The Petrine Endowment: How the Vatican Library and Archive Can Pay the Holy See's Deficit in Perpetuity — a Euclidean Proposal for Salaries, Pensions, and the 25-Year Bridge

Scope. The subject is the Holy See (D1), not Vatican City State, dioceses, religious orders, or Catholic universities, whose assets the Holy See does not own. The Vatican Museums belong to the Governorate of Vatican City State and pass their profits upward; they appear here only as a revenue line and a shock channel.

What is excluded. Nothing on this page values or proposes to release material that is closed: documents after the present opening boundary (the end of the pontificate of Pius XII, October 1958),^[19] anything under the pontifical secret or the sacramental seal, diplomatic material still sensitive, and personal data of living persons. That material is not estimated at all; it is left out of this plan entirely by postulate (P3).

Standing. This is an outside analytic proposal, written from public sources. It is not advice to any Vatican body, and it makes no claim about what any official intends.

The Church's own law was written for this subject. Book V of the Code of Canon Law (cann. 1254–1310) governs how the Church acquires, administers, and alienates temporal goods, and can. 1273 names the Roman Pontiff "the supreme administrator and steward of all ecclesiastical goods."^[32] Scripture and the Fathers supply the older grammar of stewardship; the Catechism and the social encyclicals state its present form.^[33][63] Each later section therefore closes with a Warrants box, and the sources are ranked by weight as follows.

Corollary 1.1. No cost discipline, however strict, closes a gap whose cause is the shape of the revenue function. This is why a decade of reform reduced waste and improved transparency without reaching solvency.

Corollary 1.2. There are only two exits: make assessment real, or endow the function permanently. The first pushes cost onto dioceses that are themselves strained and would contradict canon 1271's voluntary character. The second is the subject of Part III.

What follows. The Vatican cannot sell copyright in a twelfth-century register; there is none to sell. What it can offer is access and labour: high-resolution images not otherwise available, transcriptions that did not exist before, catalogue structure, and certified provenance. The Harvard team itself argues that the scarcity of well-documented data with clear provenance chains is the pressing need in AI training.^[25] This page therefore prices the Vatican corpus at €10–75 per million tokens per licence term: above the bulk floor because it is unique and verified, far below the copyright ceiling because it is not copyright. Printed public-domain books earn no licensing income on this model, because equivalent copies are already free elsewhere.

What is priced, and what is not. These figures measure the net income of a new digital licensing service, not the Library or the Archive. The manuscripts, registers, and books themselves are not priced anywhere on this page: they are patrimony held in trust (D3), not for sale (P2), and no market figure describes them.

Capping the downside. Two design rules raise the lower bound shown above. First, imaging and transcription are funded by partners, as the Colnaghi Foundation already funds restoration and digitization of more than 80,000 manuscripts,^[16] so the Holy See's share of K approaches zero. The low case then nets about €4M a year from C2, C3 and C5, worth about 4 ÷ 0.08 ≈ €0.05B. Second, commitment is staged. Only the pilot of one million pages (§19, item 4) is funded up front, at about €1–3M, and the main programme proceeds only if the pilot shows net income above cost at the tested prices. The worst outcome is therefore the cost of the pilot, not of the programme. Stopping is a real option: it cuts the downside without cutting the upside. D

The capital, K. A first phase of about 100 million priority pages costs, on this page's assumptions, €0.4–1.4 per page for imaging, handwritten-text recognition with language-model correction, and sampled human quality control: €40–140M. Precedent exists for partners funding digitization — the Colnaghi agreement is one^[16] — so the Holy See's share is taken as 20–50 percent. D

The cost side. Machine assistance in translation across the Curia's working languages, cataloguing, and first-pass document triage can save an estimated 3–8 percent of the roughly €174M personnel line, or €5–14M a year, achieved through attrition and redeployment rather than dismissal (P6). Capitalized at 5 percent: €0.1–0.28B. D

Why 3.5 percent. The Church of England's endowment, managed by the Church Commissioners, targets a long-run return of inflation plus 4 percent, returned 8 percent in 2025, and has averaged 8.6 percent nominal over ten years; it now stands at £11.6B and funds about a fifth of that Church's running costs.^[28][29] A draw of 3.5 percent real leaves a margin below such a return target, so that the principal is expected to grow in real terms across a complete cycle (CN3). The margin is what absorbs a decade of poor returns.

The order of absorption. A shock is absorbed first by vault I, whose job it is; vault III keeps paying its smoothed draw; vault II is never touched for operations. This ordering is what lets the structure survive regimes that no single asset class survives (§14).

The lesson of over-commitment. The Church Commissioners lost a large part of their assets in the early 1990s through over-commitment and poor investment decisions — later put at about £800M, a third of the fund.^[30] The bounded spending rule, the unlevered pension vault, and the independent investment committee (§16) exist to make that failure structurally hard to repeat.

Reading the table. The base case closes with a small margin. The low column pairs every pessimistic source with the lowest requirement; it falls short by about €0.75B, which means the horizon stretches to 35 years or more and the pension reform of §16 must bite earlier. The high column overshoots, and the surplus would let the draw on donations fall — returning Peter's Pence to charity rather than to administration.

Sequencing. Years 1–5: fill vault I first (safety before growth), commission the actuarial valuation, run the transcription pilot. Years 3–15: fund vault II on its schedule. Years 5–25: build vault III. Real-estate rent normalization proceeds lease by lease, since abrupt increases on resident employees would conflict with P6.

The 2020 evidence: the Museums lost about 5.6 million visitors, falling to 1.3 million, with closures running into 2021^[12]; from March 2020 APSA cut the rents of commercial tenants by 30 to 50 percent, and its contribution to the needs of the Roman Curia fell by half, from €41M to €20M^[15]; the Holy See closed 2020 with a consolidated deficit of €66.3M.^[65] Even the low liquidity vault covers a shock two to three times worse than 2020 for more than two years, longer than the closures lasted.

The Holy See was impoverished in 1870, recapitalized in 1929, and has been in structural deficit for decades. Its 2024 accounts show discipline and transparency that earlier decades lacked, but they also show that the operating gap remains and that the year was balanced by donations and one-off items. The pension liability, the largest single number, is still unmeasured in public.

The Library and the Archive can help. For the first time, machines can read a millennium of handwriting at scale, and the resulting corpus — unique, multilingual, and backed by an unbroken record of custody — can be licensed without moving or selling anything, while every person keeps the right to read it free. On this page's model, the income from licensing it has a present value of about €0.05 billion to €2.9 billion, most likely around €0.6 billion — a measure of one new service, not of the collections, which no price describes.

That is not enough to carry the Church's central government alone, and a plan that pretended otherwise would fail in the first decade of falling AI-data prices. What survives every regime is structure: one set of books, one investment mandate owned in units, a measured pension, and three vaults — liquidity, pension, perpetual — filled from six sources that do not fail together. In the base case the need is about €3.1 billion and the sources reach about €3.35 billion in 25 years. And the archive, saved rather than spent for those 25 years, pays the structural deficit on its own thereafter, for ever (Proposition 6). Like the cathedrals, it is built by one generation for the next.