Minsim Meeting Agendas and Recaps

09/29/08

For next week:

Look up information on Aspen project/failure(?) (starting point: Aspen)
Look at Chris's software and start trying to see if real-life economic problems are reproduced in the simulation
Examples:
- Tariffs, international trade
- Specialization of labor, Adam Smith
- Maximizing revenue of entrepreneurs vs total country revenue
- Inflation, Stagflation, Phillips curve
- Compare the stability of a central banking system with that of the gold standard
- Evaluate the "bailout"
Post summaries of articles up on the wiki
- Sign up for the wiki
- Edit the section "Papers on economic Agent-Based simulation"

Extension:

Went over code
Why don't prices stabilize on its own?
Broad objectives?
Figure out where supply/demand is in the code
use this to see if the simulation satisfies the properties of supply/demand
Play with parameters- More stability with more workers, etc? (keep in mind that this slows down the program)
Brainstorm things we can look at in the future
Interest rates?
Meet separately Wednesday (4:30, Friend library) to look over code
Start looking into old intro micro/macro books

Tasks

Task

eric- lifecycles of individual
tony- info flow, communication, mkt for information
stephanie- financial intruments? taxes, 401ks?
- Current banking system, politician's plans- will it work? Moratorium on mortgages. Free market vs. regulation of government
- Controversy b/w deregulation and regulation, party platforms
CVS setup

Sheri wants to look at Trend following, arbitrageurs
Tony wants to look at media- flow of information in economy where default state is that they only know local information
To simulate: make a theory, get mathematical results, then simulate to verify/revoke results
Market failure - used to justify infusion of capital, govt regulation (build model for credit default swaps - miscalculation of risk- because of lack of government regulation?)
IIASA: publications that conduct interdisciplinary studies on economy and other global forces: IIASA

Eric's professor recommends that he understand the original model in the simulation and go through literature to understand savings and lifetimes.
Tony - wants to start fundamentally on convergent prices and how information flows with pricing mechanisms.
- Zero-intelligence Agents- Economic science association conference
- JEBO- paper by Mizuta, giving agents access to order book with offers to buy and sell-- enormously stabilizing
Tony's advisor wants him to first do analytics, then do simulation
Stephanie- is going to look into interbank lending and the credit crunch, then if possible, look into CDS contracts
- Prof Steiglitz has a technical description of credit crunch by a WWS professor affililated with the Econ dept

Tony should look into auctions
Stephanie will write up API documentation on the banking code.
Framework interface is set up to define arbitrageurs like hedge fund. The entrepreneurs are arbitrage firms.
Presentation by Prof. Vichnevetsky: "World" Modeling and Simulation
Bib-tech: allows you to sort citations

Paper summaries: To migrate to the main website at minsim.cs.princeton.edu. Will create a bibliographic list, with links to popups with summaries.
Tutorial
- Should be assessable to everyone
- Getting started page
- Step by step instructions
- Eclipse cheat-sheet
- Applications to COS444 and other classes
Tony: locality problems to be looked into
Machine learning responses to bidding strategies
- Ausubel + Cramton paper

Stephanie: Interlending of banks
- Need to finish resolving the conflicts with Repast
Questions to answer
- If you make some predictions about policy, what do competing theories predict?
- After interbank lending model, can look at government effects
Further discussion: what you can include in coursework
Dan: base abstract class is done for technical trader
- Need to fill in the functions for each type
- Cannot short sell in this program
- Possible extension: Momentum trading
- plug into economy
Sheri: circuit breaking
- Bernanke's bubble laboratory(about price bubbles)
- Brunnemeier also talked about this
Tony: bilateral trades
- 2 agents - to trade, need to satisfy specific conditions
- Myerson and Sattertwaite (the homepage for COS444 has these literature references)

Dan's project
- Money supply increasing out of control that allows the runs to be stable
- This needs to be investigated before adding to the model
- 16% of runs were stable
- Where does money come from? Taxes, deficit spending (central bank)
- Why doesn't this lead to inflation in food prices?
- Bankrupt firms should die or have a new learning algorithm
Eric's progress
- Working in the cluster to run simulations
- Is there something wrong with the Q-learning algorithm?
- Overall, looking at stability
Tony's 323 project
- Models of how agents exchange products- Who trades with whom?
- Geographic proximity, an internal list, etc.
- a lot of potential for different configurations
- No central market clearing mechanism
- Distributed, pairwise trading- meeting in the marketplace
Book - Henry Hazlitt's "Economics in One Lesson"
- Look at every effect of each action
- every action affects other agents
Stephanie
- Deal with bankruptcy of bank
- Make "base" model: complete base model = free market (labor, commodity, interest, money cost)
- allow interest rate to be determined by free market, THEN introduce policy
Important points to investigate
- Bankruptcy - learning
- Double-auction market
- Short selling

Back to MinSim.