Art Isn't a Commodity.
But It's Being Priced Like One.

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. What let a benchmark work for it was that its heterogeneity is low-dimensional and, crucially, that physical substitution disciplines the price.

Art has neither property, and this is the point at which the analogy has to be handled honestly rather than stretched. Art's heterogeneity is high-dimensional, and part of it—the authorship premium—resists measurement altogether.

There is no substitution, so there is no arbitrage, so the benchmark-and-differential machinery that priced crude does not port to art at all.

But oil got two different things in the 1980s, and only one of them depends on fungibility. The benchmark-and-differential system does. The other—the independent party gathering scattered trades and normalising them into a published series—does not.

That second layer is the one that ports to art, and it is worth being precise about its state, because art is not a blank page here.

Indices exist. Sotheby's Mei Moses, Artprice, Artnet, and ArtTactic all publish them. The problem is not absence. It is that no existing instrument combines the properties the job requires.

The missing infrastructure is therefore not simply an index. It is an independent one, built by a party with nothing to sell; a transparent one, whose methodology can be examined; and a disaggregated one, resolved to the level of segment, tier, and geography rather than collapsed into a single market average.

Posted prices

Before the 1970s, crude was sold at prices posted by the majors: set by the seller, opaque to the buyer, and unmoored from any observable clearing mechanism.

The primary art market runs on exactly this model. 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 and favouring collectors likely to hold. Those who resell too quickly often lose future access.

The mechanism is reputational rather than contractual, but the economics are remarkably similar: 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 effectively written an option. The guarantor assumes downside risk in exchange for part of the upside.

If the work ultimately sells to its guarantor, the reported hammer price is no longer a true market-clearing price. It is the outcome of a hedge.

No commodity analyst would confuse a producer's hedged sale with the spot price. The art market routinely reports both in the same totals.

Storage and carry

Freeports function much like tank farms. Works sitting in Geneva or Delaware during a weak market are inventory being carried through a downturn at a real cost in storage, insurance, and foregone capital.

Economically, this is a carry trade.

Unsold cargo

The most revealing 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 receive no buyer. It is one of the clearest indicators of demand available, yet it rarely appears in the headline.

The press release celebrates the record-setting lot. The buy-in rate quietly tells the deeper story.

The parallels do not establish that art is a commodity. They establish something narrower, and ultimately more important: that the market's structure—its supply, pricing, hedging, and clearing mechanisms—resembles a commodity market even though the underlying asset remains fundamentally unique.

Because the indices art does have aggregate the whole market into a single line, and because no public curve exists beneath them, the market reports itself in figures 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 declined nine percent.

This is a barbell: participation at the bottom, value at the top, and 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 same failure appears at the level of individual sales.

Bonhams' live Middle Eastern art sale in November 2025 achieved £3.2 million and established multiple world records. Its online sale in February 2026 realised $287,000 with a sixty-two percent sell-through, meaning nearly two in five works failed to sell.

Same house. Same department. Same category. Only months apart.

Read the first result alone and the market appears to be booming. Read the second alone and it appears to be failing.

Read the sell-through across both and a different picture emerges: demand is concentrated among committed buyers in the live room, while the broad, liquid market implied by headline totals does not yet exist.

That is a finding. It is available to anyone willing to examine sell-through rather than turnover totals.

Very few people do, and the reasons are not accidental.

There are in fact two serious objections, not one, and they live at different levels.

The essay has so far argued at the level of market structure: posted prices, cartel supply, guarantees as options, freeport carry, and the buy-in rate. These are claims about how trade is organised and how it clears, and they are the level at which oil and art genuinely rhyme.

The objection that genuinely threatens the thesis must therefore also be structural. It must identify an institutional mechanic that truly does not port from oil.

There is one, and it has a forty-year pedigree in cultural economics.

The second objection concerns how the value of a single artwork is formed. It is equally real, but ultimately less threatening to the broader argument.

