Essay
Art Isn't a Special Asset Class.
It's Oil in 1975.
By Faustine Jean-Louis
Economics · Art Markets · Commodities
Apotheosis Advisory
Every conversation about art as an asset begins with a caveat. Art is different. Its units are unique and cannot be substituted for one another. It generates no cash flow. It trades rarely, through a handful of intermediaries, at prices that are mostly private. Its value is entangled with taste, reputation, and cultural authority in ways no model can capture. Art, we are told, is not really an asset class at all — and anyone who tries to analyse it as one has misunderstood what art is for.
I spent the first part of my career analysing crude oil.
Every one of those objections describes the oil market of the early 1970s. Not as a metaphor — as a matter of historical fact. Crude was heterogeneous, with grades that resisted direct comparison. It was priced by posted prices set by a small number of powerful intermediaries rather than by any public mechanism. Physical cargoes moved through relationships, not exchanges. There was no futures curve, no benchmark, no published assessment, no transparent record of what anything had actually traded for. Supply was managed by a cartel with an explicit interest in defending price. And when exchange trading was first proposed, the incumbents — the major integrated oil companies and OPEC alike — were skeptical that a commodity this politically managed and this far from standardised could be brought onto an open market at all.
The resistance was not really about whether oil could be priced transparently. It was that the people who controlled the posted prices had no interest in seeing them replaced.
Then it was.
WTI futures launched on the NYMEX in 1983. Brent developed into a traded market across the mid-to-late 1980s. And Platts — a price reporting agency, not an exchange — began turning an opaque physical trade into a public price through nothing more than disciplined editorial method.
Its reporters spent each day gathering firm, executable bids, offers, and confirmed trades from named market participants, published that information in real time so the market could test it, and at the close derived a single time-stamped assessment of value.
Crucially, they did this without waiting for a liquid exchange to exist: the method was built to work in thin, non-standardised markets, which is exactly what crude was.
The heterogeneity problem — the one everyone said was insurmountable, the fact that no two crudes are alike — was solved not by pretending the barrels were identical but by normalising them: adjusting each trade for quality, location, and timing so that a sour grade in one port could be priced as a differential to a benchmark struck in another.
No two cargoes matched. The methodology made them comparable anyway.
What made the benchmark possible, then, was not a transformation in the market. It was an independent party with no position in the trade, applying a consistent method to non-identical, thinly traded units, and publishing the result.
Oil did not become analysable because its nature changed.
Its nature did not change.
It became analysable because someone built the infrastructure.
The art market has not built that infrastructure. That is a completely different statement from "art is special," and almost everything that follows depends on telling the two apart.
The parallels are not loose
Set the two markets side by side and the correspondence is closer than it has any right to be.
Heterogeneity. Start with the obvious objection, because it is the real one: a commodity is supposed to be interchangeable, and no two artworks are. A barrel is a barrel; a painting is only ever itself. If interchangeability is the test, art fails it outright — and this is usually where the conversation about art as a commodity is allowed to end.
It should not end there, because crude fails the same test. Oil is not one substance but hundreds of grades, differing in gravity, sulphur, and location, none of them naturally substitutable for another.
Oil was not born fungible. It was made comparable — not by pretending the grades were identical, but by pricing each as a differential to a benchmark: a sour, heavy crude trades as Brent minus a quality-and-location adjustment.
Heterogeneity was never a barrier to commodity pricing. It was the specific problem the benchmark-and-differential system was invented to solve.
Art is a heterogeneous commodity whose fungibility problem is the one oil already solved, and has simply not built the layer that solves it.
There is one honest difference, and it is worth stating plainly rather than being caught on it.
Oil's heterogeneity is low-dimensional and measurable — two or three continuous axes an adjustment can capture. Art's is high-dimensional, and part of it, the authorship premium, resists measurement altogether.
Differentials handle gravity and sulphur cleanly; they handle "why this canvas and not that one" only in part. That residual is the genuine limit of the analogy, and a later section is devoted to exactly how far it bites.
But a partial differential structure is still an enormous advance on none, which is what art has now.
