Skip to main content
Top

2017 | Book

Artificial Intelligence and Economic Theory: Skynet in the Market

insite
SEARCH

About this book

This book theoretically and practically updates major economic ideas such as demand and supply, rational choice and expectations, bounded rationality, behavioral economics, information asymmetry, pricing, efficient market hypothesis, game theory, mechanism design, portfolio theory, causality and financial engineering in the age of significant advances in man-machine systems. The advent of artificial intelligence has changed many disciplines such as engineering, social science and economics. Artificial intelligence is a computational technique which is inspired by natural intelligence concepts such as the swarming of birds, the working of the brain and the pathfinding of the ants.

Artificial Intelligence and Economic Theory: Skynet in the Market analyses the impact of artificial intelligence on economic theories, a subject that has not been studied. It also introduces new economic theories and these are rational counterfactuals and rational opportunity costs. These ideas are applied to diverse areas such as modelling of the stock market, credit scoring, HIV and interstate conflict. Artificial intelligence ideas used in this book include neural networks, particle swarm optimization, simulated annealing, fuzzy logic and genetic algorithms. It, furthermore, explores ideas in causality including Granger as well as the Pearl causality models.

Table of Contents

Frontmatter
Chapter 1. Introduction to Man and Machines
Abstract
This chapter introduces this book, Artificial Intelligence and Economic Theory: Skynet in the market, and in the process studies some of the big ideas that have concerned economics and finance in the last 300 years. These ideas include Marxist thinking, the theory of invisible hand, the theory of equilibrium and the theory of comparative advantage. It, furthermore, describes methods in artificial intelligence such as learning, optimization and swarm intelligence. It sets a scene on how these theories can be better understood by using artificial intelligence techniques, thereby, setting a scene for the rest of the book.
Tshilidzi Marwala, Evan Hurwitz
Chapter 2. Supply and Demand
Abstract
The law of demand and supply is the fundamental law of economic trade. It consists of the demand characteristics of the customer which describes the relationship between price and quantity of goods. For example, if the price of a good is low the customer will buy more goods and services than if the price is high. The relationship between price and the willingness of the customers to buy goods and services is called the demand curve. The other aspect of the demand and supply law is the supply curve which relates the relationship between the price and the quantity of goods suppliers are willing to produce. For example, the higher the price the more the goods and services the suppliers are willing to produce. Conversely, the lower the price the lesser the goods and services the suppliers are willing to produce. The point at which the suppliers are willing to supply a specified quantity of goods and services which are the same as those that the customers are willing to buy is called equilibrium. This chapter studies how the law of demand and supply is changed by the advent of artificial intelligence (AI). It is observed that the advent of AI allows the opportunity for individualized demand and supply curves to be produced. Furthermore, the use of an AI machine reduces the degree of arbitrage in the market and therefore brings a certain degree of fairness into the market which is good for the efficiency of the economy.
Tshilidzi Marwala, Evan Hurwitz
Chapter 3. Rational Choice and Rational Expectations
Abstract
The theory of rational choice assumes that when people make decisions they do so in order to maximize their utility. In order to achieve this goal they ought to use all the information available and consider all the options available to select an optimal choice. This chapter investigates what happens when decisions are made by artificial intelligent machines in the market rather than human beings. Firstly, the expectations of the future are more consistent if they are made by artificial intelligent machines than if they are made by human beings in that the bias and the variance of the error of the predictions are reduced. Furthermore, the decisions that are made are more rational and thus the marketplace becomes more rational.
Tshilidzi Marwala, Evan Hurwitz
Chapter 4. Bounded Rationality
Abstract
Rational decision making involves using information which is almost always imperfect and incomplete, together with some intelligent machine, which if it is a human being is inconsistent in making a decision that maximizes utility. Since the world is not perfect and decisions are made irrespective of the fact that the information to be used is incomplete and imperfect, these decisions are rationally limited (bounded). Recent advances in artificial intelligence and the continual improvement of computer processing power due to Moore’s law have implications for the theory of bounded rationality. These advances expand the bounds within which a rational decision making process is exercised and, thereby, increases the probability of making rational decisions.
