You Can Borrow Against Stocks Without Selling. The Question Is Who Approves the Loan.

You can borrow against stocks without selling them. Once a margin account is open, the facility is simply there — each drawdown needs no application and no underwriter. That is the part everyone writes about, and it is the easy part. The hard part is what a real lender does before releasing a penny: underwrite the loan. Wealthy families call the structure a family bank. What makes it a bank is not the borrowing. It is the credit committee.

What does it mean to borrow against stocks without selling?

To borrow against stocks without selling is to pledge the securities you already own as collateral for a loan, keeping the shares — and their dividends, their compounding, and their unrealised gains — while spending the borrowed cash. In a brokerage account this is a margin loan. Elsewhere it is a securities-backed line of credit. The mechanics differ; the shape does not. You swap a sale for a debit.

Here is what that looks like on a real-shaped book. The figures below are illustrative — shaped like my own, rounded for legibility — because I run a leveraged book myself and these are the two numbers I read first: $1,880,000 of securities, against a margin loan of $606,000. No cash line — an account carrying a debit does not also hold idle settled cash; the loan is the negative balance. Divide the loan by the securities and you get 32.2%.

That number is the whole conversation, and the definition behind it is not standardised. Ours is stated plainly: margin loan ÷ gross securities value. Not net of the loan, not including outside cash. Be aware that some brokers publish a figure they also call "margin utilization" which measures something else entirely — margin used against margin available — so the same name can return a very different ratio. Say which denominator you used, every time, or the percentage is decoration: the four margin zones and how much margin is too much covers why that one choice moves a book across zone lines.

At 32.2% this book sits in Harvest — the second of four zones we govern to. It looks fine. It reads as prudent. Hold that thought.

Why the wealthy call it a family bank

The phrase is seductive and mostly accurate. A family that has assembled real capital stops selling assets to fund its life. It borrows against them, services the interest from income the assets throw off, and leaves the compounding undisturbed. Functionally, the family becomes its own lender. Hence: family bank.

What the phrase leaves out is the institution around the loan. A bank is not defined by the fact that it lends. It is defined by everything it does before and after it lends — the underwriting, the covenants, the periodic review, the person whose entire job is to say no. Strip those away and you do not have a bank. You have a debit. The same test decides between a policy and a brokerage account: which structure actually makes you keep funding it.

This is the part of the family-office model that does not travel. The borrowing travels — anyone can do it today. The governance is what stays behind: the written policy, the scheduled review, the committee that reads the file and rules on it. That layer, sitting above strategy and below execution, is what portfolio governance actually is, and it is the thing solo investors skip without noticing they have skipped it.

So the honest version of the family-bank idea is not "borrow like the wealthy." It is: you already have the lending facility. What you don't have is the risk department.

The five questions a credit committee asks

You can run the committee yourself. Not as a metaphor — as five specific questions, answered in writing, on your own numbers, this week. An institutional lender would not release funds without answers to all five. In my experience the last one is where solo books come apart.

Take fifteen minutes and a brokerage statement. This is the copy-paste artifact of this post:

Question 1 nearly everyone passes — the number is on the statement and the arithmetic is one division. Question 5 is the one I have most often seen go unanswered, and I want to be straight about why: it needs last week's reading, which means keeping a dated snapshot every week without fail. Most people do not have that, and no amount of willpower produces it retroactively. That is a real limitation of doing this by hand, and it is precisely the gap the rest of this site exists to close.

What happens when the collateral falls?

Question 2 is where the comfortable book stops being comfortable. Return to the numbers: $606,000 borrowed against $1,880,000, reading 32.2%. Now let the market take 20% off the collateral. You sell nothing. You borrow nothing further. You do not touch the account at all.

The securities are now worth $1,504,000. The loan is still $606,000, because debts do not participate in drawdowns. The reading is 40.3%.

