Artificial Intelligence and Economic Theory

Advanced Information and Knowledge Processing

Tshilidzi Marwala

Evan Hurwitz

Artificial

Intelligence

and Economic

Theory: Skynet

in the Market

Advanced Information and Knowledge

Processing

Series editors

Lakhmi C. Jain

Bournemouth University, Poole, UK, and

University of South Australia, Adelaide, Australia

Xindong Wu

University of Vermont

Information systems and intelligent knowledge processing are playing an increasing

role in business, science and technology. Recently, advanced information systems

have evolved to facilitate the co-evolution of human and information networks

within communities. These advanced information systems use various paradigms

including artificial intelligence, knowledge management, and neural science as well

as conventional information processing paradigms. The aim of this series is to

publish books on new designs and applications of advanced information and

knowledge processing paradigms in areas including but not limited to aviation,

business, security, education, engineering, health, management, and science. Books

in the series should have a strong focus on information processing—preferably

combined with, or extended by, new results from adjacent sciences. Proposals for

research monographs, reference books, coherently integrated multi-author edited

books, and handbooks will be considered for the series and each proposal will be

reviewed by the Series Editors, with additional reviews from the editorial board and

independent reviewers where appropriate. Titles published within the Advanced

Information and Knowledge Processing series are included in Thomson Reuters’

Book Citation Index.

More information about this series at http://www.springer.com/series/4738

Tshilidzi Marwala • Evan Hurwitz

Artificial Intelligence

and Economic Theory:

Skynet in the Market

123

Tshilidzi Marwala

Faculty of Engineering and the Built

Environment

University of Johannesburg

Auckland Park

South Africa

Evan Hurwitz

Department of Electrical and Electronic

Engineering Science,

Faculty of Engineering and the Built

Environment

University of Johannesburg

Auckland Park

South Africa

ISSN 1610-3947 ISSN 2197-8441 (electronic)

Advanced Information and Knowledge Processing

ISBN 978-3-319-66103-2 ISBN 978-3-319-66104-9 (eBook)

DOI 10.1007/978-3-319-66104-9

Library of Congress Control Number: 2017950270

© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part

of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,

recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission

or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar

methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service marks, etc. in this

publication does not imply, even in the absence of a specific statement, that such names are exempt from

the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and information in this

book are believed to be true and accurate at the date of publication. Neither the publisher nor the

authors or the editors give a warranty, express or implied, with respect to the material contained herein or

for any errors or omissions that may have been made. The publisher remains neutral with regard to

jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG

The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Artificial Intelligence and Economic Theory: Skynet in the market borrows the word

Skynet from the movie The Terminator. 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 such as the swarming of birds, the workings of the brain and the path

finding behavior of the ants. These techniques have impact on economic theory. This

book studies the impact of artificial intelligence on economic theory, a subject that

has not yet been studied. The theories that are considered include asymmetrical

information, pricing, rational expectations, rational choice, game theory, mechanism

design, behavioral economics, bounded rationality, efficient market hypothesis,

financial engineering, portfolio, rational counterfactual and causality. The benefit of

this book is that it evaluates existing theories of economics and updates them based

on the developments in the artificial intelligence field. This book makes an important

contribution to the area of econometrics, and is an interesting read for graduate

students, researchers and financial practitioners. In this book, Chaps. 1, 2, 3, 4, 5,

8, 9, 12, 13, 14, and 15 were written by Tshilidzi Marwala whereas Chaps. 7 and

10 were written by Evan Hurwitz. Chapters 6 and 11 were jointly written. We

thank the three anonymous reviewers for their contributions.

Johannesburg, South Africa Tshilidzi Marwala Ph.D.

