ILPA Reconciliation: Signed-Convention Formulas for LP Auditability

Abstract

This paper examines the mathematical structure of the ILPA 2025 Reporting Template (Version 2.0, released January 2025) and the auditability properties that follow from its signed-convention arithmetic. We walk through the four core reconciliation formulas — NAV roll-forward, commitment reconciliation, fee netting, and carried-interest accrual — in their full spec-canonical form and show why each one composes to the penny when the signed convention is applied consistently. We then argue that a $1 tolerance on these formulas is not a user-experience nicety but a meaningful auditability guarantee: it provides a pre-commit arithmetic check that catches import errors before they reach a board package, and a reproducible proof that the figures a GP reports are internally consistent.

The paper is written for both audiences within an institutional LP — operations teams (LP controllers, fund-administration counterparts, audit-coordination roles) for whom reconciliation is daily work, and investment teams (IC members, consultants, fiduciary counsel) for whom reconciliation discipline shapes how they read the underlying performance figures. It assumes familiarity with basic LP accounting concepts (capital calls, distributions, NAV, management fees, carried interest) but does not assume prior exposure to the ILPA template itself. Every formula and line-item enumeration in this paper is traceable to the ILPA 2025 Reporting Template Definitions (released January 2025) and the RT Supplemental Guidance Formulas Overview.

1. Why Reconciliation Matters Before You Read the Numbers

A quarterly GP report arrives. The accompanying capital account statement shows figures your analyst enters into the portfolio system. The board deck uses them in its quarterly review. Two quarters later, an auditor's routine check surfaces a discrepancy — a number that doesn't reconcile against the template's roll-forward formulas.

The error is small in isolation. But across a $200M program, even a single misinterpreted line item can produce a valuation error of several million dollars propagating through two quarterly board packages, multiple consultant reports, and at least one allocation decision. The auditor asks when the error entered the system. The answer is: the first quarter the report was processed, because nobody reconciled against the signed-convention roll-forward.

This paper is about making sure that doesn't happen. Not because anyone involved was careless — quite the opposite. LP operations teams are meticulous. But the ILPA template's mathematical structure is specific enough that a human reading the statement can arrive at an ending balance by informal mental arithmetic even while misinterpreting the sign convention, and automation that replicates that human workflow inherits the same ambiguity. The remedy is not more careful reading. It is arithmetic discipline applied to the template as a signed-convention algebra.

2. The ILPA Signed Convention

The ILPA 2025 Reporting Template is designed so that every reconciliation formula is pure addition — you never subtract. Subtraction is encoded in the sign. If every signed term is added in order, the formula composes to the reported result.

There is an important nuance that catches most teams reading the spec for the first time. The template has two different sign conventions, and which one applies depends on whether you are looking at the input cell or the formula that consumes it:

Input cells for reductions (Fee Offsets, Fee Waivers, Management Fee Rebates, Carry Paid) carry the annotation "input positive values." The user enters a positive number — e.g., 75,000 for $75K of monitoring-fee offsets.
The formula row "Less [Item]" applies the negative sign algorithmically. The formula displays the value as negative once the computation runs (e.g., -75,000 appears in the "Less Offsets to Management Fees" row).
Result: the final formula Net = Gross + (Less Rebate) + (Less Waivers) + (Less Offsets) always adds, and the signed template values compose correctly.

Ingestion systems that read the raw template must be aware of both conventions. A template exported after formula computation will show the signed (negative) values. A template exported pre-computation — or filled in manually by a GP — will show positive user-entered values awaiting sign application. Both are valid states; the reconciliation workflow must handle them consistently.

The sign convention by line-item category is as follows:

