US Venture Capital Report 2008 — The Illiquidity Nobody Had Priced

Venture capital entered 2008 looking structurally safe. It uses no leverage. Its holdings are not marked to market daily. Its companies mostly do not borrow. None of the mechanisms described in the Global Investment Outlook 2008 — funding runs, collateral spirals, counterparty freezes — appeared to touch it.

It was hit anyway, through a route the structural argument ignored: its investors.

The mechanism is worth stating precisely, because it recurs. A venture fund does not hold its investors' money. It holds commitments — legally binding promises to provide capital when called, typically over five years. The fund's asset is a claim on institutions whose own circumstances have nothing to do with venture capital.

In 2008 those institutions were in difficulty for reasons entirely unrelated to the venture portfolio:

• Their public holdings had fallen sharply, which mechanically raised the private share of their portfolios above policy limits — the denominator effect.
• Their liquidity was strained, because many held their own illiquid assets and faced their own obligations.
• Their forward commitments across all private strategies came due simultaneously, since the vintages of the preceding boom were all in their call periods at once.

So the pressure arrived at the fund as a funding problem while the portfolio was still performing. The Global Investment Outlook 2008 identifies this pattern in banks — institutions dying on timing while solvent on paper. Venture capital experienced a slower, gentler version of the same structure, and the resemblance is not a coincidence: both are maturity mismatches, one measured in days and the other in years.

The second theme is behavioural. The industry's near-universal response — extend runway, cut burn, assume no follow-on funding — was correct as general advice. It was also, for a specific minority of companies, exactly wrong, and separating the two cases is the most valuable skill the period demanded.

Hit through the liability side

The distinction between asset-side and liability-side damage is the report's organising idea, and it applies well beyond 2008.

Asset-side damage is what most risk analysis assumes: the things you own become worth less. It is visible, measurable, and shows up in a valuation.

Liability-side damage is what happens when the capital behind you becomes unavailable. Your assets are unchanged. Your ability to hold them is not.

Why venture was exposed to the second and not the first:

• Portfolio companies did not immediately deteriorate. A software company with two years of runway and growing usage was unaffected by a banking crisis in the quarter it occurred.
• Valuations were not marked in any meaningful way, so no accounting mechanism forced recognition.
• But the capital required to keep funding those companies came from institutions that had just had a very bad year in completely different assets.

The transmission ran in three steps:

1. Limited partners faced their own liquidity pressure, from public losses, capital calls across every private strategy at once, and in some cases spending obligations that do not pause.
2. Some sought to reduce or defer commitments, and a secondary market in fund interests appeared at steep discounts — an unusual event, since these positions are normally held to term.
3. General partners, reading this, slowed deployment and reserved more heavily for existing companies — which meant fewer new investments regardless of how attractive they were.

The companies were fine. The capital behind the companies was not. An asset class with no leverage and no mark-to-market was still transmitted a banking crisis, because its funding ran through institutions that had one.

The Private Equity Report 2015 develops the commitment structure in more detail, and the Secondaries Market Report 2019 covers the market that 2008 effectively created.

The denominator effect

The single most consequential mechanism for private markets in 2008 deserves its own treatment, because it is arithmetic that behaves like a policy failure.

How it works:

An institution sets an allocation policy — say, private assets should be a defined share of the total portfolio. The share is measured against total portfolio value.

When public markets fall sharply:

• The public holdings drop immediately, because they are marked daily.
• Private holdings do not, because they are valued quarterly, with a lag, using methods that smooth.
• So the private share rises without a single private transaction occurring. A portfolio at policy weight can breach it purely because the other side fell.

The institution is now over-allocated to private assets and must respond:

• It cannot easily sell private positions — that requires a secondary sale at a discount.
• It can stop making new commitments, which is the cheap lever, so that is the lever used.

The result is perfectly procyclical. New commitments are cut hardest exactly when public markets have fallen — which is to say, exactly when the vintage being formed will deploy into the lowest valuations of the cycle.

This is the opposite of what a long-horizon investor would choose, and it is not a mistake by any individual. It is what the policy framework mechanically requires.

The consequence for returns is well documented in direction if not in magnitude: vintages formed in and immediately after crises have historically been strong, precisely because less capital chased more assets. And they are the vintages institutions systematically under-commit to.

The practical fix, which some institutions adopted afterwards, is to set commitment policy on a forward-looking pacing model rather than on current allocation percentages — committing a planned amount per year regardless of where the denominator sits. The Private Markets Outlook 2016 covers how this became standard practice.

A commitment is a liability

2008 taught a generation of allocators something the documents had always said and the behaviour had never reflected.

An undrawn commitment to a fund is not an option to invest. It is a binding obligation to provide capital on demand, typically within days of a call notice, with severe consequences for default — usually forfeiture of a substantial portion of the interest already built up.

Why this had been treated casually:

• In normal conditions the calls are predictable and modest, so the obligation feels like a schedule rather than a risk.
• The capital is usually available, because the rest of the portfolio is liquid.
• And the language of the industry encourages it — "dry powder" describes uncalled commitments as though they were cash the investor holds, when in fact the investor may hold nothing of the sort.

