Pensions

LATAM holds roughly US$1 trillion in pension assets, much of it deployed abroad. Route 1%, about US$10bn, into an independently managed fund-of-funds through vetted specialist GPs under fiduciary safeguards.

The rationale

The region’s largest pool of patient capital is structurally unavailable to its longest-duration companies. Pension AUM across Chile, Mexico, Brazil, Peru and Uruguay totals roughly US$1 trillion, and much of it is invested outside the region because domestic mandates contain no investable deep tech vehicle. Trustees have nothing compliant to buy: no rated instrument, no track-record series, no carve-out in the allocation rules, so refusal never has to be a judgement about the asset class. Pension funds are under 5% of the top-50 deep tech investors by deal count across Europe and the United States, so the gap is structural rather than regional.

A fund-of-funds fixes the missing instrument rather than the missing appetite. A public-anchored vehicle takes commitments from development banks and public balance sheets first, selects specialist GPs by open competitive mandate, and issues a senior tranche that sits ahead of the equity risk trustees cannot legally take. Redirecting 1% of that US$1 trillion mobilises about US$10bn, more than double total LATAM venture capital across all sectors in 2025. What must be built before any money moves is narrow: investable vehicles that fit fiduciary mandates, the regulatory carve-outs that permit a pilot allocation, and the safeguards protecting the senior tranche.

The honest objection is that this is not public money to direct. Under the AFP and Afore systems these are privately held individual retirement accounts, and a government steering them toward a domestic policy goal is a fiduciary breach whatever the policy achieves. The answer has to be structural: participation voluntary, management independent, the public balance sheet in first-loss position, the pension tranche senior and priced. If the vehicle cannot clear an ordinary risk-adjusted return test against the alternatives a trustee already holds, it should not be built, and the coalition should say so before the first allocation.

The evidence

First moves

1. 0–6 months

   Publish the fiduciary gap analysis

   State, country by country, exactly which allocation limit, rating requirement or valuation rule blocks a pension fund from committing to a deep tech fund-of-funds today. Deliverable: one table covering five countries, each blocking rule cited to its regulation number, with the amendment that would lift it.

   Pension supervisors (Superintendencia de Pensiones Chile, CONSAR Mexico, SBS Peru, PREVIC Brazil, BCU Uruguay) with LAVCA
2. 6–18 months

   Draft the pilot carve-out and the senior tranche

   Write a capped pilot allocation, on the order of 0.1% of AUM, with a defined senior tranche funded ahead of it by development-bank capital, and put it through public consultation. Deliverable: a published term sheet and a regulatory amendment formally in consultation in at least two countries.

   Finance ministries, pension supervisors and national development banks
3. 18–36 months

   First close on competitive mandates

   Anchor the fund-of-funds with public and development-bank commitments, select GPs through open competitive mandates on the Tibi model, and publish mandate terms, fee structure and reporting standard before any pension money is drawn. Deliverable: a first close with mandate documents in the public domain.

   National development banks with IDB Invest or CAF and an independent fund manager

How we would know it worked

• Number of countries with a written pension carve-out permitting an allocation to a deep tech fund-of-funds, reported by pension supervisors, rising from zero to three
• Committed capital in the vehicle against the US$10bn one-percent benchmark, published annually with the public-to-pension ratio disclosed
• Share of GPs selected through open competitive mandate rather than negotiated appointment, published per selection round, held at 100%
• Net IRR of the vehicle reported annually against the trustee’s existing alternatives sleeve, benchmarked to the 17% deep tech net IRR McKinsey reports for 2003 to 2020 vintages

What could go wrong

Who acts

Rests on

The data behind it