Nobody knows how many authentic Rothkos are in private hands, how many bottles of ’82 Lafite remain undrunk, or how many unpolished Paul Newman Daytonas survive. Estimates exist; ledgers do not. The supply side of every collectible market on earth is an assertion, except one.
Every trading card certified by a grading company since 1991 is registered, by issue and by grade, in a publicly queryable ledger called a population report. For every card in every grade, the exact number of certified copies in existence is a published, auditable number. We call this property the census, and this paper argues that it, not nostalgia, not the celebrity auction results, is the load-bearing innovation of the asset class. Nearly everything interesting about graded cards follows from it: why returns concentrate so violently in top grades, why the market supports index construction of a kind art has attempted for decades and never achieved, why the asset is underwritable at all, and why the greatest systemic risk to the class is not a price crash but a corruption of the ledger itself.
We also argue, bluntly, that the industry’s own promotional statistics are now the primary obstacle to institutional adoption. And we close with predictions specific enough to be wrong.
Stores of value without ledgers
The passion asset - a physical object held for appreciation, whose value rests on scarcity and cultural demand rather than cash flow - has been around for centuries. Art is the canonical case: roughly $60 billion in annual transaction volume, an estimated $1.7 trillion in worldwide holdings, and indices (Mei Moses, from 1950) showing long-run compound growth around 8.5%. Wine, watches, and classic cars followed the same institutional path: auction houses, condition experts, price databases, and eventually index products - the Knight Frank Luxury Investment Index family records classic cars at roughly 20% and rare whisky at 16% annualized over their best recent decades.
Yet every one of these markets has failed to become a true institutional asset class, and they all failed for the same three reasons:
Unverifiable supply. No one can state, to the unit, how many qualifying objects exist. Scarcity, the entire basis of value, is asserted by experts, not published by a ledger.
Unstandardized condition. "Excellent condition" means different things at different auction houses on different days. Condition disputes are settled by connoisseurship, which does not scale and cannot be audited.
Episodic price discovery. A painting trades once a decade; between sales its price is a conjecture. Indices interpolate; allocators cannot mark.
Any asset class that solved all three - verifiable supply, standardized condition, continuous price discovery - would be categorically different from its predecessors. One has.
Cards before the census
Trading cards spent their first century demonstrating both halves of the scarcity equation. The 1909–11 T-206 tobacco series produced the Honus Wagner - an estimated 50–60 surviving copies, most recently trading at $7.25 million - because nearly all of its print run was destroyed by children, bicycles, and attics. Survival, not issuance, created the scarcity. The counterexample proved the same rule from the other side: the “junk wax” era of the late 1980s and early 1990s, uncontrolled overproduction into peak demand, destroyed a decade of card values so thoroughly that rookie cards of Hall of Famers from the period remain nearly worthless. Cards taught the market, twice, that supply is everything, and that until supply was measurable, the asset was uninvestable except by folklore.
Grading changed the unit of account. A graded card is not “a 1952 Topps Mantle”; it is (1952 Topps Mantle #311, PSA 8), a standardized instrument whose certified population is public. The price ladder this creates is steep: the same 1968 Topps Mantle runs roughly $200 in PSA 3 and $10,500 in PSA 9. The final piece arrived with continuous venues: eBay hosted $871 million of sports-card sales in Q1 2021 alone, and verified-comp databases (Card Ladder, PSA auction records, 130point) now publish sale-level price history free of charge. By 2021, all three missing mechanisms existed in one asset class. The 2020–21 mania and 2022–23 correction that followed were the first full cycle of a market that could, for the first time, actually be measured while it happened.
The census
The claim of this paper is that one property separates graded cards from every other passion asset. Stated precisely: an investable unit is the pair (card issue, grade). For every unit, the certified population, the number of copies ever graded at exactly that grade, is published, continuously updated, and free to query. Three consequences follow. Supply is auditable: scarcity is a database lookup, not an expert opinion. Units are comparable within grade, making verified comps meaningful. And supply growth is observable: new gradings append to the ledger, so dilution of scarcity is visible before it is priced.
Concretely: the Base Set Unlimited Charizard has been graded 101,739 times; exactly 487 came back PSA 10.
One card, one census: 1999 Pokémon Base Set Charizard #4
Reported PSA population, July 2026
Total PSA-graded
101,739
PSA 10
487 · a 0.48% gem rate
Source: PSA population report, July 2026, via Graded.World. Reported populations may include resubmissions and exclude ungraded examples.
When a PSA 10 price appears irrational next to the raw card, the census is almost always the explanation. No participant in the art, wine, or watch market can make a statement of that resolution about any object they own.
