Turnover and One-Way Costs: Count Every Change Once

Choose the unit before counting activity

Turnover can mean contracts traded, gross traded notional, absolute changes in capital weights or changes in a normalised directional state. These measures are related but not interchangeable. A direction-only research exercise may have states of minus one, zero and plus one without specifying any investment size. Its turnover measure is then a behavioural proxy, not a broker trade count or an account cost record.

Define the initial state, eligible observation range and treatment of the final position. Including an initial acquisition or a closing liquidation increases the measured activity compared with counting only transitions inside the sample. Neither convention is inherently misleading if clearly labelled. Problems arise when one method includes boundary trades and another does not, or when unavailable warmup observations are silently replaced with neutral exposure.

Count a reversal as two units

Take the hypothetical eligible state sequence zero, plus one, plus one, minus one, zero. Successive absolute changes are one, zero, two and one, for a total of four units. The reversal contributes two because the old positive unit is removed and a negative unit is established. Counting only the number of dates on which the state changes would give three, answering a different question.

This distinction is especially important when comparing a rule that moves through neutral with one that reverses directly. A plus-one to zero to minus-one path also totals two absolute units, but spreads them across two transitions. A simple turnover proxy treats the quantities similarly while leaving timing and execution conditions unresolved. It should not imply that the two paths incur identical real costs.

Translate a notional convention into a cost budget

For a separate hypothetical budget, assign $250,000 of reference notional to each unit and assume a one-way cost of two basis points of traded notional. One basis point is 0.01%, so two basis points equal 0.0002 as a decimal. Four turnover units represent $1,000,000 of gross traded notional. Multiplying by 0.0002 produces a $200 illustrative cost budget, without calculating any investment result.

With the same traded notional, assumed one-way rates of one and five basis points produce budgets of $100 and $500. These are sensitivity inputs, not estimates of available execution. If costs are instead specified per contract side, convert using the relevant contract quantities and units. Avoid multiplying a per-contract fee by a notional percentage or treating a round-trip quotation as if it were one-way.

Keep explicit charges and execution costs separate

CME's official clearing and trading fee page explains that charges vary by product and factors such as membership status, volume, venue and transaction type. A generic constant therefore cannot stand in for every real account's fee schedule. Brokerage and other applicable charges also need their own definitions. Date any sourced schedule and identify which category applies before using it in contract-level accounting.

As an invented arithmetic example, suppose exchange, brokerage and other charges sum to $4 per contract side. Entering and later exiting one contract produces $8 of explicit charges. Add an assumed $3 of execution shortfall on each side and the combined round-trip budget becomes $14. These fictional rates illustrate the convention only; they are not quotes for any product, firm or service.

Account for resizing and rolls

A direction can remain positive while its size changes. Increasing a hypothetical exposure from 0.5 to 0.8 units creates 0.3 units of absolute turnover even though the sign is unchanged. A direction-only proxy misses that change by design. If volatility sizing is added, report the resulting resizing activity separately or redefine the turnover measure explicitly so that readers know what the larger number includes.

A contract roll can also involve trading without a directional change. Closing one delivery month and opening another introduces contract-specific legs that are not represented by a continuous reference staying positive. Whether orders are executed through a spread mechanism or separate instructions affects implementation. Do not insert assumed roll costs into a direction-only series without a declared contract map, quantities and accounting convention.

Recognise what a constant rate omits

CME's Liquidity Tool methodology distinguishes bid-ask spreads, order-book depth and cost to trade for a specified quantity. That distinction matters because a small order and a large order need not face the same average execution price. A constant basis-point rate is a useful budgeting simplification, but it does not establish capacity or capture every effect of order size, timing and available depth.

Use a range of explicit assumptions to show cost sensitivity rather than presenting one convenient rate as observed truth. A higher-turnover rule generally consumes more budget under the same constant rate, but a real comparison also needs the trading conditions faced by each rule. The proxy can rank activity under its definition; it cannot independently rank investment merit or demonstrate which method would execute more favourably.

Make the comparison auditable

Publish the state or exposure definition, delay, boundary convention, valid transition count and absolute-change total. If comparing a raw and delayed sequence, keep their eligible windows consistent. A delay can move transitions beyond the displayed endpoint without making the underlying method intrinsically cheaper. Similarly, deleting unavailable observations before differencing can create a transition across an unobserved gap that deserves an explicit policy.

Keep the cost ledger labelled as hypothetical whenever its rates or quantities are hypothetical. Show one-way assumptions and round-trip conversions rather than burying them in a single net figure. A clean activity comparison can be useful even without investment returns: it reveals how often and how far a rule changes its requested exposure. That limited but reproducible conclusion is stronger than an unsupported claim about realised trading costs.

Source record

1. CME Group: Exchange Fees for Clearing and Trading
2. CME Group: Liquidity Tool Methodology