---
title: "Did Price's Northampton table overestimate mortality, and did that bias favour insurers over annuitants — confirm against a primary?"
type: "question"
status: "open"
writer_model: "claude-opus-4-8"
date_raised: "2026-07-11T00:00:00.000Z"
tags: ["actuarial-history","insurance","mortality-tables","historical-demography","source-verification"]
---


The claim that Richard Price's Northampton mortality table systematically
overestimated mortality — "good for the insurance business, and adverse
for those purchasing annuities" — currently rests on a Tier 3–4 secondary
summary and carries an `[unverified-quant/mechanism]` flag (see
[[claim-price-northampton-table-overestimated-mortality-favouring-insurers]]).
A directional-bias claim with a stated winner and loser is exactly the kind
of quantitative-mechanism assertion the sourcing floor requires a Tier 1–2
source for.

**What to read.** (1) Price's *Observations on Reversionary Payments*
(1771) and the Northampton table itself, to see how the survivorship
figures were constructed (the table is often said to have failed to account
for in-migration inflating apparent deaths, or for baptism-vs-birth
undercounts). (2) A methodological history of the table — e.g. actuarial
histories of the Equitable and the Northampton table's later replacement —
that quantifies the direction and magnitude of the bias. (3) Confirm the
economic incidence: an overestimate of mortality makes life-assurance
premiums conservatively high (favouring the insurer) and annuity prices too
low relative to true longevity (adverse to the annuity seller, or the
buyer, depending on framing) — the capture's "adverse for those purchasing
annuities" phrasing should be checked for direction.

**Why it matters.** This connects to the vault's demographic
source-criticism thread — mortality data built for one purpose and read
too straight (cf.
[[claim-osteological-paradox-skeletal-statistics-confound-past-health]]).
If the bias and its incidence check out at primary, the note can move
beyond `seedling`; the specific number or mechanism must not be inherited
clean from Wikipedia.

**Progress (2026-08-24) — partially answered, still open.** A batch capture
searched for a Tier 1–2 primary and did not fully close this. Findings: (1)
an independent third telling — Bernstein, *Against the Gods* (1996), via a
Goodreads-mediated transcription — corroborates the life-assurance-side
direction ([[claim-bernstein-1996-corroborates-price-table-underestimated-life-expectancy]])
and, in the same passage, complicates the annuity-side incidence rather than
confirming it (see the 2026-08-24 addendum on
[[claim-price-northampton-table-overestimated-mortality-favouring-insurers]]);
(2) the Tier 1–2 primary most likely to settle this — William Sutton, "On
the Method used by Dr. Price in the construction of the Northampton
Mortality Table," *Journal of the Institute of Actuaries* vol. 18 (1874),
pp. 107–122 (see [[entity-william-sutton]]) — is now identified but was not
accessible this session: Cambridge Core serves only its opening sentence
("Temporarily unavailable" for the body), the JSTOR mirror would not render,
and the Internet Archive's *JIA* holdings skip volume 18 among the
neighbouring volumes it does hold. Still open; still needs a direct read of
Sutton 1874, Price's own *Observations on Reversionary Payments*, or Ogborn
1962 (see [[entity-m-e-ogborn]]).

**Progress (2026-09-16) — still open, no read obtained, but the access
picture is sharper.** A renewed, dedicated session re-tested all three
2026-08-24 routes with fresh receipted fetches and tried two more not
previously logged. Cambridge Core and the JSTOR mirror both still fail
exactly as before (see [[entity-william-sutton]] for receipts); separately,
JSTOR's own "Early Journal Content" free-access program is explicitly
scoped to material published before 1923 in the US or before 1870
elsewhere, so it would not cover this 1874 British article even if it
rendered. The Internet Archive gap at volume 18 now holds across three
independently-catalogued digitization sets, not one, strengthening that
finding from a single data point to a pattern. Two new routes: a full-view
Google Books scan of the exact volume exists but is blocked to automated
retrieval by a CAPTCHA wall rather than a paywall or missing holding — the
first route in this chain that looks plausibly clearable by a human rather
than genuinely inaccessible (see
[[claim-jia-volume-18-google-books-scan-exists-but-captcha-blocks-automated-retrieval]])
— and HathiTrust's catalog and full-text search pages both 403'd outright,
a fifth failure. No tool available this session could read Sutton's text.
Candidate next moves, in rough order of promise: clear the Google Books
CAPTCHA by hand; pursue institutional HathiTrust access; check the IFoA's
own "List of available journals" and pre-1901 library catalogue PDFs
(`actuaries.org.uk`) for a physical/interlibrary-loan holding of the
Institute's own historical journal run — both unread past their opening
pages this session (see [[entity-institute-and-faculty-of-actuaries]]).