A quantitative framework for the second-life economy.

The global secondary market has transitioned from an informal, fragmented liquidation channel into an institutionalized, data-governed marketplace. Across consumer and commercial sectors, physical goods are increasingly priced, financed, insured and traded as durable assets with measurable depreciation curves, bid-ask spreads and yield profiles.

Aggregating mobility, industrial machinery, consumer recommerce and passion collectibles, the observable secondary physical-asset universe exceeded an estimated $2.5 trillion in annual turnover in 2025–2026. Narrow definitions of consumer recommerce — formal managed and peer-to-peer platforms — value the global market at roughly $213–215 billion in 2025, rising toward $236–258 billion in 2026, with projections near $354 billion by 2032. Broader circular-economy definitions capturing all secondhand product transactions value the market near $594 billion in 2025, expanding at a reported 13.6% compound rate toward $1.45 trillion by 2032. The pre-owned automotive sector alone is estimated at $1.83 trillion in 2026, and used heavy construction machinery at $126–134 billion in 2025. [3] [4]

These figures are third-party commercial estimates with different scopes and methods. They overlap — automotive and machinery totals include consumer channels — and must never be added together into a single global number. The suite's market-sizing workspace exists precisely to keep stock, flow and revenue measures separate.

For the tokenisation economy, the structural point is that secondary assets are acquiring the prerequisites of financial assets: standardized identity (serial numbers, digital product passports), observable transaction evidence, professional custody and enforceable claims. An asset that cannot be identified, inspected and re-sold through a documented channel cannot support a credible fractional claim — however large the headline market.

Preowned describes ownership history. Refurbished describes a restoration process. Factory seconds may be unused but fail a manufacturer's first-quality standard. Surplus describes excess inventory; it does not imply a defect. These fields answer different questions and must stay separate in any production dataset — a refurbished item may also be preowned, and a surplus item may be neither.

The distinction carries economic weight. A refurbished electronic device sells against a warranty and a testing standard; a factory-second appliance sells against a disclosed cosmetic defect; a surplus lot sells against grading yield — the share of units that turn out to be sellable at all. Each route has its own cost structure, buyer expectation and return risk, which is why the appraisal lab treats the asset route as an explicit input rather than a label.

Demand evidence is real but uneven. ThredUp's 2026 report projects global secondhand apparel at $393 billion in 2030 — a forecast for apparel, not an observed total for the whole used economy. [1] eBay reports that pre-loved and refurbished goods have represented over 40% of its gross merchandise volume since 2024, based on its platform and a survey of 10,523 shoppers and 1,707 sellers — a platform-specific result, not a universal adoption rate. [2]

Reported consumer breadth is striking — one 2026 estimate puts the share of Americans shopping online for secondhand items at 93% — but breadth of participation is not depth of wallet. Most households buy some used goods; far fewer route significant capital through them. The investment question begins where casual participation ends.

Unlike public equities, pre-owned physical assets are heterogeneous, trade at irregular intervals across fragmented venues, and show non-linear sensitivity to condition, provenance and liquidation urgency. Standardized measurement therefore begins with identity, not price.

Best-in-class operators solve SKU proliferation with deterministic taxonomies. In luxury accessories, Rebag's Clair system codifies over 10,000 bags across more than 50 designers using a five-part hierarchical code (Category.Designer.Model.Style/Material.Size) paired with computer-vision recognition. In fine wine, Liv-ex's LWIN standardizes producer, cuvée, vintage and case configuration. Automotive and heavy machinery rely on VIN decoding and Rouse Services' taxonomy covering over 100,000 equipment makes and models.

From 2026, European regulation is automating this layer. Under the Ecodesign for Sustainable Products Regulation (EU) 2024/1781, the EU Digital Product Passport registry went live in mid-2026, requiring a persistent, machine-readable unique identifier linked to a digital twin carrying verified material composition, origin, repairability scores, spare-part availability and disassembly protocols. The mandated rollout follows a statutory schedule: iron and steel in Q4 2026; EV, light-transport and industrial batteries above 2 kWh on 18 February 2027; construction products in Q2 2027; textiles, apparel, tyres and aluminium in Q3–Q4 2027; furniture in 2028; mattresses and ICT electronics in 2029–2030. [14]

For appraisal, the implication is direct: an asset with a persistent, verifiable identity record supports better comparable matching, better custody control and a stronger ownership claim. An asset without one demands a wider uncertainty band and a heavier diligence file.

Secondary assets trade across vastly different liquidity environments — from thousands of daily wholesale car auctions to once-a-decade fine art consignments — so no single index formula applies universally. Four architectures dominate, matched to asset homogeneity and trading frequency.

The Manheim methodology illustrates the engineering involved: transactions where both price and mileage fall outside 2.6 standard deviations of the model-year/make/body mean are stripped; prices within each class are adjusted for the mileage differential against a trailing 24-month average; strata are sales-weighted on a 24-month rolling volume average and seasonally adjusted. [5]

The practical lesson for an appraiser: know which architecture produced any index you cite, and never compare a repeat-sales figure with a listing-band figure as if they measured the same thing.

