The Data Space

Input data shared by every by-policy projection.

The five 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/term_life/, 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 Term_KR_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
rate_class_file	rate_class_table()	rate_class_table.csv
prem_rate_file	prem_rate_table()	prem_rate_table.csv
lapse_table_file	lapse_table()	lapse_table.csv

prem_anchor_table() is derived from prem_rate_table() rather than read from a sixth file, so it costs no extra read.

The mortality table is a construction, and there was no alternative

The industry table is the 경험생명표 (gyeongheom saengmyeongpyo, experience life table), prepared by 보험개발원 every five years from life-insurance policyholder statistics; the current edition is the 제10회, applied to new business from April 2024. It is not published. What 보험개발원 releases is the summary — 평균수명 남 86.3 / 여 90.7 and 65세 기대여명 남 23.7 / 여 27.1 — and not the rates [REG-R33] [REG-R34]; and even those four numbers reach this library through a trade newspaper, the 보험개발원 announcement itself not being retrievable, so the tilt target of step 2 below is second-hand. The 참조순보험요율 behind each carrier’s own basis is not public either [REG-R4] [R19] [R20]. This is the sharpest contrast in this repository with jplib, whose 標準生命表2018 is a free public PDF with qx by single year of age.

What is public is (i) the carriers’ 예정 경험사망률 disclosures, which the 상품요약서 must print at ages 20, 40 and 60, and (ii) the 국가데이터처 완전생명표 summary [REG-R38]. mort_table.csv is built from both and is [std] throughout:

1. A Makeham law q(x) = A + B c^x is fitted exactly to the anchor carrier’s three disclosed rates per sex — male 0.000280 / 0.000650 / 0.003390 and female 0.000200 / 0.000430 / 0.001390 at ages 20 / 40 / 60 [S12]. Three anchors, three parameters, so the fit is an interpolation and not a regression, and the shipped rows at those three ages are the disclosed rates to the digit.

2. Above age 60 the fitted law is tilted by k^(x - 60), one free parameter per sex, solved so that the complete expectation of life at 65 on the shipped table is exactly the published 경험생명표 figure — 23.7 years male and 27.1 female [REG-R33]. The solved tilts are k = 1.00989 (male) and k = 1.05928 (female), so the correction is small and upward: an unconstrained Makeham extrapolation of the three disclosed rates leaves slightly too much life at 65.

3. The resulting table sits 4.2 years (male) and 3.4 years (female) above the public 완전생명표’s own 65세 기대여명 of 19.5 and 23.7 [REG-R38]. That gap is underwriting selection, and a constructed Korean table that does not reproduce it is not an insured-lives table. Reproducing it is the one external check available.

The rows run from age 19 — below the composite’s 만19세 minimum issue age — to age 120, the terminal age of the expectation calculation, so that step 2 can be checked from the shipped file rather than taken on trust. Every row says in its provenance column whether it is an anchor or a fitted value.

Two distinctions the file does not blur. The shipped mort_rate is a pricing rate: it is a carrier’s 예정 경험사망률, which carries a margin over experience that no public document sizes, so Projection applies its own mort_be_factor to reach a best-estimate basis. And mort_rate is the 표준체 rate; the rate-class relativities live in rate_class_table.csv.

acc_mort_rate is the 예정 재해사망률, the accidental-death rate, from a different carrier [S6]. Pairing the two is defensible in a way that pairing two all-cause tables would not be: the two carriers that publish accidental rates agree to three significant figures at age 20 and to within 10% everywhere [S6] [S10], which is strong evidence that both take the 보험개발원 참조 재해사망률 almost unadjusted, where the all-cause rates are heavily adjusted and differ by a factor of 1.77 at male 40. It is used only to split the existing death decrement for the 재해사망 uplift variant, never as a decrement of its own.

The premium table is mostly sourced

prem_rate_table.csv is the one assumption file whose values are largely not standardizations, and this is the respect in which Korea is better documented than any other market in this repository. The 생명보험협회 공시실 publishes a statutory 상품비교공시 pricing 45 정기보험 products on one prescribed 대표계약 basis [S4] [S5], and each product’s 상품요약서 prints its own premium grid. Twenty cells are shipped: six 순수보장형 20-year cells and six 만기환급형 20-year cells from the anchor carrier [S12], and eight 10-year cells that are the 갱신형 ladder of the one carrier that publishes a mandatory 예상 갱신보험료 예시 [S6] [S7]. All are monthly premiums per 100,000,000 won of cover on a 표준체 basis. One asymmetry inside those eight: 흥국생명 does not sell its 1형(기본형) to women, so the four female rows are on 2형(보장추가형) and carry a 재해사망 보험금 the male rows do not [S4] [S6]. They are shipped for completeness, the row provenance says so, and no shipped model point reads one.

The is_anchor rows — male and female age 40, 20-year term, one per maturity form — are where Projection extends the scale to unpublished cells, by the ratio of mean table mortality over the term. The male 순수보장형 anchor is the cell that is simultaneously the 감독규정 기준연령 요건 [REG-R9] and the disclosure’s 대표계약 [S5], and its 15,080 won appears independently in the carrier’s own grid and in the cross-carrier disclosure, agreeing to the won [S12] [S4].

