Three 3x Funds Beat a 15x VC Return — The Math Every Endowment Should Know

Source: 20VC with Harry Stebbings | Published: 2026-09-14T13:58:31Z

Baylor Endowment CIO David Morehead ran the numbers: three consecutive 6-year 3x growth equity funds compound to 27x, outpacing a 15x VC fund over the same 18 years — because students pay tuition in dollars, not IRR.

This isn't pedantry. He's run the math: an 18-year VC fund returning 15x sounds impressive. But take three consecutive 6-year growth equity funds each returning 3x over the same 18 years, and you get 27x — twice as much. Students pay tuition in real dollars, not IRR.

Higher Education Is Drifting Into a Slow-Motion Financial Crisis

The context in which Morehead took the CIO seat matters. The US high school graduate population is declining. International student visa restrictions are tightening. Both groups happen to be the primary sources of full-price tuition. Baylor isn't an outlier — it's a case study in the long-term structural pressure bearing down on American higher education.

He started preparing five years ago. "The nice thing about demographic trends is they move slowly — you can see them coming half a decade out." As universities grow increasingly dependent on tuition revenue that's becoming harder to sustain, the annual distribution from the endowment — currently around 5% of the fund — will become an ever more critical income source. That judgment shapes the entire portfolio logic: private markets can no longer be a nice-to-have. They have to be a core part of a compounding machine.

Baylor's current allocation sits at roughly 45–47% private, 53–55% public. Morehead is explicit that real assets like real estate are being wound down, with no plans to renew them. What's staying is VC, growth equity, and buyouts — "if you're going to lock up capital, you need to be earning the highest returns."

Capital Velocity, Not Just Capital Returns

The 18-year fund math isn't just a clever talking point. The underlying logic is worth unpacking.

Traditional VC fund cycles have stretched. Ten to twelve years is ancient history — fifteen to eighteen is the new normal. The GP incentive is easy to understand: hold a great company longer, turn a 6x into a 15x, the marketing materials look better, the next fundraise goes smoother. But for an institution like Baylor that needs to distribute capital continuously to fund university operations, the math doesn't work.

Morehead's phrase for this is velocity of capital. A 3x return at year six, reinvested and compounded into another 3x — eighteen years later, that's 27x. A 15x returned at year eighteen is a smaller number, and the students see the money later. "Students can't pay tuition with an IRR. They need dollars."

This isn't an indictment of VC as an asset class — it's a critique of fund structure. Morehead is candid that Baylor hasn't cracked the top-tier VC firms (Sequoia, Benchmark, and their ilk). "It's not for lack of knocking. The door just hasn't opened." So Baylor's VC allocation currently functions primarily as diversification. The real return driver is growth equity — he mentioned their growth portfolio is running at roughly 30% annualized, far above the 8–9% portfolio benchmark.

Using Separate Accounts to Escape the Tyranny of the Average

Over the past few years, Baylor has done something relatively uncommon for its size: going directly to GPs to negotiate separate accounts — fund-of-one structures.

Morehead explains why. A commingled fund means a single GP has to serve hundreds of LPs simultaneously, so what you get is an average risk-return profile. What Baylor needs at any given moment may look nothing like that average.

He gave a concrete example: if the portfolio already has heavy Nvidia exposure and the next GP wants to add more, Baylor can see the total position and say no. Or the reverse — a GP is adding Nvidia and Baylor happens to be underweight, so they tell the GP: give us three times the normal allocation.

This kind of customization requires a large enough check to negotiate with and a clear enough view of the total portfolio to act on. He said the approach has worked "surprisingly well" over the past two to three years.

Software Dropped 50%. He Started Making Phone Calls.

In early 2026, US software stocks fell 50–60% from their October 2025 highs. The market narrative was that AI would disrupt everything and software would go to zero. Baylor bought.

Morehead's analytical entry point wasn't a DCF model. It was a phone call. He has a few friends running private family businesses with 300 to 500 employees. He called them and asked one question: if your daughter-in-law built something with vibe coding, would you rip out your CRM and replace it? The answer: absolutely not.

He also cited something the Salesforce CEO said — that around six to eight months ago, AI accuracy was running at roughly 93%, already better than many humans. But software needs 100%. The books have to close. The workflows have to run. 93% doesn't cut it.

His conclusion: in many verticals, software won't be replaced by AI — software will become the delivery vehicle for AI capabilities. SaaS companies won't sit back and watch themselves go from multi-billion dollar valuations to zero. They'll embed AI into existing products and push it out through existing customer trust relationships.

The buying was staged. He called Shawn, the growth equity portfolio manager, every day for four weeks. They went through specific company names together, discussing which ones were least likely to be disrupted by AI. The decision framework was human behavioral judgment; the stock picking was left to the professional.

Never Fire All Your Bullets

How Baylor bought into the software selloff is itself worth noting. It wasn't "software is cheap, go all in." It was a mechanical, staged entry process.

Morehead thinks about market drawdowns in 10% increments. Down 10% or less — he doesn't care, that's noise. Down 20%, he starts considering entry. Down 30%, he adds more. Down 40%, he adds again. Each tranche consumes a portion of portfolio liquidity, not the whole thing.

