Premium Financing for Foreign Nationals: What Changed When Money Stopped Costing 2%

Premium financing is a loan. A bank lends the money for the premiums on a large permanent life insurance policy, assets the family already owns are pledged to secure that loan, and the family pays interest each year instead of writing a check for the premium. The capital that would have gone into the premium stays where it is, invested. That is the appeal, and it is a real one.

It fits a narrow client. A family with US assets large enough to create a US estate tax bill, with the net worth to be lent millions of dollars against it, and with the liquidity to pay interest every year for as long as the loan is alive. It does not fit a family that needs the loan in order to have the coverage at all. That distinction reads as a technicality on the first page of a presentation. It is the entire thing.

The reason to write this now is arithmetic. The structure was designed and sold through a decade in which money cost almost nothing, most of the material a reader will find on it was written during that decade, and money has since gone back to costing what it costs. Everything this strategy produces is the gap between two rates: what the bank charges and what the policy credits. When the gap closes, the plan does not simply underperform. It runs the other way.

Start with the company that manufactures the product, writing about it in its own document.

A major US carrier keeps a separate set of underwriting guidelines for premium-financed policies. It is advisor material, not a client brochure. In it, the carrier writes that it "does not recommend, endorse, sponsor or otherwise offer" premium financing, that it is not a party to the loan agreement, and that it receives no compensation of any kind from the financing arrangement. Then it lists the risks it sees: the interest rate, additional collateral requirements, loan renewals, the possibility that the lending bank becomes insolvent, and the scenario in which the loan defaults and the insurance contract lapses.

That is the manufacturer talking about what people do with its product.

There is a second sentence, this one from a US bank that lends for these transactions, and it ends the conversation before it starts: "The ultimate source of loan repayment must be identified, which cannot be the life insurance policy's death benefit." The same bank is perfectly willing to be repaid out of the policy's cash surrender value. Out of the insured's death, no. And this is a bank that lends millions of dollars a year to pay life insurance premiums.

The presentation you were shown was probably built on the opposite assumption.

The material you will find online belongs to a different rate cycle

Search the term and most of what comes back was written by people who sell the structure, in an era when money cost 2%. Very little of it has been revised since. The mechanics it describes are still broadly right. The arithmetic is not.

What we are describing here is specific. A bank loan of several million dollars, in the United States, used to pay the premiums on a permanent policy whose death benefit will cover the US estate tax triggered by a nonresident family's American assets.

1 to 5 yearsloan term1 to 5 yearsloan termWho signs the loan, and why it is not you

The borrower is an irrevocable life insurance trust (ILIT). The trust buys the policy, signs the loan, and collects the death benefit.

That separation is what keeps the face amount out of the insured's taxable estate, and a US private bank flags it in its own published material: the trust has to be placed with an independent, non-subordinate party, to stay clear of the "incidents of ownership" rules in section 2042 of the Internal Revenue Code. For a nonresident family there is a second, statutory reason, and it comes further down.

The loan is interest only. Principal does not amortize. The published material from the banks that finance premiums puts typical loan sizes at $1,000,000 to $10,000,000, with terms of one to five years and a floating rate indexed to Prime or SOFR. Brokers experienced in these structures describe facilities of two to ten years, at the end of which the borrower has to refinance with the same lender or find another one.

Collateral comes in two layers. First, the collateral assignment of the policy's cash surrender value, the amount the policy would return if it were cancelled. Then outside collateral (cash, marketable securities, letters of credit, the cash surrender value of other policies) covering the gap between the loan balance and that cash value, already discounted by the lender. One bank that lends for these transactions requires daily monitoring of the liquid collateral.

The file is not light either. The carrier wants a personal financial statement signed by the insured and by his accountant or his attorney, identification of the assets to be pledged, the lender's term sheet, and a written exit strategy: where the principal is going to come from at maturity. The bank wants three years of tax returns, with schedules and K-1s.

And there are two approvals, not one. The insured can qualify medically for the coverage and the trust can still fail to qualify for the credit.

5.15%cost of money5.15%cost of moneyMoney costs between 5.15% and 5.90%, not 2%

Everything above is structure, and structure barely moves from one year to the next. What follows is price, and price is where this plan is decided.

