What Are the Odds an SBLOC Actually Gets Called? A Century of Market Paths

There is a question every stress test dodges.

Ask "would the 2008 crash have called my line?" and you get a clean answer: run the drawdown, check the loan-to-value, done. We built that calculation and it is genuinely useful. But it answers a question nobody actually has. Nobody is planning around a repeat of 2008 specifically. They are planning around holding a line for ten years and wondering whether, at some point in there, they get the call.

Those are different questions, and the second one is harder in a way that is easy to miss. A stress test inspects a single moment. A ten-year holding period is a path — and a portfolio can fall 40%, trigger a call, force a sale at the bottom, and be fully recovered by year eight. The stress test that only looks at endpoints never sees that event at all. It happened anyway.

So we went and measured it.

What We Did

We took every month of US market total returns from July 1926 through June 2026 — a century, 1,200 months — and resampled them into 200,000 alternative market histories per scenario, then asked how many of those histories would have breached a maintenance threshold at least once.

The resampling matters. The obvious approach is to draw each month independently from a normal distribution, which is what the textbook model does. It is also wrong in a specific way that matters enormously here: markets do not fall in isolated bad months, they fall in sustained runs. Independent draws scatter the bad months evenly and never produce a 2000-2002 or a 2007-2009. Since a collateral call is precisely what happens when bad months arrive consecutively, a model that cannot produce sustained declines cannot answer this question.

We use a stationary block bootstrap instead: rather than drawing single months, it draws runs of consecutive months, averaging about two years each. Crashes, and their recoveries, survive into the simulated paths intact.

How much does that choice matter? In our default drawn-line scenario:

Model	Ten-year call probability
Lognormal (independent returns)	1.5%
Resampled market history	7.4%

Assuming independence understated the risk roughly fivefold. Both figures are available in the calculator, and the gap between them is one of the more honest things on the page.

The Number That Matters Most Is Not the One You'd Guess

Here is what a century of paths says about a $3M portfolio backing a line, held ten years, at a 75% maintenance threshold:

Drawn against a $3M portfolio	Diversified	Concentrated single stock
$300K (10% LTV)	under 1%	4%
$1M (33% LTV)	7%	29%
$2M (67% LTV)	54%	84%

Read across the rows, not down them.

At an identical loan-to-value, pledging a concentrated single-stock position rather than a diversified portfolio takes ten-year call probability from 7% to 29%. Same loan. Same collateral value. Same expected return. Four times the risk, entirely because one position moves twice as much as the other.

This is the part standard SBLOC guidance gets wrong by omission. Advance rates are quoted against position type — lenders do haircut concentrated positions — but the risk conversation is almost always conducted in terms of how much you drew. How much you drew is the second most important variable. What you pledged is the first.

The mechanism is not complicated: collateral calls are driven by volatility, not by expected return. A concentrated large-cap position carries roughly 36% annual volatility against about 18% for a diversified portfolio, and the probability of touching a barrier is far more sensitive to how much a portfolio moves than to where it is heading on average.

That should be uncomfortable reading for exactly the person most likely to open one of these lines. The post-IPO founder, the long-tenured employee sitting on vested RSUs, the early executive with most of their net worth in one ticker — these are the people an SBLOC is marketed to hardest, because it solves a real problem for them: liquidity without a taxable sale. They are also the people whose collateral is least suited to backing a loan. And it compounds with the cure problem: a low-basis concentrated position is both the most likely to be called and the hardest to cure by selling, because the tax leak on a sale is largest exactly where the embedded gain is largest.

Capitalizing Interest Costs More Than the Interest

The calculator offers two ways to handle SBLOC interest: pay it monthly, or let it capitalize onto the balance. Capitalizing looks like a cash-flow convenience.

It is a risk decision, and here is the cleanest way to see why. A loan balance growing at 4.5% erodes your headroom at exactly the same rate as a portfolio shrinking at 4.5%. To a maintenance test, those two situations are indistinguishable — it only ever compares one number to the other. So capitalizing interest at 4.5% against an 8% expected return does not leave you growing at 8%. It leaves you growing at 3.5%, as far as your collateral call risk is concerned.

On the $1M line above, that took ten-year call probability from 7.4% to 12.3%. The interest cost is the visible price. The headroom is the one nobody quotes.

(This is also why the model needs only a single "net drift" input rather than tracking the portfolio and the balance separately — the two collapse into one number. That reduction is what makes the whole simulation cheap enough to precompute.)

Where the Simulation Can Be Wrong

Publishing a probability creates an obligation to say where it comes from and where it breaks. The input series is public — it is the Kenneth R. French Data Library's monthly factor file, with market total return taken as Mkt-RF + RF — and the method is described in full below, so the figures can be checked independently.

The concentrated regime is calibrated, not observed. This is the weakest link and worth stating plainly. Comprehensive single-stock return histories are not freely available, so rather than resample real single-stock paths, the model keeps the market factor and adds an independent shock sized to published idiosyncratic-volatility levels for large-cap US equities, landing at about 36% annual volatility. That is a fair figure for an established large-cap holding. A recent-IPO position would be materially more volatile, so treat 36% as the conservative end of the range.

There is a second, subtler conservatism: individual stocks have a lower median outcome than the market even at equal average return — most individual stocks underperform, with the index carried by a minority of big winners. Our concentrated regime reproduces the higher volatility but not that negative median skew. If anything, it understates concentrated risk.

We include the Great Depression. Starting in 1926 rather than 1950 roughly doubles measured volatility (18.3% against 15.0%) and makes every figure more conservative. That is deliberate. You do not get to exclude the worst outcome in the record because it is inconvenient, and the 83% decline of 1929-1932 is the single most informative event in the sample for a question about collateral calls.

Monthly data, continuously monitored. Return data is monthly, but a lender watches your account continuously — a portfolio can fall through the threshold mid-month and recover by month-end, and you would still have gotten the call. Comparing only month-end values would systematically understate risk, so the simulation interpolates within each month to catch those crossings. Understating risk is the wrong direction to be wrong in.

These are historical frequencies, not forecasts. The model says what fraction of resampled market histories would have breached the threshold. It does not say what the market will do. A century of data is a large sample of a single realized history, not a sample of all possible futures.

Maintenance thresholds are the lender's, not ours. We default to 75%, which is mid-range for a diversified equity portfolio. Your agreement governs, thresholds vary by collateral type, and most agreements permit the lender to change them or to liquidate without advance notice. That last clause is the one to read twice.

What To Do With This

The break-even decline — how far your portfolio can fall before a call — is available in closed form and does not require any simulation at all:

1 − (loan ÷ maintenance threshold) ÷ portfolio value

At a 75% threshold, a line drawn at 10% of portfolio value absorbs an 87% decline, deeper than anything on record. At one third, 56%. Drawn to a 70% advance rate, about 7% — an ordinary quarter.

That single number does more to determine your outcome than any assumption about markets, which is the practical takeaway: the decision that governs your risk is the size of the initial draw, and you make it once, on day one, when nothing feels risky. The simulation mostly exists to tell you what that decision was worth.

You can run your own numbers in the calculator — Scenario C reports the call probability, the typical timing, and what curing a call would cost at your own cost basis and tax rates.

Tyler Singletary is the founder of Stockstead. This is educational analysis, not investment advice, and not a recommendation to open or avoid a securities-backed line of credit. Stockstead is not a licensed investment adviser. See our disclaimers.