Template section	Line item	Input state	Signed state in formula
NAV Reconciliation	Beginning NAV (Net of Carry)	Signed as-is (typically positive)	Same
	Contributions – Cash & Non-Cash	Positive	Positive
	Distributions – Cash & Non-Cash	Negative	Negative
	Offering/Syndication Costs	Negative	Negative
	Placement Fees	Negative	Negative
	Partner Transfers	Signed (LP-specific)	Same
	Net Realized Gain / (Loss)	Signed	Same
	Net Unrealized Gain / (Loss)	Signed	Same
Fee Netting	Management Fees Gross	Negative	Negative
	Management Fee Rebate	Positive (input)	Negative (post-sign)
	Fee Waivers	Positive (input)	Negative (post-sign)
	Offsets to Management Fees	Positive (input)	Negative (post-sign)
Commitment Reconciliation	Beginning Unfunded Commitment	Positive	Positive
	(Less Contributions)	Negative	Negative
	Plus Recallable Distributions	Positive	Positive
	(Less Expired/Released Commitments)	Negative	Negative
	+/- Other Unfunded Adjustment	Signed	Same
Carry Accrual	Starting Period Balance	Negative (GP's claim = LP liability)	Same
	Carried Interest Accrued (Unrealized Profits)	Negative (accruing more GP claim)	Same
	Carried Interest Earned (Realized Profits)	Negative	Same
	Carried Interest – Paid During Period	Positive (input)	Positive — reduces liability magnitude
	Returned Clawback	Positive	Positive

Every reconciliation formula that follows is an addition of its signed terms. If the signs are read correctly, the formula composes exactly to the reported ending balance. If any term's sign is misinterpreted — even once — the formula will be off by twice that term's magnitude, because a correct negative value becomes a positive value (or vice versa), swinging the result by 2× the line item's absolute magnitude.

This is the core of auditability. A full ILPA reconciliation pass is equivalent to verifying that a set of signed numbers sums to the reported result. If the sum equals the reported figure within a $1 tolerance, the report is internally consistent. If it doesn't, there is a sign error, a missing value, or a transcription error — and you know before the data reaches a board package.

3. The Four Core Reconciliation Formulas

The ILPA 2025 template's reconciliation logic is carried by four formulas. Every other derived figure in a quarterly report is downstream of one of these. Understanding these four and reconciling against them is the core of institutional LP operations discipline.

Reconciliation	Formula (all terms added with their signs)	Diagnostic question
NAV roll-forward	`Beginning NAV + Total Cash/Non-Cash Flows + Total Net Operating Income + Total Net Realized/Unrealized Gain = Ending NAV`	Did the LP's economic claim evolve correctly?
Commitment	`Beginning Unfunded + Contributions + Recallable + Expired/Released + Other Adjustment = Ending Unfunded`	Does the remaining commitment match cash-call, recallable, release, and adjustment activity?
Fee netting	`Management Fees Gross + Rebate + Waivers + Offsets = Management Fees Net`	Is the LP's true fee burden after credits correctly stated?
Carry accrual	`Starting Balance + Accrued (Unrealized) + Earned (Realized) + Paid + Returned Clawback = Ending Balance`	Does the accrued carry position match the waterfall activity?

The sections that follow walk through each formula with a worked example sourced from the ILPA-published sample template.

3.1 NAV Roll-Forward

The NAV roll-forward in the ILPA 2025 template is a three-level composition. The top-level formula adds four aggregates; each aggregate is itself a reconciliation.

Top-level:

Ending NAV (Net of Carry) =
  Beginning NAV (Net of Carry)
+ Total Cash / Non-Cash Flows
+ Total Net Operating Income / (Expense)
+ Total Net Realized and Unrealized Gain / (Loss)

Total Cash / Non-Cash Flows is itself a sum of five line items:

Total Cash / Non-Cash Flows =
  Contributions – Cash & Non-Cash
+ Distributions – Cash & Non-Cash
+ Offering/Syndication Costs
+ Placement Fees
+ Partner Transfers

This is where many LP reconciliation implementations fall short. The simplified "Contributions + Distributions" version of the formula — which appears in casual summaries of ILPA — omits three line items that the spec explicitly includes: Offering/Syndication Costs (a direct reduction to partners' capital for fees/costs to market the fund), Placement Fees (fees paid for fundraising services), and Partner Transfers (transfers of a partner's interest that affect the LP-specific capital account). A GP template that reports any of these as non-zero will fail reconciliation against a two-term formula.

