Non-Life Reserving Workbench

R + Excel toolkit for Belgian-style non-life reserving: Best Estimate (CL / BF) → diagnostics → Mack → one-year risk (Merz–Wüthrich / CDRboot) → closing pack.

Status: Module 1 delivered (Chain Ladder + Bornhuetter–Ferguson → Excel).

---

Quick start (Module 1)

# From the repo root
install.packages(c("openxlsx"))   # once
# optional: install.packages("ChainLadder")

source("run_module1.R")

Or in a terminal:

Rscript run_module1.R

Outputs

File	Description
excel/Reserving_Pack.xlsx	Closing pack (Dashboard, Triangle, LDF, CL, BF, BE_Summary)
outputs/be_summary.csv	Ultimates / IBNR by AY
outputs/ldf.csv	Raw vs selected LDFs

Open the Excel file first — yellow cells = expert judgment inputs.

---

Learning roadmap (products)

Module	Focus	Product
0	Triangle foundations	docs/Glossary.md + data loaders
1 ✓	CL + BF Best Estimate	Reserving_Pack.xlsx
2	Diagnostics	heatmaps, LDF stability, sensitivity
3	Ultimate uncertainty	Mack + bootstrap
4	One-year / SII	Merz–Wüthrich + CDRboot
5	Governance	validation note + committee dashboard
6	Extensions	large claims, multi-LoB, risk margin

---

Repository layout

reserving-workbench/
├── R/                  # calculation engine
├── data/               # triangles (triangle_auto sample)
├── excel/              # Reserving_Pack.xlsx
├── outputs/            # CSV exports
├── docs/               # glossary + later validation note
└── run_module1.R       # one-shot Module 1 runner

---

Excel pack (Module 1)

Sheet	Role
Dashboard	KPIs for committee
Cover	Metadata / version
Triangle	Cumulative claims
LDF	Raw vs selected factors (yellow)
CL	Chain Ladder ultimates / IBNR
BF	Bornhuetter–Ferguson (a priori yellow)
BE_Summary	CL vs BF vs retained method
Totals	Portfolio aggregates

Demo retention rule: BF on the last 2 accident years, CL otherwise — change in BE_Summary / R when you have a real prior (premium × ELR).

---

Related projects

Project	Focus
CDRboot	One-year CDR bootstrap (Module 4)
mtpl-loss-model	Aggregate loss MC
EDA-GLM-RF-XGB	Claim frequency GLM vs ML
Portfolio hub	Overview

---

Author

Philippe le Hardÿ — Actuarial & Quantitative Risk Consultant