Reading	Securities	Loan	Utilization	Zone
Today	$1,880,000	$606,000	32.2%	Harvest
After a 20% fall	$1,504,000	$606,000	40.3%	Forced

Two zones, in one move you did not make. And 20% is not the threshold — it is merely a round number. The fall that puts this book exactly on the Forced line is 19.4%, which is shallower than the 20% decline conventionally used to mark a bear market. Not an everyday wobble, then — but not a crisis either, and reached without anyone doing anything wrong.

Two caveats belong beside that 40.3%, and both push the same way. First, it assumes no assignments. On a book that writes options for premium — which this one does, see the next section — a sharp fall is precisely when short puts go into the money and get assigned, and an assignment is funded by an increase in the debit. Second, short options carry margin requirements of their own that a simple loan-to-collateral ratio does not capture at all. Treat 40.3% as a floor, not a point estimate.

One more thing the ratio hides over longer horizons: margin interest accrues into the debit. If the income does not actually land, the loan grows by roughly the interest bill each year, and utilization climbs without the market moving at all.

The four zones we govern to are Clear below 30%, Harvest from 30% to under 35%, Freeze from 35% to under 40%, and Forced at 40% and above. Every boundary belongs to exactly one zone, which sounds pedantic until a book reads 40.00% and two documents disagree about what to do. The margin pillar sets out each zone and what it instructs in full.

Zone changes normally have to prove themselves. Our two-week persistence rule requires a breach to appear in two consecutive weekly readings before it drives any action, so that a single volatile Friday cannot conscript you into selling. That protection only holds while somebody is still taking the reading, which raises who runs the book if you cannot. Forced is the one exception: a book in Forced is acted on in the week it is first read. It does not get a second opinion, because the second opinion costs a week of exposure in the one zone where a week is expensive.

There is a second consequence worth naming. A book in Forced is capped in the worst band of the Governance Score — a maximum of 34 out of 100, "Exposed / At the Edge" — no matter how disciplined the rest of the operation is. Margin is weighted that heavily on purpose. It is the line item most likely to end the game.

What services the debt? The Margin Coverage Ratio

Question 4 is the one that separates a governed liability from a hopeful one, and it is the question I have most often heard skipped. A lender never assesses a loan on collateral alone. It asks what pays the interest.

We measure it with a ratio we call the Margin Coverage Ratio — the income you plan to realise, divided by the annualised interest on the loan. On this book: interest at an illustrative 6.0% on $606,000 is $36,360 a year. (Use your own broker's rate; it is on your statement, and it changes.) Against that sits the Income Engine Layer — in our taxonomy the Incomestead Stack is the whole framework and a Layer is one strategy within it, and the Income Engine is the Layer written for premium rather than appreciation. It plans $46,000 of realised income over the year.

$46,000 ÷ $36,360 = 1.27.

1.27 is Acceptable. Not comfortable — acceptable. Note what the numerator is: planned income, counted only when it is actually realised. We book option income on expiry, buyback, or assignment, on a cash basis, never marked to market. An unrealised gain has never paid an interest bill. In our own framework this is a quarterly control rather than a weekly one; run it whenever the loan or the rate moves. Carried across a lifetime holding period, the same ratio is what decides how long a borrowed book lasts.

And notice how much weight that 6.0% is carrying. Planned income of $46,000 exactly covers the interest at a rate of 7.6%. Above that, this Layer no longer services its own debit and the ratio falls below 1.0 — the band whose instruction is to reduce. Margin rates differ sharply between brokers and move with the policy rate, and that spread is easily wider than the gap between 6.0% and 7.6%. Before taking any comfort from a 1.27, put your own rate in. On this book the rate is not a detail; it is the answer.

Now put the same ratio the other way round, because this is the number that rarely gets said out loud. Interest consumes 79% of what the Income Engine Layer plans to produce. That Layer clears its share of the interest bill by $9,640 a year — before anything the rest of the book earns or loses, and while a sub-bear-market decline is enough to put the whole book in Forced under our own rules. Whether that trade is worth making is your call, not mine. But it should be made with the net number in view, not the gross one.

