On this page
Larger private funds do not reliably deliver worse market-adjusted returns for the typical fund. In the private-capital study assigned within CAIA® Level II Emerging Topics, large-fund averages are lower than small-fund averages, yet the median comparison does not show a dependable gap. The candidate’s task is to keep size, dispersion, manager history and fundraising speed distinct instead of turning one descriptive chart into a selection rule.
Know what the evidence actually measures
The CAIA curriculum overview identifies private-capital scale, scope and speed as an Emerging Topics reading. The underlying Institute for Private Capital white paper examines market-adjusted fund performance through public market equivalents, or PME. A PME asks how the fund’s cash flows compare with a specified public-market alternative. The benchmark, geography, timing of calls and distributions, and whether the value includes unrealised assets all affect interpretation. PME is useful for comparing like cases; it is not a magic adjustment for every difference in risk or liquidity.
The researchers use a global sample covering buyout, venture capital, credit, infrastructure and natural resources, and real estate funds. CAIA’s event description for the research reports 10,276 funds and 8.7 trillion US dollars in commitments. The white paper says the performance observations run through the fourth quarter of 2022. Those are the study’s scope and measurement end date, not a claim about funds launched this week or a prediction for the next vintage. Its analysis is descriptive; a relationship between a characteristic and PME does not by itself establish what caused the return.
Mean and median answer different size questions
The primary study finds that large funds have lower average PME than small funds across the strategies and geographies examined. That statement is about the arithmetic mean. The same study finds no reliable difference in median performance by size. The two findings can coexist because a few small funds with extremely high outcomes pull the small-fund mean upwards. A candidate who says “large funds underperform” without identifying the statistic has discarded the main qualification. The relevant question is whether the reader wants to describe the average dollar-neutral fund observation, the experience of a middle-ranked fund, or the possibility of an unusually high outcome.
Dispersion matters because private funds are not uniformly scalable businesses. A smaller fund may have access to a narrowly sized opportunity set and may also deliver a very weak result. The size comparison therefore needs the full distribution: centre, spread and tails. A manager’s account of its best prior fund offers a selected tail observation, not an estimate of its typical future result. Conversely, a large fund’s lower sample mean need not mean that each large fund carried an inherent handicap. The study’s median result directly limits that inference.
The older literature discussed in the assigned study note found a concave cross-sectional relation between fund size and performance in an earlier sample. A later and broader sample can produce a different conditional picture because fund types, geographies, vintages, benchmarks and data coverage differ. Treat the two findings as evidence conditioned on their samples, not as universal laws that one paper has permanently “overturned”. On an exam, an answer should name which study, measure and comparison the prompt actually supplies before claiming a scale effect.
A worked comparison using invented funds
Consider a deliberately small, invented set of market-adjusted PME results. Four small funds have PMEs of 0.8, 1.0, 1.1 and 2.1. Four large funds have PMEs of 0.9, 1.0, 1.1 and 1.2. These values are teaching numbers, not observations from the study. Each group contains a weak outcome and both include funds near parity with the public comparator. The small group also contains one very high outcome. Looking only at the best small fund would conceal how similar the middle of the two groups is.
The small-fund mean is 1.25, because its four PMEs sum to 5.0. The large-fund mean is 1.05, because its results sum to 4.2. Yet the median is 1.05 in each group: the average of the middle two ordered results. One small-fund outlier created the difference in means without changing the central observation. A real research result requires a far larger sample and proper controls, but this arithmetic shows the precise logical error in treating a lower large-fund mean as proof of a lower typical large-fund result.
Now suppose the small-fund sample came mainly from venture vintages and the large-fund sample mainly from buyout vintages. The initial mean comparison would mix strategy with size. If the public benchmarks differed, even the PME labels could conceal different comparators. If more small funds remained unrealised, their PMEs would depend more heavily on reported net asset values. The exam-relevant repair is to compare funds within sensible strategy and vintage groups, then inspect how much of each result is realised. The calculation is easy; the difficult part is deciding what the calculation is allowed to say.
The older studies ask a different question
The assigned CAIA reading also discusses research on whether a general partner’s earlier results carry into a successor fund. A early primary study reported persistence across successive buyout and venture funds in a sample drawn largely from earlier vintages. It also found that better-performing managers were more likely to raise another and larger fund. That fundraising response is different from proving that extra scale caused the next return. Strong prior performance can attract commitments, and managers who fail to raise a successor are absent from a sequence of surviving funds. A comparison must therefore separate the return signal from the capital-flow response and from selection into the next sample.
