The exit is the asset

The transaction side is larger and less visible. Card singles on eBay alone topped $2.62bn in 2025, with growth accelerating for nine consecutive quarters and driven by the number of items sold rather than by rising prices[2]. That distinction matters more than the headline: a market whose dollar volume rises because prices rise is a market getting more expensive, while one whose volume rises because more things change hands is a market getting deeper. Only the second is evidence of liquidity.

Participation is broad. An estimated 94 million Americans have bought a collectible card at least once, and more than 28 million trade or grade in a given year[6]. A Circana survey in March 2025 found 19% of US adults had bought Pokémon cards for themselves in the preceding six months, and that only about a quarter of those buyers play the game[7]. The rest collect, display or resell, which is to say they are holding an asset.

It would be easy to treat this as a reason to pick the largest number and move on. It is more useful to treat the dispersion as the finding. Estimates diverge like this when there is no central exchange to report through, no obligation on anyone to disclose a transaction, and a large share of trade happening privately between people who have no reason to tell a research firm about it. The disagreement is a direct measurement of how opaque the market is.

A narrative that treats "the card market" as one thing is therefore wrong about at least one half of it in any given year. Anyone building exposure to this category has to decide which market they mean, and an index that silently blends the two will report the average of a boom and a contraction.

This is not a detail of execution. Amihud and Mendelson established that an asset's price embeds the present value of the transaction costs its holders expect to pay in future[4]. An asset that is expensive to sell is worth less precisely because it is expensive to sell, and the discount is capitalised into the price today. Standard valuation practice puts the discount for lack of marketability at 20–30% for private companies, and 20–40% is routinely applied. The early restricted-stock studies that anchor the range found average discounts of about 35%.

If frictions of that magnitude are capitalised in cards as they are elsewhere, then a material part of what a card is worth today is a deduction for the difficulty of selling it.

Retail investors hold approximately none of it, and the machinery that would let them is arriving now for everything except this category. A 2025 survey found 56% of limited partners expect retail and wealth channels to supply at least half of new private-market inflows within two years[12], and Bain projects private-wealth alternatives tripling from $4tn to $12tn by 2034. Art has had a securitised, SEC-qualified retail route since Regulation A+ Tier 2 made it possible to raise up to $75m from non-accredited investors; one platform reached roughly 880,000 registered users and about $940m in assets by early 2025[13].

Collectible cards are, on this measure, the least institutionalised illiquid asset of any consequence, a category with millions of participants, billions in annual observable turnover, standardised condition grading, and essentially no financial infrastructure.

• —A discount removed, once. If part of a card's price is a deduction for the cost of selling it, then infrastructure that lowers that cost should raise the price. This is a one-time repricing, not a recurring return, and its size is bounded by the friction actually removed.
• —Access, ongoing. The population able to hold the asset expands from those who can buy a whole card, store it, insure it and find a buyer, to anyone who can buy a share of one. That is a demand effect and a distributional one.
• —Price discovery. A market with continuous observable pricing supports different decisions from one where the last comparable sale was fourteen months ago. This is the least glamorous of the three and probably the most consequential.

The constraint was never scarcity. Scarcity was always there. The constraint was the cost of finding the other side.

The people who opened these packs as children are now in their thirties and forties, and the grail card they could not afford at eleven is now merely expensive. That is the demand engine, and it is measurable, demographic, and slow.

Collectibles may be priced to return less

Dimson, Pukthuanthong and Vorsatz examined 110 years of returns across 13 categories of collectible and 30 distinct return series. They find a positive emotional yield in 24 of the 30 series, averaging 2.64% a year, and conclude that assets delivering emotional returns carry correspondingly lower equilibrium financial returns[5].

The implication is uncomfortable and deserves stating plainly. If the marginal buyer of a card is someone who enjoys owning it, the price already reflects a willingness to pay that a purely financial holder does not share. Buying financial exposure to such an asset means accepting a return reduced by somebody else's pleasure. This does not invalidate the liquidity argument, since removing a discount and earning a premium are different claims, but it does mean that anyone expecting collectibles to outperform equities on a risk-adjusted basis is arguing against the best available evidence.

Cohorts age out

Antique furniture, silver and porcelain were once blue-chip collectibles held by serious people for serious money. They are now difficult to give away, and the generation currently inheriting them is discovering this: the wealth transfer is accompanied by a transfer of objects that no longer command what they did[15]. Nostalgia demand is specific to the people who were children when the object was made, and those people do not live forever.

Cards may differ in three ways that matter: they are standardised and independently graded, so condition is describable rather than arguable; they are small, portable and cheap to store, which furniture is not; and the franchises behind them, Pokémon most obviously, are still producing new participants rather than only ageing existing ones. Those are reasons to think the risk is smaller here, not reasons to think it is absent.

Supply can arrive faster than demand

The category has had exactly one catastrophe, and it was a supply shock. Between roughly 1987 and 1994 manufacturers tripled output. Topps alone is estimated to have printed over a billion cards in 1986, and the vast majority of what was made in those years is now worth cents[16]. Demand did not collapse. Supply overwhelmed it.

The modern version of this risk does not require anyone to print anything. Grading converts an ungraded card into a newly tradeable unit, which means certified supply can expand sharply without a single new card being manufactured, and it did, by 32%, last year. We treat this as the most underweighted risk in the category and examine it in a separate paper.

That is a specific and bounded opportunity: to remove a discount and widen access to an asset class institutions have held in every other form for decades. It is not a claim that collectibles beat equities, and the best evidence available suggests they should be expected to do slightly worse for the pleasure of owning them.