Microeconomics of Information Systems

Abstract

Do computer science need economists and do need economists to write textbooks about ? Phd advanced
microeconomics with full demand, supply and equilibrium models will be useful for economists and not so much for
computer programmers. Computer programmers can sustain with wikipedia, manuals, documentations and
programming codes, which is not enough abstract for economists. This article try to answer this question and show
some ideas influenced by both sciences.

Introduction

There were three major engineering booms and revolutions in recent history,which dramatically affected behavior
of consumers and firms. Mechanical engineering starting in 19th century, electrical engineering in 20th century and IT at
the turn of 20th and 21th century. First two had major impact on life of the society ( e.g. urbanization, production, price
level ), but still fit in a standard microeconomics framework1 . On the other hand, information technologies could have
ambition to amend it.

First microeconomics textbooks used sectors and industries to illustrate general economics concepts. Some of
sectors has grown as a own subdiscipline of microeconomics. Good examples are financial markets, health, education,
energy, agriculture. Large and comprehensive monographs and textbooks were written for all of them in the contrast of
information technologies.

What is the reason behind ? Varian 2 published most proper paper, which correctly lays out direction and
highlights topics like ntellectual property, standards and standardization, price discrimination, games and game theory.
However, scale of the paper ( 40 pages ) looks more like notes with ideas than proper textbook. Quick growth and death
of technologies made it difficult to write monographs which would be up to date. Veryard 3 is good example of
obsolescence of a textbook. He focuses only on supply/cost side of economy and only on software as IT commodity.
In next chapter, let's try to compare publishing in economics of IT area to the other sectors of economy.

1.Microeconomics or Economy/Economics ?

Morris & Co 4 analyze economy in health care. They use basic microeconomics tools
(demand,supply,market,equity,pareto welfare,public sector,game theory) as background environment and adjust it for
health sector, they also don't forget to add health sector examples and case studies. This textbook is ideal for health
management students, who have already passed basic microeconomics subjects. Book provides basic insight into
national health markets, but no insight is offered to the structure of health sector expenditures.

Nersesian 5 describes energy market from different, business and market research view. He mentions history and
development of energy market, he works with logical market structure and sub-markets (coal, gas, nuclear energy
market etc. ). He thinks about past, current and future economic policies.
Debertin's 6 Agricultural Production Economics ( based on advanced microeconomics ) is closer to the Morris'
approach. He uses various model combining one/two inputs with one/two outputs, Cobb-Douglas production function.
He then comes with optimization process – profit maximization subject to the budget constrain. He studies the
elasticity of substitution, the demand for inputs. Finally he introduces risk/uncertainty conditions and time element with
discounting of future values. It's advanced microeconomics view, compared to the health care analysis , he stays only on
supply/production side of economy.

Two main approaches to the sector economic analysis have emerged. First ( “microeconomics” for simplicity)
works with microeconomics theory and it's mathematical tools. Sector terminology is mostly used as example in
microeconomics models ( supply/demand commodities names). Second approach ( “economy” or “business” or
“economics” ) tries to use more economic and financial thinking to create sector/industry market research, describes
history and development of market and also mentions it's real important players. It could be also use as guide for
economic or government policy.

First approach is more medium and long term by nature, second then more short term. Varian 1 illustrates short validity
of business term approach ( multiplied by fast changing IT sector ) stating “When Microsoft introduced Internet
Explorer it announced that it was free and would be always be free. This was a signal to consumers that they would not
be subject to lock-in if they adopted the Microsoft browser. Netscape countered by saying that it's products would
always be open. Each competitor played to its strength, but it seems that Microsoft had the stronger hand “. Later
changes in the browser market showed short-term truth of such statements. Both approaches are shown on picture
below.