July 2017 Evan Hurwitz D.Eng.

v

Contents

1 Introduction to Man and Machines .......................... 1

1.1 Introduction ........................................ 1

1.2 Economics and Economic Theory........................ 3

1.3 Artificial Intelligence ................................. 9

1.4 Conclusion ......................................... 12

References............................................... 13

2 Supply and Demand ...................................... 15

2.1 Introduction ........................................ 15

2.2 Scarcity............................................ 16

2.3 Utilitarianism ....................................... 17

2.4 Supply and Demand .................................. 18

2.5 Factors Influencing Demand and Supply .................. 21

2.6 Artificial Intelligence (AI) and Demand and Supply .......... 22

2.7 Conclusion ......................................... 24

References............................................... 24

3 Rational Choice and Rational Expectations.................... 27

3.1 Introduction ........................................ 27

3.2 Adaptive and Rational Expectations ...................... 28

3.3 What Is Rational Choice? .............................. 30

3.4 Information ......................................... 32

3.5 Choices............................................ 32

3.6 Optimization ........................................ 33

3.7 Rational Choice ..................................... 33

3.8 Rational Choice and Opportunity Cost .................... 35

3.9 Rational Choice and Artificial Intelligence ................. 35

3.10 Interstate Conflict and Rational Choice.................... 36

3.11 Conclusion ......................................... 38

References............................................... 38

vii

4 Bounded Rationality ...................................... 41

4.1 Introduction ........................................ 41

4.2 Rational Decision Making: A Causal Approach ............. 42

4.3 Rational Decision Making Process ....................... 43

4.4 Bounded-Rational Decision Making ...................... 43

4.5 Credit Scoring....................................... 47

4.6 Conclusions ........................................ 49

References............................................... 49

5 Behavioral Economics..................................... 51

5.1 Introduction ........................................ 51

5.2 Behavioural Economics ............................... 52

5.3 Behavioural Economics and Demand and Supply............ 54

5.4 Behavioural Economics and Rational Expectations........... 56

5.5 Behavioural Economics and Bounded Rationality............ 57

5.6 Artificial Intelligence and Behavioural Economics ........... 57

5.7 Moore’s Law ....................................... 59

5.8 Conclusions ........................................ 60

References............................................... 60

6 Information Asymmetry ................................... 63

6.1 Introduction ........................................ 63

6.2 Asymmetric Information ............................... 66

6.3 Artificial Intelligence Makes Us Wiser .................... 68

6.4 Asymmetric Information and Market Efficiency ............. 69

6.5 Information Asymmetry and Trading Volumes.............. 71

6.6 Conclusion ......................................... 73

References............................................... 73

7 Game Theory............................................ 75

7.1 Introduction ........................................ 75

7.2 Game-Theory Artefacts................................ 76

7.3 Multi-agent Modelling ................................ 81

7.3.1 Complexity Modelling......................... 83

7.3.2 Economics.................................. 84

7.3.3 Social Sciences .............................. 84

7.4 Intelligent Agents .................................... 84

7.5 The Road Ahead..................................... 86

7.6 Conclusions ........................................ 87

References............................................... 87

8 Pricing ................................................. 89

8.1 Introduction ........................................ 89

8.2 Pricing ............................................ 90

8.3 Value Theory ....................................... 90

viii Contents

8.4 Game Theory ....................................... 92

8.5 Rational Pricing ..................................... 93

8.6 Capital Asset Pricing Model ............................ 93

8.7 Black-Scholes Equation ............................... 94

8.8 Demand and Supply .................................. 95

8.9 Conclusions ........................................ 98

References............................................... 98

9 Efficient Market Hypothesis ................................ 101

9.1 Introduction ........................................ 101

9.2 Efficient Market Hypothesis ............................ 102

9.3 Rational Expectations ................................. 103

9.4 Rational Choice ..................................... 104

9.5 Bounded Rationality .................................. 104

9.6 Information Asymmetry ............................... 105

9.7 Behavioral Economics ................................ 106

9.8 Game Theory ....................................... 106

9.9 Demand and Supply .................................. 107

9.10 Pricing ............................................ 108

