Agent-based market models
Locked
12
12
 
|
Does anyone know of good examples of generic agent-based market economies? I'm thinking of something as simple as this.
A population consists of agents each of which has certain continuing needs (such as food, clothing, shelter, Internet access, etc.) to survive. As a starting point, let's assume that each agent needs one unit of each of N resources every time period. Let's also assume that each agent is specialized and is capable of creating enough of one of the needed resources to satisfy the needs of N agents. (The fact that I used the same N in both places was intentional.) To keep it simple let's assume that these acts of creation occur from scratch, i.e., that the creator doesn't need raw materials, that all that's necessary for an agent to create a needed resource is that the agent be alive. The agents presumably develop a barter economy, trading the resources they create for the resources they need to stay alive. Perhaps markets develop, and perhaps money develops. At this point the economy should be fairly stable. Each agent creates enough stuff so that s/he can trade it for what s/he needs to stay live. Perhaps some of the agents learn how to be more efficient in creating their resource and begin to accumulate "wealth" in some form. Perhaps the agents have discretionary desires, which they fill if they have enough resources left over after meeting their basic needs. Perhaps there are communal services that are paid for by taxes or memberships. This could become increasingly elaborate. It seems to me that models of this sort must have been developed -- perhaps many times. Does anyone know of any references to this sort of work? Thanks. -- Russ Abbott _____________________________________________ Professor, Computer Science California State University, Los Angeles o Check out my blog at http://bluecatblog.wordpress.com/ ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
Sorry for the (only partly) flippant answer: Simcity!
On 12/04/2009, Russ Abbott <[hidden email]> wrote: > Does anyone know of good examples of generic agent-based market economies? > I'm thinking of something as simple as this. > > A population consists of agents each of which has certain continuing needs > (such as food, clothing, shelter, Internet access, etc.) to survive. As a > starting point, let's assume that each agent needs one unit of each of N > resources every time period. Let's also assume that each agent is > specialized and is capable of creating enough of one of the needed resources > to satisfy the needs of N agents. (The fact that I used the same N in both > places was intentional.) To keep it simple let's assume that these acts of > creation occur from scratch, i.e., that the creator doesn't need raw > materials, that all that's necessary for an agent to create a needed > resource is that the agent be alive. The agents presumably develop a barter > economy, trading the resources they create for the resources they need to > stay alive. Perhaps markets develop, and perhaps money develops. At this > point the economy should be fairly stable. Each agent creates enough stuff > so that s/he can trade it for what s/he needs to stay live. > > Perhaps some of the agents learn how to be more efficient in creating their > resource and begin to accumulate "wealth" in some form. Perhaps the agents > have discretionary desires, which they fill if they have enough resources > left over after meeting their basic needs. Perhaps there are communal > services that are paid for by taxes or memberships. This could become > increasingly elaborate. > > It seems to me that models of this sort must have been developed -- perhaps > many times. Does anyone know of any references to this sort of work? > > Thanks. > > -- Russ Abbott > _____________________________________________ > Professor, Computer Science > California State University, Los Angeles > o Check out my blog at http://bluecatblog.wordpress.com/ > -- Saul Caganoff Enterprise IT Architect LinkedIn: http://www.linkedin.com/in/scaganoff ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Russ Abbott
There's quite a few papers on this at the Journal of Artificial Societies and Social Simulation: http://jasss.soc.surrey.ac.uk/JASSS.html
Also, googling "agent based market economies" gives some useful links; e.g. Tesfatsion's site about agent based computational economics seems to have some good links
Robert On Sun, Apr 12, 2009 at 8:00 AM, Russ Abbott <[hidden email]> wrote: Does anyone know of good examples of generic agent-based market economies? I'm thinking of something as simple as this. ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by scaganoff
Actually I've recently been playing SimCity 4 with my son and been discovering I'd make a rubbish mayor (you only settle in Holmesburg if you like high taxes and pollution). It got me thinking: what are the implied politics of SimCity? Seems that someone has already written about this: http://www.daquellamanera.org/files/Lobo_CityToy05LSE.pdf
Robert
On Sun, Apr 12, 2009 at 8:58 AM, Saul Caganoff <[hidden email]> wrote: Sorry for the (only partly) flippant answer: Simcity! ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Robert Holmes
The model Tom Carter presents at SFI's Complex Systems Summer School
has some overlap - and is deceptively simple: create 500 agents with 50 dollars each each timestep if an agent has any dollars, give one dollar to another agent at random That's it. What is the long-run distribution of wealth? hint: a leetle statistical mechanics :-) For answers, see Victor Yakovenko's paper here: http://arxiv.org/abs/cond-mat/0001432 Owen and JP both wrote models exploring this though I don't have links handy. JP? Owen? -Steve > > On Sun, Apr 12, 2009 at 8:00 AM, Russ Abbott <[hidden email]> > wrote: > Does anyone know of good examples of generic agent-based market > economies? I'm thinking of something as simple as this. > > A population consists of agents each of which has certain continuing > needs (such as food, clothing, shelter, Internet access, etc.) to > survive. As a starting point, let's assume that each agent needs one > unit of each of N resources every time period. Let's also assume > that each agent is specialized and is capable of creating enough of > one of the needed resources to satisfy the needs of N agents. (The > fact that I used the same N in both places was intentional.) To keep > it simple let's assume that these acts of creation occur from > scratch, i.e., that the creator doesn't need raw materials, that all > that's necessary for an agent to create a needed resource is that > the agent be alive. The agents presumably develop a barter economy, > trading the resources they create for the resources they need to > stay alive. Perhaps markets develop, and perhaps money develops. At > this point the economy should be fairly stable. Each agent creates > enough stuff so that s/he can trade it for what s/he needs to stay > live. > > Perhaps some of the agents learn how to be more efficient in > creating their resource and begin to accumulate "wealth" in some > form. Perhaps the agents have discretionary desires, which they > fill if they have enough resources left over after meeting their > basic needs. Perhaps there are communal services that are paid for > by taxes or memberships. This could become increasingly elaborate. > > It seems to me that models of this sort must have been developed -- > perhaps many times. Does anyone know of any references to this sort > of work? > > Thanks. > > -- Russ Abbott > _____________________________________________ > Professor, Computer Science > California State University, Los Angeles > o Check out my blog at http://bluecatblog.wordpress.com/ > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models
|
|
In reply to this post by Russ Abbott
A classic example is Epstein & Axtell's
"Sugerscape" model. It simulates the distribution of wealth "The rich get richer
and the poor get poorer" Probably all ABM
packages will have a Sugerscape model. Netlogo's (http://ccl.northwestern.edu/netlogo/) "Wealth Distribution" is
only one of many.
