Your Treasury Is Not a Parking Lot

The CFO of a public company and the CIO of a university endowment both manage pools of capital, but they are optimizing for fundamentally different objectives. The CFO’s pool is instrumental — cash is held as optionality for M&A, as insurance against liquidity shocks, or as dry powder for the next buyback window. Even at companies that sit on enormous cash balances, the pile is always in service of some future deployment. Nobody in the public-company world frames a cash reserve as terminal, as an end in itself, meant to exist indefinitely and throw off income forever. If they tried, the analysts would downgrade and the activists would show up before the next earnings call.

The endowment CIO operates under the opposite mandate. The corpus is the point. It is not parked, not optioned against, not waiting to be deployed. It exists permanently so that its spending yield can backstop the institution through every cycle, regardless of what happens in the operating environment. The university can lose a class of donors, weather a financial crisis, and survive a decade of bad demographics because the endowment is decoupled from the annual operating result.

That difference — instrumental versus terminal — is the real fork in the road. And for founder and employee-owned businesses, the endowment frame is one that is almost never considered. Almost nobody is running the playbook.

Nothing in the conventional playbook is wrong per se. It’s simply answering a different question. The standard corporate finance machinery assumes a shareholder base that is diversified, liquid, and time-agnostic. Those shareholders can self-insure against any single company’s volatility by holding it in a basket. They can translate their patience or impatience into action by selling. They have a hurdle rate set by the broader market — if the company can’t earn it, they want their capital back to deploy elsewhere.

Founder and employee-owned businesses have none of those properties. The owners’ equity and their livelihood come from the same firm. They cannot self-insure by diversifying, because the firm is the position. They cannot translate preferences into action through the market, because there is no market. Their hurdle rate is not set by the S&P 500; it is set by survival and continuity. Employee-owners in an ESOP are even more exposed than founders, with both wages and retirement savings anchored to a single business.

When your shareholders look like that, their objective function stops looking like a public-company shareholder’s and starts looking like a nonprofit’s. A university cannot return capital to shareholders. Its job is to fund a mission indefinitely, across market cycles and across generations. That is the actual goal structure of a closely-held operating business whose owners and workers want it to exist in fifty years. The institution matters more than the quarterly result, the continuity matters more than the optimization, and the capital exists to make the institution permanent.

Furthermore, it’s unrealistic and inefficient to expect each employee to develop a sophisticated investment strategy or even have access to all markets (like private investments), and it would be better to pool their funds together.

Once you accept the reframe, most of the conventional treasury advice either inverts or stops applying.

The practical question becomes: what are you building toward? The answer sits in two stages.

Stage 1 is the Operations Corpus. The goal is a fund large enough that a conservative spending rate (4% to 6% of a trailing three-year average, which is standard endowment practice) covers a defined operations floor. The floor is not the full budget. It is the minimum viable version of the business: core payroll for people who cannot be lost, rent, insurance, debt service, and the small number of fixed obligations that keep the lights on during a bad year. If that floor is $5 million annually, a 4% spending rule implies a corpus target of roughly $125 million. At $2 million of floor, you are building toward $50 million.

This is a multi-decade project, funded from retained earnings (accounting for dividend distributions) on a disciplined contribution schedule. It will feel slow for years and then, eventually, it will feel like a fortress. A business with a fully funded operations corpus can absorb a bad year, a lost anchor customer, a recession, a failed product launch, or a necessary strategic pivot without ever touching the operating business for survival capital. It can say no to bad contracts. It can walk away from abusive customers. It can refuse to sell at the bottom of a cycle. Most closely-held businesses cannot do any of those things, which is why most closely-held businesses end up doing what their balance sheets force them to do rather than what their owners actually want.

Stage 2 is the Pension Corpus. Once the operations floor is backstopped, subsequent retained earnings fund a defined-benefit-style retirement promise for long-tenured employees. This is not a 401(k) match. It is the kind of pension that used to be standard in American corporate life before the shift to defined contribution offloaded the risk onto workers. We will come back to why this matters beyond the firm.

The most common mistake closely-held companies make with their treasury is allocating it the way a wealthy individual would. That is a category error. A wealthy individual with a diversified portfolio is building toward an efficient frontier; they do not already have a massive concentrated bet on a single private company. You do. The operating business itself is a large, illiquid, concentrated equity position with exposure to specific end markets, specific suppliers, specific customer concentrations, and specific macroeconomic drivers.

The treasury’s job is to be what the operating business is not. If you run an industrial services company with exposure to U.S. construction cycles, the treasury should not be overweight cyclicals and real estate. If you run a software business with high exposure to enterprise IT budgets, the treasury should not be concentrated in tech equities. The correlation between the operating business and the treasury is the quantity that determines whether the corpus actually functions as a shock absorber. If they move together, it doesn’t.