The later primary research on performance persistence asks what an investor could actually know when the next fund was raising money. Final performance of the previous fund can look predictive in a retrospective study, yet it is usually unavailable at the commitment date. The previous fund may still own companies valued using interim net asset values rather than realised distributions. In the later US sample, the authors found little or no fundraising-time persistence for buyout funds, especially in post-2000 vintages, while venture-capital persistence remained visible using information available then. This does not make a prior record useless. It means a candidate must say whether the record was final or interim, identify the strategy and vintage, and resist applying a historical top-quartile rule uniformly.
Manager focus needs the same care. A growing fund might have to seek larger deals, enter unfamiliar sectors or expand geographically, but the proposed mechanism is a diligence hypothesis until evidence connects the change to performance. The later scale paper does not find a general negative scope effect in its own sample. That finding does not overturn every earlier setting in which capacity was constrained; it limits a universal claim. For exam purposes, keep three questions separate: how large the fund is relative to comparable funds, how much the same manager changed its strategy, and what information about earlier funds was available when the decision was made.
Fund growth is a different variable
A fund’s absolute size differs from its growth relative to its own predecessor. A general partner may raise a larger fund after good earlier performance, but that observed sequence creates a selection problem. The new fund’s return is not known at fundraising, and earlier success may itself have caused the larger raise. In the white paper, growth in fund size from one fund to the next was not reliably related to market-adjusted performance across the examined strategies and geographies. Therefore, a sentence such as “a rapid step-up must damage returns” overstates this evidence.
The paper also reports no reliable evidence that aggregate general-partner assets under management growth lowers fund performance. At the same time, the number of funds a manager had previously run was positively related to performance in some of its analyses, including buyout and venture. Those are separate variables. Experience can coincide with a larger organisation, but a regression coefficient on prior fund count should not be translated into a causal return to experience without considering selection and survival. Funds that never raise successors disappear from a sequence comparison; that alone is a reason to be careful with claims about later fund numbers.
Scope and speed need equally careful reading
Scope describes how broadly a manager operates across strategies or geographies. It is tempting to assume that broadening always dilutes skill. The paper’s regression analysis did not find a reliable negative effect from its scope measures; its conclusion even notes some positive associations. That result does not certify every expansion as sound. It says this sample did not support a blanket scope penalty after the researchers’ controls. A sensible exam answer should identify exactly how scope was measured and avoid converting an association into a universal operational rule.
Speed refers to the time before a manager brings a successor fund to market. The study describes a non-linear relation: the slowest returners often have poor current-fund performance, while the very fastest also show below-average performance in their most recent fund. A simple “faster is better” line misses the second side of that shape. The current fund’s performance may influence when the next fund is raised, so fundraising timing is partly an outcome of prior conditions rather than an independent lever that guarantees a return. It is a clue to investigate, not a substitute for examining cash flows and strategy.
What to write in a CAIA answer
- Specify the performance measure and the comparator before interpreting a return.
- Distinguish a lower mean from an unchanged median; use dispersion to explain how both can occur.
- Keep absolute fund size, fund-to-fund growth, manager-wide growth and fund sequence separate.
- State the study period and sample when applying a historical pattern.
- Describe scope and fundraising speed as observed associations, with plausible selection effects.
This is a more useful decision framework than remembering a slogan about big funds. If an exam vignette offers two managers, first identify the strategy, vintage, benchmark and realised share of the record. Then compare the distribution rather than one headline PME. Finally ask what the proposed fund changes in size, scope and timing, and whether the data support an inference about future performance. The CAIA program page places this analysis in the wider qualification, and Omni’s editorial methodology explains how its supporting sources are presented.
Sources
- Explore the Curriculum - CAIA Association · caia.org · Retrieved
- CAIA London: Scale, Scope, and Speed in Private Capital Funds - CAIA Association · myaccount.caia.org · Retrieved
- Scale, Scope, and Speed in Private Capital Funds - Institute for Private Capital · uncipc.com · Retrieved
- Private Equity Performance (2005): Returns, Persistence and Capital Flows · users.nber.org · Retrieved
- Has Persistence Persisted in Private Equity? · bfi.uchicago.edu · Retrieved