There is a nice description at http://www.theatlantic.com/doc/200204/rauch on this and
other related ABM models (for example, Schelling's "Segregation Model" is
another classic ABM showing how a tiny bit of prejudice can lead to segregated
neighbourhoods.)
Pieter Steenekamp
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Russ Abbott
The central property which emerges in markets is
the price. The central law which rules markets is the law of supply and demand. A basic agent-based model for markets should explain how both, the price and the law of supply and demand, emerge in competitive markets. It should be simple to extend such a model to explain bubbles and crashes. As far as I know, there is no basic model of markets in general, there are only agent based models of specific markets, for example stock markets, financial markets and electricity markets. Perhaps the law of supply and demand is too simple to justify such a model? Epstein and Axtell's Sugarscape is a model where agents begin to accumulate "wealth" in some form, but it is more about evolution of societies and less about market mechanisms and economies. It would be interesting to apply a basic agent-based market model to the society of mind. If we consider the crowd within and the battle of ideas which is going on inside our minds, can we create an agent-based model to explain the 'market of information' ? We have a demand for new or highly emotional stuff, are constantly supplied with information from the outside, and the price we have to pay to consume it is attention. -J. ----- Original Message ----- From: Russ Abbott To: The Friday Morning Applied Complexity Coffee Group Sent: Sunday, April 12, 2009 9:00 AM Subject: [FRIAM] Agent-based market models Does anyone know of good examples of generic agent-based market economies? I'm thinking of something as simple as this. A population consists of agents each of which has certain continuing needs (such as food, clothing, shelter, Internet access, etc.) to survive. As a starting point, let's assume that each agent needs one unit of each of N resources every time period. Let's also assume that each agent is specialized and is capable of creating enough of one of the needed resources to satisfy the needs of N agents. (The fact that I used the same N in both places was intentional.) To keep it simple let's assume that these acts of creation occur from scratch, i.e., that the creator doesn't need raw materials, that all that's necessary for an agent to create a needed resource is that the agent be alive. The agents presumably develop a barter economy, trading the resources they create for the resources they need to stay alive. Perhaps markets develop, and perhaps money develops. At this point the economy should be fairly stable. Each agent creates enough stuff so that s/he can trade it for what s/he needs to stay live. Perhaps some of the agents learn how to be more efficient in creating their resource and begin to accumulate "wealth" in some form. Perhaps the agents have discretionary desires, which they fill if they have enough resources left over after meeting their basic needs. Perhaps there are communal services that are paid for by taxes or memberships. This could become increasingly elaborate. It seems to me that models of this sort must have been developed -- perhaps many times. Does anyone know of any references to this sort of work? Thanks. -- Russ Abbott _____________________________________________ Professor, Computer Science California State University, Los Angeles o Check out my blog at http://bluecatblog.wordpress.com/ ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models
|
|
> Attention and action = the price to consume.
