Renesis Insights

Renesis Team

P&L Attribution for Crypto Funds: Price, Funding, Basis, Yield and Fees

P&L attribution splits a fund's return into the drivers behind it. How crypto funds separate price, basis, funding, yield, fees and FX, with a worked month.

P&L attribution splits a portfolio’s profit or loss for a period into the drivers that produced it, so the total is explained line by line instead of reported as one number. Each driver gets a bucket, the buckets add up to the change in value, and anything left over is labeled unexplained.

Banks built the practice around risk factors: delta, gamma, vega, rates and time. A crypto fund’s month looks different. In the worked example below, a hedged BTC book loses $922,310 on its spot leg, makes the same amount on its perp leg, and earns its month from funding, lending interest and a little basis, minus trading fees, a management fee and stablecoins that did not trade at exactly one dollar.

This guide covers how P&L attribution differs from performance attribution, the buckets that matter on a crypto book, the data each one needs, a worked month from 20 October to 19 November 2026, and the breaks that leave P&L unexplained.

Flow of a P&L attribution run: snapshot positions and marks at the cut-off, collect fills, fees, funding and interest, revalue each leg and assign it to one bucket, then reconcile the buckets to the NAV change

P&L attribution vs performance attribution vs P&L explain

The terms overlap, and LPs, administrators and risk teams use them differently.

  • P&L explain is another name for P&L attribution. Wikipedia describes it as an income statement with commentary that attributes the daily change in a portfolio’s value to its root causes, with columns for P&L, P&L explained and P&L unexplained. Product control in a bank produces it every day.

  • P&L predict estimates P&L from sensitivities: delta times the price move, plus gamma, vega and theta terms. As IBKR Quant explains, a full revaluation explains P&L completely by construction, while a sensitivity estimate leaves a residual that grows where convexity is high.

  • Performance attribution explains why a portfolio beat or lagged its benchmark, usually through the Brinson models’ allocation, selection and interaction effects. The CFA Institute Research Foundation’s review covers how those models developed.

Most crypto hedge funds run absolute return strategies without a meaningful benchmark, so LPs are not asking why the fund beat an index. They are asking where the return came from and whether it will repeat. That is P&L attribution. The portfolio manager reads it daily, risk uses the unexplained line as a model check, the administrator ties it to the official NAV, and LPs see it monthly, gross to net.

The buckets that matter on a crypto book

Greek-based attribution was designed for options and rates books. On a crypto fund most of the P&L arrives as cash flows or balance changes, so the buckets look like this:

  • Price (delta). Quantity times the change in mark, per leg. On a hedged book the spot and perp legs offset. Report both gross so LPs can see the hedge worked.

  • Basis. The change in a perp’s or future’s premium over its index. A short perp gains when the premium shrinks. If the perp is marked at index, basis disappears into price or into the unexplained line.

  • Funding. Periodic payments between longs and shorts on perpetual futures. On Binance they settle every 8 hours by default and pass directly between traders, with longs paying shorts when the rate is positive. Hyperliquid settles every hour. Funding is realized cash the moment it is paid.

  • Staking and lending yield. Interest and rewards from protocols. It arrives in three shapes: a balance that grows (Aave’s aTokens raise the holder’s wallet balance as interest accrues), an exchange rate that rises (sUSDS, wstETH), or reward tokens claimed separately.

  • Trading fees and rebates. Maker and taker fees, rebates, gas and borrow costs. Execution cost against the arrival price belongs here too: a TWAP or any other algo should be measured against arrival, not against its own average fill.

  • Stablecoin FX. USDT and USDC trade close to a dollar, not at it. Whether that difference shows up depends on the fund’s valuation policy.

  • Fund-level fees. Management fee accrual and performance fee accrual against the high-water mark. They turn gross P&L into net.

Realized versus unrealized is a second cut across all of these, not a bucket of its own. Funding and interest are realized when paid. A perp’s mark-to-market is unrealized until the position closes.

The data each bucket needs

Attribution is only as good as its inputs. A crypto fund needs:

  • Position snapshots per venue, wallet and protocol, all at the same cut-off, for example 00:00 UTC.