Image 1. Scientific,mathematical and microeconomics vs business,economy,economics sector view

2. Open Source and Standardization – keys to success ?

Official definitions are following – Open source software is software with source code that anyone can inspect,
modify and enhance 11. Varian 2 is more than excited stating “...open source allows programmers and “wannabe
programmers” to look inside the black box to see how the applications are assembled. This is tremendous spur to
education and innovation” and “...allowing future Edisons to build on, improve, and use existing programs- combining
them to create something that maz be quite new”. We will look on it further later.

Standardization seems to be more important and more studied by economists. Many mention it as key building
stone for the digital revolution. On other side Varian 2 recognizes that standards were important also in the history,
mentioning Bell telephony and VHS format. It's nothing new. Besen 12 incorporates game-theory and competition and
setups three forms of competition in standard setting.
1/ Standards war. Firms compete to determine the standard
2/ Standards negotiation. Both firms want a standard, but disagree about what the standard should be
3/ Standards leader. One firm leads with a proprietary standard. The other firms wants to interoperate with the existing
standard.

Libicki 9 focus on standards organizations, categorizing them ( see Table below ).

Democratic Strong Leader
By membership ISO W3C
Open to all IETF Linux open source

It's all very useful and we will use it in next paragraph ( quantification ). First, we will look on it using microeconomics
theory. Standards could be seen as regulation. Self-regulation in the case of IT sector history. There are low level–
standards in the case of IT infrastructure (tcp/ip,ftp,smtp,http) , standards for technologies, which we can consider only
as inputs to the production process (html,java,xml,c) and standards in supply to the end-users (Microsoft Windows,
ERP systems ). Corporates could be 100% positively motivated at first two levels. Build working infrastructure instead
of government is easier and make profit achievement quicker. Setup standards in programming languages leads to the
decrease of prices of inputs to the production function. More programmers can learn languages and quicker, so
corporaters can hire more programmers at lower wages. It seems logical, that hardest way to the standards will be in the
case of final products, where corporates compete each other. There can standard wars begin.

2.1. Quantification

Alex Cukierman 8 studied relations between central bank independence and inflation. He setuped many theoretical
models, but to be able to find some empirical evidence, he tried to quantify and measure independence of central banks
as written down to the law. He created methodology where national laws about central banking were transformed to the
scale from 0% to 100%. I am inspired by it and I will try to do same with “open source/price” and “standard” .

Open Source+price is composed of three components

Price ( weight 30% ) : 100% it's free and there is no legal protection to sell it, 95% it's free,70% market price, 55%
monopolistic competition price,30% oligopoly price, 0-5% monopolistic price
Change management type ( 20% ) . 100%.→ 0% Yes, No, Open to all/by membership, Democratic, Strong Leader, refer
Libicki table above
Source Code (50% ). 100% source code fully available ,75% majority available ,50% source partly and reasonable
protected ,25%-0% source code not available / black box

Standardization
Documentation ( weight 20% ) , 100% → 0% according to the documentation quality and scale
Standardization via some organization or body ( weight 30% ) , 100% → 0% Yes, No, reputation of organization or
body
Market Valuation of Standard (50%), acceptance by competitors, acceptance by third parties and industry, multi-
platform availability, deviation from standards as ( A- innovation attempt, B- differentiation of product . C- prevent/stop
entry to the market )

Based on criteria above markets and it's main subjects were classified as on image below

Image 2. Quantification of “open source” and “standardization”

2.2. Quantification as the source for regression analysis

Next step after quantification, following again Cukierman's approach, should be carry out regression analysis of
open source and standardization ( independent variables ) and market share ( dependent variable ). We could also carry
out regression for various environments : overall, science, education/university sector etc. Full market share data are not
available at the moment.

3. Games, Game Theory and Case Studies

Although IT and digital economy attracted much attention from economists, there is no reason to change
microeconomics textbooks. All types of innovation behavior, price strategies, seeking for monopolistic rent, regulations
and standards have their mirrors in the economic history. Speed of the changes is significantly higher, what put more
pressure on particular market players and it might seem that we are witnesses of something new in economic theory.
We are not, but we still can give some examples or user cases of interesting microeconomic and market situations.

