An Old way to do economics

Let us consider the following entities:

  1. A big firm
  2. A metropolitan city (Mumbai, Delhi, etc)
  3. Biological system
  4. Ecological system
  5. Social networks
  6. Climate
  7. Ant colonies
  8. Economic/Financial system (i.e. Markets)

What is common across these entities? What binds them? or the pertinent question to ask is how to study them? The answer lies in an emerging field called ‘Complexity theory‘. Complexity theory tries to study the systems like those mentioned above under a umbrella term Complex Adaptive System.

Complex Adaptive System (CAS)

What defines a CAS? Put simply, the CAS is system composed of multiple agents (capable of adapting and learning) interacting with each other at different scale. Sounds too simple? Let me list out some of the characteristics of the CAS to make it more intuitive:

  1. In CAS, the perfect knowledge of individual constituent part doesn’t give us the perfect knowledge about the system – being the aggregation of those individual components.
  2. CAS exhibits emergent behavior i.e. the system exhibits properties not present in its constituents parts. Thus, it is difficult to predict systems’ behavior from the behavior of its components. In other words, the system is not sum of its components.
  3. CAS consists of numerous heterogeneous agents with dynamic and rich interactions between them leading to emergence (see point 2). Thus, the conventional way of understanding using mathematical tools like control theory, system of differential equations, phase diagram analysis (used extensively in macroeconomics) are inadequate and if still used, will give impractical results.
  4. The interaction in CAS are highly non-linear i.e. small perturbations in initial condition/state give rise to drastic changes in the final outcome/state. This is again the reason for emphasis on point 3.
  5. The interactions are, except in few cases, local. The agent, most often, are ignorant of systems properties (not useful information for the agent to act) as a whole but possess useful local information to act and respond accordingly. This makes CAS adaptive.
  6. Any interaction can feed into other interactions or decisions of other agents (of same or different type) at different scales. Thus creating a constant feedback loop making claims of causation difficult!!!

Understanding Economics as CAS

The economic system is such a CAS. It consists of numerous agents (households/consumers, firms, banks, government, etc) interacting at different scale in an economy. The consumer uses local information (about prices, wages, rents – they don’t need to know the price index to buy their groceries, clothes, etc; they don’t need to know the unemployment rate in order to make decision about entering/leaving the labor market; they don’t need to know the GDP growth/balance of payments/deficits, money supply, etc to make any economic decision). Thus the consumers’ decision is entirely based on his/her local condition and they weigh all the alternatives in terms of its opportunity cost (cost = both monetary and non-monetary). Thus under a similar condition, out of two people (similar in terms of skills, education and facing the similar environment), one person might prefer to employ his skills in setting up a business while other might prefer a job giving a fat paycheck every month. In other words, in an economic system, consumers are sovereign and their decisions steer the markets.

Since a consumer takes decision taking into account the local environment and given that local environment consists of other agents (other consumers, firms, govt, etc), this creates the feedback loop leading to emergence of price signals in the market.

Let us go back to the two people we talked about in earlier paragraphs. Suppose the person setting up a business is Mr. A and the one doing a job is Mr. B paying him $1500 a month. Mr. A takes $1,000 credit from Bank AB and sets up a business. After a year, he has employed 50 people and his earnings are $500 a month whereas after a year, Mr. B earns $2000 a month. Now what possible things could happen at the end of the next year? From the Mr. A’s perspective, either he could expand and employ more people and his earnings could increase or his business could fail and he would have to go back to work like Mr. B or other possibilities in the spectrum. If he fails, there are impacts on the banks that has extended him the credit. From the Mr. B’s perspective, either he could continue to work as before or taking inspiration from B, start his own business. Thus, numerous combinations of initial conditions leads to numerous final outcome each having its own macroeconomic aggregates which are different – quantitatively and qualitatively. This hypothetical scenarios shows the emergence, adaptive and non-linear nature of economic system. Moreover, what Mr. A or Mr. B will do at any given point tells us nothing about the economy until their respective interactions (with each other or with other agents) materializes in effect – which is reflected in aggregate.

But how does complexity theory tries to study and model the CAS? Since we have already made the case for the inadequacy of mathematical techniques, we are left with the computational methods to study CAS. These computational methods are bottom-up approach in which a simulation environment is set up with numerous agents in form of computer code. These are initialized and allowed to run without restriction. The results, after the end of the simulation run, are analyzed. For instance, I have created two computational models based on existing literature on computational economics (Simple macro model and Baseline model – in progress). Another way to analyze such complex system is to use deductive logic to build a coherent theory from bottom. We will discuss this in another article.

Mechanical view of an economy

The status-quo, in economics, is of using mathematics to solve economic problems and derive insights from the result coming out of such mathematical analysis.

“If microeconomics is a study of two agents interacting, then macroeconomics is a study of one agent interacting”

– Unknown

The above quote sums up the current state of affairs in macroeconomics based on mathematical analysis. Let me explain the way macroeconomics is done currently. So, macroeconomists model the economy in different way – from simple 2-period inter-generational model to slightly complex endogenous growth models to more complex Dynamic Stochastic General Equilibrium (DSGE) model. But all these model have some common minimum assumptions. These assumptions are important in order for you to get any insights from these models. Broadly, there are following assumptions:

  1. These are micro-founded models i.e. all the outcomes of the model are results of consumers making optimal choices or maximizing their utility on the basis of their preference. The preference is denoted by standard utility function (is well behaved mathematically – monotonic and quasi-concave) derived from microeconomic theory. Thus, the agent is rational.
  2. There is only one consumer/household who represent every other household in an economy i.e. a representative agent. Similar assumption is for the firms.
  3. This agent has a perfect foresight (a GOD!!!) and has rational expectations (a result of famous Lucas critique which lead to emergence of New Keynesian thought). Rational expectations simply means that agents have perfect knowledge of the system they are part of and hence make expectations accordingly.
  4. Analysis is solely focused on what happens in equilibrium/steady state/transition path.

These assumptions eliminate any study of interactions, dynamics between agents and their impact on the system. Moreover, it treats an economy as a mechanical system. The implication of treating an economy like a machine are dangerous for it enables an economist (in collaboration with the government/bureaucrats) to interfere in the proper functioning of the markets and to distort the incentive structure in the economy leading to more volatile business cycles. As an expert mechanic, economists think that they can steer their beloved machine as they want and eventually end up crashing it. The emergence of fiat money, central bank, unfounded financial regulations, bailouts, etc are all result of these interventions (more on this in subsequent articles). Moreover, any model should be based on the assumption which simplifies the analysis. On the contrary, the assumption used in standard macro models ignores fundamental and defining aspects of an economic system that their removal makes the model dead. This top-down approach fails to address the issues of economic importance in contrast to bottom-up approach offered by complexity theory.

But all is not wrong in the history of economic thought. There is a school of thought which saw the economy as an emergent system while losing the ground to the mainstream economists led by Keynes. This school of thought emerged by works of Adam Smith during Scottish Enlightenment followed by the works of Austrian economists like Mises and Hayek. I think that the Austrian perspective should be blended with CAS theory to create an economic theory based on bottom-up approach infusing much needed economic realism into the discipline. In other words, we need to do economics in the old way as envisaged by Adam Smith and others.

Published by ritesh_666

A curious extremist

Leave a Reply

Create your own website, free on WordPress.com.Create your own website.
Start free