model. Theoretical contributions are arising from this line of inquiry. Gode and Sunder show that the process of coordination is an outcome of simple rules governing the buying and selling of goods. Axtell (2005) demonstrates that exchange is a process of social compuation! These theorists have created a number of innovation that other economic theorists can integrate into their own practice and understanding.

Now to bring these treasures together into a cohesive framework. A robust economic theory must explicitly identify its objects of concern, relationships between objects, and processes that govern these. Emphasis on rule-based perception and coordination is critical for understanding the core of social processes. Perception is at the foundation of agent action. Agent-based modeling provides a method to which a methodologically robust rendition of economic theory can be applied. Integration of these will provide us models that not only explain and demonstrate economic theory. The same platform also allows us to produce models whose predictions take economic process into account.

No comments:

Post a Comment