Working example from the ILPA sample template (QTD values for LP #5, in USD):

Line item	Signed value
Beginning NAV (Net of Carry)	+$45,067,000
Contributions – Cash & Non-Cash	$0
Distributions – Cash & Non-Cash	−$1,250,000
Offering/Syndication Costs	$0
Placement Fees	$0
Partner Transfers	$0
Total Cash / Non-Cash Flows	−$1,250,000
Total Net Operating Income / (Expense)	[computed from fee-netting reconciliation]
Total Net Realized and Unrealized Gain / (Loss)	[computed from realized + unrealized gains]
Ending NAV (Net of Carry)	Per GP report

The ITD (inception-to-date) column in the same sample shows Placement Fees = -$40,000 — a live example of why the formula must include all five cash-flow line items. An implementation that drops Placement Fees would report a $40K reconciliation failure on a template where the GP has included legitimate, spec-compliant data.

3.2 Commitment Reconciliation

The commitment reconciliation tracks unfunded capital through five signed flows. The ILPA 2025 formula:

Ending Unfunded Commitment =
  Beginning Unfunded Commitment
+ (Less Contributions)              [negative: reduces unfunded]
+ Plus Recallable Distributions      [positive: restores unfunded]
+ (Less Expired/Released Commitments) [negative: reduces unfunded]
+ +/- Other Unfunded Adjustment      [signed: either direction]

The three frequently-missed terms are Expired/Released Commitments and Other Unfunded Adjustment. Funds sometimes release a portion of committed capital — typically at the end of an investment period, or when a GP voluntarily releases committed but undrawn capital. Such releases reduce the LP's unfunded commitment permanently and must flow into the reconciliation. "Other Unfunded Adjustment" is a signed catch-all for corrections that don't fit the other categories — typically used for true-ups between the GP's books and the LP's records.

Working example (QTD values for LP #5):

Line item	Signed value
Beginning Unfunded Commitment	+$18,500,000
(Less Contributions)	$0
Plus Recallable Distributions	$0
(Less Expired/Released Commitments)	$0
+/- Other Unfunded Adjustment	$0
Ending Unfunded Commitment	+$18,500,000

The ITD column for the same LP shows a non-zero Other Unfunded Adjustment of -$500,000 and Recallable Distributions of +$4,000,000, demonstrating that the full five-term formula is needed in practice.

3.3 Fee Netting

The ILPA 2025 fee-netting reconciliation has four terms. Management Fees Net equals the Gross plus three signed reductions: Rebate, Waivers, and Offsets.

Management Fees – Net of Offsets, Waivers & Rebates =
  Management Fees – Gross of Offsets, Waivers & Rebates
+ Less Management Fee Rebate            [signed negative]
+ Less Fee Waivers                       [signed negative]
+ Less Offsets to Management Fees        [signed negative]

Working example from the ILPA sample (QTD):

Line item	Signed value
Management Fees – Gross	−$187,500
Less Management Fee Rebate	$0
Less Fee Waivers	$0
Less Offsets to Management Fees	−$82,600 (post-sign; input as +82,600)
Management Fees – Net	−$270,100

The total offsets applied ($82,600 in this quarter) aggregates across the twelve offset-category line items the ILPA 2025 template defines. Those categories are themselves a reconciliation — each category is reported separately, and the sum is "Total Offsets to Fees & Expenses (Applied During Period)."

The twelve ILPA 2025 fee-offset categories are:

#	ILPA 2025 Offset Category
1	Advisory & Consulting Fee Offset
2	Broken Deal Fee Offset
3	Transaction & Deal Fee Offset
4	Directors Fee Offset
5	Monitoring Fee Offset
6	Capital Markets Fee Offset
7	Arrangement Fee Offset
8	Origination Fee Offset
9	Organization Cost Offset
10	Placement Fee Offset
11	Other Offsets
12	Other Fee Offsets: 2016 ILPA Reporting Template Value

Each category has a specific spec-defined scope. A reconciliation implementation that re-categorizes during ingestion (e.g., classifying an Organization Cost Offset as a generic "Other" offset) will produce a correct total but lose the categorical fidelity the template is designed to preserve. For institutional fee analysis, this categorical fidelity matters: it distinguishes, for example, between a fund that generates substantial monitoring-fee income from portfolio companies (suggesting active post-investment management) and one that generates substantial broken-deal-fee offsets (suggesting diligence intensity relative to conversion).

A subtle point: the ILPA 2025 template also tracks an "Unapplied Offset Balance" roll-forward. Offsets recognized in a period may exceed the amount that can be applied in that period (the applied amount cannot exceed total recognized gross fees/expenses). The unapplied balance rolls forward to future periods. Full reconciliation discipline tracks both the per-category offsets and the unapplied-balance roll-forward as a separate auditable flow. Implementation note: per-category offset tracking is implemented in Meridian today; the unapplied-balance roll-forward is scheduled in the ILPA Reconciliation Enhancement Backlog (§2.2) for a future release.

