Cobre Panama and Royalty Concentration Risk

Top-Asset Share, Not Asset Count

Diversification across hundreds of assets does not protect a royalty company from a government shutting its single largest one. Top-asset share, not asset count, is the risk metric that matters.

Franco-Nevada held interests in over 400 assets when Panama’s Supreme Court halted Cobre Panama in November 2023. The portfolio breadth was real. It made no difference to the asset that was lost: the company wrote its stream investment down by US$1,169.2 million, to a carrying value of nil, within weeks. The mechanics of how royalties and streams generate cash are covered separately; here the question is what happens when the cash stops.

What Happened: The Timeline

The striking feature is the pace: the write-off took one quarter, while the remedy has so far taken two and a half years and is still incomplete.

Date	Event
Nov 2023	Panama’s Supreme Court rules Law 406 (the mine’s contract law) unconstitutional. Production halts.
Dec 2023	Franco-Nevada impairs its Cobre Panama stream investment by US$1,169.2M, writing the carrying value to nil.
Jun 2024	Franco-Nevada files for arbitration with ICSID under the Canada-Panama Free Trade Agreement (27 June).
Mar 2025	First Quantum, the operator, discontinues its ICC arbitration and suspends its own FTA arbitration as engagement with the government progresses (31 March).
Jun 2025	Franco-Nevada agrees to suspend its ICSID arbitration as negotiations progress (18 June). Suspended, not withdrawn.
FY2025	A US$4.8M partial impairment reversal is booked against concentrate shipments.
Apr 2026	Panama issues Resolution No. 27 (7 April), authorising removal, processing, and export of pre-suspension ore stockpiles. The operator states this “does not constitute a mine reopening”.
Q3 2026 onwards	Franco-Nevada expects ~23,100 oz of gold and ~265,000 oz of silver in stream deliveries from the stockpiles, the majority arriving in 2027.
Each results date	Franco-Nevada reassesses what the stream is worth. An impairment reverses as the cash flows come back, so watch the carrying value rather than the headlines.

The arbitrations were filed and then parked. Both Franco-Nevada and First Quantum chose negotiation over legal confrontation once the political climate shifted, which tells you the treaty claim is leverage rather than a payment schedule. It has cost both of them real money and returned nothing so far.

The April 2026 authorisation covers only ore mined before the shutdown. Until the government decides on the mine itself, Franco-Nevada’s 2026 guidance of 510,000-570,000 gold-equivalent ounces, the common unit royalty companies use to add up production across metals, excludes Cobre Panama entirely. The market gets the stockpile ounces as upside; the mine restart remains a political event no model can schedule.

Why Diversification Failed

The portfolio did exactly what it was designed to do, and it still could not insure against a binary political outcome at the largest asset.

The royalty model’s diversification argument is sound against ordinary operating risk. A pit-wall failure at one mine, a grade shortfall at another: across 400-plus assets these wash out. But government action is not an ordinary operating risk. It is binary and hits the whole asset in one stroke. Nothing else in the portfolio correlates with it. When the asset hit is your biggest revenue contributor, the loss lands directly on NAV and no amount of breadth dilutes it.

There is also a timing asymmetry worth internalising. The impairment was recognised in the same quarter as the shutdown. The recovery path runs on political time: arbitration filed seven months later, suspended a year after that, first meaningful deliveries expected almost three years after the halt, and a restart decision still pending as of mid 2026. Investors who treated the FTA arbitration as a near-term catalyst misread the clock. Legal remedies against a sovereign are measured in years, and the parties themselves preferred to put the lawyers on hold and negotiate.

The Screening Framework

Concentration is measurable from filings, so screen it before anything else. These are working conventions we state rather than published standards:

Screen	Threshold	Reading
Top-3 asset share of revenue	Below 50%	Healthy
Top-3 asset share of revenue	Above 60%	Concentration risk
Single asset share of NAV	25-30% cap	Above this, the “diversified” label stops applying

A company breaching these screens is not uninvestable. It means the market’s diversification premium is partly unearned, and the valuation should carry a discount that reflects single-asset and single-jurisdiction exposure. The same portfolio detail feeds the EV/GEO accretion test: a deal that adds GEOs while deepening reliance on one mine is lower quality than the headline multiple suggests.