Oil was not born fungible.
It was made comparable.
Posted prices. Before the 1970s, crude was sold at prices posted by the majors — set by the seller, opaque to the buyer, unmoored from any observable clearing mechanism.
The primary art market runs on exactly this. A gallery sets a price, does not publish it, and offers it selectively. There is no bid. There is no discovery. There is a list, and access to the list is itself the scarce good.
Cartel supply management. OPEC's central instrument is quota: withhold barrels to defend price.
The art market's equivalent is more genteel and no less deliberate. Artist estates release inventory in measured quantities to avoid flooding the market. Blue-chip galleries ration works, choosing who is permitted to buy, favouring buyers who will hold.
A collector who resells too quickly is cut off from future access — which is, structurally, precisely how a cartel disciplines a member who cheats on quota.
The mechanism is reputational rather than contractual, but the economics are identical: coordinated supply restraint in defence of price.
Hedging, and the price signals it corrupts.When an auction house or a third party guarantees a lot, they have written an option.
The guarantor takes the downside in exchange for a share of the upside. If the work then sells to its guarantor, the reported hammer price is not a market-clearing price at all — it is the exercise of a hedge.
No oil analyst would confuse a producer's hedged sale with the spot price; the distinction is elementary and universally observed.
The art market reports both in the same column, sums them, and calls the total a market.
Storage and carry. Freeports are tank farms. Works sitting in Geneva or Delaware during a soft market are inventory being carried through a downturn, at a real cost in storage, insurance, and foregone capital, in the expectation of a better price later.
That is a carry trade. It behaves like one, it responds to the same incentives as one, and it can be analysed as one.
Unsold cargo. The most honest number in any commodity market is how much of the offered supply failed to clear.
In art, that number is the buy-in rate — the proportion of lots that find no buyer — and it is the single most reliable demand signal available.
It is also almost never the headline. The press release leads with the record-setting lot.
The gap between those two framings is where most of the analytical value in this market currently sits.
What the missing infrastructure costs you
Because art has no benchmark, no differential structure, and no public curve, it reports itself in aggregates that actively mislead.
The global art market grew four percent in 2025, to roughly $59.6 billion. That is the number everyone repeats. It is also close to useless, because it sums two markets moving in opposite directions.
Works above $10 million fell forty-four percent in the first half of 2025 against the prior year, and are down seventy-two percent from the 2022 peak; not one lot sold above $50 million, against thirteen in the first half of 2022.
Over the same period, dealers selling below $250,000 reported sales up seventeen percent, while the $10 million-plus segment fell nine percent.
This is a barbell — participation at the bottom, value at the top, a hollowing middle.
No commodity analyst would publish a single aggregate for a market with that structure. They would decompose it into grades, because the aggregate destroys the information.
The aggregate destroys the information.
The same failure appears at the level of individual sales.
Bonhams' live Middle Eastern art sale in November 2025 achieved £3.2 million and set multiple world records. Its online sale in February 2026 realised $287,000, with a sixty-two percent sell-through — meaning roughly two in five works did not sell at all.
Same house, same department, same category, months apart.
Read the first alone and the market is booming. Read the second alone and it is failing.
Read the buy-in rate across both and you learn what is actually true: demand in this segment is concentrated among committed buyers who transact in the live room, and the broad, liquid market implied by the headline does not yet exist.
That is a finding. It is available to anyone who looks at sell-through instead of totals.
Very few people look, and the reasons are not accidental.
The strongest objection, and what survives it
There is a serious version of the case against everything I have just argued, and it deserves to be stated in its strongest form rather than a convenient one.
It goes like this.
In most commodity markets, the part of the price a model cannot explain — the residual — is noise around a fundamental.
In art, the residual may be the fundamental.
A Basquiat and a technically comparable canvas by a forgotten contemporary can differ in price by three orders of magnitude, and no quantity of supply data, macro context, or hedging analysis will ever recover that gap.
The gap is authorship, provenance, and social consensus about importance — and worse, it is reflexive: the high price itself helps manufacture the consensus that is meant to justify the high price.
