pyratemaking 0.1.1

Pure-Python ratemaking for P&C insurance — end to end.

pyratemaking 📊

Pure-Python ratemaking for P&C insurance — end to end.

pyratemaking is the Python implementation of the Werner & Modlin ratemaking workflow: rate level indication, on-leveling, trending, development, classification analysis, large loss procedures, and rate implementation. Built for P&C actuaries who want a reproducible pipeline from raw policies and claims to a defensible rate filing.

Why

Python has GLM libraries. It has chain-ladder libraries. It does not have an end-to-end ratemaking workflow. pyratemaking is that workflow.

Install

pip install pyratemaking            # core
pip install pyratemaking[full]      # with full ecosystem (whsmooth, actudist, etc.)

Quickstart

from pyratemaking import RatePlan
from pyratemaking.datasets import french_motor

policies, claims = french_motor.load()

plan = RatePlan(
    policies=policies,
    claims=claims,
    exposure_col="exposure",
    loss_col="claim_amount",
    ay_col="policy_ay",
)

indication = plan.indicate(method="loss_ratio", target_lr=0.65)
print(indication.summary())

plan.classify(
    rating_vars=["driver_age", "veh_power", "region"],
    backend="glum",
    family="tweedie",
    power=1.5,
)

plan.implement(cap=1.15, floor=0.85)
plan.report.filing("filing_2026.html")

Werner & Modlin coverage

Chapter	Topic	Module
5	On-leveling	pyratemaking.onleveling
6	Loss development	pyratemaking.development
7	Trending	pyratemaking.trending
8	Overall rate level indication	pyratemaking.core.indication
9-10	Classification ratemaking	pyratemaking.core.classification
11	Increased limits / large loss	pyratemaking.large_loss
12	GLMs	pyratemaking.glm
14	Implementation	pyratemaking.core.implementation

Ecosystem

Part of a 6-library actuarial Python suite:

• actudist — severity and frequency distributions
• burncost — burning cost analysis
• actuarcredibility — credibility methods
• whsmooth — Whittaker-Henderson smoothing
• pyratemaking — ratemaking workflow (this library)
• pyinsurancerating — rating engine (coming)

Roadmap

Future releases will add Bayesian hierarchical credibility, deep-learning rating models (LocalGLMnet, CANN), premium calculation principles, fairness auditing, a CLI tool, and a Streamlit exploration dashboard. See milestones.

References

• Werner, G. & Modlin, C. (2016). Basic Ratemaking (5th ed.). Casualty Actuarial Society.
• Goldburd, M., Khare, A., Tevet, D. & Guller, D. (2020). Generalized Linear Models for Insurance Rating (2nd ed.). CAS Monograph 5.
• Friedland, J. (2013). Fundamentals of General Insurance Actuarial Analysis. Society of Actuaries.
• Mack, T. (1993). "Distribution-free Calculation of the Standard Error of Chain Ladder Reserve Estimates." ASTIN Bulletin, 23(2), 213-225.

License

MIT © Isaac López