From the GP's seat, the LP base looks like a list of names and commitment amounts. From inside, it is a small number of fundamentally different institutions with different mandates, time horizons, and tolerances. A GP who treats them all the same will lose all of them.
The eight LP archetypes
Public pensions (CalPERS, NYS Common, Texas Teachers): largest slice of US institutional capital, bound by fiduciary duty and FOIA disclosure rules. Corporate pensions: similar mandate, less disclosure. Endowments (Yale, MIT, Princeton): high allocation to alternatives, long horizon, top-quartile manager selection. Foundations: smaller scale of endowment-style. Sovereign wealth funds (GIC, ADIA, Mubadala, Temasek): largest cheque sizes, often co-invest. Insurance general accounts: liability-driven, prefer credit and infrastructure. Family offices: idiosyncratic mandates, fastest decisions. Funds-of-funds and consultants: aggregators on behalf of smaller institutional or HNW capital.
Pacing models
An LP committing $500M per year to PE over a 10-year vintage cycle does not invest $5B in year one. Committed capital is called over 4–5 years; distributions start in year 4 and peak in years 6–8. A pacing model projects the timing of calls, distributions, and NAV to keep the LP at its target allocation through the cycle. After 2022, when distributions slowed, many LPs found themselves over-allocated via the denominator effect (Chapter 35) and had to slow new commitments.
Co-investment and secondaries
Sophisticated LPs supplement primary commitments with co-investments (direct stakes alongside the GP, fee-and-carry-free) and secondaries (purchasing other LPs' fund interests). A modern $50B LP might have 60% primary, 25% co-invest, and 15% secondaries. The blended fee load is dramatically lower than 100% primary — a structural alpha that drives much of the post-2010 institutional capital reorganisation.