The endowment playbook points toward broad global equity exposure, fixed income as genuine ballast, real assets for inflation protection, and some measured allocation to alternatives — with asset allocation bands and rebalancing rules written down. The discipline matters more than the specific mix. What matters most is that the mix is chosen against the operating business, not in parallel to it.

A spending rule smoothed over a trailing three-year average does work here that is easy to underestimate. It decouples the treasury’s draw from both the treasury’s and the business’s current-year volatility. When the business has a bad year and the portfolio has a bad year, the spending rule looks at a three-year average and hands over a stable number. That is the mechanism that lets the corpus actually function as insurance.

Endowments and foundations run on a written Investment Policy Statement. The IPS specifies the corpus target, the spending rule, the asset allocation bands, the rebalancing policy, the governance structure, and the review cadence. It is the artifact that turns investment practice into institutional practice.

Most closely-held businesses have nothing of the kind. The treasury policy is whatever the owner feels like that quarter, adjusted for whichever banker or advisor spoke most recently. That works for as long as the owner is around and engaged. It does not work across transitions — successor generations, ESOP transactions, key employee departures, the founder’s eventual absence. The IPS is the mechanism that makes the strategy survive the founder and become a property of the institution rather than the individual.

Writing one is not complicated. It is a five-to-ten page document. The act of writing it is the point — it forces the owners to commit in advance to decisions they would otherwise make emotionally in the moment.

This is where the strategy meets the U.S. tax code, and where most of the half-formed versions of this idea run aground. The code is navigable, but it is not indifferent. A few provisions matter enormously.

The Accumulated Earnings Tax is the first real constraint for C-corps. A 20% penalty applies to earnings retained “beyond the reasonable needs of the business.” The doctrine has substantial case law and is less scary than it sounds, but it is also not automatic. The IRS recognizes several legitimate categories of retained earnings: working capital needs (often calculated via the Bardahl formula from the 1965 Tax Court case), business expansion plans, product liability reserves, self-insurance reserves, retirement of debt, and employee retirement obligations. The endowment framing is not on the list by that name, but several of its components are. The practical requirement is documentation. Board minutes that tie retained earnings to specific recognized categories — particularly employee retirement obligations and self-insurance reserves — go a long way. What you cannot do is retain earnings, invest them passively, and tell the IRS the plan is “indefinite accumulation.”

The Personal Holding Company tax is a separate 20% trap that catches companies where too much of the income has become passive. If 60% or more of adjusted ordinary gross income is passive — dividends, interest, rents — and five or fewer individuals own more than 50% of the stock, the PHC rules apply. A corpus large enough to matter will throw off enough investment income to push a smaller operating company over the threshold. The structural answer is not to hold the corpus inside the operating entity.

Pass-through entities (S-corps and LLCs) face a different problem. Retained earnings are taxed to the owners personally whether distributed or not. Retaining large sums inside the entity creates personal tax bills without corresponding personal cash flow, which forces the owners to either distribute enough to cover the tax (reducing the retention), accept the drag, or restructure. For a pass-through building toward a substantial corpus, the usual answer is to move the corpus into a parallel structure.

The parallel holding company structure solves most of the above cleanly. The operating company up-streams excess cash to a holding company, and the holdco holds the corpus. This separates the operating entity’s character from the investment pool, which addresses both the AET and PHC concerns. It also simplifies succession, insulates the corpus from operating-company liabilities, and creates a cleaner platform for future transactions. For ESOP-owned companies this is the natural structure anyway: the ESOP trust owns the holdco, the holdco owns both the operating company and the corpus.

Captive insurance companies are a legitimate tool for self-insuring real business risks while building a tax-advantaged reserve. Premiums up to the annual statutory limit (currently $2.85 million, indexed) are deductible to the operating company and received tax-free by the captive, which pays tax only on investment income. The corpus inside the captive grows against retained underwriting profit. The IRS has been aggressive about abusive micro-captives, and rightly so — many were thinly disguised tax shelters with no real risk transfer. The legitimate version requires real insurable risks that are currently uninsured or underinsured, real actuarial work to price the premiums, a real claims process, and diversification of risk. When those conditions are met, a captive is one of the most efficient corpus-building vehicles available.

Nonqualified deferred compensation and SERPs are the mechanism for the Stage 2 pension promise above the limits of qualified plans. The corporate side holds an unfunded promise to pay, which can be informally funded via a rabbi trust holding the actual investment assets. Rabbi trust assets remain subject to the company’s creditors in bankruptcy, which is the feature that keeps the arrangement tax-deferred to the employee. This is the actual plumbing for promising a key employee a defined retirement benefit.