In 2008 all three assumptions failed at once:

• Calls accelerated in some funds, as managers supported existing companies.
• Liquid assets to meet them had fallen in value, so meeting a call meant selling something at a loss.
• And the obligation could not be deferred.

This is a maturity mismatch in the classic sense — a demandable obligation funded by assets that cannot be sold quickly at a fair price. It is the same structure as the bank funding problem in the Global Investment Outlook 2008, running at a different speed.

The durable lesson for any private markets investor is that uncalled commitments must be modelled as a liability with an uncertain call date, and the liquidity to meet them held against a stress scenario in which everything else has fallen and calls have accelerated. The two are correlated, which is the entire difficulty.

When the exit window closes

A venture fund's returns are realised only when companies are sold or listed. 2008 closed both routes substantially, and the timing consequence is more severe than it first appears.

Why closure matters more than valuation:

• A fund has a finite life, typically ten years with limited extensions. The clock does not pause for market conditions.
• A company that cannot exit must be funded longer, consuming reserves that would otherwise back new investments.
• And the fund's reported return is a function of time as well as multiple. A holding that eventually returns three times invested capital produces a very different annualised return over six years than over eleven.

The mechanical consequences through 2008–2009:

• Holding periods extended across the industry, structurally rather than temporarily.
• Reserve ratios rose — a larger share of each fund was set aside for follow-on investment in existing companies rather than new ones.
• Down rounds and recapitalisations became common, and the terms attached to them mattered enormously, per the structural analysis in the US Venture Capital Report 2022.
• The secondary market for company shares grew, initially as a workaround for employees and early investors with no other route to liquidity — the origin of a market that became structurally important later, as the US Venture Capital Report 2013 describes.

The generalisable point: in any closed-end fund structure, exit conditions are a return driver on par with entry valuation, and they are far less controllable. A fund that buys well into a closed window can still underperform one that buys expensively into an open one.

Correct advice, wrong for the exception

The industry's guidance in late 2008 was direct and nearly unanimous: assume no further funding is available, cut costs to reach profitability or a long runway, and survive.

As base-rate advice this was right. Most companies could not raise, most that tried failed, and most that cut hard enough survived to raise later on better terms.

But the advice was general and the situation was not. A specific and identifiable minority faced the opposite problem.

The distinguishing features of the exception:

• Genuine demand that was growing rather than cyclical — usage, retention or revenue improving through the downturn, not despite marketing spend but because the product was becoming more necessary.
• A cost structure that fell with scale rather than rising with it, meaning growth improved rather than strained the finances.
• Competitors who were cutting, which meant hiring, customer acquisition and market share were all available at unusually low cost.
• Sufficient capital to act, which was the binding constraint and the reason most of the minority could not exploit the position.

For these companies, cutting was the expensive choice. The cost of hiring an engineer, acquiring a customer or entering a market fell sharply in 2008–2009, and the companies that spent into that were buying inputs at a cyclical low.

The advice to cut was right for the base rate and wrong for the tail. Almost all of the decade's returns came from the tail.

This is the core epistemic problem in venture capital, and it is why the asset class resists general rules: returns are concentrated in outcomes that are by definition unusual, so guidance calibrated to the typical case is calibrated to the cases that do not drive returns. The US Venture Capital Report 2009 develops what the resulting vintage produced.

What an allocator could act on

Model uncalled commitments as liabilities with uncertain timing. They are demandable obligations, not options, and the liquidity to meet them must survive a scenario where the rest of the portfolio has fallen and calls have accelerated — the two are correlated.

Set commitment pacing on a forward plan, not on current allocation percentages. Policy measured against a falling denominator forces the deepest cuts into the strongest vintages, which is procyclical by construction rather than by error.

Distinguish asset-side from liability-side damage. An asset class with no leverage and no mark-to-market can still be transmitted a crisis through the institutions that fund it. Ask who your manager's investors are and what else they own.

Treat exit conditions as a return driver equal to entry price. Fund life is fixed, so a closed window converts a good multiple into a poor annualised return, and no amount of entry discipline compensates.

Expect general advice to be wrong for the tail. In an asset class where returns concentrate in the unusual case, guidance calibrated to the median company is calibrated to the companies that will not drive the fund.

Note that input costs fall in a downturn. Engineering talent, customer acquisition and market share were all cheap in 2008–2009 — a company with capital and genuine demand was buying at a cyclical low, which is the mirror image of the denominator effect hitting its investors.

What 2008 established

• Venture capital is exposed through its liability side, and no amount of asset-side conservatism protects against it.
• The denominator effect makes institutional allocation procyclical, cutting commitments hardest into the best vintages.
• Uncalled commitments are demandable liabilities, a maturity mismatch of the same family as bank funding.
• A closed exit window damages returns through time, independently of valuation.
• The secondary market emerged — for fund interests and for company shares — as a response to illiquidity that later became permanent infrastructure.