is it possible to have an idea / new information and not in some way, internally or externally, react to it? Memes spread. Humans by design infect each other on all levels. The Complex seems like a great example: by definition intends to do this. Idea / attention / action-reaction / experience / new idea. And so forth. Am enjoying the bouncing of ideas and information in this particular string... Tory On Apr 12, 2009, at 7:04 AM, Jochen Fromm wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models & creative messiness
|
|
I'm surprised that I want to respond to this thread, but it touches on
some lifelong personal interests and attempts at clearer understandings. Jochen very clearly lays out some of the areas of exploration that are central to complexity research and applications. There is no disagreement. It represents an approach to understanding. As such, it is reductive. It will result in reductive systems and models understandings. The evolving nature of life, of societies and their actions, and of individuals living in dynamic states of fear, of confusion, of love, and more, is messy. There are messy ways towards understanding, as well. Blending reductive understandings with deep experiential involvements in messiness, inspires some to pursue lives of creative, caring acts, influences and understandings. RL Quoting Victoria Hughes <[hidden email]>: > > Attention and action = the price to consume. > > is it possible to have an idea / new information and not in some way, > internally or externally, react to it? > Memes spread. Humans by design infect each other on all levels. > The Complex seems like a great example: by definition intends to do this. > Idea / attention / action-reaction / experience / new idea. > And so forth. > Am enjoying the bouncing of ideas and information in this particular > string... > Tory > > On Apr 12, 2009, at 7:04 AM, Jochen Fromm wrote: > >> The central property which emerges in markets is >> the price. The central law which rules markets is >> the law of supply and demand. A basic agent-based >> model for markets should explain how both, the >> price and the law of supply and demand, emerge >> in competitive markets. It should be simple to >> extend such a model to explain bubbles and crashes. >> As far as I know, there is no basic model of >> markets in general, there are only agent based models >> of specific markets, for example stock markets, >> financial markets and electricity markets. Perhaps >> the law of supply and demand is too simple to justify >> such a model? >> >> Epstein and Axtell's Sugarscape is a model where >> agents begin to accumulate "wealth" in some form, >> but it is more about evolution of societies and less >> about market mechanisms and economies. >> >> It would be interesting to apply a basic agent-based >> market model to the society of mind. If we consider the >> crowd within and the battle of ideas which is going on >> inside our minds, can we create an agent-based model >> to explain the 'market of information' ? We have a >> demand for new or highly emotional stuff, are constantly >> supplied with information from the outside, and the price >> we have to pay to consume it is attention. >> >> -J. >> >> ----- Original Message ----- From: Russ Abbott >> To: The Friday Morning Applied Complexity Coffee Group >> Sent: Sunday, April 12, 2009 9:00 AM >> Subject: [FRIAM] Agent-based market models >> >> Does anyone know of good examples of generic agent-based market >> economies? I'm thinking of something as simple as this. >> >> A population consists of agents each of which has certain continuing >> needs (such as food, clothing, shelter, Internet access, etc.) to >> survive. As a starting point, let's assume that each agent needs one >> unit of each of N resources every time period. Let's also assume >> that each agent is specialized and is capable of creating enough of >> one of the needed resources to satisfy the needs of N agents. (The >> fact that I used the same N in both places was intentional.) To keep >> it simple let's assume that these acts of creation occur from >> scratch, i.e., that the creator doesn't need raw materials, that all >> that's necessary for an agent to create a needed resource is that >> the agent be alive. The agents presumably develop a barter economy, >> trading the resources they create for the resources they need to >> stay alive. Perhaps markets develop, and perhaps money develops. At >> this point the economy should be fairly stable. Each agent creates >> enough stuff so that s/he can trade it for what s/he needs to stay >> live. >> >> Perhaps some of the agents learn how to be more efficient in >> creating their resource and begin to accumulate "wealth" in some >> form. Perhaps the agents have discretionary desires, which they >> fill if they have enough resources left over after meeting their >> basic needs. Perhaps there are communal services that are paid for >> by taxes or memberships. This could become increasingly elaborate. >> >> It seems to me that models of this sort must have been developed -- >> perhaps many times. Does anyone know of any references to this sort >> of work? >> >> Thanks. >> >> -- Russ Abbott >> _____________________________________________ >> Professor, Computer Science >> California State University, Los Angeles >> o Check out my blog at http://bluecatblog.wordpress.com/ >> ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Victoria Hughes
Thanks for all the responses. What prompted my question was thinking about the "end state" of the current financial crisis. The question is: what does a successful stable economy look like? Only secondarily was I asking about markets, wealth distribution across society, etc., which is what many of the existing models focus on.
In attempting to describe the foundations of a stable economy I was making the following assumptions.
I must say that I'm surprised at the number of things I had to write down to explain what I am after. I thought it would be simpler. And I want something as simple as possible. Any thoughts/suggestions? -- Russ On Sun, Apr 12, 2009 at 9:14 AM, Victoria Hughes <[hidden email]> wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models
|
|
Russ,
Your posting encapsulated the subject of Hazel Henderson's 1988 book, The Politics of the Solar Age; Alternatives to Economics. She and many others have gone further since. Your encapsulation also touches on the roots of the long emerging field of Ecological Economics. RL Quoting Russ Abbott <[hidden email]>: > Thanks for all the responses. What prompted my question was thinking about > the "end state" of the current financial crisis. The question is: what does > a successful stable economy look like? Only secondarily was I asking about > markets, wealth distribution across society, etc., which is what many of the > existing models focus on. > > In attempting to describe the foundations of a stable economy I was making > the following assumptions. > > - We live off the gift of free energy -- the sun. So in some sense a > stable economy requires little more than harvesting that energy. > - As human beings we have the innate (i.e., also free) ability to > transform that energy into other forms that we need. For example, > most of us > on this list are capable of writing code. I'm taking skills like that as > more free gifts from nature. Yes, it requires, education, study, practice, > etc., but ultimately it grows out of our being alive, a free gift. > - In talking about "free gifts" from nature, I'm not trying to be > spiritual, new age-y, etc. It just seems to me to be the way the world > works. So I'm willing to build it into the model. > - But, even though we live off these free gifts, we all need energy to > survive. If you look at my "Reductionist blind spot" paper, you see that we > are what I call dynamic entities. Dynamic entities need to import > externally > supplied energy to persist -- in our case food, at least. So even though we > find ourselves in a world in which there is externally supplied energy > available to be imported (from the sun and its stored versions), we > are each > still required to do something to