The Data Space

Input data shared by every by-policy projection.

The nine 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/pflegerentenversicherung/, 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 Pflege_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
benefit_scale_file	benefit_scale_table()	benefit_scale_table.csv
mort_table_file	mort_table()	mort_table.csv
incidence_file	incidence_table()	incidence_table.csv
care_file	care_table()	care_table.csv
lapse_file	lapse_table()	lapse_table.csv
surrender_file	surrender_table()	surrender_table.csv
expense_file	expense_table()	expense_table.csv
basis_file	basis_table()	basis_table.csv

Every file but model_point_table.csv carries a final provenance column, one tag per row — the library’s second ruling, asserted by tests/test_model_conventions_de.py. A model point is a configuration rather than an assumption, which is why it is the single exemption.

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

DAV 2008 P, the German market’s standard multi-state *Pflegetafel*, is the property of the Deutsche Aktuarvereinigung, is not public, and is not redistributed by this library. Neither is DAV 2008 T, DAV 2004 R or any other DAV table. They are cited by name in sources.md and nothing here is derived from them. What ships instead is a set of shaped proxies, and the anchors below are what a substitute must reproduce for the technical notes’ worked example to close to the cent:

• mort_table.csv — active-life mortality, a Gompertz proxy 1 - exp(-B c**age) by sex over ages 18 to 108, with mort_rate = 1.0 at age 109. Anchored at q(65) = 1.35 % and q(85) = 10.5 % male, q(65) = 0.75 % and q(85) = 7.0 % female, which fix B_M = 1.47884e-05, c_M = 1.110680, B_F = 4.76290e-06, c_F = 1.119962. A replacement must preserve those four anchor rates.

• incidence_table.csv — the rate at which an active life enters any Pflegegrad, min(I0 exp(g (age - 65)), 0.50) with I0_F = 0.0110, g_F = 0.1400, I0_M = 0.0085, g_M = 0.1380. The slope is anchored on the one shape the research file states with confidence — prevalence roughly doubling every five years of age above 75, a growth rate of ln 2 / 5 = 0.1386.

• care_table.csv — the whole in-care basis in five rows. mort_mult is a multiple on the force of active mortality at the same age, not on the rate.

A replacement table must preserve four properties: (a) incidence by attained age, sex and grade of entry, because a stroke or a fracture enters directly at grade 3 or 4; (b) deterioration dominating recovery above age 75; (c) mortality in care as a grade-increasing multiple of active mortality; and (d) transition probabilities out of each state summing, with the stay probability, to one. Dropping a licensed or company table in place of these changes the basis with no formula change.

The remaining tables are standardizations of a different kind: lapse_table.csv and surrender_table.csv encode a shape argued from the Zillmerung and the § 169 VVG five-year spread rather than from data — no lapse rate and no Rückkaufswert for a German Pflegerente at any duration was established — and expense_table.csv carries five placeholder levels whose only cited element is the 25 ‰ ceiling the acquisition charge is set exactly at. basis_table.csv holds the Rechnungszins, the terminal age, the unisex mix, the recovery-damping shape, the Beitragssumme convention, the check tolerance and the five first-order prudence margins; only the Rechnungszins is cited.

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 fourteen model points, read from model_point_table.csv.

The only input file without a provenance column: a model point is a configuration — one policy’s own terms — rather than an assumption, and that is the single exemption from the library’s second ruling.

benefit_scale_table()

The Leistungsstaffel by schedule and Pflegegrad, from benefit_scale_table.csv.

Two schedules ship. delib_std is 0 / 30 / 50 / 75 / 100 % — the flatter, higher shape a Pflegerente aimed at the residential funding gap tends to use, and a [std] choice within an observed range. bahr is the statutory 10 / 20 / 30 / 40 / 100 % minimum grid of § 127 SGB XI, the only Leistungsstaffel fixed by German statute; it is carried for comparison, since a Pflegerente written by a Lebensversicherer cannot be a geförderter Tarif.

mort_table()

Annual active-life mortality by sex and attained age, from mort_table.csv.

A [std] Gompertz proxy over ages 18 to 108 with the limiting-age rate of 1.0 at age 109, not a fitted or standard table; see the Space docstring for the four anchor rates a replacement must preserve. In-care mortality is not tabulated: it is this rate’s force multiplied by mort_mult from care_table.csv.

incidence_table()

Annual incidence into any Pflegegrad by sex and age, from incidence_table.csv.

A [std] exponential proxy capped at inc_cap. It is the rate of leaving the active state for care; the grade actually entered is drawn from entry_share in care_table.csv, because entry is not uniformly at the lowest grade.

care_table()

The whole in-care basis in five rows, from care_table.csv.

entry_share is the distribution of the grade first entered and sums to 1.00; det_rate the annual deterioration rate to the next grade up, zero at grade 5; rec_rate the annual recovery rate to the grade below — to the active state from grade 1 — before the age damping; and mort_mult the multiple on the force of active mortality at the same age.

entry_share is deliberately not the stock distribution of about 9 / 44 / 27 / 14 / 6 %: entrants skew lower than the stock because deterioration moves people up over a spell, and using the stock as the entry mix is a listed pitfall.

lapse_table()

Annual lapse rates from the active state by policy year, from lapse_table.csv.

Policy years 1 to 40; year 40’s rate applies to every later year. A [std] shape: no lapse rate for a German Pflegerente at any duration was established, and the profile is argued from the Zillmerung — the Rückkaufswert is near zero for the first years, so an early lapse is expensive to the policyholder.

surrender_table()

The guaranteed Rückkaufswert as a fraction of premiums paid to date.

Read from surrender_table.csv, by Versicherungsjahr y(t) = t // 12 + 1, 1 to 40: year k’s ratio applies throughout policy year k, t = 12(k - 1) … 12k - 1, and year 40’s applies thereafter, intermediate years interpolated in the shipped file. It is the current policy year, not the completed one — a surrender at t = 24, two completed years, already takes year 3’s ratio. The scale-free form, which is the form a German contract states. The shape encodes two cited facts — the 25 ‰ Zillmerung allowance, which is why the first two years are zero, and the § 169 Abs. 3 VVG five-year spread floor, which is why it turns positive in year three — and no cited level, because none was established.

expense_table()

The five expense assumptions and their units, from expense_table.csv.

acq_permille (‰ of Beitragssumme, charged once at t = 0), admin_prem_pct (fraction of each Beitrag collected), admin_mth_pp (EUR per policy in force per month at t = 0 prices), claim_expense_pp (EUR per annuity payment made) and expense_infl (annual). All five levels are [std]; only the 25 ‰ ceiling the first is set exactly at is cited.

basis_table()

The scalar bases and conventions, from basis_table.csv.

The Rechnungszins; the terminal age omega_age; the unisex pricing mix; the rec_age_ref / rec_age_decay recovery damping; the inc_cap shape device; the beitragssumme_cap_age convention the Zillmerung base is struck on; the roll_fwd_tol every check_* identity uses; and the five first-order prudence margins inc_margin, det_margin, rec_margin, care_mort_margin and act_mort_margin. Only the Rechnungszins is cited; everything else is [std].