One fifth of Tradeweb’s Q3 volume earned 25 cents per million while cash credit earned 459 times more

Tradeweb processed $42.995 trillion of “other rates derivatives” in the third quarter—20.4% of its reported volume—but disclosed an average variable fee of only $0.25 per million dollars traded. Cash credit produced $872.5 billion of volume, just 0.4% of the total, at $114.85 per million. The fee-density gap was 459 times. This does not mean one product was mispriced. It reveals why Tradeweb’s $210.3 trillion headline volume cannot be used as a proxy for revenue, commercial importance or customer cost without first separating products.[1][2]

On a simple multiplication of the disclosed volumes and fee rates, other rates derivatives would contribute roughly $10.7 million of variable transaction fees, while cash credit would contribute about $100.2 million—more than nine times as much from one fiftieth of the volume. The calculation is diagnostic because published average fees can be rounded and exclude fixed fees or other revenue. It nevertheless captures the central mechanism: a vast notional flow can be economically thin, while a much smaller market can carry a much richer fee per unit.[1][2]

Tradeweb itself warns readers through its product tables rather than a headline. Rates derivatives, government bonds, credit, equities and money markets use different protocols, durations, fee schedules and notional conventions. A one-million-dollar swap notional is not economically equivalent to one million dollars of corporate bonds traded. Adding them is useful for showing system throughput, but the resulting total is not a common denominator for profitability, liquidity quality or end-investor exposure.[1][2]

A fifth of the volume carried almost no fee density

Other rates derivatives represented one dollar in every five of Tradeweb’s total reported volume. Yet at 25 cents per million, $4 billion of notional would be required to generate $1,000 of variable fees at the published average. Cash credit required about $8.7 million of volume to generate the same amount. The enormous difference reflects product economics: high-frequency or compression-like derivative flows, short-dated instruments and protocol design can produce high notionals with low marginal fees, while cash credit trading can command more per million.[1][2]

The volume measures also carry different duplication and duration characteristics. Derivative notional records contractual reference amounts, not cash exchanged, market value or risk capital. A swap with a large notional may have a much smaller replacement value. Cash bonds involve principal amounts changing hands. Comparing the two without explaining those differences can turn Tradeweb’s aggregate into a misleading “size of market” claim. The platform is large, but its combined notional does not represent money invested, assets held or revenue earned.[1]

Tradeweb’s total average variable fee per million fell to $2.09 from $2.16. That decline could be read as broad pricing pressure. The company’s product table supplies a different explanation: a greater share of low-fee other rates derivatives depresses the blended ratio. Tradeweb reported that excluding that category, average variable fees per million increased to $2.57 from $2.54. The aggregate moved down while the adjusted measure moved up because the denominator mix changed.[1][2]

Notional growth can dilute the apparent take rate

When a low-fee product grows faster than a high-fee product, Tradeweb’s total volume can accelerate while revenue per million falls. That is arithmetic, not necessarily deterioration. Conversely, volume can slow while revenue remains resilient if the mix shifts toward cash credit or other higher-fee products. This is why monthly platform-volume releases should not be translated directly into earnings forecasts. The missing bridge is product-level fee economics, fixed fees, data revenue and the precise mix of trading protocols.[1][2]

The same issue affects comparisons with MarketAxess, Bloomberg, LSEG, CME or interdealer brokers. Their “average daily volume” series may cover different products, count one or both sides, include compression or post-trade activity, or use cash versus notional measures. Even within fixed income, government bonds, corporate bonds, repos and swaps cannot be ranked on a common dollar denominator without methodological adjustments. A platform with lower notional can generate higher transaction revenue because the serviced activity is economically denser.[1]

For customers, fee per million is not total transaction cost. Dealer spreads, market impact, information leakage, clearing, settlement, data, connectivity and internal workflow costs remain outside the published platform fee. A higher Tradeweb fee could accompany better execution or worse execution; the rate alone cannot answer. The investigation therefore does not rank the two products’ value. It shows that the aggregate volume obscures radically different monetisation and should not be presented as if each dollar were comparable.[1][2]

The original finding is a revenue bridge, not a criticism of the platform

Tradeweb did not hide the fee dispersion. It published the product volumes and average variable fees needed to reproduce it. The analytical work is to connect those rows and resist the intuitive but wrong assumption that 20% of volume should contribute roughly 20% of variable fee revenue. At the disclosed averages, other rates derivatives appear to contribute about 2.4% of the combined diagnostic variable fees of the two selected categories, despite representing 98.0% of their combined volume.[1][2]

That two-product comparison is intentionally narrow. It does not allocate all Tradeweb revenue, because fixed transaction fees, subscription fees, market data and other products are outside the calculation. Nor does it account for rounding of fee rates or intra-quarter mix. The approximate $10.7 million and $100.2 million figures should be treated as independently reproduced diagnostics, not reconciled GAAP revenue lines. Their purpose is to expose the order-of-magnitude difference using the company’s own units.[1][2]

A fuller investigation would obtain protocol-level pricing, client-tier schedules and monthly product volumes, then reconcile the result to transaction revenue. It could test whether the low-fee derivative flow is highly automated, whether it creates cross-selling or data value, and how much infrastructure it consumes. The public filing does not disclose those economics. It therefore cannot establish whether 25 cents per million is unusually profitable, loss-making or strategically subsidised.[1]

The finding would be falsified if the stated average fee for other rates derivatives used a materially different volume denominator from the published $42.995 trillion, or if the table’s unit were not dollars per million of the same volume. Tradeweb presents the metrics together, supporting the calculation, but rounding means the exact implied revenue can differ from actual recognised amounts. The 459.4-times fee-density ratio is more reliable than claiming the calculated revenue to the nearest dollar.[1][2]

For market reporting, the practical rule is simple: name the product mix whenever quoting Tradeweb’s total volume. A $1 trillion increase driven by other rates derivatives could carry a very different revenue implication from a $1 trillion increase in cash credit. Volume remains an important liquidity and adoption signal, but only after its notional, fee and protocol definitions are understood. Without that, the largest number in the release becomes the least informative.[1][2]

The dispersion also changes how “record volume” should be communicated. If a new low-fee derivatives protocol adds tens of trillions of notional, it can dominate growth charts while contributing only modest variable fees. That may still be strategically valuable through network effects, data or customer retention, but those benefits require separate evidence. A responsible headline should identify the product and avoid converting aggregate notional into an implied revenue windfall.[1][2]

The reverse error is also possible. Cash-credit volume may look small beside swaps, yet its fee density means a relatively modest change can materially affect transaction revenue. Monitoring only the largest notional category would therefore miss the products most relevant to earnings. Tradeweb’s own disaggregated table provides the necessary correction.[1][2]

The practical audit check is to recalculate the blended fee after removing the largest low-fee category and to preserve both results. That simple sensitivity test prevents mix from being mistaken for a broad price change and can be repeated whenever Tradeweb adds a protocol or revises a volume definition.[1][2]

Report an error or submit an update