[Reader-list] For Economists and Python Programmers

[Sunil Abraham]

http://www.kuro5hin.org/story/2003/7/29/17659/1721

Do check out the demo at http://216.36.193.92/optimaes/optimaes.cgi

Open Project To Investigate Money And Economic Systems (Politics)

By interstar
Thu Jul 31st, 2003 at 12:06:49 PM EST 	
This article introduces the "Open Project To Investigate Money And
Economic Systems" and invites anyone interested to join in.

OPTIMAES is a way for people to explore alternative economics through
computer simulation. It came from one of the founders' desire to get a
better understanding of politically contentious questions about markets,
globalization and property. And to explore more exotic ideas like local
currencies and interest-free money. 

It's an amateur effort, run more like an open-source project than a
piece of academic research. And the hope is to grow a community of
interested and argumentative people to explore these issues; to "fact
check" the politicians, mainstream economists and radical protest
groups; and to get more discussion and understanding of these issues on
the net.


When I was young, I was a socialist. It was obvious that there were poor
and disadvantaged people in the world. And someone should do something
about it! Government seemed the obvious instrument for doing that
something. People who told me that there were problems with this, who
worried about government inefficiency and the stifling of the
entrepreneurial spirit, seemed merely venal. Apologists for selfishness.

But I was also a computer geek. I spent my time on the net, reading
about technology and it's attendant philosophies. I was a
techno-optimist : a believer that innovation and progress came from 
openness and freedom; from deregulated, "stupid" networks; from
decentralization, self-organization and the bazaar rather than the
cathedral. 

Online, I read libertarians who claimed to want a better world too. And
that freedom and property rights could take us there. I began to wonder
if, given enough freedom, the poor, the third-world, the disenfranchised
could escape poverty, using the same mechanisms that the rich had
already pioneered. 

I was being converted.

And yet, some of my socialist friends reminded me, things weren't so
good. The libertarians promised that free-trade would benefit everyone.
But many in the third world were getting poorer. The working class in
the first world were getting poorer, working longer hours, and with less
security. Studies of scale-free networks showed that the rewards of
economic freedom wouldn't be spread evenly, but distributed according to
a power-law. If that was true, it was even possible that opening up
trade with a previously closed impoverished country, could suck wealth
out of it, rather than help it develop.

As I looked around, I found  many competing political positions. And I
started to notice that they were grounded in rival presumptions about
economics. Many of the ideas, from libertarian to marxist, conservative
to eco-feminist, sounded at least plausible. The more I read, the less
certain I became. My intuition and common-sense reasoning weren't enough
to make up my mind.

I thought of studying economics. But, friends who were economists were
disillusioned with the subject. And I didn't have the maths.

Nevertheless I'm have enough hacker genes to believe that I can
understand something; if I can build it (or at least a scale model) in
software. I realized there was only one way I'd really sort things out
to my own satisfaction : computer simulation. 

The Dialogue of Models

I could also see that we couldn't expect computer models to give a
definitive answer to the questions. All models are simplifications and
abstractions. And all models are biased by the preconceptions of the
person building them. 

But models are great ways of sharpening up intuitions and clarifying
assumptions. To implement something, you have to make all those
prejudices explicit. And that makes it  possible for opponents to
identify and call attention to them.

For example, as a left-wing coder, I could undoubtedly make a simulated
economy where communistic redistribution worked better than the free
market. 

And a right wing libertarian would dismiss it as bogus. 

But the real value would be if the right-winger took the code,
identified and pointed out the assumptions that bias the model, and
changed them to demonstrate his claims. The option would then be open to
another leftist to respond the same way, to refine the model yet
further. So I decided my simulation had to be part of an open research
project with the models as stages in a critical debate. A "dialogue of
models". 

Ideally, this exploration should be with people who start with different
intuitions and political positions. The more diverse the opinions, and
the harder people are willing to fight for their corner, the more work
can get done. 

Why not leave it to the experts?

There are some economists and social scientists who use agent-based
models. There's a long list of them here. 

I wanted to learn from these people. But I still felt dissatisfied.
Papers are written in academic language for other academics. This is a
strain to read. I'd rather write and read code than read academic
jargon. And my project was intended to make both doing the research, and
knowledge of the results, available to a wider community than just the
academic one. 

Having said that, I certainly welcome criticisms and suggestions from
academic researchers in the same area.

In practice

So what does the Optimaes project consist of?

At the moment it has two parts : 


an embryonic computer model of an economy
      * 
        the optimaes wiki.
      * The model is a seed. Something that I hope can grow. The idea
is, if you want to ask a question or demonstrate or falsify a
hypothesis, you can modify the code to embody that hypothesis and watch
what happens.

We're still at the beginning, but we do have gift giving agents,
bartering agents, and a first draft of a money using agent (one that
holds local auctions in it's social neighbourhood). Agents live in
social networks that can be connected randomly or locally. (And we hope
to add small-worlds soon.) With our initial minimal assumptions, the
economy of gift giving agents is far more productive than those based on
trade. But I'm sure critics will soon discover why, and try to fix that
;-) 

The code is written in Python; because it's simple, runs on many
platforms, and a C++ / Java / Perl / Smalltalk programmer can get into
it in a couple of hours. Our first priority is to maximize the
inclusiveness of the project and accessibility to the model.

The wiki is a place to document and discuss experiments with the code.
In general I aim for one page describing a problem, the assumptions and
the results and conclusions. If it involves writing new classes of
economic agent, these are documented on other pages. We use  Use-Mod
wiki, so it's possible to break the description into finer detail using
the subpage notation.

The idea is, that if you want to criticize an existing model, it's
possible to create a new page linked from that model's page. In this
way, the wiki structure captures the structure of the arguments. 

What are we investigating?

At the moment, the founders are working towards three investigations :

an investigation of interest free money and local currencies. See this
overview of many of these ideas.  
      * 
        an investigation of gift economies. Can they be as efficient at
distributing resources as exchange economies? Can they use  reputation
management to protect themselves against exploitation by selfish agents?
And what is a good simplified model of human psychology, anyway?
      * 
        questions about globalization and free trade : what happens when
you allow free trade between two previously shielded economies? Is it
good for everyone, or can the stronger economy end up pulling wealth
from the weaker?
      * 
        But we are interested in other questions that might be
suggested. One such, currently, is to look into digital goods and
non-scarce resources. Other people might be interested in modeling
incentives for innovation, or various kinds of redistributive taxation,
or the fine detail of auctioning, bartering or gift giving strategies,
or ... 

How far are we?

About four man-weeks in. We have a working first draft. Six people have
tried out the code, asked questions and made suggestions. We have some
on-wiki discussions about refactoring; our first bug-fix suggested by a
non-founder; and interested people in the UK, Italy, Portugal and Brazil
(where the project started). 

Now we are excited ...



-- 
Sunil Abraham, sunil at mahiti.org http://www.mahiti.org
MAHITI Infotech Pvt. Ltd.'Reducing the cost and complexity of ICTs'
314/1, 7th Cross, Domlur Bangalore - 560 071 Karnataka, INDIA
Ph/Fax: +91 80 4150580. Mobile: 98455 12611
"If you have an apple and I have an apple and we exchange apples 
then you and I will still each have one apple. 
But if you have an idea and I have one idea and we exchange these ideas,
then each of us will have two ideas" George B. Shaw