A reserve strategy allocates capital for possible follow-on investments. The percentage reserved is an output of the strategy, not evidence that it improves returns.
To judge the policy, an LP or manager needs to know what the reserved capital is expected to earn, what alternative investments it displaces and how decisions will be made when the portfolio develops differently from the plan.
Separate observations from simulations
Laura Thompson's 2022 Sapphire analysis combines observations about selected funds with hypothetical reserve scenarios. It usefully shows that follow-ons can reduce a blended multiple even when they produce positive proceeds. But a hypothetical model is not an estimate of how often a particular policy succeeds across the industry. Sapphire's reserves analysis.
The distinction also applies to observed winners. Studying which investments generated returns in successful funds can reveal concentration. It does not establish that managers could have identified those eventual winners accurately when follow-on decisions were made.
A policy must work with the information available at the time. Perfect hindsight is an outcome description, not an implementable selection rule.
Audit the comparison before accepting the conclusion
Four checks make reserve research easier to interpret:
1. Capital held constant: does the model reallocate one budget, or compare a smaller fund with a larger one? 2. Information held realistic: does follow-on selection use contemporaneous evidence or eventual outcomes? 3. Costs treated consistently: are fees, carry and unused capital included on the same basis? 4. Timing made visible: when is money called, invested and returned?
A larger fund can produce more absolute profit while a lower multiple reaches LPs. That is not an arithmetic contradiction. It means the model has changed both the capital amount and its productivity.
Similarly, a gross deal multiple and a net fund multiple cannot be compared as if they measure the same thing.
Compare the next £1m on a common basis
Consider two illustrative uses of £1m available for investment today. The outcomes are scenarios, not forecasts or historical data.
| Use | Assumed outcomes | Total proceeds | Gross multiple |
|---|---|---|---|
| Five new £200,000 investments | One returns 10x, one 3x, three zero | £2.6m | 2.6x |
| Two £500,000 follow-ons | One returns 4x, one 1x | £2.5m | 2.5x |
On these assumptions, new investments produce £100,000 more. But the table alone is insufficient. The likelihood of the outcomes, concentration, time to receipt and effect on existing holdings may differ.
If the £2.6m arrives after eight years and the £2.5m after four, with no intermediate cash flows, the annualised gross returns are approximately 12.7% and 25.7%. That does not make the second path certain or universally preferable. It shows why a slightly lower multiple can still be economically attractive.
Price the follow-on at its own terms
The original cheque bought exposure at an earlier price. A follow-on buys additional exposure at today's price and security terms.
Ask what the new money changes: a commercial milestone, solvency, ownership, preference ranking or simply time until another financing. Separate the expected return on the new cheque from the hope of recovering the old one.
If participation protects value in the existing holding, include that effect explicitly in scenarios. Do not bury it in a claim that the new investment is independently attractive. This makes the decision easier for an investment committee or LP to understand.
Translate the policy into a decision process
List expected follow-on dates, likely amounts and the evidence required to consider participation. Revisit the list as the portfolio changes. A reserve is available capital, not a promise to every company.
Set concentration limits and an escalation process for exceptions. Record cases where the manager declines despite having capital. If every existing company qualifies for support, the stated selection policy may not be doing much work.
Where follow-ons depend on raising separate vehicles, distinguish potential investor interest from committed capital. The associated allocation and conflict rules also need to be clear.
The evidence can identify trade-offs and expose weak assumptions. It cannot supply a universal reserve percentage without a portfolio, prices and decision rule. A credible policy shows how each additional cheque earns its place against the alternatives available at that time.
FAQ
What percentage should a VC fund allocate to reserves?
Build the amount from expected follow-on opportunities, cheque sizes, concentration limits and decision rules. Compare those uses with new investments and retain room for the portfolio to develop differently. No universal percentage follows from the evidence reviewed here.
What is a good reserve ratio for a first-time manager?
One that fits an executable portfolio at the capital actually raised. Model a smaller close, delayed milestones and fewer attractive follow-ons. Explain how the policy changes ownership and deployment. Borrowing an established firm's ratio does not reproduce its access, investment prices or operating capacity.
Can you do follow-ons without holding reserves?
Potentially, by raising new capital through a suitable vehicle or involving other investors. That money is conditional until committed, and the arrangement can introduce allocation and conflict questions. Do not describe prospective SPV fundraising as equivalent to a reserve already available within the fund.
Do reserves improve fund returns?
They can, but that depends on what the additional investments earn relative to the alternatives, their timing and their effect on existing holdings. A hindsight model selecting eventual winners does not establish an implementable policy. Judge the information and decisions available when each follow-on is considered.