Importance sustains the price; the price certifies the importance.
There is no exogenous fundamental to anchor to.
A barrel of crude is not worth more because it sold for more. A painting can be.
That is a real disanalogy with oil, and anyone who pretends otherwise is selling something.
In art, the residual may be the fundamental.
I think this objection is correct.
I also think it proves far less than it appears to.
Grant the whole of it — grant that the authorship premium cannot be derived from theory alone.
It does not follow that the market is unanalysable, for the simple reason that no illiquid asset's price can be derived from theory alone, and we analyse those markets anyway.
Nobody prices a trophy vineyard, a distressed credit with no comparable trades, or a controlling stake in a private company from first principles.
What the unexplained component in art actually contains is two different things, and they fail in different ways.
Part of it is persistent — the premium attaching to an artist's name, stable enough across sales that it drops out the moment you look at changes rather than absolute prices.
You cannot say why the premium is the size it is.
You can observe perfectly well that it compressed twenty percent over eighteen months.
The rest is idiosyncratic: the lot-level surprise that carries one work to more than four times its high estimate while a comparable work by a comparable artist fails to sell in the same season.
That part does not cancel.
For any single work it is irreducible.
But irreducible is not the same as unanalysable.
Idiosyncratic variance is precisely the kind that averages out across a portfolio — an insurer does not predict which house burns down, it prices the distribution across the book.
A lender against one painting is fully exposed to it; a lender against fifty is exposed to a fraction of it.
So the honest conclusion is that valuing a single work is the one thing this market genuinely resists, and that almost every question an institution actually brings — about a segment, a collection, a direction of travel — sits where the unexplained component either differences away or averages away.
And the disanalogy with commodities is narrower than it looks, because commodities carry unmodelable premia too — we have simply given them respectable names.
A large and highly variable share of the crude price is a geopolitical risk premium that no supply-demand model recovers cleanly; analysts book the residual and analyse it, rather than declaring oil beyond economics.
Gold is almost pure premium — near-zero industrial fundamental, priced on belief, reflexive in exactly the way art is — and it supports an entire rigorous analytical industry.
A large, socially constructed, self-referential premium does not exempt a market from economics.
It relocates where the economics does its work.
Art's premium is bigger and stranger than gold's.
That is a difference of degree, not of kind.
The premium is unmodelable.
The market around the premium is not.
So here is the honest position, which is stronger than the maximal one.
The art market is not special in the way it claims.
Concede the one thing that is genuinely true: the authorship premium on an individual work resists modelling, as it does in gold and in every trophy asset.
But the market has taken that narrow, real exception and used it to license a general escape from rigour it has not earned.
The premium is unmodelable.
The market around the premium — its supply management, its liquidity, its hedging distortions, its cyclicality, and its capital flows — is not.
Confusing the two is not a description of the art market.
It is the art market's most successful piece of misdirection.
Where the oil frame stops being a metaphor
In the Gulf, the parallel stops being an analogy and becomes a causal chain.
Gulf sovereign cultural spending is downstream of hydrocarbon revenue.
The United Arab Emirates has committed something on the order of $5.3 billion to arts and culture infrastructure.
Christie's reports that the value of the modern Middle Eastern art it sells tripled between 2020 and 2024.
Contemporary Middle Eastern sales in London rose eighty-nine percent in the first quarter of 2025.
Art Basel launched a fair in Doha this year — its first new fair in years, and a considered institutional bet on where the next collector geography lies.
None of this is disconnected from the oil market.
It is financed by the oil market.
The capital buying Middle Eastern art, building museums on Saadiyat Island, and underwriting a new fair in Doha is hydrocarbon capital, deployed on a cycle that tracks hydrocarbon revenue with a lag.
An analyst who can model that relationship is holding a forward indicator on a segment of the art market that most researchers cannot see at all.
The capital buying art in the Gulf is hydrocarbon capital.
And they cannot see it because the Gulf has data but no price infrastructure.
Public results do exist — Bonhams holds two live Middle Eastern sales a year, Christie's has a Dubai auction history, Sotheby's transacts periodically — and this essay has leaned on them throughout.