His framing is vivid: in practice, they have never fully deployed their dry powder before the market bottom. Usually the market starts recovering while they still have room to buy. Yes, this means leaving some bottom gains on the table. But what it buys is never being caught in the psychological trap of "I'm fully committed and it's still falling."

"You have to get scarred by a trade to really understand it — you think you're right, you go heavy, and it keeps going down. That's a very uncomfortable place to be." The mechanical buy structure is designed to strip emotion out of the decision.

Cash is an active variable in the portfolio. When he can't find opportunities returning more than 8–9% annualized — he prices the opportunity cost of cash at roughly 8.5%, blending 3.5% in interest with a 5% market alternative — cash accumulates. When he can find something, cash gets deployed. He said their cash weighting ran as high as 15–16% during 2017–2019, but has stayed very low since. "Because we keep finding things running at 20–30% annualized."

Meaningful Exposure Means $3 Million per Company

Morehead described the moment that changed how he thinks about position sizing. A GP sent a notice: a portfolio company had been sold, 7x return, Baylor would receive $400,000.

His first reaction: what's the point?

Since then, Baylor has sized individual fund commitments from the bottom up, starting with the underlying company. They target roughly $2.5 to $3 million of exposure per portfolio company. If a fund holds 10 companies, that's a $30 million commitment. If one of those 10 companies delivers 5x, Baylor gets back $15 million. That's what "meaningful to the fund" looks like.

He has no patience for the alternative math — "we're 10% of the fund, so we have 1% indirect exposure to each company" — because it's still thinking in percentages instead of actual dollars. "Students can't pay tuition in percentages."

The Crack Between GP Incentives and LP Math

Morehead is blunt about this structural misalignment. The VC incentive system is designed to optimize for the manager's business, not for LP compounding.

Hold a great company an extra three to five years, turn a 3x into a 6x, and the next fund's marketing deck looks better. But for Baylor, those three to five years of locked-up capital, if returned and redeployed, might compound at a higher rate. He calls this "GP incentives and LP math not being aligned."

He doesn't dismiss VC's value — Baylor holds roughly 2.5% of the endowment in Anthropic, though he quickly adds that this was a manager's call, not his. His view: VC is primarily a diversification tool for Baylor, not a source of expected excess returns. The excess return engine is growth equity.

Move Positions Without Telling Me, You're Done

Morehead has a baseball analogy for managing external managers, and it makes a very specific point.

He casts himself as a team's general manager: he hires a third baseman, a shortstop, and a second baseman, each with a role, because the overall roster needs that structure. If he walks into the ballpark one day and finds two players on second base and nobody covering third, the third baseman is fired — regardless of how well he's been playing at second.

In practice: if a GP originally described their mandate as "companies with proven product-market fit," and then starts backing two-person teams with unvalidated ideas, that GP gets cut from Baylor. If they drift from late Series A to early Series B, fine — that's the same thing.

The logic: before Morehead can execute asset allocation properly, he needs each manager to stay in the slot he designed. He's accountable for the whole portfolio, not for each individual manager's opportunistic judgment calls. "If you think you can do something different, come tell me first."

Underlying this framework is a stated priority: Baylor spends more time on asset allocation than on manager selection. He acknowledges that's a somewhat contrarian stance in the industry.

Private Credit Is a Bad Business

His view on private credit is simple: pass.

The risk-return structure is broken — on the downside, private credit carries equity-like exposure, but on the upside, it doesn't earn equity-like returns. "We prefer equity." He didn't elaborate further, but the position was clear: private credit is popular right now primarily because it's easy to raise capital for, not because the returns justify it.

Data Center Permits Have Become Scarce Assets

On AI infrastructure, Morehead has a very specific observation drawn from his own portfolio.

Baylor holds data center land assets. Six months ago, those assets appreciated 50%. Not because of any technological breakthrough. Because of permits.

He described a shift that's been playing out: the critical factor in data center site selection used to be land. Then it became land with power. Now it's land with power and permits — and the latter is becoming acutely scarce. Data centers require massive amounts of electricity and water, and local residents are pushing back. Yard signs. Pressure on permit boards. Permit board members who want to get re-elected start saying no.

The response from energy companies, he said, is telling: they're proactively calling data centers that already have permits, offering to deliver power ahead of schedule.

He's more pessimistic about Europe. "We have a permitted data center site in the UK. That alone makes it worth a lot." He sees Europe broadly falling behind on AI infrastructure buildout, compounded by geopolitical and defense restructuring. His current European positioning is primarily long-short, with a macro hedge on European equity indices.

Over the Next Decade, Biotech Beats AI

When asked what excites him most, Morehead didn't circle back to AI. He said biotech.

His view: biotech's impact over the next ten years will exceed what it achieved in the prior twenty. The reason is a fundamental shift in what scientists are doing — moving from treating symptoms to solving diseases. An added advantage is relatively low market correlation; biotech's scientific progress doesn't track macro bull and bear cycles the way most asset classes do.

Baylor already has meaningful biotech exposure, but he said they're actively considering adding more. He planned to attend a biotech conference that week, with another in October. This isn't a distant thesis — it's a position they're actively building.