some of this should be a grant. some of it should be a hold

One question shows up the same week the shares vest, and it is the wrong first question: should I give this away, or should I invest it?

That is the wrong first question, and answering it costs people real money. The better one is narrower: which job is this particular dollar doing? Some dollars have to go first, absorb the risk nobody else will price, and never come back. Other dollars can sit still for twenty years and should expect something in return for the waiting. Those are two different jobs. A donor-advised fund is very good at the first one and legally barred from the second. Your taxable account is the reverse.

If your equity just became liquid, you are about to build a stack whether you mean to or not. Under US tax rules, this post is about not asking one tool to do both jobs.

two jobs, not one decision

Job one: go first. In any protection deal, something has to be paid for before anything is financeable — the feasibility assessment, the option payment, the title and mineral-rights work, the steward's payroll, two years of monitoring, a deposit against a closing date set by a seller who does not care about your thesis. None of that has a coupon. Some of it fails. This is what charitable capital is actually for, and it is the scarcest capital in conservation.

Job two: own duration. A different kind of dollar can carry a cost basis on a real asset for a long time, tolerate illiquidity, and be paid over that period out of someone else's premium or rent. This dollar is not brave. It is patient, and it wants underwriting.

Most bad allocations of new liquidity are one tool asked to do both jobs. A grant asked to behave like an investment becomes reporting theater. An investment asked to behave like a grant becomes a capital loss you never priced — and not a charitable deduction.

the tools

Read the last three columns together. A grant returns nothing. A recoverable grant returns to the DAF. A PRI returns to the foundation. An endowment or MRI returns to whichever balance sheet owns it. Only the taxable hold returns to you. The tool that returns nothing is the one that can take risks the others cannot.

what a program related investment actually is

A program related investment is an investment a private foundation makes primarily to advance its charitable purpose rather than to make money. Under Internal Revenue Code §4944(c), three tests apply: the primary purpose must be charitable, neither income nor appreciation may be a significant purpose, and the funds may not be used for lobbying or political activity. Clear those and the Internal Revenue Service does not treat it as a jeopardizing investment.

Two mechanics matter more than the definition:

• A PRI counts as a qualifying distribution in the year it goes out. It satisfies payout the way a grant does.
• Repaid principal comes back with a string attached. Return of PRI principal works like a refunded grant: it increases the foundation's required payout in the year of repayment. The recycling is real, but it does not let capital sit.

PRIs are also rarer than the literature suggests. Candid's own guidance notes that only a small percentage of US grantmaking foundations make them at all, and only about one in three PRI funders makes them on an annual basis. This is a specialist instrument, not a default.

Now the part that matters if you are reading this with a brokerage statement open: you probably cannot make a PRI. The PRI is a private-foundation instrument. A donor-advised fund account is not a private foundation — it is an account at a sponsoring public charity, which holds legal control of the assets while you hold advisory privileges. The label is not yours to use. If you stand up a private foundation, PRIs join your menu along with the 5% payout, the excise taxes, and the staffing. If your charitable money lives in a DAF, your nearest analogs are narrower and worth knowing. Some sponsors also run impact-investment sleeves inside the DAF (Vanguard Charitable via CapShift, ImpactAssets, a number of community foundations). Returns stay in the fund. That is still not a PRI, and it is still not money back to you.

the recoverable grant is the DAF's version

A recoverable grant is a grant with a documented expectation that the recipient may return the funds if the funded program hits agreed financial and impact milestones. Fidelity Charitable offers them to donors in its Private Donor Group; Vanguard Charitable and Morgan Stanley GIFT run their own versions. Terms are usually principal only — the stated aim is return of the granted amount with nothing on top.

Two features keep people honest about what it is. First, it is not a loan. As Morgan Stanley's donor FAQ puts it plainly, unlike lending there is no legal recourse if the funds are not returned; if the program misses its marks, the nonprofit simply keeps the grant and you keep the deduction you already took. Second, and this is the sentence people misread: the money returns to the donor-advised fund. It does not return to you.

the DAF cannot hand you a security

This is not a policy preference or our house rule. Internal Revenue Code §4967 imposes an excise tax of 125% of the benefit on a donor, donor-advisor, or related person who advises a DAF distribution that results — directly or indirectly — in a more than incidental benefit to that person, and a further 10% on a fund manager who knowingly agrees to it. Section 4958 covers the more obvious case of the sponsor paying a donor directly.

So the following are unavailable no matter how good the place is or how well-intentioned everyone in the room happens to be. A DAF grant cannot send you back a security, a share, a partnership interest, a token, a certificate you own, or a contractual right to future cash flow. Your deduction generally happened when you funded the sponsor. The grant that follows is the sponsor's money moving to a public charity. If it also handed you an asset, it would be a purchase with a deduction stapled to it — the precise transaction the statute exists to prevent.

Treat that as a feature. The DAF does the giving job well because it cannot do the owning job at all. The moment you want an asset back, you have left the charitable stack, and you should say so out loud rather than engineer around it.

first-loss, in one paragraph

Blended finance means capital taking different positions on the same deal rather than different feelings about it. Concessionary money goes in where the risk actually sits — the first dollar of loss — so commercial money will show up behind it at a price it can accept. Grant and PRI dollars buy the part with no coupon. The money that finances the purchase itself then gets paid over time out of a premium: the recurring payment a downstream dependent (a utility, a municipality, an insurer) makes for the place's protection. Two payors, one place. That is the blend.

the okefenokee did both jobs

Trail Ridge is the line of mineral-rich sand dunes along the eastern edge of the Okefenokee, the largest blackwater swamp in North America and a national wildlife refuge of nearly half a million acres, more than 350,000 of them designated wilderness. Twin Pines Minerals spent six years pursuing an 820-acre mine site on the ridge, and held thousands more acres behind it. Mining there threatened the water table of the refuge and the St. Marys River watershed that drains its eastern side.

