The Data Space

Input data shared by every by-policy projection.

The seven input CSVs are read here, once per model, and referenced from Projection as data. Projection is parameterized by point_id, so each Projection[N] is a separate ItemSpace with its own cells cache; if the readers lived there, every model point would re-read every file. Holding them in an unparameterized Space reads each file once no matter how many policies are projected.

Inputs are external files: plain CSVs in the model folder’s parent directory, products/assurance_emprunteur/, rather than data stored inside the model. The model folder therefore holds nothing but formulas — no _data/, no IOSpec, no embedded values — so a diff of the model shows logic changes only. This follows annuallife.TradLife_A; contrast basiclife.BasicTerm_S, which keeps its inputs inside the model through modelx’s IOSpec machinery.

The consequence worth knowing: the model is not portable on its own. Copying the ADE_FR_S folder without its parent’s CSVs produces a model that reads and then fails on first evaluation.

input_dir() resolves the directory from _model.path.parent at run time, so the model works wherever the repository is checked out. Each table has a filename Reference and a reader Cells:

Reference	Cells	File
model_point_file	model_point_table()	model_point_table.csv
mort_table_file	mort_table()	mort_table.csv
itt_inception_file	itt_inception_table()	itt_inception_table.csv
itt_termination_file	itt_termination_table()	itt_termination_table.csv
franchise_file	franchise_table()	franchise_table.csv
lapse_table_file	lapse_table()	lapse_table.csv
crd_rate_file	crd_rate_table()	crd_rate_table.csv

There is no loan schedule file, and that is deliberate. The échéancier is deterministic given the capital, the rate and the term, so Projection.crd computes it and Projection.check_crd asserts it closes. A pasted amortisation table is the notes’ first-listed pitfall: it cannot be checked against the roll-forward, and the whole product hangs off the capital restant dû.

The decrement files are separate because they are separate objects with different parameterizations, exactly as a real basis would publish them. Mortality and ITT inception are keyed by sex and pivot age and interpolated; ITT terminations are keyed by claim duration year and split three ways — recovery, transition to IPT, death in claim — because ITT has three competing exits and a real disability basis publishes them apart. The franchise factor is its own one-column file because it is a property of the deferred period rather than of the life, and the CRD premium scale is its own file because it is a tariff and not a decrement.

Every shipped rate table carries a provenance column saying in words that it is a [std] proxy. No French decrement, incidence or termination table for this product was retrieved: insurer rate cards are proprietary, and TH 00-02 / TF 00-02 are cited by name but are not redistributable. A licensee replaces the files and changes no formula.

Cells Descriptions

input_dir()

The directory holding the input CSVs: the model folder’s parent.

Inputs are external files, not data stored inside the model, so the model folder is pure formulas. The path is resolved at run time from where the model was read, following annuallife.TradLife_A.

model_point_table()

The model point table, read from model_point_table.csv.

Twelve single-life cells. Point 1 is the notes’ worked example configuration; the others carry the two premium bases, the two indemnity bases, the two IPT benefit bases, a quotité below 1, the franchise menu, a cell whose ITT/IPT cover ends long before its loan, and one claim in payment in each of the two disabled states.

mort_table()

Healthy-life annual mortality by sex and pivot age, from mort_table.csv.

A [std] proxy. The homologated French tables for a non-annuity contract are TH 00-02 / TF 00-02 with the annexed décalage d’âge, cited by name in the notes but not redistributable, so the shipped values are shaped from INSEE population data — the only freely redistributable French series. Population mortality is heavier than medically-selected insured experience, so the proxy overstates death cost for a standard-risk book and understates it for a book written without medical selection. The female rows are 0.60 x male, a pick; one sampled insurer is unisex instead.

itt_inception_table()

ITT claim-payment inception rates, read from itt_inception_table.csv.

Annual rates per life in healthy, by sex and pivot age. Shaped as a claim payment inception rate — the quantity a real disability basis publishes per deferred period — so a sickness spell that recovers inside the franchise never leaves healthy and no separate “sick, not yet in payment” state is needed. The shipped rates are the franchise 90 days column; franchise_table() scales them to the other four. Every value is [std].

itt_termination_table()

The three ITT exit rates by claim duration year, from itt_termination_table.csv.

Annual recovery, transition-to-IPT and death-in-claim rates. Three columns rather than one because ITT has three competing exits and they move in opposite directions with duration: recovery falls 0.55 to 0.15 across the three duration years while the IPT transition rises 0.02 to 0.12. Short claims mostly recover and long claims mostly consolidate, which is the qualitative structure of any disability termination basis and the reason the in-claim population needs a duration dimension at all. The values are [std] proxies with no French anchor.

franchise_table()

The franchise multipliers on the inception rate, from franchise_table.csv.

Keyed by franchise_days over the sourced 30 / 60 / 90 / 120 / 180 menu. A longer deferred period admits fewer spells to payment, so the factor falls 1.60 to 0.65 with 1.00 at the 90-day pick the inception table is written on. The menu is sourced; the factors are [std].

lapse_table()

Résiliation rates by policy year, read from lapse_table.csv.

The loi Lemoine substitution decrement, and the behavioural heart of the product: 4 % in year 1, 12 % in years 2 and 3, 10 %, then a 7 % ultimate. Materially higher than a classic protection lapse because the cover does not stop, it moves — the borrower may cancel à tout moment and the insurer must remind them annually. The shape is a reading of the statutory mechanics rather than of data: the published French series are counts of substitution requests, not portfolio lapse rates, so the whole table is [std].

crd_rate_table()

The CRD-basis premium scale by attained age, from crd_rate_table.csv.

Annual rates applied to the capital restant dû at each policy anniversary, used only when premium_basis = capital_restant_du. A tariff, not a decrement, and [std]: it is calibrated so its present value over the worked-example cell matches the level 0.84 % capital initial scale to about 0.11 %.