On July 30, 2026, SOFR stood at 3.65%, according to the Federal Reserve Bank of New York. The bank prime loan rate was 6.75% and the effective federal funds rate 3.63%, according to the Federal Reserve Board's H.15 release published on July 31 with data from that same July 30.

The spread banks charge over SOFR on this product runs, according to brokers experienced in these structures, between 150 and 225 basis points. Add it up: roughly 5.15% to 5.90% a year. That sum is our arithmetic on verified data, not a quote, and no bank is obliged to offer you anything like it.

That same carrier caps what it will tolerate on a policy it issues: SOFR plus 3.50%, or Prime plus 2.50%. Today, 7.15% and 9.25%.

Rate locks are sold in three, five, or seven year terms. The people who review plans already in force write that those locks have expired or will expire shortly, and that interest costs will have more than doubled.

Compare that with the cost figures still circulating online. Much of that material was written on the explicit assumption that low rates would last, and the numbers it quotes belong to that era. It is not bad faith. It is text from another monetary decade that was never updated and is still the first thing anyone finds when they search the term.

Consider how fast this moved. On July 1, 2022, SOFR was 1.53%. On October 3, 2023, it was 5.32%. At a 150 basis point spread, the loan went from costing 3.03% to costing 6.82% in fifteen months.

6.50%illustrated credit6.50%illustrated creditA 2019 illustration, with 2026 money laid on top

The analysis that follows was published on May 30, 2019. A 50 year old man, preferred non-tobacco, an indexed universal life (IUL) policy illustrated at 6.50%, initial death benefit of $10,516,478. Annual premium borrowed of $562,128 for seven years, $3,934,896 in total. Assumed loan rate between 4.06% and 4.67%.

Out-of-pocket interest starts at $24,902 in the first year, peaks at $181,792, and totals $1,644,253 over thirteen years. Buying the same coverage with cash, through a guaranteed universal life policy, cost $4,636,850 in total outlay to age 100. Outside collateral behaves like this:

Year 6 is the crossover: cash surrender value ($3,406,342) passes the loan balance ($3,372,768) and the outside collateral is released in full. The loan is retired out of the policy's own value in years 13 and 14, in two withdrawals of $1,967,448.

It is an honest illustration for its date. And the whole thing rests on a loan at roughly 4.4% against an illustrated credit of 6.50%. Two points of positive spread.

Today that spread inverts. With a loan between 5.15% and 5.90%, and with what the harshest analysis published in the trade press this year computes as a realistic net credit inside a capped, dividend-free IUL (3.82% to 4.30% off an 8% S&P 500 return), the spread goes negative by a point or two. The crossover moves out. Or it never arrives, and the outside collateral is never released.

That critique fits in a single line: a product earning 4% cannot outrun a loan costing 7%. The 7% is the author's number, not today's. Put 5.5% in its place and the sign of the answer does not change.

If the presentation you were handed uses a 4% loan rate, it is not a projection. It is a historical document.

3%guaranteed cap3%guaranteed capThe cap the carrier can cut without asking you

Two products get used, whole life and indexed universal life, almost always in high early cash value versions, because that value is the collateral. At least one major carrier sells an early cash value rider built for precisely this use.

Whole life credits a declared dividend, slow and predictable. Brokers experienced in these structures warn that the dividend may not keep pace with the loan when rates rise. And the people who defend the strategy in public concede that dividend scales have been sliding for two decades.

IUL credits the movement of an index, without dividends, subject to a cap the carrier sets unilaterally. Those caps were 11% to 13% in 2018 and 2019. Today they run between 6% and 11%, and in most contracts the carrier keeps the right to cut them to 3%, which is the guaranteed cap. Some brokers call IUL a structured note inside a life insurance policy. Our own reading goes one step further: the borrower is short an option the lender does not pay him for.

Neither product is safe under leverage. The uncomfortable number comes from the same side: across 2023 and 2024 the S&P 500 rose 56% cumulatively while many proprietary indexes credited between 1% and 3%, and some credited 0%. With a loan on top, a zero year is not neutral. The interest runs anyway.

