How GPs Report IRR and Why LPs Are Starting to Distrust the Number

Why private equity IRR reporting is harder to read than it looks

The core problem with IRR as a PE performance metric is not the math. Any fund administrator can calculate IRR correctly given a set of cash flows and dates. The problem is that several legitimate choices made before that calculation begins can produce dramatically different outputs from the same underlying investment performance.

The most documented example is methodology selection. Academic research in Venture Capital, Private Equity, and the Financing of Entrepreneurship shows directly: switching from the calendar-time method to the time-zero method on the same portfolio data can materially inflate reported returns — research using Burgiss database data has documented significant gaps on identical portfolios depending on the methodology applied and decreases it from 43% to 34% in another. The direction of the swing is unpredictable. The magnitude is material. And until the ILPA 2026 template, there was no requirement to disclose which method was being used, let alone to calculate both.

The implication is direct: two GPs running identical portfolios, generating identical cash returns, can report meaningfully different IRRs depending entirely on their internal methodology choice. An LP comparing fund performance across their private markets portfolio may be comparing figures computed under different rules, without any way to tell from the quarterly report.

The methodology question extends beyond calendar-time versus time-zero. GPs also make choices about how to handle management fee offsets, how to treat transaction costs at the fund versus investment level, and whether to calculate gross IRR from fund-to-investor cash flows or from fund-to-investment cash flows. Each of these choices is defensible. None of them is standardized across the industry. A GP that calculates gross IRR using fund-to-investment cash flows will typically report a higher gross figure than one using fund-to-investor cash flows, because the former excludes fees from the denominator. The difference is not fraud. It is discretion. And discretion, applied consistently in one direction, produces performance figures that look better than they would under a more conservative convention.

The deeper problem is that LPs often do not know which conventions their GPs are using. Quarterly letters report IRR. They rarely report the methodology behind it. Fund LPAs (limited partnership agreements) sometimes specify calculation methodology, but the provisions are frequently vague and the interpretation is left to the GP. Academic research using data from the Burgiss database, covering 997 buyout funds and 1,074 venture funds, found documented evidence of return misreporting during fundraising periods, with underperforming managers more likely to manipulate reported returns and less likely to raise a successor fund. The manipulation is not always intentional. The point is that the reporting framework has historically made it easy, and the ILPA 2026 template is the first major structural attempt to close that gap through standardization rather than regulation.

Fund performance dashboards display IRR figures, but the calculation methodology behind those figures is rarely disclosed alongside them. The same underlying portfolio can produce different reported returns depending on which method a GP applies.

The subscription credit line effect: a concrete illustration

Subscription credit lines, also called capital call facilities, let GPs borrow against LP commitments rather than calling capital directly. A GP that uses a subscription line to fund an investment in January, then calls LP capital to repay it in June, has effectively shortened the LP's measured investment period by six months. Since IRR is time-sensitive, a shorter investment period on a given return produces a higher reported IRR.

This is not fraud. The investment may have genuinely performed well. The IRR figure is mathematically accurate under the chosen measurement convention. But an LP comparing this fund's IRR against a fund that called capital directly in January is comparing figures computed under different timing assumptions. The fund that used the subscription line will show a higher IRR even if the underlying investment returns are identical. The difference can run to several percentage points on a fund level.

The research is clear on the scope of this. The use of subscription credit facilities has expanded substantially across the industry. A significant share of reported fund-level IRRs in recent vintages include the timing benefit from delayed capital calls. This is not a fringe practice by a few aggressive managers. It became standard procedure across many mid-market and large buyout funds through the 2015-2022 period, and many LPs only realized the scope of the effect when they started asking for calculations with and without the subscription line separately.

The effect can be quantified. A fund that delays the average capital call by six months through a subscription line and generates a 2x return over a five-year investment period will show a materially higher IRR than the same fund calling capital directly. The six-month delay shortens the measured investment period from the LP's perspective while the actual investment timeline is unchanged. On a five-year hold, this timing adjustment can inflate the reported IRR by several percentage points. On a three-year hold, the effect is proportionally larger because the delay is a bigger fraction of the total period.