  • Marks from the valuation policy’s price source at that cut-off, with each perp’s mark price and its index stored separately so basis can be measured.

  • Every fill with its fee and fee currency.

  • The funding payment history from each venue. Rebuilding funding from rates and position sizes drifts from what was actually paid.

  • Protocol state: token balances, vault share prices and reward claims, captured at the same cut-off.

  • Stablecoin prices, transfers between venues, and the administrator’s fee accruals.

Then comes the test. Sum the buckets, compare the total with the change in NAV after subscriptions and redemptions, and the gap should be close to zero. A persistent gap is a missing data source, not a market event, and it is the same break that shows up in a shadow NAV reconciliation.

Worked example: one month on a hedged $10M book

The book is illustrative. The market data is real. On 20 October 2026 at 00:00 UTC, a fund with $10.0M NAV:

  • buys 48.85 BTC spot at $81,875, which is $4.0M;

  • sells the same quantity of Binance BTC/USDT perpetuals against $1.0M of USDT margin;

  • supplies $5.0M of USDC to Aave v3 on Ethereum.

It holds everything until 19 November at 00:00 UTC, when BTC is at $62,997, down 23.1%. Funding is the 90 Binance payments in that window, which averaged 7.11% annualized (Renesis market data, captured 19 November 2026 at 05:53 UTC). The Aave rate is the daily supply APY from DefiLlama, which averaged 4.49%. BTC, USDT and USDC prices come from the DefiLlama coins API at each cut-off. Illustrative assumptions: the perp traded 5 bps above index at the start and 1 bp at the end, opening both legs cost 4 bps all-in, and the management fee is 2% a year.

Bar chart of annualized rates from 20 October to 19 November 2026: Binance BTC perp funding 7.11%, OKX 6.68%, Bybit 6.52%, Aave v3 USDC supply 4.49%, 3-month T-bill 4.33%

Bucket

P&L (USD)

bps of NAV

BTC spot, price

−922,310

−922.3

BTC perp, price

+922,310

+922.3

Basis

+1,692

+1.7

Funding

+21,857

+21.9

Lending yield

+18,472

+18.5

Trading fees

−3,200

−3.2

Stablecoin FX

−3,029

−3.0

Gross P&L

+35,792

+35.8

Management fee accrual

−16,438

−16.4

Net P&L

+19,354

+19.4

Unexplained

0

0.0

Four things stand out.

  1. Price nets to zero. Both legs hold the same quantity, so a 23% fall in BTC moves $922,310 from one line to the other and nothing to the total. Reporting only the net would hide that the hedge absorbed almost $1M of offsetting P&L.

  2. Funding is the largest line, and smaller than its rate. At 7.11% on the starting $4.0M notional, 30 days would pay $23,375. The fund received $21,857, because funding is paid on the position’s current value and that value fell with BTC.

  3. The stablecoins cost $3,029. USDT slipped from $0.99925 to $0.99896 and USDC from $0.99991 to $0.99956. That cost $296 on the margin, $1,742 on the Aave deposit and $983 because the perp’s gains were paid in USDT worth slightly less than a dollar. A policy that marks stablecoins at par turns this into a NAV difference with the administrator instead.

  4. Gross carry matched T-bills. $35,792 on $10M over 30 days is 4.35% annualized, against 4.33% on the 3-month T-bill (FRED, 17 November). Net of the illustrative fee it is 2.35%. That is the comparison an allocator will make, and attribution is what makes it visible.

The unexplained line is zero because this example is closed-form. On a live book it is the check that every data source above was captured. For the mechanics of the trade itself, see delta-neutral strategies in crypto.

Common breaks that leave P&L unexplained

  • Funding booked as fees. Some exchange exports put funding payments in the same ledger as trading fees, or net them into realized P&L. The funding line vanishes and the fee line looks wrong.

  • Basis hidden in price. Marking the perp at index instead of its own mark price folds basis into price P&L. On a hedged book it shows up as a small drift between the legs that nobody can explain.

  • DeFi yield counted twice. When an aToken balance grows, the extra tokens are yield. A system that also revalues the larger balance as a price gain books the same interest twice. Exchange-rate tokens such as wstETH need the same care: the rise in the rate is yield, and only the underlying asset’s move is price.

