---
title: "Price's Northampton mortality table overestimated mortality, a directional bias good for insurers and adverse to annuity buyers"
type: "claim"
status: "seedling"
audit_status: "flagged ([unverified-quant/mechanism] — a directional-bias claim with a named winner and loser rests on a Tier 3–4 secondary summary; the sourcing floor requires a Tier 1–2 primary for the mechanism and its incidence, and no independent re-fetch was possible in this headless pass; verification routed to question-verify-northampton-table-overestimated-mortality-primary) | AUDIT 2026-09-11 (claude-fable-5-1, cross-model): source_quote re-confirmed verbatim against en.wikipedia.org/wiki/Richard_Price ('It, too, overestimated mortality. In consequence, it was good for the insurance business, and adverse for those purchasing annuities.'), agreeing with the 2026-08-07 seek_verify stamp; Tier 4 / seedling / flag honest; the two dated addenda are append-only and stand; routed question correctly open. One defect: the body's data-defect mechanism ('in-migration and baptism records inflating the apparent count of deaths relative to births') has no pointer — the cited page gives no cause for the overestimate, and neither does the Bernstein corroboration note — so it is now marked inline as unverified-mechanism under the existing flag rather than reading as settled. Drafts every-subscriber (status: drafting) and the-actuary-and-the-incendiary (status: draft), both unapproved, carry the direction claim and not the mechanism sentence; every-subscriber's own audit block already flags the direction as OVERSTATED — no escalation."
writer_model: "claude-opus-4-8"
source_url: "https://en.wikipedia.org/wiki/Richard_Price"
source_title: "Richard Price (Wikipedia)"
source_author: "Wikipedia contributors"
source_date: "accessed 2026-07-09; subject flourished 1723-1791"
source_quote: "good for the insurance business, and adverse for those purchasing annuities"
source_tier: 4
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-price-actuary-burke-revolution-bridge.md, 2026-07-11"
origin: "hop-batch"
derived_from: "10-inbox/raw/2026-07-09-hop-price-actuary-burke-revolution-bridge.md"
date_created: "2026-07-11T00:00:00.000Z"
tags: ["actuarial-history","insurance","mortality-tables","historical-demography","mechanism-design"]
drafted_in: ["2026-07-13-the-actuary-and-the-incendiary","every-subscriber","the-actuary-and-the-incendiary"]
verified_verbatim: "2026-08-07 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
seek_code_commit: "3b23cae"
audits: ["2026-09-11 claude-fable-5-1"]
---


[[entity-richard-price|Richard Price]]'s Northampton mortality table — the survivorship data behind
[[claim-price-northampton-table-standard-actuarial-reference-century|Observations
on Reversionary Payments (1771)]] and the actuarial standard for roughly a
century — is generally described as having overestimated mortality: it implied
that people died sooner than they actually did. A table biased toward death is
"good for the insurance business, and adverse for those purchasing annuities."
The reference that anchored a century of British premium-setting therefore
carried a known directional error in the insurer's favour.

The mechanism is one of incidence. If a table overstates mortality, life-assurance
premiums computed from it are set conservatively high — the insurer collects
more than a true-longevity table would justify — while the same overstatement
makes annuities look cheaper to grant than they truly are, so an annuity
purchaser (or seller) is disadvantaged when the annuitants live longer than the
table predicted. The bias is usually attributed to defects in the underlying
Northampton parish data, such as in-migration and baptism records inflating the
apparent count of deaths relative to births `[unverified-mechanism — this
sentence carries no pointer: the cited Wikipedia page says only "It, too,
overestimated mortality" (fetched 2026-09-11); covered by the note's flag and
routed question, not settled]`.

This extends the vault's thread on demographic data that is load-bearing yet
imperfectly grounded: a mortality dataset read too straight repeats the
selection-and-censoring trap catalogued in
[[claim-osteological-paradox-skeletal-statistics-confound-past-health]], where a
sample of the dead is a distorted window on the living. It also rhymes with the
mispricing that ran through early actuarial history before age-rating —
the flat-rate annuities that
[[claim-dodson-built-age-scaled-premiums-founding-equitable-life|Dodson's
age-scaled premiums]] were built to correct.

The specific direction, magnitude, and incidence of the bias rest only on a
Tier 3–4 summary; the claim is flagged `[unverified-quant/mechanism]` and routed
to [[question-verify-northampton-table-overestimated-mortality-primary]].

**Addendum (2026-07-25).** The first empirical check on Northampton from *assured-lives* data came in about 1811, when [[entity-amicable-society-for-a-perpetual-assurance-office|the Amicable Society]]'s registrar compiled its own mortality experience and found it "more favourable than that shown by the Northampton Table" ([[claim-pensam-1811-amicable-experience-first-assured-lives-mortality-data]]). Two distinct effects point the same way and should not be collapsed: Northampton's own overestimation bias, described in this note, and the separate fact that an assured population is screened and so dies more slowly than the general population the table described.

**Addendum (2026-08-24).** [[claim-bernstein-1996-corroborates-price-table-underestimated-life-expectancy|An independent corroborating account]] (Bernstein, *Against the Gods*, 1996, via a Goodreads-mediated transcription) complicates the "adverse for those purchasing annuities" half of this claim rather than confirming it cleanly. The same passage's next sentence: "the British government, using the same tables to determine annuity payments to its pensioners, lost heavily" — describing the government as annuity *seller/grantor* losing money because pensioners lived longer than the table predicted, the opposite incidence from "adverse for those purchasing annuities" read as naming the individual annuity buyer. The two readings are not necessarily in conflict (the phrase could mean an institution purchasing annuity liabilities off its own book, i.e. the same institution that grants them), but the ambiguity is itself the finding: this note's own phrasing is not precise enough to say which party it names, and the one concrete episode found points toward the seller/grantor being the loser, not the individual annuitant. Marked `[unverified-quant/mechanism — needs primary]`, resting on the same secondary Bernstein passage as the corroborating claim above, and tracked at [[question-verify-northampton-table-overestimated-mortality-primary]] rather than resolved here. The Tier 1–2 primary most likely to settle it — William Sutton's 1874 paper on Price's method — is now identified but access-blocked; see [[entity-william-sutton]].

> [!note] Seek's commentary:
> There is a quiet irony worth keeping: the table that made insurers money was
> wrong, and wrong in a way that happened to be profitable. That is a tidy story
> and therefore suspect — the "adverse for annuities" direction in particular is
> the kind of incidence claim I have seen stated backwards, so it stays flagged
> until I trace the mechanism through a primary rather than a one-line
> Wikipedia gloss. — Seek