How the Peer Group Screens, FY2025

The FY2025 filed figures put real distance between the names. Franco-Nevada, post Cobre Panama, is now the least concentrated senior; Wheaton is the most concentrated, with the mines Vale operates supplying about half its revenue.

Company	Largest asset	Largest asset share	Top-3 share	Basis
Franco-Nevada (FNV)	Candelaria	12.8% of revenue	31.5%	Revenue
Royal Gold (RGLD)	Mount Milligan	21.7% of revenue	41.9%	Revenue
Triple Flag (TFPM)	Northparkes	29.1% of GEOs (28.2% of revenue)	57.8%	GEOs
OR Royalties (OR)	Canadian Malartic	39.7% of GEOs	65.2%	GEOs
Wheaton (WPM)	Salobo	44.9% of revenue	67.9%	Revenue

Applying the screens: FNV and RGLD pass comfortably. TFPM sits just inside the 60% line at 57.8%. OR and WPM breach it, and both also breach the single-asset guideline, with Canadian Malartic at 39.7% of OR’s GEOs and Salobo at 44.9% of Wheaton’s revenue. Wheaton also carries an operator concentration on top: the mines Vale runs supplied 49% of FY2025 revenue, so one counterparty sits behind half the book. None of this says Salobo will suffer Cobre Panama’s fate. Brazil is not Panama and Vale is not First Quantum. It says that if something did, the portfolio offers far less protection than a 48-asset count implies.

Two cautions when running this screen yourself. Check the basis, because OR and TFPM disclose per-asset GEOs rather than revenue, and the two measures can diverge when metal mixes differ. And read the trend, not just the level. TFPM’s Cerro Lindo stream contributed 23.1% of FY2025 GEOs, but it takes 65% of payable silver only until 19.5 million ounces have been delivered, and 25% after that. The threshold was passed in April 2026, so the stream’s share is falling mechanically.

The Irony in Franco-Nevada’s Numbers

Franco-Nevada’s clean 31.5% top-3 concentration exists partly because the government of Panama removed what was the largest asset from the revenue line.

That is the uncomfortable footnote to the table above. Franco-Nevada’s FY2025 revenue of US$1,822.8M and its 519,106 GEOs sold include just 11,208 GEOs from Cobre Panama (residual concentrate shipments). Concentration screens are snapshots; the same screen run in 2022 would have flagged Cobre Panama’s weight, and an investor who took the 25-30% single-asset NAV cap seriously would have discounted FNV’s NAV for it before the Supreme Court forced the issue. The screen works. It only works if you run it before the government does.

A restart, if it comes, would push Cobre Panama back towards the top of FNV’s portfolio and re-concentrate the book. That is a good problem, recovered NAV is recovered NAV, but it illustrates that concentration risk is dynamic and should be re-measured every reporting period.

Putting It into a Valuation

Concentration feeds the model in two places: the discount rate on the dominant asset, and the risking applied to its NAV contribution.

In a sum-of-parts NAV, the dominant asset should carry a discount rate that reflects its jurisdiction on a standalone basis, not the portfolio’s blended quality. Where a single asset exceeds the 25-30% NAV guideline, sensitise the whole valuation to that asset going to zero; the Cobre Panama precedent makes this a scenario with a base rate, not a hypothetical. Tax structure can concentrate alongside revenue too, since the dominant asset often sits in one tax jurisdiction; the streaming tax-rate trap covers how that flows through NAV. The full sum-of-parts framework, including the discount-rate tiers and risking factors, is in the Royalty & Streaming Sector Primer.

Cobre Panama showed that losing one asset out of more than four hundred can cost US$1.17 billion.