Qualified plans, defined benefit pensions, cash balance plans, and ESOPs, are the tax-advantaged side of the pension corpus. Defined benefit plans have fallen out of corporate fashion, which is precisely why they are underrated for closely-held companies with stable cash flow. A DB plan generates large deductible contributions, ERISA-protected assets separate from the company, and a genuine pension promise to employees. Cash balance plans are a DB variant that looks more like a 401(k) to participants while retaining the deductibility advantages for the sponsor. For a closely-held company in Stage 2, some combination of DB, cash balance, 401(k) match, and nonqualified SERP is the standard stack.

ESOP-specific considerations deserve their own paragraph. An S-corporation owned entirely by an ESOP pays no federal income tax on the ESOP-owned share of earnings. Not deferred — not paid. This is the single largest tax advantage available to an operating business in the U.S. code, and it is almost perfectly aligned with the endowment strategy. The tax savings are the funding source. An S-corp ESOP with a disciplined retention policy can build an operations corpus and then a pension corpus essentially out of the federal tax liability it would otherwise owe. The strategy was subtly designed for this.

None of this is exotic. All of it requires a competent tax advisor, ERISA counsel, and an investment advisor who has worked with foundations or endowments rather than only private wealth clients. The tools exist. The playbook is just not widely run.

Stage 2 is where the strategy earns its moral weight, and it connects to a policy conversation that deserves more serious engagement than it has received, and one of my personal passion projects.

The current debate over AI and automation has settled into two main responses. The first is retraining, which has a weak track record. The second is universal basic income, which is politically stuck and fiscally heavy, and which severs the link between work and income that most people actually want to preserve, even though studies have shown that when given basic income, people actually work more, not less. Both responses treat the problem as a federal problem to be solved through federal redistribution.

While, I’ve developed a more efficient alternative to UBI that could work at the federal level, there is a third answer, and it is already working wherever closely-held businesses have taken it seriously. Firms that capture the surplus from automation and productivity gains can retain it inside permanent-capital structures that fund real pensions for the people who helped build the firm. This is not new. It is what American corporate capitalism looked like for most of the twentieth century, before the shift to defined contribution plans offloaded the retirement risk onto individual workers and the shift to shareholder primacy made retained-earnings pension funding look like capital that should be returned.

The endowment-style treasury puts the mechanism back, but embeds it in the firm rather than outsourcing it to capital markets. The operations corpus makes the firm survivable. The pension corpus makes retirement for long-tenured employees survivable. Together they turn a business into something closer to an institution — a guild, a mutual, a trust — that outlives the founder and provides for the people who contributed to building it even when people are no longer needed to run most of it.

The argument for this as a response to labor-share compression is straightforward. The modern incarnation of Western Capitalism cannot survive AI. If AI and automation do what their proponents and critics both expect, the firms that capture the after-tax surplus will have a choice about where it goes. They can return it to diversified capital, in which case the gains accrue to whoever happens to own the right index funds…the already wealthy. Or they can retain it inside permanent-capital structures that fund the workers who built the firm, thereby converting a portion of the automation dividend into pensions earned at the site where the surplus was generated.

This is a private, voluntary, firm-level response to the same problem UBI tries to address at the federal level. It has some advantages worth naming. It preserves the link between work and income. It is funded by the specific surplus the workers helped create, not by redistribution from unrelated parties. It does not require political consensus to begin; any individual owner or ESOP board can start next quarter. And it produces a pension, which is a fundamentally different promise than a subsistence transfer. A pension is earned, it belongs to the worker, and it carries the dignity that goes with both of those things.

This is not a full substitute for the federal conversation about automation and labor. Many workers are not in firms that could credibly run this playbook, and many of those firms will not choose to. But at the scale of the closely-held and employee-owned sector, which employs a meaningful share of the American workforce, the aggregate impact of widespread adoption would be substantial. It is a strategy that can be pursued one firm at a time, by owners and boards who do not need anyone’s permission. Policymakers at the state and local level along with community leaders and labor unions can create policies to support employee-owned businesses that implement this strategy.

An operating business paired with a real endowment is a permanent capital vehicle in the truest sense of the term. No redemption rights. No forced sellers. No quarterly reporting. No pressure to dividend or buy back. The corpus funds the perpetuity of the institution and, eventually, the retirement of the people who built it.

The tools exist. The tax code largely accommodates the strategy for owners willing to work within its structure. The playbook is available to any founder-owned or employee-owned business with the discipline to retain earnings on a schedule and the governance to write down what they are doing and stick to it.

What is missing is the framing. Treasury management in closely-held businesses has been treated as a technical function borrowed from the public-company world, where the job is to minimize the cost of capital and keep the balance sheet efficient. For a business whose owners intend to exist in fifty years and whose workers have staked their livelihoods on it, that is the wrong job. The right job is the endowment’s: build the permanent capital, protect the institution, fund the mission, pay the people.

That is what permanent capital actually means.

If you need 1:1 help developing a treasury policy, fill out this form to get started.