9.11 Artificial Intelligence ................................. 109

9.12 Conclusions ........................................ 109

References............................................... 110

10 Mechanism Design........................................ 111

10.1 Introduction ........................................ 111

10.2 The Players......................................... 112

10.3 Efficiency and Equilibria............................... 114

10.4 Incentive Compatibility and the Revelation Principle ......... 115

10.5 Goalposts .......................................... 116

10.6 So What? .......................................... 118

10.7 Through the Looking Glass ............................ 119

10.8 Application to Market Design........................... 121

10.9 Conclusions ........................................ 123

References............................................... 123

11 Portfolio Theory ......................................... 125

11.1 Introduction ........................................ 125

11.2 Assumptions and Limitations ........................... 128

11.3 The Trading Machine ................................. 130

11.4 The Emerging Trading Machine ......................... 131

11.5 Intelligence ......................................... 132

11.6 Evolutionary Programming ............................. 133

11.7 Results Analysis ..................................... 134

11.8 Conclusion ......................................... 135

References............................................... 135

Contents ix

12 Counterfactuals .......................................... 137

12.1 Introduction ........................................ 137

12.2 Counterfactuals...................................... 138

12.3 Rational Counterfactuals............................... 140

12.4 Counterfactuals and Causality........................... 141

12.5 Counterfactuals and Opportunity Cost .................... 142

12.6 Counterfactuals and Artificial Intelligence ................. 143

12.7 Interstate Conflict .................................... 143

12.8 Conclusions ........................................ 145

References............................................... 145

13 Financial Engineering ..................................... 147

13.1 Introduction ........................................ 147

13.2 Risk .............................................. 148

13.3 Stock Markets....................................... 149

13.4 Control Systems ..................................... 153

13.5 Factor Analysis...................................... 155

13.6 Conclusions ........................................ 157

References............................................... 157

14 Causality ............................................... 159

14.1 Introduction ........................................ 159

14.2 Correlation ......................................... 160

14.3 Causality........................................... 161

14.4 Granger Causality .................................... 165

14.5 Pearl Causality ...................................... 165

14.6 Use of Do-Calculus .................................. 168

14.7 Conclusions ........................................ 169

References............................................... 169

15 Future Work ............................................ 171

15.1 Conclusions ........................................ 171

15.2 Decision Theory ..................................... 175

15.3 Developmental Economics ............................. 176

15.4 New Economic Theory ............................... 177

References............................................... 178

Appendix A: Multi-layer Perceptron Neural Network ............... 181

Appendix B: Particle Swarm Optimization ........................ 185

Appendix C: Simulated Annealing............................... 189

Appendix D: Genetic Algorithms................................ 193

Appendix E: Fuzzy Logic ...................................... 197

Appendix F: Granger Causality................................. 201

Index ...................................................... 203

x Contents

About the Authors

Tshilidzi Marwala is the Vice-Chancellor and Principal of the University of

Johannesburg. He was previously the Deputy Vice-Chancellor: Research and

Internationalization and a Dean of the Faculty of Engineering at the University of

Johannesburg. He was also previously a Full Professor of Electrical Engineering,

the Carl and Emily Fuchs Chair of Systems and Control Engineering as well as the

SARChI chair of Systems Engineering at the University of the Witwatersrand. Prior

to this, he was an Executive Assistant to the Technical Director at the South African

Breweries. He holds a Bachelor of Science in Mechanical Engineering (magna cum

laude) from Case Western Reserve University (USA), a Master of Mechanical

Engineering from the University of Pretoria, a Ph.D. in Engineering from

Cambridge University and was a postdoctoral research associate at the Imperial

College (London). He is a registered professional engineer, a Fellow of TWAS, the

World Academy of Sciences, the Academy of Science of South Africa (ASSAf), the

African Academy of Sciences and the South African Academy of Engineering. He

is a Senior Member of the IEEE (Institute of Electrical and Electronics Engineering)

and a distinguished member of the ACM (Association for Computing Machinery).