3.4 Carried-Interest Accrual

The carried-interest reconciliation is the most conceptually subtle of the four, because carry is a GP claim against the LP's capital. From the LP's statement perspective, the carry balance is typically represented as a negative value — the LP's liability to the GP. The ILPA 2025 template's formula:

Accrued/Earned/Paid Carried Interest – Ending Period Balance =
  Starting Period Balance                [typically negative: LP's liability]
+ Carried Interest Accrued (Unrealized Profits)    [negative: accruing more liability]
+ Carried Interest Earned (Realized Profits)       [negative: realizing accrued liability]
+ Carried Interest – Paid During the Period        [positive: reduces liability magnitude]
+ Returned Clawback                                [positive: reduces liability magnitude]

The two middle terms are often summed into a single "Carried Interest Accrued/Earned (Total)" line, but the spec distinguishes them because they represent different underlying economics:

Accrued (Unrealized Profits): Change in the GP's share of any unrealized profits from portfolio investments. Computed as if all investments were realized at fair market value at quarter-end. Volatile and reversible.
Earned (Realized Profits, Inclusive of Amount Held in Escrow): Change in the GP's share of realized profits, including amounts held in escrow pending milestone satisfaction. More permanent than accrued.

Paid is the amount of already-earned carry that has been collected by the GP in the period. Returned Clawback is carry that had been paid out in prior periods but has been returned to the fund — typically because aggregate performance over the fund's life has fallen below the hurdle, triggering a clawback provision.

Working example (QTD for LP #5 from the ILPA sample):

Line item	Signed value
Starting Period Balance	−$4,750,000
Carried Interest Accrued (Unrealized Profits)	−$250,000
Carried Interest Earned (Realized Profits)	−$50,000
Carried Interest – Paid During the Period	+$50,000
Returned Clawback	$0
Ending Period Balance	−$5,000,000

Reading this sample: the GP's accrued claim against the LP grew by $300K in the quarter ($250K unrealized + $50K realized), of which $50K was collected (paid). The net effect is a $250K increase in the LP's liability to the GP — starting at −$4.75M and ending at −$5.0M.

A common anti-pattern in LP accounting is to track carry only in terms of a single "accrual delta" per quarter (net of paid), losing the distinction between accrued/earned/paid. When a clawback event occurs, the history of accrual vs. payment matters materially — the clawback is computed against amounts paid, not amounts accrued. An implementation that has netted these components over time cannot reconstruct the clawback calculation without archaeology.

4. What a $1 Tolerance Actually Guarantees

Having established the four reconciliation formulas, we can now state precisely what a $1-tolerance reconciliation pass guarantees and what it does not.

What $1 tolerance does guarantee	What $1 tolerance does not guarantee
Arithmetic consistency — every reported figure composes from its constituent signed terms. Sign flips produce discrepancies thousands to millions of dollars larger than $1, so a clean pass means signs were read correctly.	Valuation accuracy — the reported ending NAV composes from reported components. It does not mean the underlying portfolio companies are correctly valued. Valuation review is a separate discipline.
Categorical completeness — every fee offset, rebate, waiver, recallable distribution, expired/released commitment, other adjustment, accrued/earned/paid carry component, and returned clawback has been captured. A missing line item causes the formula to fall short by the missing amount.	Fee reasonableness — reconciling fees to within $1 means gross minus rebate minus waivers minus offsets equals reported net. It does not mean the net fee is reasonable versus peers, benchmarks, or transparency expectations.
Reproducibility under audit — the check is itself a reproducible calculation. An auditor can independently compute the sum of signed terms and verify the same result without access to the LP's internal systems.	Correct period allocation — distributions received on the last day of a quarter and reported in the following quarter's statement will reconcile within each quarter but may produce timing mismatches across periods.

The distinction matters. Institutional LP operations spend material time on valuation review and fee reasonableness — the judgment-heavy parts of oversight. A clean reconciliation does not replace that work. What it does is ensure that the work is being done against correctly captured numbers. Every hour spent debating whether a 15.3% IRR is good is wasted if the underlying cash flows that generated the figure were transcribed with sign errors.