import it for ourselves. That is, we must > each work for a living. > - Presumably a standard economic model says that we live at a subsistence > level if the amount of work we do matches our needs. If I spend all my > working hours doing nothing more than earning what I need to > survive, that's > subsistence. > - I want to allow specialization so I suggested that each agent > specializes in producing some resource but that the aggregate amount > produced exactly satisfies the needs of the producers. That is, each agent > produces N units of his specialty resource; there are N agents; N resources > are needed for survival. So if the aggregate production is evenly > distributed everyone has a subsistence living. > - A society/economy grows "richer" when it can produce enough resources > for itself without requiring everyone to work on producing those resources. > In that case, the available excess labor can produce discretionary items, > like art, pop music, software, soft drinks, philosophy, etc. > - As technology advances a smaller and smaller percentage of the > population is needed to produce the necessities. The rest can spend their > time producing discretionary items. > - So at this point we are all well fed, and we are all willing to spend > our extra money supporting our local NPR stations. > - But what happens when there is shock to the economy and people find > that they have less "wealth" than they thought they had? Presumably they > will spend less on discretionary items, and the people who create those > items will be out of work. Where do we go from there? > > So I am looking for a model to explore these kinds of issues. > > I must say that I'm surprised at the number of things I had to write down to > explain what I am after. I thought it would be simpler. And I want something > as simple as possible. > > Any thoughts/suggestions? > > -- Russ > > > On Sun, Apr 12, 2009 at 9:14 AM, Victoria Hughes > <[hidden email]>wrote: > >> > Attention and *action = *the price to consume. >> is it possible to have an idea / new information and not in some way, >> internally or externally, react to it? >> Memes spread. Humans by design infect each other on all levels. >> The Complex seems like a great example: by definition intends to do this. >> Idea / attention / action-reaction / experience / new idea. >> And so forth. >> Am enjoying the bouncing of ideas and information in this particular >> string... >> Tory >> >> On Apr 12, 2009, at 7:04 AM, Jochen Fromm wrote: >> >> The central property which emerges in markets is >> the price. The central law which rules markets is >> the law of supply and demand. A basic agent-based >> model for markets should explain how both, the >> price and the law of supply and demand, emerge >> in competitive markets. It should be simple to >> extend such a model to explain bubbles and crashes. >> As far as I know, there is no basic model of >> markets in general, there are only agent based models >> of specific markets, for example stock markets, >> financial markets and electricity markets. Perhaps >> the law of supply and demand is too simple to justify >> such a model? >> >> Epstein and Axtell's Sugarscape is a model where >> agents begin to accumulate "wealth" in some form, >> but it is more about evolution of societies and less >> about market mechanisms and economies. >> >> It would be interesting to apply a basic agent-based >> market model to the society of mind. If we consider the >> crowd within and the battle of ideas which is going on >> inside our minds, can we create an agent-based model >> to explain the 'market of information' ? We have a >> demand for new or highly emotional stuff, are constantly >> supplied with information from the outside, and the price >> we have to pay to consume it is attention. >> >> -J. >> >> ----- Original Message ----- From: Russ Abbott >> To: The Friday Morning Applied Complexity Coffee Group >> Sent: Sunday, April 12, 2009 9:00 AM >> Subject: [FRIAM] Agent-based market models >> >> Does anyone know of good examples of generic agent-based market economies? >> I'm thinking of something as simple as this. >> >> A population consists of agents each of which has certain continuing needs >> (such as food, clothing, shelter, Internet access, etc.) to survive. As a >> starting point, let's assume that each agent needs one unit of each of N >> resources every time period. Let's also assume that each agent is >> specialized and is capable of creating enough of one of the needed resources >> to satisfy the needs of N agents. (The fact that I used the same N in both >> places was intentional.) To keep it simple let's assume that these acts of >> creation occur from scratch, i.e., that the creator doesn't need raw >> materials, that all that's necessary for an agent to create a needed >> resource is that the agent be alive. The agents presumably develop a barter >> economy, trading the resources they create for the resources they need to >> stay alive. Perhaps markets develop, and perhaps money develops. At this >> point the economy should be fairly stable. Each agent creates enough stuff >> so that s/he can trade it for what s/he needs to stay live. >> >> Perhaps some of the agents learn how to be more efficient in creating their >> resource and begin to accumulate "wealth" in some form. Perhaps the agents >> have discretionary desires, which they fill if they have enough resources >> left over after meeting their basic needs. Perhaps there are communal >> services that are paid for by taxes or memberships. This could become >> increasingly elaborate. >> >> It seems to me that models of this sort must have been developed -- perhaps >> many times. Does anyone know of any references to this sort of work? >> >> Thanks. >> >> -- Russ Abbott >> _____________________________________________ >> Professor, Computer Science >> California State University, Los Angeles >> o Check out my blog at http://bluecatblog.wordpress.com/ >> >> >> ============================================================ >> FRIAM Applied Complexity Group listserv >> Meets Fridays 9a-11:30 at cafe at St. John's College >> lectures, archives, unsubscribe, maps at http://www.friam.org >> >> >> >> ============================================================ >> FRIAM Applied Complexity Group listserv >> Meets Fridays 9a-11:30 at cafe at St. John's College >> lectures, archives, unsubscribe, maps at http://www.friam.org >> > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Russ Abbott
Russ writes:
> there are N agents; N resources are needed for survival. So if the > aggregate production is evenly distributed everyone has a > subsistence living. I don't think this is necessarily true. Your mentioning a gift economy is further a reason to look at a model as simple as the Tom Carter/Victor Yakovenko model I mentioned where agents are giving a dollar to another random agent (ie gifting). This model, with minimal assumptions, generates non-equitable distribution of wealth where very few agents end up holding most of the wealth (Boltzmann distribution) even though the average wealth in the model is unchanged at $50/agent. The important constraint in the model is that agents can only have positive amounts of cash/goods -- they can't gift a dollar they don't have. Just as energy distributes itself as a boltzmann distribution as particles can not have less than zero energy after a collision (transaction). Basically we have diffusion against a wall in both models. Of course we can introduce debt and progressive taxes to the economic model to mess with the dynamics, but I think it's important to recognize the non-equity is there from the beginning with very minimal assumptions. Note that the wealthy agents do turnover in this model so the definition of what is equitable is important. -Steve On Apr 12, 2009, at 11:40 AM, Russ Abbott wrote: > Thanks for all the responses. What prompted my question was > thinking about the "end state" of the current financial crisis. The > question is: what does a successful stable economy look like? Only > secondarily was I asking about markets, wealth distribution across > society, etc., which is what many of the