  • Stablecoin depegs. Marking stablecoins at par hides a loss until redemption, then books it in one day.

  • Cut-off mismatches. A funding payment at 00:00 UTC belongs to one day or the next. Venues snapshotted minutes apart create P&L that reverses the following day.

  • Transfers counted as P&L. Assets in transit between an exchange and a wallet drop out of one snapshot before they reach the other.

Each of these is a data problem, which is why perp P&L is harder than it looks. Hyperliquid is a good example: calculating P&L there correctly means combining fills, hourly funding and transfers into one consistent ledger.

Reporting P&L attribution to LPs

LPs read attribution to judge whether the return will repeat. The monthly report should:

  • show each bucket in dollars and basis points of NAV, gross to net;

  • keep the same buckets month to month and disclose any change in method;

  • show hedged legs gross, not only the net;

  • separate carry (funding, basis and yield) from directional P&L;

  • tie the total to the administrator’s NAV and to the performance fee accrual.

A fund that runs a basis or funding strategy is selling carry. If the attribution shows most of the return came from price, the LP will ask why.

Where Renesis fits

Attribution breaks when its inputs live in different places: funding in an exchange’s payment history, lending interest inside a token balance that grows by itself, fills in an execution system and prices in a spreadsheet. Renesis values, reconciles and reports any balance sheet that holds digital assets, so every fill, funding payment and yield accrual lands in the fund’s books as its own line. The month’s P&L then splits into price, basis, funding, yield, fees and FX with nothing left over, and that split flows straight into NAV and the investor report in the Renesis portfolio management system.

Frequently asked questions

What is P&L attribution?

P&L attribution is the breakdown of a portfolio’s profit or loss over a period into the drivers that caused it, such as price moves, funding, yield and fees. The buckets add up to the total, and the remainder is reported as unexplained P&L. It is also called P&L explain.

How do you calculate P&L in trading?

For a position, P&L is the end value minus the start value, plus cash received such as funding and interest, minus fees, adjusted for any capital added or withdrawn. For a fund, the same calculation runs across every position and venue at one cut-off time. Attribution then splits that total by driver.

What is P&L explain?

P&L explain is the bank term for P&L attribution: a daily report from product control that attributes the change in a trading book’s value to market moves, time, new trades and other causes. Its unexplained column is a check on the risk models and the data behind them.

What is the difference between realized and unrealized P&L?

Realized P&L comes from closed positions and cash already received, such as a funding payment, or a fee already paid. Unrealized P&L is the mark-to-market change on positions still open. Both belong in NAV and in attribution. The split matters for tax and for how much of the return is locked in.

Still not sure if Renesis is right for you?

Ask ChatGPT, Claude or Perplexity what they have to talk about us. Click below to ask your favorite AI about us:

Built by builders.
For builders.

We're a DeFi-native team shipping fast. No enterprise sales cycles, no bloated pricing. Start free, talk to us when you're ready.

Renesis Insights

Renesis Team

P&L Attribution for Crypto Funds: Price, Funding, Basis, Yield and Fees

P&L attribution splits a fund's return into the drivers behind it. How crypto funds separate price, basis, funding, yield, fees and FX, with a worked month.

P&L attribution splits a portfolio’s profit or loss for a period into the drivers that produced it, so the total is explained line by line instead of reported as one number. Each driver gets a bucket, the buckets add up to the change in value, and anything left over is labeled unexplained.

Banks built the practice around risk factors: delta, gamma, vega, rates and time. A crypto fund’s month looks different. In the worked example below, a hedged BTC book loses $922,310 on its spot leg, makes the same amount on its perp leg, and earns its month from funding, lending interest and a little basis, minus trading fees, a management fee and stablecoins that did not trade at exactly one dollar.

This guide covers how P&L attribution differs from performance attribution, the buckets that matter on a crypto book, the data each one needs, a worked month from 20 October to 19 November 2026, and the breaks that leave P&L unexplained.

Flow of a P&L attribution run: snapshot positions and marks at the cut-off, collect fills, fees, funding and interest, revalue each leg and assign it to one bucket, then reconcile the buckets to the NAV change

P&L attribution vs performance attribution vs P&L explain

The terms overlap, and LPs, administrators and risk teams use them differently.