5. Building a Reconciliation Workflow

A reconciliation workflow that achieves $1 tolerance consistently has five architectural components. Whether implemented in spreadsheets, a bespoke system, or an integrated platform, these components are distinct and each one fails in characteristic ways.

Component	Requirement	Characteristic failure mode
Signed ingestion	Preserve ILPA signs exactly as exported (post-sign) or preserve raw inputs plus sign metadata (pre-sign). Store the signed source; derive unsigned views at presentation time, not ingestion time.	Tool assumes one convention (post-sign) but receives templates in the other (pre-sign), silently flipping signs at ingestion. External reconciliation against the template becomes impossible.
Formula-level validation	Run each of the four formulas as a distinct validation step. Pass/fail each independently. Include every spec-defined line item; do not drop infrequently-populated items.	Simplified formula (e.g., "Contributions + Distributions" instead of the full five-term cash flows) reconciles correctly for most funds but fails on funds with non-zero placement fees, partner transfers, expired commitments, or clawback activity.
Tolerance policy	Document and consistently apply an absolute-value tolerance across all funds. $1 is defensible; the specific threshold matters less than consistency and documentation.	Tolerance applied to signed differences. A +$5 and −$3 discrepancy in the same fund net to +$2 but represent two separate errors; signed tolerance hides this.
Exception workflow	Classified (sign error / missing item / timing / rounding / GP error), owned (LP analyst / GP relationship / data provider), logged, with a resolution deadline.	Exceptions become inbox items or sticky notes. Known misstatements accumulate as "we'll handle that next quarter" and corrode the entire reconciliation discipline.
Audit trail	Every step produces an immutable record: template ingested, formulas run, tolerance results, exceptions raised, resolutions applied, final state. Re-runs create new audit trails rather than overwriting.	Audit trail is a notebook or Slack thread. Reconciliation depends on individual analysts rather than institutional memory; staff departures take the record with them.

5.1 On Signed Ingestion

The first requirement is that data ingestion preserves ILPA signs and handles both post-sign and pre-sign template states. The ILPA template's "input positive values" convention for Rebates, Waivers, Offsets, and Paid Carry means the raw data value and the formula-computed value will differ by sign. An LP tool that ingests the raw cell value without applying the formula's sign transformation will reconcile incorrectly by 2× the line-item magnitude — the same signature as a sign flip.

The discipline: ingest the spec-compliant signed state, and document whether the source is pre-sign or post-sign. Keep a single source of truth — the signed template — and generate any required alternative views from it.

5.2 On Tolerance Policy

A tolerance policy should be documented and consistently applied. Some institutions use $1; some use $10 or $100 for larger funds; some use basis-point tolerances relative to reported balances. The specific threshold is less important than consistency. What matters is that every fund is reconciled against the same tolerance policy, every quarter, and that the policy is documented in operations procedure.

The tolerance should apply to the absolute value of the discrepancy, not the signed value. A $0.50 excess and a $0.50 shortfall are both sub-tolerance. A $5 excess and a $3 shortfall in the same fund are both above tolerance even though the net is $2 — because they represent two separate errors, not a single offsetting one.

5.3 On Audit Trail

The audit trail is the mechanism by which reconciliation discipline becomes institutional rather than individual. An LP organization whose reconciliation depends on the specific analyst who processes quarterly reports has succession risk; an organization whose reconciliation is backed by a reproducible, auditable trail has a process that outlasts any individual contributor.

Immutability matters specifically for regulatory and audit defense. Once a quarterly reconciliation is closed, the audit trail should be frozen at that state. Legitimate re-runs (for example, during annual audits) should produce a new audit trail on top of the old one, not overwrite it. A trail that can be silently modified is not an audit trail; it is a notebook.

6. Common Reconciliation Failures and Their Diagnoses

Over a large sample of ILPA-template reconciliations, certain failure patterns recur. This section catalogs the most common, with diagnostic signatures.