But a handful of observations a year is not a price series.
There is no benchmark, no differential structure, no continuity from one sale to the next, and no independent party doing the work of normalising heterogeneous results into something comparable.
What is public also samples a thin and unrepresentative slice of a market whose real volume moves through sovereign acquisition programmes and dealer relationships that leave no record at all.
This is the oil market before Platts.
The response is not to declare the market unanalysable.
It is to build the best available signal from indirect evidence — the partial price series that does exist, sovereign acquisition disclosures, fair activity, and the macro drivers of the capital doing the buying — and to publish the methodology so a reader can weigh the conclusions against the evidence instead of taking them on trust.
The same discipline applies to African and Global South art, where the significant development of the past two years is structural rather than a price move.
The auction houses have begun folding these artists into their main international sales instead of segregating them into regional catalogues.
Sotheby's April 2025 African art sale reached £2.4 million, with forty-six percent more bidders than the year before and eleven new artist records — against $1.7 million a year earlier, when the equivalent sale fell short of its low estimate.
Integration is the mechanism, because it exposes these artists to direct price comparison with global contemporaries for the first time.
In oil terms: they are being priced as a differential to the benchmark rather than in a separate market with its own posted prices.
If they hold parity, that is a re-rating, not a rally — and an allocator should respond to those two things very differently.
Why nobody has built the benchmark
Oil's price infrastructure was not built by the majors.
It was built by exchanges and by price reporting agencies — independent parties with no barrel of their own to sell — whose entire value came from transparency.
And it was resisted, for a long time, by the incumbents whose margins depended on the opacity.
The benchmark was possible precisely because the people who built it had no position in the trade.
That independence was not incidental to the assessment's credibility.
It was the source of it.
Independence was not incidental.
It was the source of credibility.
The institutions that produce the art market's flagship research also lend against art, manage collections, insure works, and audit art businesses.
Their commercial relationships are with the very participants whose behaviour the research describes.
This does not make their work dishonest.
It makes it structurally constrained.
There are conclusions available in the data that such an institution has no incentive to look for, and every incentive not to publish if found.
Nobody whose revenue depends on transaction volume leads with a buy-in rate.
Nobody who guarantees lots publishes a guarantee-adjusted total.
Nobody who lends against ultra-contemporary art is going to be first to point out that roughly a third of the segment's turnover sits with ten artists whose demand could rotate within a season.
These are not conspiracies.
They are incentives, and you can read them in the output.
The claim, stated plainly
Art is not a special asset class — not in the way it claims to be.
Concede the one thing that is genuinely true: the authorship premium on an individual work resists modelling, as it does in gold and in every trophy asset.
Everything else is ordinary.
Art is a heterogeneous, opaque, thinly traded physical commodity with an unusually rich demand function and an unusually primitive price infrastructure.
It is roughly where crude oil sat in 1975: real capital, real price formation, real supply management — and no public mechanism through which to observe any of it.
Art is where crude oil was in 1975.
Oil was not rescued from that condition by someone deciding it was less special than it seemed.
It was rescued by benchmarks, differentials, published assessments, and analysts who insisted on decomposing an aggregate into its grades.
Art will get there.
It will be dragged there, as oil was, by the people who need the market to be legible in order to lend against it, insure it, or allocate to it — and it will be resisted, as oil was, by the people whose margins depend on it remaining obscure.
The tools already exist.
Supply analysis. Price formation under illiquidity. Hedging and the distortions it introduces. Cycle positioning under data scarcity. Capital flow tracking.
None of this needs to be invented for art.
It needs to be applied to art, by someone with no position in the outcome.
The claim that art lies beyond economics has always served the people who benefit from art not being examined too closely.
It is a superb piece of public relations.
It has never been an economic argument.
About the Author
Faustine Jean-Louis is an art market economist and the founder of Apotheosis Advisory, an independent research and advisory practice applying commodity market economics to the art market. Her work focuses on market structure, price formation, institutional behaviour, and the intersection of cultural assets with global capital markets.