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/berufsunfaehigkeit/, 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 BU_DE_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
inception_file	inception_table()	inception_table.csv
claim_duration_file	claim_duration_table()	claim_duration_table.csv
mortality_file	mortality_table()	mortality_table.csv
occupation_file	occupation_table()	occupation_table.csv
lapse_file	lapse_table()	lapse_table.csv
freq_loading_file	freq_loading_table()	freq_loading_table.csv

Every file but model_point_table.csv carries a per-row provenance column, which is delib’s second ruling: the citation discipline reaches the data files and is machine-checked rather than merely practised. The model point table is the single exemption, because a model point is a configuration and not an assumption.

The biometric tables are [std] proxies, and what a replacement must preserve

Nothing in this directory is a DAV table. DAV 1997 I (Invalidisierungswahrschein- lichkeiten), DAV 1997 RI (Reaktivierungswahrscheinlichkeiten), DAV 1997 TI (Sterbewahrscheinlichkeiten der Invaliden) and DAV 2008 T are the property of the Deutsche Aktuarvereinigung, are not published and are not redistributed by delib. They are cited by name in the product’s sources.md and the tables shipped beside this model are anchored [std] constructions that reproduce their shape and nothing more.

Each proxy is anchored so that the technical notes’ worked example reproduces exactly. The three anchors, and what a replacement built from company or DAV data must preserve:

inception_table.csv

inc_rate(30) = 0.001100 exactly, on a two-slope Gompertz form 0.00110 x 1.06^(min(x,45) - 30) x 1.13^(max(x,45) - 45). What must be preserved is the age shape — nearly flat to 30, moderate through the forties, sharply accelerating from the mid-forties, so that the last decade before the Endalter dominates the liability — and the declaration of whether the table is gross or net of declinature. The shipped table is gross, and accept_factor = 0.80 sits on top of it; a table already net of declinature must be used with that factor at 1.00 or the Anerkennungsquote is counted twice.

mortality_table.csv

mort_rate_actv(30) = 0.000350 exactly, with the disabled-lives rate at 4.00x it, both on 1.095^(age - 30). What must be preserved is the excess of disabled over active mortality — never one rate for both states — and the fact that the active table is an insured-lives Todesfall-character shape rather than a population table.

claim_duration_table.csv

recov_rate(1) = 0.250 falling to 0.006 ultimate, with disabled-mortality select factors 3.0 / 2.0 / 1.6 / 1.4 / 1.3 and 1.2 from claim year 6. What must be preserved is the duration shape: reactivation is concentrated in the first one to two claim years and is close to zero after about five, and disabled-lives mortality is itself select on claim duration. A flat reactivation rate is a modelling error, not a simplification, and it is worth roughly a factor of two on projected benefit.

The remaining three files are market conventions rather than biometric bases: occupation_table.csv carries the five-class Berufsgruppen cut with BG1 at 1.00 and BG4 at 3.00 as its anchors, lapse_table.csv a low German BU Stornoquote falling from 4.0 % to a 2.0 % ultimate, and freq_loading_table.csv the Ratenzahlungszuschlag ladder with the months between instalments. All three are [std]: no German insurer publishes any of them.

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.

Thirteen policies indexed by point_id, each carrying the twenty contractual attributes the projection reads. The one input file with no provenance column, and the only exemption from delib’s second ruling: a model point is a configuration — one policy’s own terms — rather than an assumption, so tagging it row by row would repeat the same provenance once per policy while saying nothing about any assumption.

inception_table()

The annual Invalidisierungswahrscheinlichkeit by attained age, ages 18-66.

Read from inception_table.csv. A [std] two-slope Gompertz proxy for the reference occupational class BG1, unisex and gross of declinature, anchored at inc_rate(30) = 0.001100; DAV 1997 I is DAV property and is not shipped. See the Space docstring for what a replacement must preserve.

claim_duration_table()

The claim-duration table by claim year, from claim_duration_table.csv.

Two columns on one index, because both quantities are select on the same duration: recov_rate, the annual Reaktivierungswahrscheinlichkeit, and mort_dis_sel_factor, the multiplier on disabled-lives mortality. Rows 1-10 are claim years 1-10 and row 11 is the ultimate, applied to claim year 11 and beyond. Both columns are [std]; DAV 1997 RI and DAV 1997 TI are not shipped.

mortality_table()

The two annual mortality rates by attained age, ages 18-70.

Read from mortality_table.csv: mort_rate_actv for lives in the aktiv and run-off ledgers and mort_rate_dis for lives in claim, before the claim-duration select factor. Both are [std] Gompertz proxies anchored at mort_rate_actv(30) = 0.000350, with the disabled rate at 4.00x the active one; DAV 2008 T and DAV 1997 TI are not shipped. Using one rate for both states is a numbered pitfall, which is why the file carries two columns rather than one.

occupation_table()

The occupational loading by Berufsgruppe, from occupation_table.csv.

occ_factor multiplies the inception rate and reaches the premium only through the equivalence — it is not the Risikozuschlag, which loads the premium alone. The label column carries the class description so that a model point’s berufsgruppe is readable without the product specification open. [std]: one base table with occupational loadings is how German BU pricing works, but no insurer’s Berufsgruppenverzeichnis was retrievable and carrier classifications are not comparable with one another.

lapse_table()

The annual Stornoquote by policy year, from lapse_table.csv.

Rows 1-5 are policy years 1-5 and row 6 is the ultimate, applied to policy year 6 and beyond. [std] and low by the standards of every other delib product, which is a real product fact rather than a modelling choice: once health has changed the cover cannot be replaced, so an insured with a claimable impairment cannot rationally lapse. Lapse selection is not modelled; the direction of the resulting error is stated in the technical notes.

freq_loading_table()

The payment frequency table, from freq_loading_table.csv.

prem_mode_months is the number of months between instalments — 12, 6, 3, 1 — and freq_load the Ratenzahlungszuschlag on the tariff premium, 1.00 / 1.02 / 1.03 / 1.05. The loading scales the Bruttobeitrag and the Beitragsverrechnung together, so the quoted Zahlbeitrag ratio is untouched by it. [std]: the ladder is the recalled German market convention and no retrieved document confirms it.