And the honest part, which is the reason a committee exists rather than a formula. If the collateral falls 20%, what happens to that $46,000? Two forces pull against each other: a smaller book supports fewer and smaller premium-writing positions, which cuts the number — while a falling market usually brings higher volatility, which raises premium per contract. The direction is genuinely contested and I cannot state the net precisely; I would distrust anyone who hands you a point estimate for it. That uncertainty is the finding. You are servicing a fixed obligation with a variable stream whose stressed value you cannot state. A bank would not lend against that without a covenant. I lent against it for years without noticing there was a question.

What would the committee tell you to do?

Suppose the fall happens and the book reads 40.3%. What does it take to get back inside Harvest — under 35% — at the new, lower valuation? Do this one by hand once, because the shape of it is counter-intuitive and it is where I see most plans quietly fail.

First, the trap. At $1,504,000 of collateral, a 35% reading corresponds to a loan of $526,400. You are at $606,000, so the gap looks like $79,600. It is not.

You have no cash — the loan is the negative balance — so the only way to repay from inside the account is to sell securities. And a sale takes money off both sides of the ratio: the loan falls, but so does the collateral standing behind it. Sell $79,600 and you are left with a $526,400 loan against $1,424,400 of securities, which reads 37.0%. Still Freeze. You have done the work and not moved zones.

That shortfall is not particular to this book — it is a fixed multiple of the gap, and the general form is worked in full in the margin-zones guide.

Solve it properly. With x the amount sold, (606,000 − x) ÷ (1,504,000 − x) = 0.35 breaks even at x = $122,462 — better than half again as much as the gap suggested. That is the sale that lands you exactly on the line, which is not where you want to be.

Then round up, away from the edge, to an increment you would actually transact in. On a book this size, call that $25,000. That takes you to $125,000 — which lands at 34.9%, inside Harvest, with $1,650 of headroom — loan dollars — before the Freeze line. Technically correct, and not good enough: from there the collateral need only slip 0.34%, a single ordinary session, to put the reading back on 35%. Reject it and take the next step that leaves durable room — $150,000, a loan of $456,000, a reading of 33.7%, and $17,900 of headroom.

Why solve back to Harvest rather than merely out of Forced? Because stopping at 39.9% leaves you one ordinary session from re-entering the zone you just acted to leave — and on the way back in, Forced does not grant you two weeks to think about it.

Three things had to be stated to make that instruction reproducible: the direction (up, away from the edge), the increment ($25,000), and the precision it was checked at (one decimal place). Drop any one and a different person running the same rule on the same book gets a different answer — which is the definition of a rule that isn't one. And note what those three do not settle: applied strictly they return $125,000. Rejecting it for thin headroom was my judgement, not the rule's output. A minimum-headroom threshold is a fourth parameter, and until it is written down this last step is not reproducible either. I would rather say that than pretend the rule did the work.

One cost this arithmetic does not carry: selling to deleverage is a taxable event. After a drawdown some lots may be at a loss, and realising those can be useful rather than costly; on appreciated lots the bill can rival or exceed a year's interest. Which case you are in is a question for your accountant and not for me — and it is a reason to settle it in advance rather than while the book is already in Forced.

To be clear about what is mine and what is yours: the $150,000 above is your arithmetic, not an instruction from me or from any engine. What a governance engine does with a Forced reading is narrower and duller — it states the reduction required and ranks the three positions whose sale would move the number most, each with its identifier, its value, the proceeds, and the margin impact. Then it stops. You choose whether, and which, and when.

Is this the same as a margin call?

No, and the distance between the two is important enough that I will not let the previous section imply otherwise.

Every zone in this post is self-imposed. FINRA's baseline is that "generally, a customer's equity in a margin account that holds margin stocks only must not fall below 25 percent of the current market value of the long securities." For such an account, equity and utilization are complements on the same base, so that floor is a 75% ceiling on our own measure — putting our Forced line 35 percentage points of utilization below it. Read that as an order of magnitude rather than an equivalence: the rule is scoped to accounts holding margin stocks only, and this book writes options, whose margin requirements sit outside a loan-to-collateral ratio entirely. Nothing here describes a book near a regulatory margin call — but the two conditions below can put a real account closer to a house call than its utilization suggests.