  • P&L explain is another name for P&L attribution. Wikipedia describes it as an income statement with commentary that attributes the daily change in a portfolio’s value to its root causes, with columns for P&L, P&L explained and P&L unexplained. Product control in a bank produces it every day.

  • P&L predict estimates P&L from sensitivities: delta times the price move, plus gamma, vega and theta terms. As IBKR Quant explains, a full revaluation explains P&L completely by construction, while a sensitivity estimate leaves a residual that grows where convexity is high.

  • Performance attribution explains why a portfolio beat or lagged its benchmark, usually through the Brinson models’ allocation, selection and interaction effects. The CFA Institute Research Foundation’s review covers how those models developed.

Most crypto hedge funds run absolute return strategies without a meaningful benchmark, so LPs are not asking why the fund beat an index. They are asking where the return came from and whether it will repeat. That is P&L attribution. The portfolio manager reads it daily, risk uses the unexplained line as a model check, the administrator ties it to the official NAV, and LPs see it monthly, gross to net.

The buckets that matter on a crypto book

Greek-based attribution was designed for options and rates books. On a crypto fund most of the P&L arrives as cash flows or balance changes, so the buckets look like this:

  • Price (delta). Quantity times the change in mark, per leg. On a hedged book the spot and perp legs offset. Report both gross so LPs can see the hedge worked.

  • Basis. The change in a perp’s or future’s premium over its index. A short perp gains when the premium shrinks. If the perp is marked at index, basis disappears into price or into the unexplained line.

  • Funding. Periodic payments between longs and shorts on perpetual futures. On Binance they settle every 8 hours by default and pass directly between traders, with longs paying shorts when the rate is positive. Hyperliquid settles every hour. Funding is realized cash the moment it is paid.

  • Staking and lending yield. Interest and rewards from protocols. It arrives in three shapes: a balance that grows (Aave’s aTokens raise the holder’s wallet balance as interest accrues), an exchange rate that rises (sUSDS, wstETH), or reward tokens claimed separately.

  • Trading fees and rebates. Maker and taker fees, rebates, gas and borrow costs. Execution cost against the arrival price belongs here too: a TWAP or any other algo should be measured against arrival, not against its own average fill.

  • Stablecoin FX. USDT and USDC trade close to a dollar, not at it. Whether that difference shows up depends on the fund’s valuation policy.

  • Fund-level fees. Management fee accrual and performance fee accrual against the high-water mark. They turn gross P&L into net.

Realized versus unrealized is a second cut across all of these, not a bucket of its own. Funding and interest are realized when paid. A perp’s mark-to-market is unrealized until the position closes.

The data each bucket needs

Attribution is only as good as its inputs. A crypto fund needs:

  • Position snapshots per venue, wallet and protocol, all at the same cut-off, for example 00:00 UTC.

  • Marks from the valuation policy’s price source at that cut-off, with each perp’s mark price and its index stored separately so basis can be measured.

  • Every fill with its fee and fee currency.

  • The funding payment history from each venue. Rebuilding funding from rates and position sizes drifts from what was actually paid.

  • Protocol state: token balances, vault share prices and reward claims, captured at the same cut-off.

  • Stablecoin prices, transfers between venues, and the administrator’s fee accruals.

Then comes the test. Sum the buckets, compare the total with the change in NAV after subscriptions and redemptions, and the gap should be close to zero. A persistent gap is a missing data source, not a market event, and it is the same break that shows up in a shadow NAV reconciliation.

Worked example: one month on a hedged $10M book

The book is illustrative. The market data is real. On 20 October 2026 at 00:00 UTC, a fund with $10.0M NAV:

  • buys 48.85 BTC spot at $81,875, which is $4.0M;

  • sells the same quantity of Binance BTC/USDT perpetuals against $1.0M of USDT margin;

  • supplies $5.0M of USDC to Aave v3 on Ethereum.