existing models focus on. > > In attempting to describe the foundations of a stable economy I was > making the following assumptions. > • We live off the gift of free energy -- the sun. So in some sense > a stable economy requires little more than harvesting that energy. > • As human beings we have the innate (i.e., also free) ability to > transform that energy into other forms that we need. For example, > most of us on this list are capable of writing code. I'm taking > skills like that as more free gifts from nature. Yes, it requires, > education, study, practice, etc., but ultimately it grows out of our > being alive, a free gift. > • In talking about "free gifts" from nature, I'm not trying to be > spiritual, new age-y, etc. It just seems to me to be the way the > world works. So I'm willing to build it into the model. > • But, even though we live off these free gifts, we all need energy > to survive. If you look at my "Reductionist blind spot" paper, you > see that we are what I call dynamic entities. Dynamic entities need > to import externally supplied energy to persist -- in our case food, > at least. So even though we find ourselves in a world in which there > is externally supplied energy available to be imported (from the sun > and its stored versions), we are each still required to do something > to import it for ourselves. That is, we must each work for a living. > • Presumably a standard economic model says that we live at a > subsistence level if the amount of work we do matches our needs. If > I spend all my working hours doing nothing more than earning what I > need to survive, that's subsistence. > • I want to allow specialization so I suggested that each agent > specializes in producing some resource but that the aggregate amount > produced exactly satisfies the needs of the producers. That is, each > agent produces N units of his specialty resource; there are N > agents; N resources are needed for survival. So if the aggregate > production is evenly distributed everyone has a subsistence living. > • A society/economy grows "richer" when it can produce enough > resources for itself without requiring everyone to work on producing > those resources. In that case, the available excess labor can > produce discretionary items, like art, pop music, software, soft > drinks, philosophy, etc. > • As technology advances a smaller and smaller percentage of the > population is needed to produce the necessities. The rest can spend > their time producing discretionary items. > • So at this point we are all well fed, and we are all willing to > spend our extra money supporting our local NPR stations. > • But what happens when there is shock to the economy and people > find that they have less "wealth" than they thought they had? > Presumably they will spend less on discretionary items, and the > people who create those items will be out of work. Where do we go > from there? > So I am looking for a model to explore these kinds of issues. > > I must say that I'm surprised at the number of things I had to write > down to explain what I am after. I thought it would be simpler. And > I want something as simple as possible. > > Any thoughts/suggestions? > > -- Russ > > > On Sun, Apr 12, 2009 at 9:14 AM, Victoria Hughes <[hidden email] > > wrote: > > Attention and action = the price to consume. > > is it possible to have an idea / new information and not in some > way, internally or externally, react to it? > Memes spread. Humans by design infect each other on all levels. > The Complex seems like a great example: by definition intends to do > this. > Idea / attention / action-reaction / experience / new idea. > And so forth. > Am enjoying the bouncing of ideas and information in this particular > string... > Tory > > On Apr 12, 2009, at 7:04 AM, Jochen Fromm wrote: > >> The central property which emerges in markets is >> the price. The central law which rules markets is >> the law of supply and demand. A basic agent-based >> model for markets should explain how both, the >> price and the law of supply and demand, emerge >> in competitive markets. It should be simple to >> extend such a model to explain bubbles and crashes. >> As far as I know, there is no basic model of >> markets in general, there are only agent based models >> of specific markets, for example stock markets, >> financial markets and electricity markets. Perhaps >> the law of supply and demand is too simple to justify >> such a model? >> >> Epstein and Axtell's Sugarscape is a model where >> agents begin to accumulate "wealth" in some form, >> but it is more about evolution of societies and less >> about market mechanisms and economies. >> >> It would be interesting to apply a basic agent-based >> market model to the society of mind. If we consider the >> crowd within and the battle of ideas which is going on >> inside our minds, can we create an agent-based model >> to explain the 'market of information' ? We have a >> demand for new or highly emotional stuff, are constantly >> supplied with information from the outside, and the price >> we have to pay to consume it is attention. >> >> -J. >> >> ----- Original Message ----- From: Russ Abbott >> To: The Friday Morning Applied Complexity Coffee Group >> Sent: Sunday, April 12, 2009 9:00 AM >> Subject: [FRIAM] Agent-based market models >> >> Does anyone know of good examples of generic agent-based market >> economies? I'm thinking of something as simple as this. >> >> A population consists of agents each of which has certain >> continuing needs (such as food, clothing, shelter, Internet access, >> etc.) to survive. As a starting point, let's assume that each agent >> needs one unit of each of N resources every time period. Let's >> also assume that each agent is specialized and is capable of >> creating enough of one of the needed resources to satisfy the needs >> of N agents. (The fact that I used the same N in both places was >> intentional.) To keep it simple let's assume that these acts of >> creation occur from scratch, i.e., that the creator doesn't need >> raw materials, that all that's necessary for an agent to create a >> needed resource is that the agent be alive. The agents presumably >> develop a barter economy, trading the resources they create for the >> resources they need to stay alive. Perhaps markets develop, and >> perhaps money develops. At this point the economy should be fairly >> stable. Each agent creates enough stuff so that s/he can trade it >> for what s/he needs to stay live. >> >> Perhaps some of the agents learn how to be more efficient in >> creating their resource and begin to accumulate "wealth" in some >> form. Perhaps the agents have discretionary desires, which they >> fill if they have enough resources left over after meeting their >> basic needs. Perhaps there are communal services that are paid for >> by taxes or memberships. This could become increasingly elaborate. >> >> It seems to me that models of this sort must have been developed -- >> perhaps many times. Does anyone know of any references to this sort >> of work? >> >> Thanks. >> >> -- Russ Abbott >> _____________________________________________ >> Professor, Computer Science >> California State University, Los Angeles >> o Check out my blog at http://bluecatblog.wordpress.com/ >> >> >> ============================================================ >> FRIAM Applied Complexity Group listserv >> Meets Fridays 9a-11:30 at cafe at St. John's College >> lectures, archives, unsubscribe, maps at http://www.friam.org >> > > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
I'll have to look at the paper. But I'm surprised at the result. It seems counter intuitive. If an agent has nothing, he can't give anything away. But he has as good a chance to receive a gift as anyone else. So agents will tend not to go below a basic minimal level. Agents with lots of wealth are always giving it away. Yet they don't have a better change of getting any new wealth than any other agent. So why should their wealth grow? It would seem that wealthy agents would get wealthier only if wealth increases the probability of getting a gift. But the model didn't seem to suggest that.