Failure pattern	Discrepancy signature	Typical cause	Fix
Single sign flip	Discrepancy equals exactly 2× the magnitude of one line item	Ingestion logic sign-flipped an input-positive item (rebate, waiver, offset, carry paid) while treating it as post-sign	Identify whether source is pre-sign or post-sign; apply formula sign transformation consistently
Missing line item	Discrepancy equals the magnitude of one specific line item, often Placement Fees, Partner Transfers, Expired/Released Commitments, or Returned Clawback	Implementation uses simplified formula missing spec-defined line items	Expand ingestion schema to the full ILPA 2025 line-item taxonomy
Category misclassification	Multiple small discrepancies across related offset line items summing to a larger aggregate error	Offsets classified into wrong ILPA category (e.g., Organization Cost Offset reclassified as Other Offsets) — correct signs, wrong buckets	Map offsets to the ILPA-defined 12-category taxonomy precisely; do not re-categorize during ingestion
Carry-component netting	Carry balance reconciles within the quarter; annual accruals don't match GP's annual disclosure; clawback reconstruction requires archaeology	Accrued/Earned/Paid tracked only as a net per-quarter delta rather than as distinct signed flows	Capture all five carry line items separately: starting balance, accrued, earned, paid, returned clawback
Period-boundary drift	Individual quarters reconcile; annual totals disagree	A distribution with report date in Q4 and cash date in Q1 is booked by the GP and LP in different periods	Document and enforce a single period-end convention (report date or cash date) across all funds
Currency-conversion drift	Non-USD funds reconcile within $1 in early quarters; drift accumulates to tens or hundreds of dollars over time	Different FX rates applied to different cash flows within the same quarter, compounded by internal rounding differences	Adopt a single FX policy per fund (use GP's reported USD values directly, or re-convert at a uniform period-end rate)

Each of these patterns has a characteristic arithmetic signature that points directly at its source. Reconciliation-as-diagnostic-tool is only valuable if the failure output is this specific; a workflow that produces only "did not reconcile" messages loses the diagnostic value entirely.

Worked example: diagnosing a 2× signature

A fund's NAV reconciliation reports the following signed values:

Line item	Signed value as ingested
Beginning NAV	+$24,500,000
Contributions	+$2,100,000
Distributions	+$3,400,000 (mis-signed — source template has −$3,400,000)
Offering/Syndication Costs	$0
Placement Fees	$0
Partner Transfers	$0
Net Operating Income (computed)	+$130,000
Net Realized & Unrealized Gain	+$600,000
Computed sum	+$30,730,000

The GP-reported ending NAV is $23,930,000. Discrepancy: +$6,800,000.

$6,800,000 ÷ 2 = $3,400,000 — which equals the distribution amount exactly. This is the 2× signature. The ingestion captured distributions as positive when the template signs them as negative. The fix is to preserve the template sign at ingestion; the reconciled sum then becomes $23,930,000 and the discrepancy collapses to zero.

This is what good reconciliation tooling produces: not a binary pass/fail, but an arithmetic error whose magnitude and sign directly identify the source.

7. Implications for Fiduciary Practice

Reconciliation discipline is not a back-office concern. It is foundational to fiduciary practice in private-markets investment. The fiduciary duties of prudence and loyalty — ERISA for U.S. retirement plans, parallel frameworks for endowments, foundations, and sovereign funds globally — require fiduciaries to act with the care that a prudent expert would apply. A prudent expert, at minimum, ensures that the numbers used in investment decisions are the numbers the GP reported, correctly interpreted.

Consider a concrete scenario. A pension fund's investment committee reviews quarterly performance across its private-equity program. One fund shows a 14.7% IRR; another shows 11.2%; a third shows a loss. The committee allocates additional capital to managers in the top quartile of the pension fund's reference set. That allocation decision depends on the reported figures being correct. If one fund's IRR was computed against cash flows with a sign error in distribution reporting — or against a simplified NAV formula that missed the fund's placement fees — the fund's true IRR could be materially different, and the allocation decision was made on wrong information.

The fiduciary exposure is real. If challenged — by beneficiaries, by regulators, by plaintiff's counsel in an ERISA-breach case — the fiduciary is asked to document prudent process. "We reviewed the GP's quarterly reports" is weak evidence. "We reviewed the GP's quarterly reports, we reconciled each one against the full ILPA 2025 signed-convention formulas including all spec-defined line items, we documented a $1-tolerance audit trail, and the trail is reproducible" is evidence that survives challenge.