Tshilidzi Marwala, Evan Hurwitz
Chapter 5. Behavioral Economics
Abstract
Behavioural economics is an approach to economics which takes into account human behavior. In his book “Thinking fast and slow”, which is based on the work he did with Tversky, Kahneman describes human thought as being divided into two systems i.e. System 1 which is fast, intuitive and emotional, and System 2, which is slow, rational and calculating. He further described these systems as being the basis for human reasoning, or the lack thereof, and the impact of these on the markets. Some of the findings are the inability of human beings to think statistically, called heuristics and biases, the concept of Anchoring, Availability effect, Substituting effect, Optimism and Loss aversion effect, Framing effect, Sunk costs and Prospect theory where a reference point is important in evaluating choices rather than economic utility. With the advent of decision making using intelligent machines, all these effects and biases are eliminated. System 1, which is intuitive, is eliminated altogether. System 2 becomes the norm, as advances in artificial intelligence are made. System 2 becomes fast because contemporary computational intelligent machines work fast. If one considers Moore’s Law, which states that computational power doubles every year, System 2 next year is faster than System 2 this year, thus making machines “Think Fast and Faster”.
Tshilidzi Marwala, Evan Hurwitz
Chapter 6. Information Asymmetry
Abstract
Often when human beings interact to make decisions, one human agent has more information than the other and this phenomenon is called information asymmetry. The fact that information asymmetry distorts the markets won Akerlof, Stiglitz and Spence a Nobel Prize. Generally, when one human agent is set to manipulate a decision to its advantage, the human agent can signal misleading information. On the other hand, one human agent can screen for information to diminish the influence of asymmetric information on decisions. With the dawn of artificial intelligence (AI), signaling and screening are easier to achieve. This chapter investigates the impact of AI on the theory of asymmetric information. The simulated results demonstrate that AI agents reduce the degree of information asymmetry and, therefore, the market where these agents are used become more efficient. It is also observed that the more AI agents that are deployed in the market, the less is the volume of trades in the market. This is because of the fact that for trades to occur, asymmetry of information should exist, thereby, creating a sense of arbitrage.
Tshilidzi Marwala, Evan Hurwitz
Chapter 7. Game Theory
Abstract
Game theory has been used quite extensively in economics. In game theory agents with rules interact to obtain pay-off at some equilibrium point often called Nash equilibrium. The advent of artificial intelligence makes intelligent multi-agent games possible. This enriches the ability to simulate complex games. In this chapter, intelligent multi-agent system is applied to study the game of Lerpa.
Tshilidzi Marwala, Evan Hurwitz
Chapter 8. Pricing
Abstract
Pricing theory is a well-established mechanism that illustrates the constant push-and-pull of buyers versus consumers and the final semi-stable price that is found for a given good. Embedded in the theory of pricing is the theory of value. This chapter studies various pricing models and, in particular, how they are changed by the advances in artificial intelligence (AI). The first pricing model studied is game theory based pricing where agents interact with each other until they reach a Nash equilibrium price. Multi-agent systems are found to enhance this pricing model. The second is rational pricing and here when pricing the amount of arbitrage is minimized and AI is found to improve this model. The third is capital asset pricing model, which is also improved by the advent of evolutionary programming. Then the fourth is the Black-Scholes pricing model, which is impacted by the use of fuzzy logic to model volatility. The last one is the law of demand and supply, and it is found that the advent of AI within the context of online shopping infrastructure results in individualized pricing model.
Tshilidzi Marwala, Evan Hurwitz
Chapter 9. Efficient Market Hypothesis
Abstract
The efficient market hypothesis (in its varying forms) has allowed for the creation of financial models based on share price movements ever since its inception. This chapter explores the impact of artificial intelligence (AI) on the efficient market hypothesis. Furthermore, it studies theories that influence market efficiency and how they are changed by the advances in AI and how they impact on market efficiency. It surmises that advances in AI and its applications in financial markets make markets more efficient.
Tshilidzi Marwala, Evan Hurwitz
Chapter 10. Mechanism Design
Abstract
In game theory, players have rules and pay-off and they interact until some point of equilibrium is achieved. This way, we are able to see how a game with sets of rules and a pay-off reaches equilibrium. Mechanism design is the inverse of that, we know what the end-state should look like and our task is to identify the rules and pay-off function which will ensure that the desired end-state is achieved. This is done by assuming that the agents in this setting act rationally. However, these agents are bounded rationally because the degree of rationality is limited. This chapter also discusses how artificial intelligence impacts mechanism design.