I guess what might happen is that if wealth becomes concentrated, fewer transactions (gifts) occur at each time step -- since only those agents with any resources give gits. I.e., if lots of agents have nothing to give, fewer gifts occur at any time step, and each agent has a smaller probability of receiving anything. So if the distribution of wealth is such that the expected value of an agent's wealth stash decreases at every time step (as long as he has anything) then agent wealth stashes will tend toward zero. But since the money doesn't disappear, it will accumulate in a few agents. So I guess the population will tend to a distribution in which the percentage of agents with more than 0 units of wealth becomes constant at the level of the probability that any agent will receive a gift. (Or something like that.) That is, if every agent has one unit to give away, the expected value of the number of agents who will have a unit after the exchange is less than the total number of agents since some agents may receive two gifts. If only one agent has something to give away (plus something in reserve), the expected number of agents who will have something after the exchange will be two since the receiving agent and the gifting agent will each have something. So there must be a point at which the number of agents with some wealth will be stay more or less constant. I'm afraid I don't know how a Boltzmann distribution comes about. But I guess it's something like that. -- Russ On Sun, Apr 12, 2009 at 11:16 AM, Stephen Guerin <[hidden email]> wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
P.S. In model, though, there are no gifts, only exchanges. And exchanges are made when one agent has something another agent needs. So I don't think that model applies, at least not in its simple form.
-- Russ Abbott _____________________________________________ Professor, Computer Science California State University, Los Angeles o Check out my blog at http://bluecatblog.wordpress.com/ On Sun, Apr 12, 2009 at 11:47 AM, Russ Abbott <[hidden email]> wrote: I'll have to look at the paper. But I'm surprised at the result. It seems counter intuitive. If an agent has nothing, he can't give anything away. But he has as good a chance to receive a gift as anyone else. So agents will tend not to go below a basic minimal level. Agents with lots of wealth are always giving it away. Yet they don't have a better change of getting any new wealth than any other agent. So why should their wealth grow? It would seem that wealthy agents would get wealthier only if wealth increases the probability of getting a gift. But the model didn't seem to suggest that. ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
|
|
In reply to this post by Stephen Guerin
All -
I tend not to think of it as a "gifting" economy, but rather as a model where the "instrumentation" is such that only certain things are visible (i.e., "dollars"), but other things are invisible (i.e., "what the dollars are purchasing"). So, an agent who has "zero dollars" doesn't necessarily have "nothing to sell" (e.g., their labor . . .). The model tracks the flow of dollars . . . You will get the same Boltzmann distribution if you just let the agents engage in a "dollars random walk" with a reflecting barrier. Also, allowing "debt" doesn't change the distribution as long as there is a debt ceiling of some sort (i.e., as long as debt is bounded). tom On Apr 12, 2009, at 11:16 AM, Stephen Guerin wrote: > Russ writes: >> there are N agents; N resources are needed for survival. So if the >> aggregate production is evenly distributed everyone has a >> subsistence living. > > I don't think this is necessarily true. > > Your mentioning a gift economy is further a reason to look at a > model as simple as the Tom Carter/Victor Yakovenko model I mentioned > where agents are giving a dollar to another random agent (ie > gifting). This model, with minimal assumptions, generates non- > equitable distribution of wealth where very few agents end up > holding most of the wealth (Boltzmann distribution) even though the > average wealth in the model is unchanged at $50/agent. > > The important constraint in the model is that agents can only have > positive amounts of cash/goods -- they can't gift a dollar they > don't have. > > Just as energy distributes itself as a boltzmann distribution as > particles can not have less than zero energy after a collision > (transaction). Basically we have diffusion against a wall in both > models. Of course we can introduce debt and progressive taxes to the > economic model to mess with the dynamics, but I think it's important > to recognize the non-equity is there from the beginning with very > minimal assumptions. Note that the wealthy agents do turnover in > this model so the definition of what is equitable is important. > > -Steve > > > > On Apr 12, 2009, at 11:40 AM, Russ Abbott wrote: > >> Thanks for all the responses. What prompted my question was >> thinking about the "end state" of the current financial crisis. >> The question is: what does a successful stable economy look like? >> Only secondarily was I asking about markets, wealth distribution >> across society, etc., which is what many of the existing models >> focus on. >> >> In attempting to describe the foundations of a stable economy I was >> making the following assumptions. >> • We live off the gift of free energy -- the sun. So in some sense >> a stable economy requires little more than harvesting that energy. >> • As human beings we have the innate (i.e., also free) ability to >> transform that energy into other forms that we need. For example, >> most of us on this list are capable of writing code. I'm taking >> skills like that as more free gifts from nature. Yes, it requires, >> education, study, practice, etc., but ultimately it grows out of >> our being alive, a free gift. >> • In talking about "free gifts" from nature, I'm not trying to be >> spiritual, new age-y, etc. It just seems to me to be the way the >> world works. So I'm willing to build it into the model. >> • But, even though we live off these free gifts, we all need >> energy to survive. If you look at my "Reductionist blind spot" >> paper, you see that we are what I call dynamic entities. Dynamic >> entities need to import externally supplied energy to persist -- in >> our case food, at least. So even though we find ourselves in a >> world in which there is externally supplied energy available to be >> imported (from the sun and its stored versions), we are each still >> required to do something to import it for ourselves. That is, we >> must each work for a living. >> • Presumably a standard economic model says that we live at a >> subsistence level if the amount of work we