Model03: Schelling’s Segregation Model

Jiin Jung, Ph.D.

Lehigh University

jiin.jung@lehigh.edu

Schelling’s Segregation Model

• One of first agent-based models
• Tom Schelling – Nobel Prize winner for his work on cold war.
• Also called Schelling’s Tipping Model
• Originally published in JMS (1971)
• A book chapter in his Micromotives and Macrobehavior (1978)

​

​

How much prejudice does it take to observe residential segregation?

Running Segregation Model

Agent Properties

• Agents have two attributes (orange, blue)
• Agents represent people in houses�
• Happy agents are square and unhappy agents are cross
• An agent is happy if at least 30% of it neighbors have same color
• An agent is unhappy if neighbors with same color are less than 30%

Interaction Rules

• If an agent is happy, it stays.
• If an agent is not happy, it moves to a random open spot.

​

What does this mean?

• If I am a majority or large minority, I stay. In other words, I will be happy even though 70% of my neighbors are different from me.
• If I am a small minority, I move.

​

Running the model

Simulation Results

• With small biases toward people who are similar to us, we get segregation.
• It doesn’t take large prejudice to observe residential segregation.
• A segregated society (macro behavior) can emerge from individuals’ low bias (micro motive) (Schelling, 1978)

• At that time when he run this model (or thought simulation), he didn’t call it agent-based model.
• It was with pennies and dimes on a checkerboard.

​

Social Simulation Experiments

Initial Density

• Do density influence the process?
• Let’s decrease density

​

Initial Density

• Let’s decrease density
• As density decreases, segregation goes down little bit but not so much.

​

Similarity required to be happy

• Let’s decrease similarity required to be happy.
• With 20%, we don’t get segregation (50% 🡪 55%) as much as 30%
• 30% = tipping point