That comparison holds for a diversified book. It is not unconditional. Firms set their own "house" requirements above the regulatory floor, and they set them per security — a concentrated or volatile position can carry a requirement many times the baseline, occasionally to the point of contributing no collateral value at all. A book that fails Question 3 can be closer to a house call than its account-level utilization suggests. Check the requirement on the position, not the average.

What is worth reading carefully is how little warning the real thing carries. Firms may set requirements of their own, "often called 'house' requirements — that can be higher than the margin requirements under Reg T or the rules of FINRA and the exchanges," and — in FINRA's words — "may increase the house requirements at any time and aren't required to provide you with advanced written notice." Further: a firm "isn't required to notify you if your account equity drops below the minimum maintenance equity," and firms "don't have to issue a margin call before selling securities in your margin account to meet a margin call and may sell enough securities to completely pay off your margin loan, not just meet the margin call." (FINRA, Margin Calls.)

Which is the actual argument for governing to a self-imposed line so far above the trigger. You are not managing the call. You are managing the distance from it — and you are doing it on a weekly cadence, which will never see an intraday move. A weekly reading is a governance instrument, not a margin-call defence. Anyone who tells you a weekly process protects you from a fast liquidation is selling you something.

One more caveat that belongs beside every one of these numbers: leverage amplifies losses as readily as gains, and the collateral concentration behind the loan matters as much as its size. A book where one winner has quietly become the whole portfolio is worse collateral than the same dollar value spread across many names, and no utilization percentage will tell you that on its own.

The bank you already have

The reason a family office can borrow against its assets for decades without a disaster is not superior access to credit. It is that somebody reads the file every week, against rules written down when nobody was frightened, and has standing authority to say the word no.

You have the lending facility already. Every one of the five questions above is answerable with a statement and fifteen minutes — except the fifth, which requires last week's answer too, and the week before that. Which is why the weeks you skip are the dangerous ones: a committee that meets when it remembers is not a committee. It is an intention.

That is the whole gap between borrowing like a wealthy family and merely borrowing. Not the loan. The file.

Frequently asked questions

Can you borrow against stocks without selling them?

Yes. Pledging securities you already own as collateral for a margin loan or a securities-backed line of credit lets you access cash while keeping the shares, their dividends and their unrealised gains. Opening the account requires a signed margin agreement and the broker's approval, but once it is open each drawdown needs no application and no underwriting — which is convenient, and is also why the borrower has to supply the underwriting themselves.

What is a brokerage family bank?

It is the practice of treating your own portfolio as the lender for your household — borrowing against assets rather than selling them, and servicing the interest from income the assets produce. The borrowing is the simple half. What makes it a bank rather than a debit is the governance around it: a written policy, a scheduled review, a stress test, and a standing authority to decline.

How much can you safely borrow against a stock portfolio?

There is no universal number, and nobody honest will give you one for your account. What we can share is the convention we govern to: margin loan divided by gross securities value, with our Clear zone sitting below 30% and our Forced zone at 40% and above. Those are self-imposed limits set far more conservatively than any broker or regulatory trigger, and they say nothing about what is appropriate for your circumstances.

What happens to a margin loan if the market falls?

The loan does not fall with it; the collateral does. Because utilization is the loan divided by the value of the securities, a falling market raises the percentage without any action by you. On the book in this post, a 20% decline moves the reading from 32.2% to 40.3% — from Harvest to Forced — with no trade placed, and a 19.4% decline is enough to reach the Forced line. Repaying afterwards costs more than the gap suggests, because selling securities reduces the collateral as well as the loan.

Stefano Starkel is a practitioner, not a licensed adviser, and runs a leveraged portfolio under the governance process described here. This article is education, not personalised financial, tax or legal advice. Margin borrowing amplifies losses as well as gains and can result in the forced sale of your securities without prior notice; in a sharp decline it is possible to lose more than you invested and to owe the broker the difference. Figures are illustrative and computed for this article; the interest rate shown is an assumption, not a quoted rate. Nothing here is a recommendation to borrow any particular amount, or at all.

Incomestead recommends. You decide.