The census also explains the asset’s signature behavior - grade convexity. Because population thins geometrically toward the top grade while demand concentrates there, price multiples over a raw near-mint copy climb steeply through the top of the scale:
Typical ranges across widely traded cards
PSA 8
~1×
PSA 9
1.5–3×
PSA 10, modern
3–10×
PSA 10, vintage
10–30×
Dark bar = low end of range · grey = high end · 18px ≈ 1×
Source: GradingMetric, raw-vs-graded price-multiplier analysis.
The top grade of a census-scarce card is not a nicer version of the same asset. It is a different asset, with different supply, different holders, and a different return history. Most confusion about this market, including most of its bad statistics, comes from averaging across that boundary.
What the census enables
Treat the census as a primitive and a stack of institutional machinery follows from it, most of it already under construction.
Indexation. An index requires defined constituents, verifiable supply, and repeatable pricing. Cards are the first passion asset to satisfy all three. Early families already exist: PWCC's Top 100/500 (verified transactions since 2008), Card Ladder's CL50, and set-level composites priced daily from sold comps with census-derived "gem market caps" - price × population, a statistic that is impossible to compute for any other collectible.
Underwriting. Population bands make eligibility rules mechanical: a buyer can require, ex-ante, a census between defined bounds, a minimum sales history, and a maximum spread, and audit every criterion. This converts collecting judgment into screening, the same transition that converted stock-picking folklore into factor investing.
Marks. Comparability within grade makes verified-comp marking defensible: recent arm's-length sales of the same card-and-grade pair are a real mark, not an appraisal. Art has never had this; every art-fund mark is an argument.
Supply surveillance. Because gradings append to a public ledger, dilution is watchable: a population spike in a "scarce" modern card is visible months before it fully reprices.
Eventually, passive products. Once a benchmark exists and custody is institutional (insured vaulting is already commodity infrastructure), tracking products are an engineering problem, not a conceptual one. Our predictions section puts a date on this.
The evidence: returns live at the census boundary
The return evidence, read carefully, is not a story about “cards going up.” It is a story about where in the market returns occur, and the answer, consistently, is: in the top grades of census-scarce, culturally durable cards, and nowhere else.
PWCC transaction-built indices vs. S&P 500 price return over the same window
Index constituents selected by transaction volume; excludes transaction costs.
Source: PWCC Marketplace, Business Wire, Jan 2019. S&P figure over the comparable window as cited in Filipek, “Investigating Sports Cards as an Asset Class” (2023).
Card Ladder's CL50 broad composite is up over 400% since 2016, through a drawdown and plateau in 2022–23 that the index reports honestly. As of July 2026 the live picture is strong: the Graded 100 composite (PSA-10 comps, eBay and Fanatics) rose +23% since January, with single sets moving +81% (Fossil) and −6.6% (Jungle) in thirty days - dispersion that rewards selection and punishes averaging.
The census boundary is easiest to see in a single card:
Approximate annualized return by census position
One card is an illustration, not a proof, but the monotonic relationship is the general pattern.
Source: AInvest (2026), from public sales records.
One more point on benchmarks. Cards should be measured against passion assets, not equities. Against that peer set they are competitive at the top of the market - the Knight Frank luxury composite turned $1 million (2005) into roughly $5.4 million versus $5.0 million for the S&P 500 - while offering what no peer offers: a published supply ledger. The honest pitch is not “cards beat stocks.” It is: among scarcity assets, cards are the only one you can actually measure, and measurability is what lets a disciplined buyer capture the segment the indices describe instead of the average the critics describe.
The market, 2026
Estimates of market size vary with the boundary drawn - cards alone versus cards plus memorabilia, sports versus TCG. We present the range rather than the most flattering point in it. The dispersion is itself a datum: this is a young market measured with unstandardized definitions.
Base-year estimates, USD — scopes overlap and are not additive
Forecast CAGRs range from 5.2% (TCG) to 18.2% (memorabilia + cards); grading itself at 12%.
The census layer is scaling faster than the market it measures. Industry-wide, more than 26 million cards were graded in 2025, 19.26 million of them by PSA, up 26% year over year. June 2026 set a monthly record at 3.5 million, up 65% year over year.
Two structural shifts complete the picture. First, consolidation: Fanatics (collectibles revenue approaching $5 billion) now controls the major league licenses and, through Fanatics Collect, an integrated vault-marketplace with a Sotheby's channel for six-figure cards; Blackstone backs CGC, the fastest-growing grader. Second, the TCG inversion: trading-card games - Pokémon, One Piece, Magic - overtook sports cards inside the grading pipeline.
Year-over-year change in cards graded, by category
TCG is now roughly 71% of all graded volume. Pokémon was the #1 U.S. toy brand of 2025 at $2.5B retail, +87% YoY.
Sources: GemRate aggregations via Card Atlas (2026) and Sports Illustrated (July 2026).
The inversion is powered by the 30-year nostalgia cycle now reaching the Pokémon generation, compounding through globally distributed IP with structurally larger audiences than any single sport. We regard it as permanent, not cyclical, and say so in the predictions below.
