Francesco Caprioglio

Computer Science and Simulation for Economics

Project work on

"Lemons and fine cars"


The applet requires Java 1.5 or higher. It will not run on Windows 95 or Mac OS 8 or 9. You may obtain the latest Java plugin from Sun's Java site.

created with NetLogo

view/download model file: lemons2.nlogo


This model wants to simulate the economic effects of asymmetric information on market, that were analysed by G. A. Akerlof (1970) in the paper: "The market for lemons: quality uncertainty and the market mechanism". If buyers haven't informations on goods they will buy much more lemons.
The aims of model are to see the effects of reputation and bankruptcy. Can reputation cancel asymmetric information consequences or not? What are the effects of bankruptcy on market transactions?
On the market there are two kinds of agents: "turtles" (buyers) that want to buy cars and "patches" (sellers) that sell good or bad cars. Through two graphs the simulation can provide economic implications, and answer to the above questions. The first gives information about number, average price and average quality of cars; the second compares number of lemons and fines cars buyed.


There are two kinds of patches (sellers): if the patch is active set his pcolor green and set sell price as random 1000, ifelse set his pcolor black. Turtles (buyers) interact only with green patches, in this way it's possible to modify the number of sellers on the market. Half of green patches sell fine cars and half sell lemons, on the market there are the same quantity of car kinds. If the buyer price is higher seller price the transaction is made, ifelse turtle find another seller.

If there is information turtles see the quality of cars and offer to sellers a price higher (random 900) for fine cars and lower (random 400) for lemons.
If there isn't information buyers don't see the quality of cars and offer the same price (random 500) to all sellers, they will buy much more lemons.
If there isn't information but there is reputation, turtles offer to seller a random price (random 500) plus a random quantity multiply for reputation of seller (random 900 * reputation of patches-here). In this way turtles offer an higher price to seller that have a good reputation.

When there is the possibility of bankruptcy the variable "Tfailure" of single patch is modify at every time. If "Tfailure" became >= number of maximum failure transactions (fixed with switch "Bankruptcy") the patch became black and go out from market.


The "setup" button create Turtles (Buyers) and Patches (Sellers) on the market.
The "search" button make interact the agents that try to buy/sell cars. There are two graphics that represent some important elements of market (average price; quality; etc...).

User can modify:

number of buyers through slider: "Buyers"
number of active patches (sellers) through slider: "%Sellers"
number of patch sale after that buyers can observe seller reputation through slider: "Delay"
number of failure transactions after that seller make bankruptcy through slider: "Bankruptcy".

User can add/delete on market:

Informations to buyers through switch "Information"
---- If switch is on buyers know the value of goods, they offer more cash to patches
---- that sell fine cars, Ifelse they offer the same random price to all sellers.
Sellers reputation through switch "Reputation"
---- After fixed number of single seller sales, buyers can see judgment of others buyers, that create seller reputation.
---- They add to price a random premium multiply for seller reputation.
Bankruptcy of sellers through switch "Bankruptcy"
---- After fixed number of failure transactions sellers become black and go out by
---- market. Only green patches sell cars in the model.


It's interesting modify market conditions while agents are interacting, this action allow to compare two situation in the same graphics. For example if in market without information we add reputation time series levels grow up. This action can be repeat from market with reputation to market with information, to see the effects. To know all economic implication of model is important to pay attention in the graphic with quantity of fine and lemon car sold.


The value of variable called "Delay" can be interpreted as time to create seller reputation in market. With high delay number time series levels will slowly grow up.
Through the variable "Bankruptcy" we can modify the probability that sellers make bankruptcy and the speed with they go out by market.
It's possible to change sellers and buyers number but don't provide strong economic implications.


To improve the model there are two ways:

Change the pcolor of bad sellers to see what kind of sellers make bankruptcy earlier. Reputation can reduce the number of good sellers bankruptcies or not?

Introduce the possibility to modify buyers trust on sellers reputation to see the effects on the market. What is the link with sellers reputation and buyers trust?


On the web it's possible to see others NetLogo models about
"The market for lemons". The models sites are:

Model name: "Asymmetric_information_in_the_market_for_lemons"

Model name: "Lemon"


This model was created as a project work for the course of "Computer Science, Simulation and Economics" (2005-2006) by Professor Pietro Terna, Faculty of Economics, University of Turin. The author of the model is F.Caprioglio (2006).