Rethinking IRR: Understanding PE's Key Metric

The headline numbers that won’t die

KKR still touts a 25.5 % gross “since-inception” IRR and Apollo waves a 24 % net badge. Impressive – until you remember that an IRR is not a compound annual growth rate (CAGR). It is simply the single discount rate that equates all past cash flows with the current NAV. Because early distributions dominate the equation, the figure locks-in early and barely budges afterwards. That’s why KKR’s first $31 m fund could “explain” trillions of value today without breaking the model.

Reality check from independent benchmarks

Cambridge Associates’ pooled 20-year net IRR for 1,600+ US buy-out and growth funds is 14.5 %; the 10-year figure is 15.0 %. Good, but a far cry from the 20–30 % marketing slogans – and only ~400 bp over the MSCI World PME.

Bain & Co. show that the 10-year horizon IRR has drifted lower since 2021 as higher rates bit into exits.

Why the metric is mis-used

• Re-investment fallacy – IRR assumes every distribution is reinvested at the same lofty rate.

• Non-additivity – you cannot average IRRs across deals or funds.

• Gaming 101 – draw bridge loans to avoid calling LP capital, sell the early winners fast, and your IRR soars even if later deals drag.

As a former junior in PE I saw how debt-funded “stub exits” in year 2 made the quarterly letter sparkle while the real work had barely begun.

Implications for UK investors

The Mansion House Accord nudges DC schemes to push 10 % of assets into private markets. Yet today DC plans put just 0.5 % into PE versus 5 % in Australia.

Before trustees chase eye-watering vintage IRRs, they should:

Toolkit	What to ask for	Why it matters
Net TVPI / MOIC	Multiple on invested capital after fees & carry	Hard cash back, no reinvestment assumption
Horizon IRR (5/10/15 y)	NAV-to-NAV money-weighted returns	Scrubs out the 1980s exits
Public-market equivalent (PME)	CA mPME, KS PME, etc.	Benchmark vs. low-cost trackers
Value-creation bridge	EBITDA growth vs. multiple expansion vs. leverage	Tests GP’s operating alpha

MY TAKE

• IRR isn’t evil, just mis-sold. Treat it like EPS before share-based comp – useful trend info, but never the whole story.

• 12-15 % net over two decades is still solid. But paying “2 and 20” for an illusion of 25 % is value-destructive.

• Liquidity is king in 2025. With $3.9 tn of dry powder and exits still sluggish, GPs that can generate cash – not just mark-ups – will stand out.

• UK schemes should walk, not run. Start with small secondaries or co-invest sleeves, insist on transparent cash-flow data, and align carry hurdles with public benchmarks plus an illiquidity premium (~300 bp feels fair in today’s rate regime).

Bottom line

Don’t be dazzled by a metric that was never designed for cross-fund comparison. Focus on multiples, horizon returns and cash distributions. In a market where 14 % really is the new 25 %, disciplined underwriting and fee scrutiny will be the difference between genuine alpha and expensive theatre.