4 to 5 timescollateral required4 to 5 timescollateral requiredCollateral is not asked for once

The interest cost is the number clients ask about. The collateral is the number that ends these plans.

This is where most presentations get it wrong. The client is told that outside collateral is an upfront requirement that disappears around year 5 or 6. The attorneys who litigate these cases describe it the other way around: collateral calls repeat, and they are the precise moment the exit narrative falls apart. Some clients took that call right after being assured there would never be another one. And one advisory firm that reviewed more than a hundred of these plans found that required collateral frequently rises to four or five times what was illustrated.

Meanwhile, those assets sit frozen. You cannot touch them. The securities account that was going to fund the next acquisition, or the apartment, or a capital call already on the calendar, is pledged, watched, and unavailable for as long as the loan is alive.

When the plan fails, the sequence is documented: the lender takes the policy and surrenders it; it applies the surrender value, which can be less than the cash value shown on the statement; it applies the pledged collateral next; and the shortfall becomes the personal liability of whoever signed the guaranty. The insured is left with no coverage. At that age, and in whatever health he is in by then, replacement may not exist.

When rates moved against the plan, these arrangements produced disputes between clients and the people who sold them. A tally published in the trade press this year identifies at least three dozen premium-financed IUL cases in active litigation, and its own author warns that the real number is unknowable.

Brokers point to something structural as well. Institutional clients who use premium financing hire actuaries to stress test the plan at inception and every year after. That level of scrutiny does not show up in family estate planning.

$10,000,000net worth required$10,000,000net worth required$10,000,000, a US LLC, and a Form 4506-C

Everything to this point applies to any borrower. What follows applies only because the family is not American.

Yes, a foreign family can get this done. The underwriting guidelines of one of the carriers that issues these policies say in so many words that US citizenship is not required. What the same document does require:

• US ownership of the policy: a US LLC or a US trust.
• A US lender bank.
• Verifiable net worth of $10,000,000 or more for foreign nationals, against $5,000,000 for domestic clients.
• A minimum face amount of $2,000,000.
• A signed premium finance disclosure and acknowledgement letter.

That is what one carrier asks. Each sets its own standard. We found none that publishes a willingness to accept a foreign trust as the owner of the policy.

There is one concession almost nobody mentions. That same carrier requires full recourse as a rule, but for policies owned by a US LLC or a US trust that are fully collateralized, it will accept the arrangement without the insured's personal guaranty. That is the difference between an obligation limited to the pledged assets and one that reaches the entire family balance sheet. If nobody has put that question to the bank, put it yourself.

There is friction too. The carrier asks for IRS Form 4506-C, which authorizes it to pull the insured's federal returns for the two most recent years. A nonresident who has never filed in the United States does not have them, so the file gets assembled out of substitute documentation and takes longer. That is an observation from practice, not a written rule.

Why the structure works for a nonresident family fits on one line of section 2105(a) of the Internal Revenue Code: the amount receivable as insurance on the life of a nonresident who is not a citizen is not treated as property situated in the United States. The death benefit sits outside the net. The policy exists to pay the tax on what is inside it.

Two corrections have to be made almost every time, because presentations get them backwards. The interest is not deductible: section 264(a)(4) disallows it, and the key person exception in 264(e)(1) is capped at $50,000 of indebtedness per insured, which covers none of this. And if you are offered the version with a foreign carrier, section 4371(1) imposes a 1% federal excise tax on those premiums. A policy issued by a US carrier does not trigger it.

On gift tax, a US private bank writes that the funds the trust borrows to pay annual premiums and interest expense do not usually give rise to gift tax. That is the bank's formulation, not a guarantee, and it is the exact point where your attorney should put an opinion in writing before anything is signed.

One tension is worth naming. The carrier permits interest to accrue instead of being paid precisely for foreign nationals with net worth of $10,000,000 or more. Brokers experienced in these structures recommend, in almost every case, paying the interest each year and not letting it accrue, because accrual multiplies the plan's sensitivity to rates. The flexibility extended to the cross-border family is exactly the feature that the people who have watched the most of these plans fail warn against.

Do not accrue.