The window
Hard-to-hold assets institutionalize in a sequence with a known payoff structure. First the infrastructure arrives: standardization, custody, price data. Then capital arrives to use it. Then an access product arrives, and the repricing that follows accrues overwhelmingly to whoever owned the asset before the product existed. Gold spent millennia as a store of value, but the seven years after the first mainstream gold ETF launched in November 2004 roughly quadrupled its price, as a vault-and-assay asset became a brokerage-account line item. Bitcoin repriced around its January 2024 spot ETFs by the same mechanism. Art, instructively, never completed the sequence: with no census, no benchmark could be built that an allocator would accept, and the asset class stalled permanently at stage one.
Locate graded cards on that sequence and the situation becomes uncomfortable to look at directly. The infrastructure is finished: the census is published, verified comps are free, insured vaulting is commodity infrastructure, and the grading pipeline is registering 26 million new units a year. The access product does not exist: there is no benchmark index recognized by allocators, no tracking product, and institutional allocation to the asset class is, to a first approximation, zero. Cards are today the only scarcity asset that has completed stage one and not yet begun stage three.
The repricing has started without the institutions. The Graded 100 composite is up +23% in the first half of 2026; WOTC-era vintage +33% over six months. The largest pool of potential buyers has not yet placed a single bid.
Supply in the segment that matters is fixed and visibly thin. Vintage populations do not grow; gem rates at the top of the census run below half a percent; and every year of record grading volume re-confirms, in public data, that the top grades stay scarce.
The close of the window is predictable, and dated. This paper commits to it below: a recognized benchmark and a passive product by July 2031. When that product launches, the census-scarce segment gets marked by the marginal institutional dollar instead of the marginal hobbyist dollar. The interval between now and then is the window.
Every prior asset class made its early owners wait for the infrastructure. Cards inverted the order: the measurement arrived before the money.
What could kill this
An asset class is only as credible as its treatment of its own weaknesses. Each concern below is stated at the strength its best critics give it.
The marketing problem. The most-circulated pitch for cards claims +3,261% over twenty years against ~421% for the S&P 500. Financial commentators have publicly dismantled this figure, and they are right: it averages only graded, top-condition, headline cards - survivorship bias by construction - compares that biased average to a broad index while ignoring dividends (SPY's actual 20-year total return through July 2026 was ~510%), and ignores every friction of ownership. Our position is unambiguous: this number, and numbers like it, are now the primary obstacle to institutional adoption. The industry does not have a returns problem; it has a benchmark-integrity problem. The remedy is the census itself - indices with disclosed constituents, verified comps, and published supply.
The census monopoly. Everything in this paper rests on the integrity of population reports. That ledger is now controlled, to a degree no analogous market would tolerate, by a single for-profit company: Collectors Holdings - PSA plus SGC plus, since December 2025, Beckett - roughly 80% of U.S. grading. The past twelve months supplied the stress test: a "buyback" controversy in which cards repurchased by PSA at low grades were later regraded higher; an antitrust class action (April 2026) seeking to unwind the acquisitions; every grading tier under $80 paused behind a ~12-million-card backlog. A price crash would injure this asset class; a credible corruption of the census would invalidate it, the equivalent of one firm controlling the gold assay and occasionally assaying its own inventory. What survives: the census layer is redundant in principle (four independent ledgers exist; Blackstone-backed CGC grew 121% in 2025), cross-grader arbitrage polices standards at the top of the market, and the equilibrium we expect is a duopoly, the structure serious participants should actively prefer.
Illiquidity and frictions. A realistic seller recovers 75–85% of quoted value before fees in normal markets; a six-figure card can take months to place; all-in frictions on high-value cards can exceed 25%. Nothing here can be engineered away, only priced. The census asset is priceable daily but not sellable daily, and any allocation framework that confuses the two will fail.
Cyclicality and regime-dependent correlation. The broad Pokémon market cooled 30–50% from its 2021 peak before the current recovery; thin floats reprice sets by double digits monthly. Our own estimates for the investable segment assume 22–30% volatility, severe-scenario drawdowns of 40–55%, and equity correlation of 0.25–0.45 in normal regimes rising toward ~0.65 in stress, because cards are bought with discretionary wealth, and the marginal bid withdraws exactly when equity portfolios draw down. In a liquidity shock this is a luxury discretionary asset, not a safe haven, and no honest paper on the class may claim otherwise.
A sterile asset. Cards produce no cash flows, and carry - storage, insurance - is negative. This is unfixable, and it dictates the class's economics: entry price is the entire game. An asset with no yield forgives nothing; the buyer's discount to fair value at acquisition is the only cushion the asset will ever provide.
Predictions that can be checked
A thesis that risks nothing asserts nothing. The following are stated so that they can be wrong.