It holds everything until 19 November at 00:00 UTC, when BTC is at $62,997, down 23.1%. Funding is the 90 Binance payments in that window, which averaged 7.11% annualized (Renesis market data, captured 19 November 2026 at 05:53 UTC). The Aave rate is the daily supply APY from DefiLlama, which averaged 4.49%. BTC, USDT and USDC prices come from the DefiLlama coins API at each cut-off. Illustrative assumptions: the perp traded 5 bps above index at the start and 1 bp at the end, opening both legs cost 4 bps all-in, and the management fee is 2% a year.

Bar chart of annualized rates from 20 October to 19 November 2026: Binance BTC perp funding 7.11%, OKX 6.68%, Bybit 6.52%, Aave v3 USDC supply 4.49%, 3-month T-bill 4.33%

Bucket

P&L (USD)

bps of NAV

BTC spot, price

−922,310

−922.3

BTC perp, price

+922,310

+922.3

Basis

+1,692

+1.7

Funding

+21,857

+21.9

Lending yield

+18,472

+18.5

Trading fees

−3,200

−3.2

Stablecoin FX

−3,029

−3.0

Gross P&L

+35,792

+35.8

Management fee accrual

−16,438

−16.4

Net P&L

+19,354

+19.4

Unexplained

0

0.0

Four things stand out.

  1. Price nets to zero. Both legs hold the same quantity, so a 23% fall in BTC moves $922,310 from one line to the other and nothing to the total. Reporting only the net would hide that the hedge absorbed almost $1M of offsetting P&L.

  2. Funding is the largest line, and smaller than its rate. At 7.11% on the starting $4.0M notional, 30 days would pay $23,375. The fund received $21,857, because funding is paid on the position’s current value and that value fell with BTC.

  3. The stablecoins cost $3,029. USDT slipped from $0.99925 to $0.99896 and USDC from $0.99991 to $0.99956. That cost $296 on the margin, $1,742 on the Aave deposit and $983 because the perp’s gains were paid in USDT worth slightly less than a dollar. A policy that marks stablecoins at par turns this into a NAV difference with the administrator instead.

  4. Gross carry matched T-bills. $35,792 on $10M over 30 days is 4.35% annualized, against 4.33% on the 3-month T-bill (FRED, 17 November). Net of the illustrative fee it is 2.35%. That is the comparison an allocator will make, and attribution is what makes it visible.

The unexplained line is zero because this example is closed-form. On a live book it is the check that every data source above was captured. For the mechanics of the trade itself, see delta-neutral strategies in crypto.

Common breaks that leave P&L unexplained

  • Funding booked as fees. Some exchange exports put funding payments in the same ledger as trading fees, or net them into realized P&L. The funding line vanishes and the fee line looks wrong.

  • Basis hidden in price. Marking the perp at index instead of its own mark price folds basis into price P&L. On a hedged book it shows up as a small drift between the legs that nobody can explain.

  • DeFi yield counted twice. When an aToken balance grows, the extra tokens are yield. A system that also revalues the larger balance as a price gain books the same interest twice. Exchange-rate tokens such as wstETH need the same care: the rise in the rate is yield, and only the underlying asset’s move is price.

  • Stablecoin depegs. Marking stablecoins at par hides a loss until redemption, then books it in one day.

  • Cut-off mismatches. A funding payment at 00:00 UTC belongs to one day or the next. Venues snapshotted minutes apart create P&L that reverses the following day.

  • Transfers counted as P&L. Assets in transit between an exchange and a wallet drop out of one snapshot before they reach the other.

Each of these is a data problem, which is why perp P&L is harder than it looks. Hyperliquid is a good example: calculating P&L there correctly means combining fills, hourly funding and transfers into one consistent ledger.

Reporting P&L attribution to LPs

LPs read attribution to judge whether the return will repeat. The monthly report should:

  • show each bucket in dollars and basis points of NAV, gross to net;

  • keep the same buckets month to month and disclose any change in method;

  • show hedged legs gross, not only the net;

  • separate carry (funding, basis and yield) from directional P&L;

  • tie the total to the administrator’s NAV and to the performance fee accrual.

A fund that runs a basis or funding strategy is selling carry. If the attribution shows most of the return came from price, the LP will ask why.