And one thing no illustration we have seen models: the loan is in dollars, the collateral has to be in dollars and held in the United States, and the family's income is usually in another currency. A sharp devaluation at home is, in practice, a collateral call. It arrives in the year the family can least afford it, because the same devaluation is doing everything else it does. That is our reading, not a figure from the sources.

5 yearstypical maturity5 yearstypical maturityThe exit is the whole deal

There are four documented exits: the policy's cash surrender value; other assets, ideally a liquidity event already on the calendar such as the sale of a business; a GRAT whose remainder feeds the trust; or a term policy bought at inception for the projected loan balance, to cover the gap if the insured dies first. The last two come out of 2008 planning material and have to be rebuilt with today's numbers before anyone proposes them.

Death is not an exit. The attorneys who litigate these cases put numbers on it: these loans commonly mature in five years or less and almost never line up with life expectancy. Experienced brokers accept death as a source of repayment only when the insured is within roughly ten years of his.

On the first exit, the one everyone assumes, those same brokers are blunt: in most of the reprojections they run there is not enough cash surrender value to repay the loan and cover the policy's remaining expenses as well.

The people who promote the strategy hold that cash value retires the principal usually within fifteen to twenty years. That is the seller's number. It is also twenty years of paying interest without missing a single renewal.

Who should do this, and why almost nobody

Someone who could write the check for the premium and chooses not to. That is the entire filter.

The side that defends the strategy sets out its own conditions: the client can pay the premiums out of pocket; no interest is accrued; the death benefit and the premium are reasonable against income and net worth; the plan is stress tested against genuinely adverse scenarios; and there is an exit that does not consist of hoping the policy pays for itself. Use that list as a test. You will see how many transactions fail it.

It does not work for anyone who cannot pay the interest out of pocket every year, indefinitely. Nor for the business owner or the real estate investor with 80% or 90% of net worth locked up. That same analysis calls it a trap that multiplies risk, because when the plan fails the liquid holdings have to be sold at the worst possible moment. Nor for anyone who cannot absorb a collateral requirement that quadruples.

Some advisory firms have barred their advisors from premium-financed IUL, and from that side the estimate is that if suitability standards were seriously applied, the universe of appropriate clients would be vanishingly small. The other side answers that the problem is execution, not structure. Both positions ran this year in the same trade publication. We have both on file, and both are worth reading in full before deciding.

If someone pitched this to you as "free insurance," notice who rejects the phrase. The bank that lends the money calls it a myth built on overly optimistic assumptions about policy performance and interest rates. The people who defend the strategy in public call it flatly wrong. The two sides agree on that much.

What to ask for before you sign

None of this requires you to understand insurance. All of it is answerable in writing, and an arrangement that cannot answer it in writing has told you something.

• Ask for the plan rerun at a loan rate materially above the one illustrated, with the policy crediting materially less, both in the same document. If nobody will produce that page, you have learned something about how the plan behaves when rates move.
• Ask which page of the illustration shows the assumed crediting rate, then read the guaranteed cap next to it. Those are two different numbers, and the second one is the one the carrier can fall back to.
• Ask for the collateral formula in writing: how the requirement is calculated, how your pledged assets are discounted before they count toward it, how often it is recomputed, and who decides. Then ask what it becomes at four or five times the illustrated figure, and whether you could post that without selling anything.
• Ask to read the written exit strategy. The carrier already requires one, naming where the principal comes from at maturity. If it names the policy's own cash value, ask for the reprojection that supports it rather than the illustration that assumes it.
• Ask whether the obligation is full recourse, and if it is, whether a fully collateralized structure owned by a US entity can be done without the personal guaranty. Nobody will ask that on your behalf.
• Get it into the file that interest will be paid every year and not accrued, whatever the carrier is prepared to permit.

Every figure in this piece comes from a document we have identified and keep on file, and we will go through it with anyone who asks.

When a family asks us about this, we ask for two pages before we give an opinion: the lender's term sheet and the page of the illustration that shows the assumed loan rate. If that page says 4%, the conversation is short.

This piece is general information for educational purposes and does not constitute legal or tax advice or an offer of any product. Every situation turns on its own facts and should be reviewed with your attorney and your accountant.