This is why reconciliation belongs in the platform, not the spreadsheet. A spreadsheet-based reconciliation can be done well by a careful analyst, but the evidence it produces is the analyst's personal work product. It is not institutionally reproducible, does not survive staff transitions without documentation archaeology, and cannot be audited without pulling the analyst's files. An institutional reconciliation workflow produces evidence that is the institution's, not any individual's — and that evidence is what fiduciary defense is built on.

8. What to Look For in Reconciliation Tooling

Institutions evaluating reconciliation tooling — whether bespoke internal builds, third-party software, or integrated platforms — should hold the tooling to specific functional standards. The following are minimums, not stretch goals.

Standard	Requirement	Red flag if missing
Template fidelity	Ingest the full ILPA 2025 template in its native structure, preserving signs, field names, and the complete line-item taxonomy — not a simplified subset.	"The system covers the major line items." Simplified formulas fail silently on templates where minor line items are non-zero.
Formula transparency	Each reconciliation formula readable as its formula. Inputs, signed terms, and the arithmetic visible to any reviewer.	"The system tells you it matched." Opaque calculation blocks defeat independent spot-check review.
Sign-convention handling	Handle both pre-sign and post-sign template exports. Document which convention the tool expects.	"The template must be in X format." Real-world GP exports vary; the tool must accommodate both.
Exception workflow integration	Classified, owned, logged, aged. Exceptions escalate automatically when beyond policy thresholds.	Exceptions flow into email or spreadsheet tabs. Known discrepancies age into permanent misstatements.
Audit trail immutability	Completed reconciliations are frozen. Re-runs create new trails on top of old ones, never overwriting.	Historical reconciliations can be modified silently. The trail is no longer evidence; it's a working document.
Temporal reproducibility	Reconciliation run today against Q1 2025 data produces the same result as the reconciliation closed at Q1 2025. Template-version updates and source-system changes do not retroactively change historical results.	Template-version auto-updates applied to historical data. Historical reconciliations drift without audit.

9. Summary and Next Steps

The ILPA 2025 Reporting Template's signed-convention arithmetic provides a foundation for institutional-grade LP reconciliation. Its four core formulas — NAV roll-forward (with its five cash-flow line items, operating income aggregate, and realized/unrealized gain aggregate), commitment reconciliation (with its five terms including expired/released and other adjustment), fee netting (with its four terms including rebate), and carried-interest accrual (with its five components including the accrued/earned split and returned clawback) — each compose to the penny when signed terms are added correctly.

A $1-tolerance reconciliation discipline applied across all four formulas — at their full spec-canonical depth, not a simplified subset — produces an auditability property that survives regulatory review, investment-committee scrutiny, and succession of individual analysts.

The discipline is not glamorous. It does not produce the insights that allocation decisions turn on. But it is the precondition for those insights being defensible. An LP organization that reconciles with rigor has a foundation on which everything else — performance attribution, fee analysis, manager selection, board reporting — can be built. An organization that does not reconcile has a foundation that will eventually crack, usually during an audit, usually in a way that embarrasses the institution publicly.

For LPs reviewing their current reconciliation practice, the following questions are diagnostic:

#	Question	If the answer is uncertain...
1	Can your team reconcile any fund's quarterly report against the full ILPA 2025 signed-convention formulas today — including placement fees, partner transfers, expired/released commitments, and the carry accrued/earned split?	Reconciliation is implemented against a simplified subset, and will fail on real-world templates
2	Is your tolerance (whether $1 or otherwise) documented as policy, or is it an informal convention carried by specific analysts?	The policy lives in people's heads and leaves when they do
3	Does your reconciliation handle both pre-sign and post-sign template exports correctly?	A real-world GP change in export convention will silently break reconciliation
4	Does your reconciliation produce an audit trail that an auditor could independently reproduce?	Reconciliation evidence is the analyst's work product, not the institution's
5	Do reconciliation exceptions flow into a workflow with ownership and resolution deadlines?	Known discrepancies accumulate as "we'll fix that next quarter" and corrode the discipline
6	When a senior analyst leaves, does the institution's reconciliation discipline persist?	The discipline is a property of the individual, not the institution

If the answer to any of these is uncertain, the reconciliation practice is not yet institutional. It is the practice of an individual supported by tools. Closing that gap — making reconciliation a property of the institution rather than a property of its people — is the work of LP operations maturity.