Where Renesis fits

Attribution breaks when its inputs live in different places: funding in an exchange’s payment history, lending interest inside a token balance that grows by itself, fills in an execution system and prices in a spreadsheet. Renesis values, reconciles and reports any balance sheet that holds digital assets, so every fill, funding payment and yield accrual lands in the fund’s books as its own line. The month’s P&L then splits into price, basis, funding, yield, fees and FX with nothing left over, and that split flows straight into NAV and the investor report in the Renesis portfolio management system.

Frequently asked questions

What is P&L attribution?

P&L attribution is the breakdown of a portfolio’s profit or loss over a period into the drivers that caused it, such as price moves, funding, yield and fees. The buckets add up to the total, and the remainder is reported as unexplained P&L. It is also called P&L explain.

How do you calculate P&L in trading?

For a position, P&L is the end value minus the start value, plus cash received such as funding and interest, minus fees, adjusted for any capital added or withdrawn. For a fund, the same calculation runs across every position and venue at one cut-off time. Attribution then splits that total by driver.

What is P&L explain?

P&L explain is the bank term for P&L attribution: a daily report from product control that attributes the change in a trading book’s value to market moves, time, new trades and other causes. Its unexplained column is a check on the risk models and the data behind them.

What is the difference between realized and unrealized P&L?

Realized P&L comes from closed positions and cash already received, such as a funding payment, or a fee already paid. Unrealized P&L is the mark-to-market change on positions still open. Both belong in NAV and in attribution. The split matters for tax and for how much of the return is locked in.

Renesis Insights

Renesis Team

P&L Attribution for Crypto Funds: Price, Funding, Basis, Yield and Fees

P&L attribution splits a fund's return into the drivers behind it. How crypto funds separate price, basis, funding, yield, fees and FX, with a worked month.

P&L attribution splits a portfolio’s profit or loss for a period into the drivers that produced it, so the total is explained line by line instead of reported as one number. Each driver gets a bucket, the buckets add up to the change in value, and anything left over is labeled unexplained.

Banks built the practice around risk factors: delta, gamma, vega, rates and time. A crypto fund’s month looks different. In the worked example below, a hedged BTC book loses $922,310 on its spot leg, makes the same amount on its perp leg, and earns its month from funding, lending interest and a little basis, minus trading fees, a management fee and stablecoins that did not trade at exactly one dollar.

This guide covers how P&L attribution differs from performance attribution, the buckets that matter on a crypto book, the data each one needs, a worked month from 20 October to 19 November 2026, and the breaks that leave P&L unexplained.

Flow of a P&L attribution run: snapshot positions and marks at the cut-off, collect fills, fees, funding and interest, revalue each leg and assign it to one bucket, then reconcile the buckets to the NAV change

P&L attribution vs performance attribution vs P&L explain

The terms overlap, and LPs, administrators and risk teams use them differently.

  • P&L explain is another name for P&L attribution. Wikipedia describes it as an income statement with commentary that attributes the daily change in a portfolio’s value to its root causes, with columns for P&L, P&L explained and P&L unexplained. Product control in a bank produces it every day.

  • P&L predict estimates P&L from sensitivities: delta times the price move, plus gamma, vega and theta terms. As IBKR Quant explains, a full revaluation explains P&L completely by construction, while a sensitivity estimate leaves a residual that grows where convexity is high.

  • Performance attribution explains why a portfolio beat or lagged its benchmark, usually through the Brinson models’ allocation, selection and interaction effects. The CFA Institute Research Foundation’s review covers how those models developed.

Most crypto hedge funds run absolute return strategies without a meaningful benchmark, so LPs are not asking why the fund beat an index. They are asking where the return came from and whether it will repeat. That is P&L attribution. The portfolio manager reads it daily, risk uses the unexplained line as a model check, the administrator ties it to the official NAV, and LPs see it monthly, gross to net.

The buckets that matter on a crypto book

Greek-based attribution was designed for options and rates books. On a crypto fund most of the P&L arrives as cash flows or balance changes, so the buckets look like this:

  • Price (delta). Quantity times the change in mark, per leg. On a hedged book the spot and perp legs offset. Report both gross so LPs can see the hedge worked.