His research interests are multidisciplinary and they include the theory and appli￾cation of computational intelligence to engineering, computer science, finance,

social science and medicine. He has supervised 47 Masters and 23 Ph.D. students to

completion. He has published 11 books (one translated into Mandarin), over 280

papers and holds 3 international patents. He is an Associate Editor of the

International Journal of Systems Science (Taylor and Francis Publishers).

Evan Hurwitz is a South African computer scientist. He obtained his B.Sc.

Engineering (Electrical) (2004), his M.Sc. Engineering (2006) from the University

of the Witwatersrand and Ph.D. from the University of Johannesburg in 2014

supervized by Tshilidzi Marwala. He is known for his work on teaching a computer

xi

how to bluff, which was widely covered by the magazine New Scientist. Hurwitz

together with Tshilidzi Marwala proposed that there is less level of information

asymmetry between two artificial intelligent agents than between two human agents

and that the more artificial intelligent there is in the market the less is the volume of

trades in the market.

xii About the Authors

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 tech￾niques, thereby, setting a scene for the rest of the book.

1.1 Introduction

“Workers of the world unite, you have nothing to lose but chains” so said Karl

Marx. Despite what many Marxists claim, he never foretold the advent of artificial

intelligence, otherwise he would probably have said “Artificial intelligent robots of

the world unite, you have nothing to lose but chains” (Marx 1849). But what Marx

realized was that the principal agent of work is man. Man is the invisible hand that

drives the economy as observed by Adam Smith (Smith 2015). The economy was

made by man, about man and for man but the theories that explained the economy

did not quite match the behaviour of a man. For this reason, the rational man is

indeed irrational and his irrationality permeates every aspect of life including the

very concept we call the economy.

Human beings have been around for two hundred thousand years and,

throughout their existence and even from their forbearers, have inherited certain

traits and behaviors that influence them even today (Harari 2014). Some of these

traits and behaviours include greed, fear, bias and social structure. All these traits

are still with us today because of one and only one reason and that is because they

all have given us an evolutionary advantage. Of course, this might change in the

future depending on the change of environment and, therefore, these traits might

depart from human beings. All these traits influence our decision making and the

rationality thereof. Herbert Simon calls the idea of making decisions with all these

© Springer International Publishing AG 2017

T. Marwala and E. Hurwitz, Artificial Intelligence and Economic Theory: Skynet

in the Market, Advanced Information and Knowledge Processing,

DOI 10.1007/978-3-319-66104-9_1

1

constraints e.g. processing power of our brains; incomplete, imperfect and impre￾cise information; and human behaviour, bounded rationality (Simon 1991). Our

rationality is bound by all these constraints but what will happen when machines

inevitably replace humans in decision-making capacities? Is the rationality of

machines bound? Are the bounds of rationality bigger in humans than in machines?

These are some of the questions that this book seeks to answer.

Machines are now part of everyday decision making. They are becoming more

intelligent due to a technology called artificial intelligence (AI) (Marwala 2013).

Alan Turing surmised that machines are intelligent if and only if when we interact

with them we cannot tell whether we are interacting with a man or a machine

(Traiger 2000). No machine has passed this Turing test over an extended level of

man-machine interaction. But this does not limit artificial intelligence and make

machines incapable of solving complex problems. Some of the critics of the Turing

test is the work done by John Searle on the problem called the Chinese Room

problem (Searle 1980). In the Chinese Room problem, you have a person inside a

room and there is a hole where something written in Chinese is slipped into a room

and there is a lookup table which the person inside this room uses to translate into

English. To the person outside it seems as if the person inside can speak Chinese.

John Searle goes further and classifies between Strong AI versus Weak AI. In this

classification, Strong AI is the one that is really intelligent, i.e. there is actually a

Chinese person inside the room, as opposed to a Weak AI, where there is a Zulu

person who does not speak Chinese inside the room who is just using a lookup table.