do matches our needs. If >> I spend all my working hours doing nothing more than earning what I >> need to survive, that's subsistence. >> • I want to allow specialization so I suggested that each agent >> specializes in producing some resource but that the aggregate >> amount produced exactly satisfies the needs of the producers. That >> is, each agent produces N units of his specialty resource; there >> are N agents; N resources are needed for survival. So if the >> aggregate production is evenly distributed everyone has a >> subsistence living. >> • A society/economy grows "richer" when it can produce enough >> resources for itself without requiring everyone to work on >> producing those resources. In that case, the available excess labor >> can produce discretionary items, like art, pop music, software, >> soft drinks, philosophy, etc. >> • As technology advances a smaller and smaller percentage of the >> population is needed to produce the necessities. The rest can spend >> their time producing discretionary items. >> • So at this point we are all well fed, and we are all willing to >> spend our extra money supporting our local NPR stations. >> • But what happens when there is shock to the economy and people >> find that they have less "wealth" than they thought they had? >> Presumably they will spend less on discretionary items, and the >> people who create those items will be out of work. Where do we go >> from there? >> So I am looking for a model to explore these kinds of issues. >> >> I must say that I'm surprised at the number of things I had to >> write down to explain what I am after. I thought it would be >> simpler. And I want something as simple as possible. >> >> Any thoughts/suggestions? >> >> -- Russ >> >> On Sun, Apr 12, 2009 at 9:14 AM, Victoria Hughes <[hidden email] >> > wrote: >> > Attention and action = the price to consume. >> >> is it possible to have an idea / new information and not in some >> way, internally or externally, react to it? >> Memes spread. Humans by design infect each other on all levels. >> The Complex seems like a great example: by definition intends to do >> this. >> Idea / attention / action-reaction / experience / new idea. >> And so forth. >> Am enjoying the bouncing of ideas and information in this >> particular string... >> Tory >> >> On Apr 12, 2009, at 7:04 AM, Jochen Fromm wrote: >> >>> The central property which emerges in markets is >>> the price. The central law which rules markets is >>> the law of supply and demand. A basic agent-based >>> model for markets should explain how both, the >>> price and the law of supply and demand, emerge >>> in competitive markets. It should be simple to >>> extend such a model to explain bubbles and crashes. >>> As far as I know, there is no basic model of >>> markets in general, there are only agent based models >>> of specific markets, for example stock markets, >>> financial markets and electricity markets. Perhaps >>> the law of supply and demand is too simple to justify >>> such a model? >>> >>> Epstein and Axtell's Sugarscape is a model where >>> agents begin to accumulate "wealth" in some form, >>> but it is more about evolution of societies and less >>> about market mechanisms and economies. >>> >>> It would be interesting to apply a basic agent-based >>> market model to the society of mind. If we consider the >>> crowd within and the battle of ideas which is going on >>> inside our minds, can we create an agent-based model >>> to explain the 'market of information' ? We have a >>> demand for new or highly emotional stuff, are constantly >>> supplied with information from the outside, and the price >>> we have to pay to consume it is attention. >>> >>> -J. >>> >>> ----- Original Message ----- From: Russ Abbott >>> To: The Friday Morning Applied Complexity Coffee Group >>> Sent: Sunday, April 12, 2009 9:00 AM >>> Subject: [FRIAM] Agent-based market models >>> >>> Does anyone know of good examples of generic agent-based market >>> economies? I'm thinking of something as simple as this. >>> >>> A population consists of agents each of which has certain >>> continuing needs (such as food, clothing, shelter, Internet >>> access, etc.) to survive. As a starting point, let's assume that >>> each agent needs one unit of each of N resources every time >>> period. Let's also assume that each agent is specialized and is >>> capable of creating enough of one of the needed resources to >>> satisfy the needs of N agents. (The fact that I used the same N in >>> both places was intentional.) To keep it simple let's assume that >>> these acts of creation occur from scratch, i.e., that the creator >>> doesn't need raw materials, that all that's necessary for an agent >>> to create a needed resource is that the agent be alive. The agents >>> presumably develop a barter economy, trading the resources they >>> create for the resources they need to stay alive. Perhaps markets >>> develop, and perhaps money develops. At this point the economy >>> should be fairly stable. Each agent creates enough stuff so that s/ >>> he can trade it for what s/he needs to stay live. >>> >>> Perhaps some of the agents learn how to be more efficient in >>> creating their resource and begin to accumulate "wealth" in some >>> form. Perhaps the agents have discretionary desires, which they >>> fill if they have enough resources left over after meeting their >>> basic needs. Perhaps there are communal services that are paid for >>> by taxes or memberships. This could become increasingly elaborate. >>> >>> It seems to me that models of this sort must have been developed >>> -- perhaps many times. Does anyone know of any references to this >>> sort of work? >>> >>> Thanks. >>> >>> -- Russ Abbott >>> _____________________________________________ >>> Professor, Computer Science >>> California State University, Los Angeles >>> o Check out my blog at http://bluecatblog.wordpress.com/ >>> >>> >>> ============================================================ >>> FRIAM Applied Complexity Group listserv >>> Meets Fridays 9a-11:30 at cafe at St. John's College >>> lectures, archives, unsubscribe, maps at http://www.friam.org >>> >> >> >> ============================================================ >> FRIAM Applied Complexity Group listserv >> Meets Fridays 9a-11:30 at cafe at St. John's College >> lectures, archives, unsubscribe, maps at http://www.friam.org >> >> ============================================================ >> FRIAM Applied Complexity Group listserv >> Meets Fridays 9a-11:30 at cafe at St. John's College >> lectures, archives, unsubscribe, maps at http://www.friam.org > > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org > ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models (p.s.)