In June 2025, The Conservation Fund bought all of it — Twin Pines' roughly 7,700 acres on and around Trail Ridge, plus the underlying mineral rights — for just under $60 million. No market buyer was going to pay $60 million for the right to leave a mineral deposit in the ground. The dollars that moved first were charitable, they were exposed on a closing date, and they bought the one thing nobody else would finance.

The takeout came later, and from someone else. In 2026, Georgia's Department of Natural Resources board approved the first phase — roughly 2,900 acres for just over $8 million — with another tranche slated for 2027, forming a new Alachua Trail Wildlife Management Area of about 4,000 acres. The remaining acreage is intended for the federal refuge, with the mineral rights retired so the ground cannot be mined again.

That sequence is the first-loss half of this post, with a public agency in the duration seat: a charitable balance sheet takes the position a market would not price, a longer-duration public holder buys portions out, the charitable capital recycles toward the next one, and the place lands in permanence. The state's first tranche — about $8 million for roughly 2,900 acres — sits well below the $60 million outlay. The gap is the first-loss, absorbed by charitable dollars. The Conservation Fund is not us and was not using any of our instruments. We are pointing at it because it is the clearest recent example of the split, and because the ecology of the place is one we have already written about — the trembling earth of Okefenokee.

where the hold side actually is

The living system is the thing being protected. Watersheds, forests, soils, wetlands, and species exist whether or not anyone funds them; ensurance is how the funding and the record get organized, not what the place is.

On the hold side, two objects are real today. One is titled natural real property — the cost basis someone has to carry while a place moves from at risk toward permanent protection. That dollar comes back, if it comes back, the ordinary way: a later sale to a permanent holder, and sometimes a premium during the hold from a downstream dependent. The land is the security. The other is a certificate: a record tied to one named natural asset, routed to the account that belongs to that asset. Today it is a funded record, not an instrument with a coupon. It is not a deed and it does not make you the owner of a marsh.

If a grant ever funds a certificate, the recipient holds the record, not you. If you buy one from your taxable account, it is a purchase — there is no charitable deduction.

Behind both sits arithmetic. We run a natural capital accounting engine, RealValue, that prices a place's ecosystem service flows per acre per year against what the real asset costs. One 83-acre forested wetland we have written up carries a ratio near 493% — 493%: the return rate hiding in a swamp. That ratio is a bridge, not a verdict and not a promised return. The 493% accrues to everyone downstream of the wetland. It is not a return to whoever holds the paper.

Two neighbors, so you land in the right place:

• If you are the chief investment officer or the board, the endowment version of this argument is already written: how foundations can invest endowments in nature — without buying land. That post covers PRI-eligible structures and portfolio construction for an institution.
• If you want a coupon you can underwrite rather than a gift, the allocator's door is nature as infrastructure: the search for non-correlated yield.

This post is the individual's version: one person, one new balance sheet, two jobs.

Our stage, stated plainly. Live volumes are small. We do not have a shipped "grant from your DAF" button, and being DAF-receivable is a design target rather than a completed one. For the grant side, the practical route today often runs through a land trust, conservancy, or fiscal sponsor that already holds the right status — the mechanics are in how to grant from a DAF to a place.

how to split your own stack

Not advice — just the sequence that keeps the two jobs from contaminating each other.

1. Size the first-loss bucket before you pick a place. Decide what you can send out and never see again. That number, not your enthusiasm, determines what kind of risk you are able to absorb.
2. Ask your sponsor two questions. Do you offer recoverable grants, and what is the minimum? The answer moves a real fraction of your charitable pool from one-way to recyclable — back to the DAF, not to you.
3. Decide whether you need a private foundation at all. PRI access and control come bundled with a 5% payout, excise taxes, filings, and staff time. For a charitable pool in the low millions, the answer is often no — ask your adviser where the break-even sits for your pool.
4. Put the hold in the taxable account, and underwrite it like an investment. Ask what secures the cost basis, who pays the premium, over what period, and what happens if the steward walks. If those answers are vague, it is a gift wearing an investment's clothes.
5. Write the sentence you want still to be true in ten years, and name the place in it. "I supported conservation" does not survive a decade. "The mineral rights on that ridge are retired" does.

take the next step

If you sit on the institutional side of this — a foundation, an endowment, an investment committee — start with how foundations can invest endowments in nature. It is the same split with a fiduciary attached.

If it is your own money and you want to talk through which dollars do which job, start a giving conversation. Bring the amount, the place if you have one, and whether the capital needs to come back.

This guide is educational. It is not tax, legal, or investment advice. What you can actually do depends on your sponsor's rules, your own facts, and your counsel — confirm all of it with qualified advisers before moving money.

the series

Twelve posts for people about to have more money than plan, most of it headed for a donor-advised fund. Read in order or jump.

1. the gift isn't finished when the deduction posts
2. a donor-advised fund is a parking lot
3. give the shares, not what's left after the tax
4. effective altruism forgot the factory
5. nature is infrastructure. a DAF can fund it
6. how to grant from a DAF to a place
7. longtermism still needs a living present
8. pick a natural asset the way you'd pick a GiveWell charity
9. the warehouse is the product unless you choose otherwise
10. the model is trained on a basin
11. some of this should be a grant. some of it should be a hold — you are here
12. what you actually hold after the grant