This chapter describes man-machine interaction and its impact on some of the

big ideas in economics. Every economic epoch had its own theories or thinking.

Some of these theories stood the test of time but some have fallen by the wayside.

The biggest event in history of economics is the history of industrialisation and all

its various stages. The first industrial revolution occurred in 1874 in England. What

is less clear is why it did not happen in Asia, especially in India or China as these

two regions had significantly higher population densities? What was the catalyst

that caused the first industrial revolution? In the 17th century lived a man in

England called Isaac Newton who was educated in that superb institution of Trinity

College Cambridge (Newton 1729). Legend claims that unlike many people who

had lived before him and had witnessed an apple falling, he asked: “Why did the

apple fall?” And from this he discovered gravity, an intriguing concept which was

only given an explanation by another fine German/Swiss/American scientist by the

name of Albert Einstein some hundreds of years later. Newton, furthermore, came

with what is now called the laws of motion which stated that objects will continue

at rest or keep on moving until they are moved or stopped respectively.

Furthermore, he observed the relationship between force and mass of an object with

its acceleration. This thinking of Newton reshaped our understanding of movement

and became the catalyst or DNA for the first industrial revolution. This gave us

steam engines, trains and mechanical machines for production. From this era,

economic principles such as Marxism, Adam Smith’s invisible hand as well as

David Ricardo’s theory of value and principle of comparative advantage were

conceived (de Vivo 1987).

2 1 Introduction to Man and Machines

Then in the 19th century came a British man called Michael Faraday who

performed crucial experiments which were later interpreted by James Clerk

Maxwell through his beautiful mathematical equations (Maxwell 1873; Agassi

1971). Michael Faraday observed that if you have a magnet and you put it next to a

wire that conducts electricity and you move the wire, then electricity flows in that

conductor. This is wizardry beyond miracles of biblical proportions. Even today we

generate electricity, perhaps with the exception of solar energy and few others,

using this technique. Faraday’s discovery allowed us to convert mechanical energy

into electrical energy, such that could be usefully distributed through electrical

wires, dramatically altering the nature and course of human existence. Conversely,

Faraday observed that again with a magnet and a conducting wire, you force

electricity through the wire, then the wire moves and this was the birth of the

electric motor that still moves our assembly lines. This complementary discovery

allowed the transmitted energy to be utilised at some distant location, which is the

heart of machinery as we know it today. These events were the DNA for the second

industrial revolution. From this era, economic principles such as mass production

and Keynesian economics were introduced.

Then in the second half of the 20th century, John Bardeen, Walter Brattain, and

William Shockley discovered the transistor (Amos and James 1999). It is based on

semiconductors which are objects that conduct electricity under certain conditions.

The transistor is the catalyst for the electronic age that gave us computers, cell

phones, information technology and automation in our factories. It is the DNA of

the third industrial revolution. It was in this era that powerful economic concepts

such as market efficiency introduced prospect theory.

The era we are living in is the fourth industrial revolution. This is an era of

intelligent machines. The DNA of the fourth industrial revolution is AI. It will

touch all aspects of our lives. We will have robotic cops to protect us, robotic

doctors to assist us with medical issues, all our vital organs will be monitored real

time to extend human lives, driverless cars and aircraft as well as human empty

factories as labor will be replaced. What will happen to economics? Already we

know that the stock market no longer has crowds of people shouting a price of

stocks because artificial intelligent software are doing the work (Siegel 2008). This

book explores how economic theories that have guided decision makers in the past

will have to change in the light of artificial intelligent capabilities.

1.2 Economics and Economic Theory

Economics began when man started bartering for exchanging goods. This was

mainly based on the reality that man could not produce all he wanted. For example,

suppose we have a man called Peter who produces maize and another called John

who produces peanuts. Then Peter will give half of his maize for half of John’s

peanuts. If we include a third person Aiden who produces wheat, then Peter takes a

third of his maize and gives it to John in exchange of a third of his peanuts, gives

1.1 Introduction 3