By July 2031, the asset class has an institutional benchmark and a passive product. At least one graded-card index with disclosed constituents, census-based eligibility, and verified-comp pricing will be recognized as a benchmark by institutional allocators, and at least one investable product tracking such an index will exist. The precondition, the census, already exists; art never had it and never got there. Cards will.
Grading resolves to a functional duopoly. PSA's certified-volume share falls below 65% (from ~72% in 2025) as CGC scales into a genuine second census pillar, or an independent census-audit layer emerges. Either outcome hardens the ledger; a self-interested near-monopoly over the asset class's foundational data is the one equilibrium that cannot persist.
The TCG inversion is permanent. Trading-card games' share of graded volume never again falls below 50% for a full calendar year.
Survivorship-era marketing disappears from serious materials. Within five years, no institutional-facing document in this industry leads with an aggregate "cards vs. the S&P" statistic. Census-based language - populations, gem rates, verified comps, disclosed-methodology indices - becomes the standard register, for the same reason quoted spreads replaced stock tips: the audience changed.
The first passion asset with a ledger
Graded trading cards are commonly analyzed as if they were small art: a nostalgia asset with good recent returns and questionable statistics. This paper has argued the opposite framing. Cards are the first passion asset in which supply is a published fact rather than an expert assertion, and that single property, the census, is what converts a collectible into an asset class: it makes scarcity auditable, comps meaningful, indices constructible, and dilution observable.
The class’s liabilities are equally clear, and none of them is hidden here: it is sterile, cyclical, illiquid in stress, marketed with statistics that insult its best audience, and dependent on a census layer currently controlled by a self-interested near-monopoly. Every one of these is either priceable, reformable, or structurally resolvable, and the predictions above commit this paper to dates on which those claims can be judged.
The passion assets that came before - art, wine, watches, cars - spent decades building institutions to compensate for facts they could not verify. Cards begin with the facts. The asymmetry deserves to be stated once, plainly: every claim in this paper can be verified today, by anyone, from public ledgers, and almost no allocated capital has yet bothered to look.
If you’re underwriting this space, reach out. john@tash.cards · tash.cards
Sources
Market Decipher, "Sports Memorabilia and Trading Cards Market to Reach $393 Billion by 2036," PR Newswire, 29 June 2026. Link
Research & Markets, "Trading Cards Market Size, Share & Trends Analysis Report, 2026–2033," April 2026. Link
Filipek, C., "Investigating Sports Cards as an Asset Class" (2023) - historical development, grading price ladders, PWCC/CL50 index discussion, art-market comparison.
Sports Collectors Daily, "$871 Million Worth of Sports Cards Sold on eBay in Q1," July 2021.
Fanatics Inc., "Fanatics Collect" - marketplace, vaulting, Sotheby's partnership. Link
BonVision, "From the Playground to the Portfolio," 2026 - 30-year nostalgia cycle. Link
Card Atlas, "PSA at a Crossroads," 2026 - GemRate grading volumes, market shares, backlog, buyback controversy, antitrust timeline. Link
AInvest, "Pokémon Cards Beat the S&P 500 by 2.5x - But That Return Is a Survivorship Trap," 2026. Link
PWCC Marketplace, year-end index performance, Business Wire, 16 January 2019. Link
Card Ladder - index construction methodology. Link
Graded.World, "Graded Card Market Report: July 2026" - Graded 100 composite, population and gem-rate data. Link
Sports Illustrated, "PSA Leads Record-Breaking June as Card Grading Reaches New All-Time High," July 2026. Link
GradingMetric, "Raw vs Graded Card Value" - grade-tier price multipliers. Link
24/7 Wall St., "Pokémon Cards Beat the S&P 500 by 2.5x, But the Math Is a Lie," 18 July 2026. Link
Knight Frank, The Wealth Report - Luxury Investment Index. Link
Historical market data: SPDR Gold Shares (GLD) launched 18 Nov 2004 (~$440/oz gold); ~$1,900/oz by Sept 2011. U.S. spot bitcoin ETFs approved 10 Jan 2024 (~$46,000); >$100,000 by Dec 2024. World Gold Council; public exchange data.
tash Cards Inc., research estimates, July 2026 - volatility, drawdown, and correlation ranges for the investable graded segment.
This document is a statement of research opinion by tash Cards Inc., provided for informational purposes only. It is not an offer to sell or a solicitation of an offer to buy any security, and is not investment, legal or tax advice. Trading cards are physical, illiquid assets whose values are estimates drawn from comparable sales and can move sharply; the collectibles market has moved through pronounced boom-and-correction cycles. Figures described as estimates are assumptions and market references, not guarantees or projections. Market data is drawn from third-party sources believed reliable but not independently verified. The predictions herein are opinions held as of the publication date and may prove incorrect. Past segment performance does not predict future results; capital is at risk.