  • Basis. The change in a perp’s or future’s premium over its index. A short perp gains when the premium shrinks. If the perp is marked at index, basis disappears into price or into the unexplained line.

  • Funding. Periodic payments between longs and shorts on perpetual futures. On Binance they settle every 8 hours by default and pass directly between traders, with longs paying shorts when the rate is positive. Hyperliquid settles every hour. Funding is realized cash the moment it is paid.

  • Staking and lending yield. Interest and rewards from protocols. It arrives in three shapes: a balance that grows (Aave’s aTokens raise the holder’s wallet balance as interest accrues), an exchange rate that rises (sUSDS, wstETH), or reward tokens claimed separately.

  • Trading fees and rebates. Maker and taker fees, rebates, gas and borrow costs. Execution cost against the arrival price belongs here too: a TWAP or any other algo should be measured against arrival, not against its own average fill.

  • Stablecoin FX. USDT and USDC trade close to a dollar, not at it. Whether that difference shows up depends on the fund’s valuation policy.

  • Fund-level fees. Management fee accrual and performance fee accrual against the high-water mark. They turn gross P&L into net.

Realized versus unrealized is a second cut across all of these, not a bucket of its own. Funding and interest are realized when paid. A perp’s mark-to-market is unrealized until the position closes.

The data each bucket needs

Attribution is only as good as its inputs. A crypto fund needs:

  • Position snapshots per venue, wallet and protocol, all at the same cut-off, for example 00:00 UTC.

  • Marks from the valuation policy’s price source at that cut-off, with each perp’s mark price and its index stored separately so basis can be measured.

  • Every fill with its fee and fee currency.

  • The funding payment history from each venue. Rebuilding funding from rates and position sizes drifts from what was actually paid.

  • Protocol state: token balances, vault share prices and reward claims, captured at the same cut-off.

  • Stablecoin prices, transfers between venues, and the administrator’s fee accruals.

Then comes the test. Sum the buckets, compare the total with the change in NAV after subscriptions and redemptions, and the gap should be close to zero. A persistent gap is a missing data source, not a market event, and it is the same break that shows up in a shadow NAV reconciliation.

Worked example: one month on a hedged $10M book

The book is illustrative. The market data is real. On 20 October 2026 at 00:00 UTC, a fund with $10.0M NAV:

  • buys 48.85 BTC spot at $81,875, which is $4.0M;

  • sells the same quantity of Binance BTC/USDT perpetuals against $1.0M of USDT margin;

  • supplies $5.0M of USDC to Aave v3 on Ethereum.

It holds everything until 19 November at 00:00 UTC, when BTC is at $62,997, down 23.1%. Funding is the 90 Binance payments in that window, which averaged 7.11% annualized (Renesis market data, captured 19 November 2026 at 05:53 UTC). The Aave rate is the daily supply APY from DefiLlama, which averaged 4.49%. BTC, USDT and USDC prices come from the DefiLlama coins API at each cut-off. Illustrative assumptions: the perp traded 5 bps above index at the start and 1 bp at the end, opening both legs cost 4 bps all-in, and the management fee is 2% a year.

Bar chart of annualized rates from 20 October to 19 November 2026: Binance BTC perp funding 7.11%, OKX 6.68%, Bybit 6.52%, Aave v3 USDC supply 4.49%, 3-month T-bill 4.33%

Bucket

P&L (USD)

bps of NAV

BTC spot, price

−922,310

−922.3

BTC perp, price

+922,310

+922.3

Basis

+1,692

+1.7

Funding

+21,857

+21.9

Lending yield

+18,472

+18.5

Trading fees

−3,200

−3.2

Stablecoin FX

−3,029

−3.0

Gross P&L

+35,792

+35.8

Management fee accrual

−16,438

−16.4

Net P&L

+19,354

+19.4

Unexplained

0

0.0

Four things stand out.

  1. Price nets to zero. Both legs hold the same quantity, so a 23% fall in BTC moves $922,310 from one line to the other and nothing to the total. Reporting only the net would hide that the hedge absorbed almost $1M of offsetting P&L.