The 10-year rows and the 20-year rows come from different carriers, and the model never mixes them: the shipped 갱신형 model points reach published cells only, and the extension for an unpublished cell runs off the 20-year anchor. That the two carriers are at the same level is checkable — on the disclosure basis the 흥국생명 비갱신형 20-year premium is 15,000 won against the anchor’s 15,080 [S4] — which is why the composite can carry both.

The lapse file holds three endpoints, not a curve

lapse_table.csv is three rows because the Korean lapse basis is disclosed as three numbers and a prescribed shape, not as a table. The 적용해지율 (jeogyong haejiyul, the pricing lapse rate) is published in the 상품요약서 wherever a 무해지 form is sold — 「납입기간 이내에 대하여 경과기간별로 연 0.1%~4.6%, 납입기간 이후에 대하여 경과기간별로 연 0.7%~1.6%」 at the anchor carrier [S12], 「연 0.1%~8.4%, 납입기간 이후 연 0.8%」 at another [S1] — and the shape between those endpoints is supervisory rather than chosen: the 2024 IFRS17 계리가정 가이드라인 makes a 로그-선형 model converging to 0.1% the 원칙모형 for 무·저해지 business and sets a post-완납 ultimate of 0.8% [REG-R27]. The chain from supervisory guideline to disclosed pricing parameter runs end to end, which is unique in this repository, so the file ships the endpoints and Projection.lapse_rate interpolates the prescribed shape between them. One qualification travels with it: the guideline’s values are verified from the 보도자료 and its 별첨 was never converted from HWP, so the functional form is [unverified] at instrument level [REG-R27].

To swap in a company basis, replace a CSV with a same-schema file, or point its *_file Reference at a different name, and clear the cache. No formula changes.

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.

Indexed by point_id; point_id = 1 is the anchor cell of the technical notes’ worked example — male, 보험나이 40, 20년만기 전기납, 1억원, 표준체, 순수보장형 해약환급금 미지급형 — which is both the 감독규정 기준연령 요건 [REG-R9] and the disclosure’s 대표계약 [S5].

mort_table()

The 표준체 pricing mortality rates by sex and age, from mort_table.csv.

mort_rate is a [std] construction of the 예정 경험사망률 basis: a Makeham law fitted exactly to one carrier’s three disclosed rates [S12] and tilted above age 60 so that the table reproduces the 제10회 경험생명표’s published 65세 기대여명 exactly [REG-R33]. The industry table itself is not published [REG-R34]. These are pricing rates carrying a margin over experience, which Projection.mort_be_factor removes, and 표준체 rates, which rate_class_table scales.

acc_mort_rate is the 예정 재해사망률 [S6], used only to split the death decrement for the 재해사망 uplift variant. See the Space docstring for why the two columns may come from different carriers.

rate_class_table()

The rate-class relativities by class and sex, from rate_class_table.csv.

Both columns are computed from one carrier’s disclosures at age 40 [S12]. mort_ratio is the ratio of the class’s 예정 경험사망률 to the 표준체 rate; prem_ratio the same ratio taken on the published premium grid. They differ, and not always in the same direction: 0.583 against 0.586 at male 40, where the premium ratio sits above the mortality one as a loading that does not scale with the risk implies, but 0.856 against 0.846 at female 40, where it sits below. Nothing retrieved explains the reversal, so both columns are shipped and neither is derived from the other.

Korea is the only market in this repository that publishes the mortality behind its preferred classes, so these are sourced ratios rather than a standardization. Held flat across ages [std]: the disclosures are at three ages and the ratios move little between them.

prem_rate_table()

The published premium cells, read from prem_rate_table.csv.

Indexed by (form, sex, issue_age, term_y). prem_mth_per_100m is the monthly office premium per 100,000,000 won of cover on a 표준체 basis; is_anchor marks the one row per form and sex from which Projection.prem_rate_mth extends the scale to unpublished cells. Twelve rows are the anchor carrier’s 20-year grid [S12] and eight are the 갱신형 ladder of the carrier that publishes an 예상 갱신보험료 예시 [S6] [S7].

prem_anchor_table()

The is_anchor rows of prem_rate_table(), indexed by form and sex.

Derived from the table already in memory rather than read from a file of its own, so it costs no extra read. Each anchor is the age-40 20-year cell of its maturity form, which for the 순수보장형 is the doubly prescribed cell [REG-R9] [S5].

lapse_table()

The three disclosed 적용해지율 endpoints, read from lapse_table.csv.

Indexed by segment: in_payment_start (4.6% in the first policy year, t = 0), in_payment_end (0.1% at 납입완료) and post_payment (0.8% thereafter). The endpoints are disclosed [S12] [S1] and the log-linear shape between them is prescribed [REG-R27]; Projection.lapse_rate builds the curve.