|
|
In reply to this post by Stephen Guerin
All -
Apropos some of this, there are notes from lectures I've given recently on some of these topics here: Math Club lecture (including some discussion of the "wealth model," and some "power laws" discussion) tom On Apr 12, 2009, at 11:16 AM, Stephen Guerin wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models (p.s.)
|
|
All -
The real "random walks" link: tom On Apr 12, 2009, at 3:58 PM, Tom Carter wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models (p.s.)
|
|
Hey Tom,
Very nice work. Do you have anything that would help build intuition about how it all works? Also, what I'm after is an understanding of how an economy works. Your model says that an economy that satisfies certain conditions will (inevitably) wind up with a certain wealth distribution. But it doesn't tell me what's happening in the economy -- only that some people will become rich and a lot more will become poor. I'd like to develop a feeling for what's actually going on in the economy. For example, in my original model in which (a) each agent needs to consume one unit of each of N different resources at each time step (b) there are N agents each of which produces N units of a particular resource i (different from each agent) and (c) agents trade (barter style) units of resource one for one as needed, it seems to me that there will be no wealth accumulation. Everything that is produced is consumed, and as long as agents are able to trade with each other on a one-for-one basis, every agent survives but no agent accumulates any excess. So that doesn't quite fit the model. Furthermore, what happens if some agents (as a result of technological advances) are able to produce N units of a resource in less than one time step? Then those agents will have time on their hands and can produce something other than the resource that they contribute to the community. So they start producing resources that no one needs -- but that perhaps some agents want. The agents with time on their hands then exchange these discretionary resources among each other. Again this might be modeled as a barter system -- although some convenient resource could come to serve as a unit of measure and medium of exchange. If you take this to the limit so that every agent is producing nothing but discretionary resources (i.e., the "needs" are so cheap to produce that there is no point keeping track of them), then what? In this model, every agent creates N units of value every time step. So there will never be any agents with no resources. In addition, since every transaction involves the exchange of units of resource of the same value, it seems like there is no room for accumulation of wealth. (This is assuming that one acquires a resource in order to consume, i.e., destroy, it.) So when two agents exchange resources, they are both left poorer since they each consume the received resource. In this model an agent grows wealthy only by not exchanging resources and keeping all the resources it produces. It then has the ability eventually to go on a shopping spree, pig out, and consume lots and lots of resources at one time. But what else is that wealth good for in this economy? So all this doesn't quite seem to fit the model you are describing. -- Russ On Sun, Apr 12, 2009 at 4:31 PM, Tom Carter <[hidden email]> wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models (p.s.)
|
|
Russ -
Thanks . . . As far as thinking about how economies "really" work . . . what follows is not really a response to your questions, but a source for thinking about transitions from subsistence/barter to a "monetized" system . . . A clever scheme: 1.) Find a region where there is a subsistence/barter "economy" (with, typically, tribal/communal shared land). 2.) With "high minded" fanfare, institute "land reform." That is, divide the land up into small plots, and assign "ownership" of the plots to individuals. 3.) Now, the clever part: establish a "land tax," but make sure that the "land tax" must be paid in currency. Note that an "income tax" won't work, because then people could remain in their subsistence/barter mode, avoid having any accountable "income," and thus avoid the "taxation" trap . . . 4.) Before long, everyone will have to either switch to "cash crops," or engage in "wage labor" (or, more likely, both) in order to get currency to pay the "land tax." 5.) You will then have a "monetized" economy, together with a ready pool of "wage laborers," and "consumers" who purchase goods, since they have moved to "cash crops," away from subsistence farming. A quick Google search yields this essay (the Google search was: colonial land taxation cash crop): A quote from the essay: "The history of direct taxation in colonial capitalism also has some wider theoretical implications. It shows, for example, “that ‘monetization’ did not spring forth from barter; nor did it require ‘trust’—as most stories about the origins of money claim” 15 (Wray, 1998, p. 61). In the colonial capitalist context, money was clearly a “creature of the state” " Anyway . . . tom On Apr 12, 2009, at 5:07 PM, Russ Abbott wrote: Hey Tom, ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
Re: Agent-based market models (p.s.)
|
|
Nice point about money. I believe that even barter exchanges are construed as income. So there is no escape from the taxation trap except, perhaps, complete self-sufficiency.
-- Russ
On Sun, Apr 12, 2009 at 8:48 PM, Tom Carter <[hidden email]> wrote:
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org |
«
Return to Friam
|
1 view|%1 views
| Free forum by Nabble | Edit this page |