3.1. First ever single whole-world market with no trade barriers ( legal and also physical ) ?

It's possible to buy any software ( or derived service ) online and pay immediately using various payment
instruments. Good ( or service ) is delivered also immediately and only constrain is speed of internet connection. There
are yet no customs duties on software and software translation costs are minimal and can be taken as fixed costs, so
expansion to the new territory or region is almost costless. Is it stable game-theoretical output ? Governments were able
to put quotas on local movies or music ( similar world-wide markets ) , so it seems to be win-win solution, but still
under risk.

3.2. Time paradox of efficiency

First cars were not efficient and pressure on fuel consumption came hand in hand with it's mass expansion. First
TV's and PC monitors were big and wide, gradually becoming flat and small. Generally prototypes brings innovation,
but time is needed to make product perfect. First programming codes had challenged difficulties like low memory and
CPU settings, so first programming codes have been written “heavenly”. They had been written in low-level
programming languages and required high labor skills. Later whole industry needed more and cheaper labor ( computer
programmers ) and current high level languages and robust databases forgive many mistakes. Quality and efficiency
decrease.

3.3. Ford mass production

Computer programmers and all IT engineers belong to the best paid labor force. Smooth standardization could
have roots in the industry interest to attract additional and cheaper workers as I have already mentioned. But market
pressures could go further, applications like Wordpress, Weebly etc. are examples of mass production in computer
science. No web developer is needed to develop new website. Computer science contributed to the automation in other
industries, but the next prey could be the predator itself and automation and AI could replace some of computer
programmers jobs as well.

3.4. Visible vs. Invisible markets

Besen 12 and Varian 2 have considered Netscape vs Microsoft battle as standard war. Same attention was paid to the
fight between Microsoft and Google Chrome, European Commission's Microsoft competition case was also in the
spotlights. Social network Facebook is similar case. Effect of browsers war on setting HTML standard is overestimated
and Facebook brings more impulses to the psychology than to the economy or technology. Key point is that social
media and browsers have billion users and every news , including technical, has to be considered as PR product. There
are many other, mostly B2B , markets, where more interesting and important things occur out of public view.

4. Artificial Intelligence and Machine Learning

AI has been pure scientific concept for long time, but recently it's easier to start with some economics analysis
because we can see now many real applications on the markets. There could be again two views on AI – “pure
scientific”, where almost everything is achievable and “game theoretical” where AI solution enters real life with all it's
risks. Later I will try to open hypothesis that AI alone can work perfectly in finite environments where number of inputs
is predictable and limited. Otherwise, AI should be controlled by humans.
 I joined some AI seminar at University of Manchester in winter 2019. Lecturer mentioned all achievements ,
autonomous cars, possibility of autonomous planes etc. He presented hacking as the risk for autonomous cars. But
hacking is still among “pure scientific risks” from my point of view. As you can physically pierce the tire by car or stole
the car, you can hack its AI module and damaged it or hack it to control it. He also mentioned a fact, that autonomous
cars have to, for example, recognize child on the road and react. I suggested possibility, when somebody could use
posters or videos with children to make an accident happen. It's “game-theoretical” risk of using AI. This example is
hypothetical, but there are more real ones.

BBC 7 reported that the pound had plummeted in Asian trading in early hours on Friday, October 7th, 2016 (refer
to image below).What actually happened ? Overnight value of the pound fell dramatically by 6%. There were no clear
reason for it as Forex trading has no single market and central clearing house, but two hypothesis had been taken as
possible. Human error of trader or computer algorithm reaction to a news story about French President Hollande
reaction over Brexit. Although it was not possible to proof it was AI inappropriate reaction, this story showed that AI is
already far away from labs and is built-in the real life and thus also vulnerable not only to standard security(hacking)
threats, but also to the new “game-theoretical” threats.