  2. Funding is the largest line, and smaller than its rate. At 7.11% on the starting $4.0M notional, 30 days would pay $23,375. The fund received $21,857, because funding is paid on the position’s current value and that value fell with BTC.

  3. The stablecoins cost $3,029. USDT slipped from $0.99925 to $0.99896 and USDC from $0.99991 to $0.99956. That cost $296 on the margin, $1,742 on the Aave deposit and $983 because the perp’s gains were paid in USDT worth slightly less than a dollar. A policy that marks stablecoins at par turns this into a NAV difference with the administrator instead.

  4. Gross carry matched T-bills. $35,792 on $10M over 30 days is 4.35% annualized, against 4.33% on the 3-month T-bill (FRED, 17 November). Net of the illustrative fee it is 2.35%. That is the comparison an allocator will make, and attribution is what makes it visible.

The unexplained line is zero because this example is closed-form. On a live book it is the check that every data source above was captured. For the mechanics of the trade itself, see delta-neutral strategies in crypto.

Common breaks that leave P&L unexplained

  • Funding booked as fees. Some exchange exports put funding payments in the same ledger as trading fees, or net them into realized P&L. The funding line vanishes and the fee line looks wrong.

  • Basis hidden in price. Marking the perp at index instead of its own mark price folds basis into price P&L. On a hedged book it shows up as a small drift between the legs that nobody can explain.

  • DeFi yield counted twice. When an aToken balance grows, the extra tokens are yield. A system that also revalues the larger balance as a price gain books the same interest twice. Exchange-rate tokens such as wstETH need the same care: the rise in the rate is yield, and only the underlying asset’s move is price.

  • Stablecoin depegs. Marking stablecoins at par hides a loss until redemption, then books it in one day.

  • Cut-off mismatches. A funding payment at 00:00 UTC belongs to one day or the next. Venues snapshotted minutes apart create P&L that reverses the following day.

  • Transfers counted as P&L. Assets in transit between an exchange and a wallet drop out of one snapshot before they reach the other.

Each of these is a data problem, which is why perp P&L is harder than it looks. Hyperliquid is a good example: calculating P&L there correctly means combining fills, hourly funding and transfers into one consistent ledger.

Reporting P&L attribution to LPs

LPs read attribution to judge whether the return will repeat. The monthly report should:

  • show each bucket in dollars and basis points of NAV, gross to net;

  • keep the same buckets month to month and disclose any change in method;

  • show hedged legs gross, not only the net;

  • separate carry (funding, basis and yield) from directional P&L;

  • tie the total to the administrator’s NAV and to the performance fee accrual.

A fund that runs a basis or funding strategy is selling carry. If the attribution shows most of the return came from price, the LP will ask why.

Where Renesis fits

Attribution breaks when its inputs live in different places: funding in an exchange’s payment history, lending interest inside a token balance that grows by itself, fills in an execution system and prices in a spreadsheet. Renesis values, reconciles and reports any balance sheet that holds digital assets, so every fill, funding payment and yield accrual lands in the fund’s books as its own line. The month’s P&L then splits into price, basis, funding, yield, fees and FX with nothing left over, and that split flows straight into NAV and the investor report in the Renesis portfolio management system.

Frequently asked questions

What is P&L attribution?

P&L attribution is the breakdown of a portfolio’s profit or loss over a period into the drivers that caused it, such as price moves, funding, yield and fees. The buckets add up to the total, and the remainder is reported as unexplained P&L. It is also called P&L explain.

How do you calculate P&L in trading?

For a position, P&L is the end value minus the start value, plus cash received such as funding and interest, minus fees, adjusted for any capital added or withdrawn. For a fund, the same calculation runs across every position and venue at one cut-off time. Attribution then splits that total by driver.

What is P&L explain?

P&L explain is the bank term for P&L attribution: a daily report from product control that attributes the change in a trading book’s value to market moves, time, new trades and other causes. Its unexplained column is a check on the risk models and the data behind them.

What is the difference between realized and unrealized P&L?

Realized P&L comes from closed positions and cash already received, such as a funding payment, or a fee already paid. Unrealized P&L is the mark-to-market change on positions still open. Both belong in NAV and in attribution. The split matters for tax and for how much of the return is locked in.