The used vehicle market — an estimated $1.74 trillion in 2025 and $1.83 trillion in 2026 — functions as the primary mobility market for price-sensitive households. Post-pandemic supply distortions have hollowed out the affordable tier: the share of US used vehicles priced below $20,000 contracted from 53% in 2019 to roughly 30% in 2025, and by September 2026 average used retail listing prices reached a reported $27,235, with sub-$15,000 inventory down 16.2% year-over-year at an ultra-tight 34 days' supply. [3]

Cross-referencing 3-year European fleet data against 5-year US retail studies exposes a genuine paradox. Over 36 months, EV depreciation has stabilized at a reported 38–42%, nearly converging with petrol vehicles at 35–40%; fleet telemetry from Geotab's 2025 study shows modern EV battery packs degrading at just 2.0% annually, retaining 92–94% state-of-health at years three to four. [7] Yet over 60 months, iSeeCars' 2026 study finds the average EV loses 57–59% of original value against roughly 45% for conventional vehicles and about 40% for hybrids — with high-ticket luxury EVs at 60–65%. [6]

The divergence is driven not by physical battery failure but by range-and-charging architectural obsolescence and primary-market price incentives. The forward supply model adds a stress test: battery electric vehicles were a reported 6.9% of US lease maturities in January 2026 — more than double the 2.8% share a year earlier — and are scheduled to reach 18.7% of maturities by January 2027. [8] As 2023–2024 leased EVs re-enter wholesale lanes simultaneously, remarketers without battery-health certification face residual compression.

Appraisal implication: a vehicle's depreciation curve is a scenario input, not a constant. Powertrain, lease-cycle position and the incentive stance of the manufacturer all belong in the file.

The used construction equipment market — an estimated $126–134 billion in 2025 — shows an exceptionally deep secondary-to-primary ratio: roughly 2.24 million used machines transacted globally in 2025 at an average transaction value near $56,100, more than double the approximately 1.08 million new units sold. Asia-Pacific anchors demand with a reported 48–51% revenue share. Excavators function as the blue-chip currency of industrial remarketing at 43–46% of secondary value, while cranes are the fastest-appreciating niche. [4]

Two structural shifts are redefining valuation: digital auction migration — digital channels captured a reported 42% of global secondary equipment volume in 2025, with market leader RB Global generating $4.7 billion in quarterly gross transaction value in Q2 2026 — and telematics-verified provenance, where machines with logged duty cycles and hour-meter accuracy transact faster at measurable premiums. Conversely, assets older than eight years face tightening credit, and urban zero-emission mandates are accelerating obsolescence for older diesel fleets.

In enterprise IT, the asset-disposition market — an estimated $24–28 billion in 2025, growing at a reported 10.5–11.9% compound rate — has evolved from data-sanitization into a critical secondary hardware supply chain. Its most volatile sub-sector is pre-owned data-center AI compute, where depreciation is governed by silicon release cycles rather than physical wear. Empirical tracking describes a four-stage retention curve: months 0–12 hold 70–90% retention on current-generation scarcity; months 12–18 moderate to 50–70% as next-generation architectures are announced — the optimal liquidation window; months 18–24 see values plunge 50–60% within six months as new silicon reaches volume; months 24–36 stabilize at 20–35% as budget inference and academic workloads absorb supply.

Appraisal implication: for productive equipment, operating hours and verified telemetry are the mileage of the sector; for technology, the product roadmap of the manufacturer is a depreciation schedule in disguise.

Over the 20 years to end-2024, a $1.0 million allocation tracking the Knight Frank Luxury Investment Index reportedly grew to $5.4 million, modestly ahead of the S&P 500's $5.0 million — before storage, insurance and transaction fees. The rebased 2026 report, splitting art into six sub-segments and wine into three, recorded a -0.4% 12-month dip against +2.2% over five years and +38.6% over ten: return leadership rotates sharply across holding horizons. [10]

Over one year, fine art led the 2025 recovery: global art sales rose a reported 4% to $59.6 billion across 41.5 million transactions, led by the US at $26.0 billion, with Impressionist art up 13.6% within the index while Contemporary lagged at -6.0%. [9] Over five years, classic cars led all tracked classes at +17.4%, with the market bifurcating between a soft driver-condition underbelly and surging late-1980s to early-2000s poster cars. Over ten years, rare whisky (+111.9%) and Burgundy (+105.8%) dominated despite a multi-year rate correction — fine wine exhibits a lifecycle J-curve, frequently underperforming immediately after release before accelerating as cases are consumed. By early 2026, key Liv-ex benchmarks had turned positive year-over-year, pre-signaled by the Liv-ex 100 Bid:Offer ratio climbing to 0.7 — the threshold at which buyer-initiated liquidity absorbs post-correction supply. [11]

Appraisal implication: the same object can be a one-year loss, a five-year winner and a ten-year laggard. The holding period is not an administrative detail; it is the thesis. Gross index returns must also be adjusted for holding costs — vault storage, climate control, insurance and maintenance — which are substantially higher for automobiles than for vaulted wine or art.