I remembered this story from the year 2016, but I was not able to find it later in the 2019 to be able to use it as
example in this article. I have used Google search and various combination of keywords “FX,Forex,Foreign Exchange“
and keywords “AI,mistake,error,trading,algorithm trading,BBC etc.“. I have used almost 50 combinations and always
first 50 google results were adds on Forex companies offering their services. I would dare to paraphrase Varian 2 idea
„The fact that my father got a great deal on Lotus 123 or Wordperfect may be of little solace to me when I have to pay a
high price for their successors.” to state “The fact Mr. Varian used Yahoo or Alravista search engine may be of little
solace to me when I get polluted result from google search. Varian used this quote when analysing supply-side
economies of scale and time consistency problem. There is gain in surplus that consumers receive in competition phase
to acquire monopoly power (great deal on Lotus 123, great free search by Yahoo or Google ), but this gain may offset
the losses incurred in monopoly phase ( expensive Office suit, add polluted Google search).

Image 3. Pound night crash as result of possible AI error.

IBM very well understood that sport area is suitable for marketing their own ability to produce AI solutions with
no risk. ( compare to the risk in Forex or autonomous cars ). IBM super-computer Deep Blue played six-games chess
match against world-champion and one of best ever players Garry Kasparov in 1996. Kasparov won 4-2 and rematch
was played in New York in 1997 and Deep Blue won 3 and half : 2 and half. This was great success and demonstration
of possible power of AI. Two lessons should be learnt from it. Chess is finite game and although it has billions
combination, rules (inputs ) are limited. Such condition are convenient for AI. Kasparov lost second game, but was able
to adapt quickly and showed so called anti-computer tactics in the second game of second series (game-theoretical
approach). It involves playing conservatively for a long-term advantage that the computer is unable to find in its game
tree search, it also means selecting moves that appear sub-optimal in the short term in order to exploit known weakness
in the way computer players evaluate positions 10.

IBM further worked as official web, data and data-analysis provider for 4 tennis Grand Slam tournaments. Tennis
is a game where various game statistics (serve success rate, winners and unforced errors etc/) are correlated with the
game result and could be taken as explanatory variables. IBM went further and have tried to find key elements of every
set and game, elements which decided about game result. In other words, which statistical differences between both
players were more significant ( see image below ). There are not only above mentioned statistics, but also other like at
which position at the court is player in the key moments etc. As chess AI was great success, tennis AI looks more like
marketing tool. Why ? Tennis is not finite game, inputs and factors are not limited. There could be always factors like
wind, sun, referees errors, ball spins, efficiency of player style against other player etc.

Image 4. IBM AI tries to analyze tennis match

5. Programming languages, The Evolution and Intellectual Property

Let's start again with microeconomics look on programming languages. Programming languages is key input in
various economic production processes, but it's very uncommon good itself. This good is produced only once ( if we
include updates procedures than few times ). Although such productions process is exclusive, usual price of
programming language is 0. Also if there is some non-zero price set by corporations then it's mostly price for tools
which help programmers or for support, not for programming language itself. It makes economic sense. If production
process runs just once and good is later used “million” times, than price would be close to 0. Difference between “close
to zero” and “zero” can be explained as the rent, which programming language creators (corporations or programmers
associations or individuals ) receive in the form of higher skill, which they can later monetize on the labor market.
It's economical aspect of production process, Ghezzi 13 describes technological part of programming language creation
brilliantly.

Let's now look on different aspect of programming languages, their evolution. It's not dissimilar to the human
languages. Human languages evolution took centuries and was affected by geography, demography, economy and wars.
Programming languages evolution started in 50's and has lasted decades, in its intensive form 4 or 5. One human
generation could be a witness of this. Genealogy list of programming languages 14 could be ( and surely is ) discussed in
the separate book. Opposite to the human languages, there should be no constrain in our new evolution. As we already
discussed, it's only input to the production function of economic agents, so they should have no interest to delay and
block this process. Varian 2 see critical concern about intellectual property, copyright law and patent law. There is no
other way than to agree, but regarding to the zero valuation of programming languages, intellectual property works as
the constrain in the evolution process. Once again,this statement is valid only for programming language as sub-
segment of information technologies. Image 2 confirms this hypothesis as open source is becoming standard in this sub-
market.