What made this practice defensible for so long is that subscription credit lines serve legitimate operational purposes. They allow GPs to move quickly on investments without waiting for LP capital call mechanics. They reduce the administrative friction of calling small amounts of capital from large LP bases. A fund with 100 LP investors cannot realistically call capital in 48 hours; a subscription line bridges the gap. The problem is that a tool designed for operational efficiency became a tool for performance presentation, and the two uses are hard to distinguish from outside the GP's books.

Classifying each capital call transaction by funding source (LP capital vs. subscription line drawdown) is the core data requirement of the ILPA 2026 Performance Template. AI agents can work through capital call notices and fund administration records to produce this classification at scale.

What happens when an LP runs the numbers themselves

An increasing number of sophisticated LPs no longer accept the IRR in a quarterly letter at face value. They run their own calculations. This is not a sign of adversarial relationships; it is a sign of a more analytically mature LP base, and GPs who understand this are better prepared for the conversations it produces.

The LP calculation process is straightforward in principle. The LP takes the cash flow history they have received over the fund life: capital calls, distributions, and their estimated share of remaining net asset value. They compute IRR from those cash flows using their own system, using their own methodology. If their number matches the GP quarterly report, the conversation is over. If their number does not match, the conversation begins.

The divergences LPs commonly find fall into three categories. First, the subscription line timing effect: the LP's calculation starts the clock from when they received a capital call, not from when the underlying investment was made. If the GP's reported IRR starts the clock from when the investment was made (before the LP was called), the two calculations will differ. Second, methodology differences: the GP may be using a different IRR calculation convention than the LP's system defaults to. Third, NAV timing: if the GP marks the portfolio at a different date or using a different methodology than the LP assumes for the residual value, the terminal cash flow input differs.

None of these divergences is necessarily evidence of wrongdoing. But each requires an explanation, and a GP that cannot provide a clear account of why their reported IRR differs from an LP's independent calculation has a relationship problem on their hands. The ILPA 2026 Performance Template addresses this directly by providing the cash flow table that allows LPs to replicate the GP's calculation from first principles. When both parties start from the same transaction-level data, the source of any divergence becomes visible immediately rather than requiring weeks of back-and-forth between fund administrators.

What LPs are actually asking for now

The shift in LP performance priorities is measurable, not anecdotal. McKinsey's 2026 Global Private Markets Report found that 2.5 times as many LPs ranked DPI as a most critical metric compared to three years ago. Private Equity International went further, publishing a piece titled "DPI: the three-letter metric that will drive fundraising in 2026." The language is unambiguous.

The reason is structural, not cyclical. Five-year rolling DPI as a share of AUM for buyout funds hit its lowest recorded level in 2025. Distributions as a percentage of AUM fell to approximately 6% in the six months ending June 2025, according to McKinsey's 2026 Global Private Markets Report,, against a ten-year average of 14%. LPs who had been patient through 2022 and 2023, reassured by paper IRRs that looked healthy, found that those IRRs were not converting into actual cash. The metric that was supposed to measure return was measuring something else.

Wrong.

The shift to DPI is not a rejection of IRR as a concept. IRR measures the rate of return on invested capital over time, which is genuinely useful for comparing investments with different timing profiles. The problem is that IRR as reported in PE has accumulated so many discretionary inputs that it has lost much of its comparability. LPs are not saying IRR is useless. They are saying they no longer trust it as a standalone number, and they want cash in hand to confirm the story the IRR is telling.

Return misreporting is also documented in the academic literature. Research using data from the Burgiss database covering 997 buyout funds and 1,074 venture funds found that some underperforming managers manipulate reported returns during fundraising, and that those managers are less likely to raise a next fund. The manipulation is not always intentional. Some is a product of the discretion inherent in the methodology. But the outcome is the same: an LP reading a quarterly report cannot always tell whether the IRR reflects what the fund actually achieved or what the fund chose to report.