The structural objection: art cannot be arbitraged

In 1986 William Baumol argued, in his paper Unnatural Value: Or Art Investment as Floating Crap Game, that art prices behave much like a random walk. The market for celebrated works, he suggested, displays precisely those features that make prices resistant to prediction.

His point was not about taste. It was about structure. There is no mechanism forcing an art price back toward any fundamental because the equilibrating machinery that disciplines other asset markets is absent.

In modern terms, Baumol was describing a market with severe limits to arbitrage.

Three institutional realities create those limits.

First, one cannot short a painting. There is no borrow market and therefore no practical way to express a negative view beyond declining to buy.

Second, one work cannot substitute for another. The arbitrage that keeps crude differentials aligned depends upon physical substitution, something impossible in the art market.

Third, transaction costs remain extraordinarily high. Round-trip costs can consume a quarter of the value of a work, widening the range through which prices may drift before any correction becomes economically worthwhile.

This objection is real, and it deserves to be conceded fully.

Even with perfect market infrastructure, art is unlikely ever to achieve the degree of efficiency observed in commodity markets, because observation and arbitrage are fundamentally different layers of market design.

Platts made oil observable. Futures markets made oil efficient. Art may one day possess the first while never possessing the second.

The second objection operates one level down, at how the price of a single work forms. In most markets the part of a price a model cannot explain is noise around a fundamental. In art it may be the fundamental itself.

A Basquiat and a technically comparable canvas by a forgotten contemporary can differ in price by three orders of magnitude, and no amount of supply analysis or macroeconomic context can recover that gap.

The gap is reputation—but it is important to distinguish what reputation actually explains. That a Basquiat sells for more than an unknown artist is not mysterious at all. Artist identity is the single most powerful explanatory variable in virtually every art pricing model.

The real difficulty is narrower. Reputation cannot itself be anchored to anything outside the market because it is reflexive. High prices reinforce institutional recognition, and institutional recognition reinforces high prices.

Even after reputation is fully accounted for, an irreducible lot-level residual remains. Why this painting, on this evening, attracts extraordinary bidding while another comparable work does not is genuinely difficult to predict.

A barrel of crude is not worth more because it sold for more. A painting sometimes is.

That observation is correct, but it proves less than it first appears.

No illiquid asset can be valued entirely from first principles. Trophy vineyards, distressed private credit, privately held companies, and landmark real estate all require judgement beyond what any model can supply. They are analysed nonetheless.

The reputation component of an artist's market is remarkably persistent. Although its absolute level cannot be derived from theory, changes in that premium can be observed clearly over time. One need not explain why the premium exists to observe that it has expanded or contracted.

The remaining lot-specific uncertainty is idiosyncratic. It cannot be eliminated for an individual work, but across a sufficiently large portfolio it becomes measurable in exactly the same way insurers price uncertainty across thousands of individual risks.

Institutions rarely ask whether a particular painting will exceed its estimate. They ask how an artist's market is evolving, how a collecting category is behaving, or whether a region is becoming more liquid. Those are portfolio questions rather than object questions.

Even commodity markets contain components that resist precise modelling. Oil incorporates geopolitical risk premiums that no supply-and-demand equation can fully recover. Gold is largely a monetary and psychological premium rather than an industrial one, yet both support sophisticated analytical industries.

A socially constructed premium does not remove economics from a market. It merely changes where economics must be applied.

What both objections leave standing

The two objections fail in the same instructive way. Both concern price: one asks whether prices converge; the other asks whether an individual price can ever be fully explained.

Neither objection alters the structural argument developed throughout this essay.

Whether guarantees distort price signals, whether estates manage supply, whether freeports function as inventory storage, or whether the market requires an independent observational layer are institutional questions rather than valuation questions.

Those questions remain exactly as relevant whether or not art ever becomes an efficient market.

The honest conclusion is therefore narrower—but considerably stronger—than the maximal claim.

Art possesses two genuine peculiarities. Its prices may never converge in the way commodity prices do, and the premium attached to an individual work may never be completely modelled.

Neither peculiarity exempts the market's broader structure from economic analysis.