6. Human Thinking and Computer Science mutual influencing

Industrial revolution and recent production process automation led to the replacement of manual physical labor by
the machine work. Patterns of production process remained almost same, but machines can produce faster, cheaper and
in larger volumes. Mechanical engineering replaced human working procedures and added some, which would not
possible to reach using manpower. Electrical engineering added energy,drive and accuracy.The influence is one-sided
only. Mechanical engineers and designers can be inspired by human work, but working humans are not able to change
their procedures by watching machines.

Information technologies as third big engineering are mostly (excluding hardware) product of human thinking and
creativity. We have already mentioned role of standards, but at higher level than standards, we can see various general
principles. Computer programming can be object-oriented, functional or procedural. Databases can be relational, graph
or object-oriented. Does it mean that some of these principles are more suitable to the way in which humans think ? Or
are we again in the process of the evolution and some of these principles will win and some will become obsolete ? Or
can different people prefer different patterns of thinking, so all these principles can coexist concurrently ? And finally to
connect on first paragraph, can the principle which computer programmers use on other side affect the way humans
think ?
Conclusion

There are two main principles where economics could enrich information technologies. First is ability to switch and
move between details and general principles. It's necessary to be specialized in current IT labor market, but it will be
not much useful to be specialized without understanding where do you stay in the overall picture and how the overall
picture is changing. Second principle is economics statute of social science and know-how in analyzing human behavior
and psychological aspects of economic agents activities. Tools like game-theory in it's wider meaning and economic
modeling which has not be mathematically exact (natural science), but should show all inputs, outputs, powers and
dynamics and so ceteris paribus. Economists were successful to get Google CEO job (Hal Varian) and Amazon is
biggest employer of Economics Phd students, before all universities, using them to improve it's auctioning system. It's
amazing, but does it mean that role of economists in IT ends up with monetization of technology and has nothing to say
about technologies itself ? So please take this article as appeal to contribute more and show whole power of economics.
References

1/ Hal R. Varian. Microeconomic Analysis, Third Edition.

Viva Books, 2009.

2/ Hal R. Varian. Economics of Information Technology.

University of California, Berkeley, 2001

3/ Richard Veryard. The Economics of Information Systems and Software.

Butterworth-Heinemann, 1991

4/ Stephen Morris, Nancy Devlin, David Parkin. Economic Analysis in Health Care. John Wiley & Sons, 2007.

5/ Roy L. Nersesian. Energy Economics Markets, History and Policy. Routledge, 2016

6/ Debertin, David L., "Agricultural Production Economics" (2012). Agricultural Economics Textbook Gallery. 1.
https://uknowledge.uky.edu/agecon_textbooks/1

7/ BBC. What's happening to the pound?. 7.10.2016.

https://www.bbc.co.uk/news/business-37584727

8/ Alex Cukierman, Central Bank Strategy, Credibility and Independence: Theory and Evidence, The MIT
Press,Cambridge, MA, 1992

9/ Martin Libicki, James Schneider, Dave R. Frelinger, and Ann Slomovic. Scaffolding the New Web: Standards and
Standards Policy for the Digital Economic. RAND, Santa Monica, CA, 2000.
http://www.rand.org/publications/MR/MR1215/.

10/ https://wikivisually.com/wiki/Anti-computer_tactics

11/ https://opensource.com/resources/what-open-source

12/ Stanley Besen and Joseph Farrell. Choosing how to compete: Strategies and tactics in standartization. Journal of
Economic Perspectives, oages 117-131. 1994

13/ Carlo Ghezzi, Mehdi Jazayeri. Programming language concepts, third edition. John Wiley & Sons, 1996.

14/ Wikipedia https://en.wikipedia.org/wiki/Generational_list_of_programming_languages