At the high-velocity end sit cultural commodities traded on electronic order books. The global sneaker resale market — an estimated $10.8 billion in 2025, projected at a 10.2% compound rate toward $23.5 billion by 2034 — rotates rapidly: StockX's 2025–2026 reports show Saucony sales up 239% overtaking Asics as the fastest-growing brand, Vans average resale prices jumping from $102 to $180, and Supreme resale premiums expanding from +11% to +41%. [12]

The trading-card vertical — an estimated $2.01 billion in 2026, projected at a 23.5% compound rate to $13.44 billion by 2035 — is the fastest-expanding liquid asset class on secondary exchanges. GemRate reports professional authenticators graded a record 26.8 million cards in 2025, up 32% year-over-year, with a structural inversion: trading-card games overtook sports cards, surging 95% to 16.8 million cards while sports grading contracted 12%. Grading market share is heavily concentrated — PSA at a reported 72%. [13]

Grading economics are a stochastic expected-value problem. Because a single grade-point drop from gem-mint to mint can destroy 50–70% of a card's market value, net grading profit depends on the empirical probability of each grade, the value at each grade, the marketplace take rate and the grading and shipping fees. For modern high-population cards, only the top grade is consistently profitable after fees; for vintage low-population cards, the top grade can command five to twenty times the raw price. The Rwannie Trading Cards suite models this card-by-card; the point here is that velocity and grading throughput create data — and data creates appraisability.

Appraisal implication: high-velocity cultural assets offer the best price evidence and the worst demand durability. Rotation can reverse within a single season; a premium over retail is a sentiment reading, not a floor.

A foundational finding across every module: gross price appreciation frequently diverges from net investor return because of marketplace transaction friction. Platforms employ radically different fee-incidence architectures depending on whether their asset class is supply-constrained or demand-constrained.

1 · Buyer-premium shifting. In fine art and trophy collectibles, auction houses compete for scarce consignments by compressing seller commissions — down to 0% or negative via guaranteed advances for trophy lots — while buyers pay record premiums: from 2026, 28% on hammer up to $2.0 million, 22% to $8.0 million and 15% above at the major houses. With sales tax and specialized crating, a $40,000 hammer artwork can require roughly $56,400 out of pocket — a 41% uplift — meaning the asset must appreciate 40–50% in hammer value merely to break even on a short round trip.

2 · The zero-seller-fee supply magnet. In high-volume consumer fashion, Vinted scaled past €10 billion in GMV charging sellers nothing while buyers pay a protection fee — maximizing inventory depth against seller-commission rivals charging 10–20%.

3 · Managed consignment take-rate compression. Full-service marketplaces that physically receive, authenticate and fulfill inventory need high gross takes to cover unit labor: The RealReal's sliding scale pays consignors as little as 20% on items under $100 and up to 85–90% on high-value watches, with its blended take rate compressing to a reported 36.4% in Q1 2026 as mix shifted upmarket. Sellers demanding immediacy accept instant-buyout discounts of 25–40% instead.

4 · Value-added service layering. In certified refurbishment and industrial auctions, operators expand take through services rather than core commission: Back Market generated a reported €440 million revenue on €3.5 billion GMV in 2025 — a 12.6% blended take — layering quality-assurance fees, warranties and sponsored placement over a 10% refurbisher commission, while RB Global operates near a 20% service-revenue take on industrial GTV.

Appraisal implication: the fee architecture of the exit channel is as important as the price. The same asset sold through auction, managed consignment, instant buyout or order-book exchange can differ by 30 percentage points of net proceeds. The appraisal lab's selling-fee input is where this research becomes arithmetic.

Build comparable groups by model, age, condition, geography, sale date and sale channel. Prefer completed sales over listings; record shipping, buyer premiums and whether a warranty is included. A median with a sample count and a spread is more useful than an unexplained single price. This suite has no transaction dataset connected: its appraisal output is scenario arithmetic, not a comparable-sales valuation.

Local demand, freight, import costs, servicing networks and currency exposure can dominate margins. A Jamaica-focused analysis should use local completed transactions and verified landed costs rather than importing US resale assumptions unchanged. The same principle applies everywhere: the market you can actually sell into is the market that matters.

An operating asset can earn income before resale — include only net operating cash after maintenance, insurance, downtime and operator expenses, and test idle periods and major repairs separately. Household savings are not rental revenue. On circularity: extending useful life can reduce demand for new production, but savings depend on substitution, transport, refurbishment and energy use. This edition makes no quantified carbon-saving claims; such estimates require category-specific lifecycle evidence.

Finally, the boundaries that protect every number in this suite: stock is what is owned, flow is what trades during a period, and revenue is what a platform earns — never add them together, and never add overlapping marketplace volumes into a global total. Uncertainty bands are user-defined scenario ranges, not statistical confidence intervals. The resale model excludes tax, inflation, financing, irregular cash-flow timing and probability-weighted failure; it does not calculate IRR. These are deliberate limits, and every export carries them.