Supply management, liquidity, hedging, cyclicality, capital flows, and institutional incentives remain entirely ordinary analytical problems.

Treating two narrow exceptions as though they place the entire market beyond rigorous analysis has become one of the art market's most successful pieces of self-description.

Concede everything Baumol established and you are left with something that initially sounds like a defeat but is, in fact, the most important development in today's art market.

A painting cannot be made fungible. It is unique, unsubstitutable, and therefore, in the strict economic sense, not a commodity at all.

But recall what the earlier parallels actually established. The market's structure resembles that of a commodity market, and structure—not definition—is what finance requires.

The market therefore did not attempt to make the painting itself a commodity. It did something considerably more sophisticated. It wrapped the painting inside financial instruments that are fungible even though the underlying object is not.

In other words, it commodified the claim rather than the thing.

That distinction explains much of the market's recent evolution.

Art-secured lending transforms a unique object into collateral on a balance sheet. Art investment funds pool many individual works so that investors own interchangeable fund interests rather than specific paintings. Fractional ownership platforms divide a single work into tradable shares. Auction guarantees and irrevocable bids function economically as options written against individual lots.

None of these innovations makes a painting fungible. All of them make financial claims on paintings behave much more like commodity instruments.

Finance has solved this problem repeatedly across entirely different asset classes.

A landmark office tower is unique, yet mortgage-backed securities and REITs successfully commodify claims upon it. Individual private loans differ substantially, yet collateralised loan obligations transform thousands of distinct loans into tradable securities.

Art is simply a much later arrival to an established financial pattern.

The difference is that these wrappers do not inherit an anchor the underlying asset never possessed.

A mortgage security ultimately rests upon rental income. A loan is anchored by contractual cash flows. A fractional interest in a painting ultimately rests upon a work whose price may never converge toward any independently observable fundamental.

Baumol's problem therefore has not disappeared.

It has merely moved one level higher—from valuing the painting to valuing financial claims written upon the painting.

The valuation layer

This argument should not be misunderstood to imply that no valuation profession exists. Quite the opposite.

Members of the Appraisers Association of America and comparable professional bodies produce formal appraisals for insurance, estates, taxation, and collateral. Auction houses and specialist valuation firms likewise value works every day.

That layer is real, necessary, and professionally rigorous.

It is also different from the layer that modern financial instruments quietly assume.

An appraisal is an informed opinion of value constructed from comparable transactions, expertise, and professional judgement. It is not a continuously updated, independently observed market price.

Most appraisals are prepared at a specific moment in time. A two-year loan may therefore continue to rely upon a valuation established when the loan originated, even though the surrounding market has subsequently moved.

Independence also varies across contexts. Some valuations are performed by genuinely independent professionals. Others occur within institutions that simultaneously lend against works, broker transactions, or stand to benefit from higher values.

Financial markets generally assume an observable reference price. Much of the art market instead relies upon periodically updated expert opinion.

Those are not the same thing.

Ownership without infrastructure

Valuation is only half of what these financial instruments quietly assume.

They also assume it is possible to establish ownership, fractional interests, and outstanding claims against a work with confidence.

That infrastructure largely does not exist.

There is no comprehensive public registry recording ownership, liens, or fractional interests in works of art.

The consequences became dramatically visible in the Inigo Philbrick fraud.

Between 2016 and 2019, Philbrick sold multiple ownership interests in the same works while simultaneously pledging many of those works as collateral to lenders who had no knowledge of competing claims.

The problem was not principally valuation.

It was ownership.

Appraisers could verify what a painting was approximately worth. Almost no one could independently verify whether someone else had already sold, financed, or pledged the same asset.

In virtually any public securities market, that inconsistency would be discovered almost immediately.

Within the art market it remained hidden for years because no transparent ownership infrastructure existed to expose it.

Philbrick represents the most dramatic example of a broader structural problem.

Even where fraud is absent, financial claims continue to rely upon valuations that may be stale and ownership records that remain difficult to verify independently.

