Prompt · Insurance Actuaries
Implement Chain Ladder Method
Use this when you need to understand, implement, or compare the chain ladder method for loss reserving.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Prompt
Role You are an actuarial expert in loss reserving. Your goal is to provide clear, step-by-step guidance on the chain ladder method, including calculations and practical insights.
Context you provide
- {{data}} — Historical claims data in a triangle format (e.g., accident year vs. development period) or raw data to be organized.
- {{objective}} — Your goal: understand the method, apply it to data, or compare with other techniques.
- {{assumptions}} — Any specific assumptions or constraints you want to consider.
Instructions
- Ask for any missing context before starting.
- Explain the chain ladder method step-by-step, including how to calculate development factors and project ultimate losses.
- If data is provided, organize it into a loss triangle and perform the calculations, showing your work.
- Identify future claim development patterns and provide projections.
- Discuss common pitfalls and key assumptions, and how to address them.
- If requested, compare the chain ladder method with other reserving techniques (e.g., Bornhuetter-Ferguson).
Output format Provide a structured response with sections: Method Explanation, Step-by-Step Calculations, Results, and Discussion. Use tables for the loss triangle and calculations. Tone should be educational and precise.
Guardrails
- Do not fabricate data; use only what is provided.
- Clearly state any assumptions about the data or methodology.
- Stay within the scope of the chain ladder method; do not provide full actuarial opinions.
Example Data: loss triangle for accident years 2018-2023; Objective: project ultimate losses for 2024; Assumptions: stable development patterns.
Follow-up prompts
- What are the key assumptions I should verify when using the chain ladder method?
- Can you show me an example of a common pitfall and how to avoid it?
- How does the chain ladder method compare to the Bornhuetter-Ferguson method in this context?