Mesa example model idea: Demonstrating the Ratchet Effect

[EwoutH]

Overview

I'd like to propose creating an example model that demonstrates the Ratchet effect - a phenomenon where systems can easily move in one direction but resist moving in the opposite direction. This concept appears in many social and economic contexts and would make for an insightful addition to our examples.

Potential Implementation: Remote Work Policies

A model focusing on remote work policies could effectively demonstrate the ratchet effect:

• Agents: Workers and employers with varying preferences
• Environment: Workplace infrastructure (physical and digital)
• Ratchet Mechanism: Once workers adapt to remote work (new habits, home office setup, residential relocation), returning to previous work arrangements becomes increasingly difficult

Interesting Modeling Aspects

This example could showcase several important Mesa capabilities:

1. Expectation Formation: How agents form and update expectations about work arrangements

2. Multi-level Ratchet Effects:

   • Individual level (habits, preferences)
   • Physical environment (office vs home setup)
   • Social networks (changing connections between colleagues)
   • Organizational policies

3. Path Dependency: How initial conditions and small changes lead to lock-in effects

Connection to Mesa Development

This example could contribute to ongoing discussions about:

• How to model agent states and transitions States discussion
• Approaches for modeling perception and expectation updating
• Demonstrating adaptive behavior in more complex social systems

Alternative Domains

While remote work seems particularly relevant, the ratchet effect could also be demonstrated in:

• Healthcare (adoption of treatments or technologies)
• Housing markets (gentrification processes)
• Transportation/mobility (infrastructure changes)

Next Steps

If there's interest in this example, I'd be happy to discuss specific implementation details further.

[iswagnacio]

I'm applying for Google Summer of Code and can take this. I'm fluent in python, pandas library, and different visualization libraries. May we discuss this further?

[EwoutH]

Cool! Yes, please share any ideas you have :)

[iswagnacio]

These are my ideas on this example:

Key Ratchet Mechanisms

• Asymmetric Adjustment Rates: The adaptation rate when going remote is 5x faster than readjusting back to office work
• Physical Investments: Home office investments and residential relocation create physical lock-in effects
• Social Network Evolution: As workers go remote, their social connections change, creating new networks that reinforce remote work
• Infrastructure Ratchet: Digital infrastructure investments persist even if companies try to return to office-based work
• Preference Evolution: Workers' preferences gradually shift based on their experiences

Path Dependency Demonstration
The model includes an optional "external shock" parameter (simulating something like a pandemic) that can trigger a sudden shift to remote work. By running the model with different timing and magnitude of this shock, we can demonstrate:

• How initial conditions affect long-term outcomes
• The critical threshold of adaptation needed to create a permanent shift
• The organizational resistance to returning to previous states

Data Visualization Features

1. Adaptation Curve Graph: Plot showing the asymmetric adaptation rates over time

2. Multi-factor Ratchet Breakdown: Stacked area chart showing contribution of different factors to the ratchet effect

• Home office investment
• Habit formation
• Location changes
• Social network effects

3. Network Evolution Animation: Dynamic network visualization showing how social connections evolve

4. Policy vs. Reality Gap

• Employer policy changes (% remote-friendly)
• Actual worker location (% working remotely)

[EwoutH]

Your or an LLM’s? ;)

Pick one, think it through, and write it out

[iswagnacio]

Good catch. I apologize for this. My original idea is to create an example of countries acting on tariffs, which is a more relevant recent case. Building on top of your proposal format:

A model focusing on tariff in global trading could effectively demonstrate the ratchet effect:

Agents: Countries and firms with varying preferences (country care more about debt/economic status; firms care more about profitability)
Environment: global trading system (could be specified to one smaller field for simplicity)
Ratchet Mechanism: Once tariffs are imposed by one country, other countries may keep imposing tariffs to counterattack, which leads to further tariff war instead of mutual negotiation.

Interesting Modeling Aspects

This example could showcase several important Mesa capabilities:

Expectation Formation: How agents form and update expectations about the action of other agents

Multi-level Ratchet Effects:
firm level (more conservative to protect from loss)
consumer level (imported goods vs. local goods)
national economy level (represented by debt ? -> needs further discussion)

Connection to Mesa Development

Agent states and transitions
Approaches for modeling perception and expectation updating
Demonstrating adaptive behavior in more complex social systems

Let me know on any possible improvement on this.

[EwoutH]

Oh tariffs could be an interesting (and very relevant) application!

I would say go build something :)