Tshilidzi Marwala, Evan Hurwitz
Chapter 11. Portfolio Theory
Abstract
The basis of portfolio theory is rooted in statistical models based on Brownian motion. These models are surprisingly naïve in their assumptions and resultant application within the trading community. The application of artificial intelligence (AI) to portfolio theory and management have broad and far-reaching consequences. AI techniques allow us to model price movements with much greater accuracy than the random-walk nature of the original Markowitz model. Additionally, the job of optimizing a portfolio can be performed with greater optimality and efficiency using evolutionary computation while still staying true to the original goals and conceptions of portfolio theory. A particular method of price movement modelling is shown that models price movements with only simplistic inputs and still produces useful predictive results. A portfolio rebalancing method is also described, illustrating the use of evolutionary computing for the portfolio rebalancing problem in order to achieve the results demanded by investors within the framework of portfolio theory.
Tshilidzi Marwala, Evan Hurwitz
Chapter 12. Counterfactuals
Abstract
The concept of rational counterfactuals is an idea of identifying a counterfactual from the factual (whether perceived or real), and knowledge of the laws that govern the relationships between the antecedent and the consequent, that maximizes the attainment of the desired consequent. In counterfactual thinking, factual statements like: ‘Greece was not financially prudent and consequently its finances are in tatters’, and with its counterfactual being: ‘Greece was financially prudent and consequently its finances are in good shape’. In order to build rational counterfactuals, artificial intelligence (AI) techniques are applied. The interstate conflict example considered uses AI to create counterfactuals that are able to maximize the attainment of peace.
Tshilidzi Marwala, Evan Hurwitz
Chapter 13. Financial Engineering
Abstract
Financial engineering has grown with the advent of computing and this growth has accelerated in the last decade with the advances in artificial intelligence (AI). This chapter explores how subjects such as evolution, deep learning and big data are changing the effectiveness of quantitative finance. This chapter explores the problem of estimating HIV risk, simulating the stock market using multi-agent systems, applying control systems for inflation targeting and factor analysis. The results demonstrate that AI improves the estimation of HIV risk, makes stock markets homogeneous and efficient, is a good basis for building models that target inflation and enhances the identification of factors that drive inflation.
Tshilidzi Marwala, Evan Hurwitz
Chapter 14. Causality
Abstract
Causality is a powerful concept which is at the heart of markets. Often, one wants to establish whether a particular attribute causes another. As human beings, we have perceived causality through correlation. Because of this fact, causality has often been confused for correlation. This chapter studies the evolution of causality including the influential work of David Hume and its relevance to economics and finance. It studies various concepts and models of causality such as transmission, Granger and Pearl models of causality. The transmission model of causality states that for causality to exist, there should be a flow of information from the cause to the effect. Simple example of the study on the link between circumcision and risk of HIV are used in this chapter.
Tshilidzi Marwala, Evan Hurwitz
Chapter 15. Future Work
Abstract
This chapter concludes this book and summarizes the general direction of artificial intelligence and economics. It summarizes all the key concepts addressed in this book such as rational expectations and choice, bounded rationality, behavioral economics, information asymmetry, game theory, pricing, efficient market hypothesis, mechanism design, portfolio theory, rational counterfactuals, financial engineering and causality. Additionally, it evaluates how all these ideas are influenced by the advent of artificial intelligence. This chapter also studies the concept of decision making which is based on the principles of causality and correlation. Then it proposes a way of combining neoclassical, Keynesian and behavioral economics together with artificial intelligence to form a new economic theory. Furthermore, it postulates on how the interplay between advances in automation technologies and human attributes that can be automated determine a limit of the extent of automation in an economy or a firm.
Tshilidzi Marwala, Evan Hurwitz
Backmatter
Metadata
Title
Artificial Intelligence and Economic Theory: Skynet in the Market
Authors
Prof. Tshilidzi Marwala
Dr. Evan Hurwitz
Copyright Year
2017
Electronic ISBN
978-3-319-66104-9
Print ISBN
978-3-319-66103-2
DOI
https://doi.org/10.1007/978-3-319-66104-9

Premium Partner