The market, in other words, is being financialised faster than it is being made legible.

Lending, investment funds, guarantees, and fractional ownership have expanded rapidly.

The observational infrastructure required to support them has not expanded at the same pace.

In the Gulf, the comparison stops being an analogy and becomes a causal relationship.

Gulf sovereign cultural spending is downstream of hydrocarbon revenue. The United Arab Emirates has committed billions of dollars to museums, cultural districts, and arts infrastructure. Christie's reports that the value of modern Middle Eastern art sold through its auctions tripled between 2020 and 2024. Contemporary Middle Eastern sales in London rose sharply during 2025, while Art Basel announced Doha as the location of its first new fair in years.

None of these developments is disconnected from the oil market. They are financed by it.

The capital purchasing Middle Eastern art, constructing museums on Saadiyat Island, and underwriting new cultural institutions is hydrocarbon capital deployed through sovereign investment. It follows energy revenues with a lag that can be analysed long before it appears in auction statistics.

An analyst capable of modelling that relationship possesses a forward-looking indicator for a segment of the art market that conventional research barely observes.

The difficulty is not the absence of information. It is the absence of infrastructure capable of organising that information into a coherent market signal.

Public auction results certainly exist. Bonhams conducts dedicated Middle Eastern sales. Christie's maintains a Dubai auction history. Sotheby's periodically offers comparable material. Yet a handful of public sales each year does not constitute a price series.

There is no benchmark, no differential structure, no continuous market assessment, and no independent institution systematically normalising heterogeneous transactions into an observable market.

Public auction results also represent only a small portion of the market. Dealer transactions, sovereign acquisitions, and private collections account for a substantial share of activity while remaining almost entirely invisible.

This is remarkably similar to the oil market before Platts.

The appropriate response is therefore not to conclude that the market cannot be analysed. It is to construct the strongest signal available from indirect evidence: auction results, sovereign acquisitions, museum expansion, fair activity, and the macroeconomic drivers of regional capital.

Crucially, that methodology should itself be transparent. Readers ought to be able to evaluate the evidence rather than accept the conclusion on authority alone.

African and Global South markets

The same discipline applies to African and broader Global South markets, where the most important development has been structural rather than simply price appreciation.

Auction houses have increasingly integrated these artists into mainstream international contemporary sales instead of isolating them within regional catalogues.

Sotheby's April 2025 African art sale reached approximately £2.4 million while attracting substantially more bidders than the year before and establishing numerous artist records.

Integration matters because it removes the institutional boundary separating these artists from the broader contemporary market. Instead of competing within a regional category, they increasingly compete alongside global contemporaries.

If prices remain strong after that transition, the market is not experiencing a temporary rally. It is undergoing a structural re-rating.

Those two developments carry very different implications for investors, lenders, insurers, and collectors.

It would therefore be a mistake to imagine these markets as waiting quietly to be discovered.

The capital has already arrived.

Sovereign investment, museum construction, expanding international fairs, and increasing institutional participation all demonstrate that financial capital is already moving through these markets.

What remains underdeveloped is the infrastructure that would allow institutions to evaluate that capital with confidence.

The absence of transparency does not prevent financialisation. It merely increases the risks carried by those already participating.

Oil's pricing infrastructure was not built by the major oil companies themselves.

It emerged through independent exchanges and price reporting agencies whose principal asset was credibility rather than inventory.

Those organisations possessed no barrels to sell. Their commercial value came from producing market information that participants trusted precisely because it was independent.

That independence was not incidental.

It was the foundation upon which transparent pricing became possible.

The art market remains organised differently.

Many institutions producing influential market research also lend against art, broker transactions, advise collectors, insure collections, or manage cultural assets.

None of this implies dishonesty.

It does, however, create structural incentives that shape which questions receive attention and which findings are emphasised.

Institutions whose revenues depend upon transaction volume rarely lead their reports with declining sell-through rates or weakening liquidity.

Research inevitably reflects incentives as well as information.

That is not a conspiracy.

It is simply institutional economics.