Israel's Invoice Law in 2026: The Threshold Drop and What Gets You Paid

What is Israel's invoice allocation number, and why is the threshold dropping again?

An allocation number (מספר הקצאה) is a 9-digit code the Israel Tax Authority issues in real time, through its SHAAM platform, before a business finalizes a tax invoice that meets three conditions: it exceeds the current threshold (before VAT), it includes a VAT component, and it's issued to a registered dealer (a B2B invoice, not a consumer receipt). It sits behind Section 38(a1) of the Value Added Tax Law, 5736-1975, added as part of the "Invoice Israel" reform, and its entire purpose — stated plainly by the Tax Authority and by the accounting firms tracking the reform — is fighting the black-market trade in fictitious invoices.

The mandate was originally supposed to phase in through 2028, starting with a high ₪25,000 threshold so only the largest transactions were affected in the pilot year. It has been accelerated since: ₪20,000 through 2025, ₪10,000 from January 1, 2026, and ₪5,000 from June 1, 2026 — and industry coverage of the Tax Authority's own announcements suggests further reductions are still coming after that. The practical effect for a business: transactions that were completely untouched by this rule in 2024 will need an allocation number by mid-2026, and a growing share of your invoices are now exactly the size where this stops being someone else's problem.

Is the crackdown actually catching anything — and why is Israel doing this at all?

The stated goal isn't paperwork for its own sake. The State Comptroller's report on how the Tax Authority handles fictitious invoices, published in January 2024, put the state's estimated annual damage from fictitious invoices at around ₪2.5 billion. A more recent Comptroller review, covered by Calcalist, found the Invoice Israel system rejected roughly 62,000 allocation-number requests in 2025 alone, worth a combined ₪42 billion — had those invoices gone through, the same coverage estimates they'd have cost the state about ₪7.1 billion in improper VAT credits, on top of an estimated ₪10–15 billion more in avoided income and corporate tax.

That same 2025 window shows exactly why the threshold keeps falling instead of holding steady: separate Tax Authority research into the reform's first full year found businesses shifting an estimated ₪12.2–16.4 billion in transactions below the reporting threshold — splitting a single sale into several smaller invoices, or under-declaring its value, specifically to keep each piece under the number that triggers an allocation-number check — at an estimated cost to the state of roughly ₪2.2–3 billion in lost VAT for the year alone. It's also, per the Comptroller review, an increasingly technical fight: reported cases include cybercriminals hijacking a legitimate business's own Invoice Israel account through identity theft, with the Tax Authority in every reviewed case only catching it after the business itself filed a complaint.

Does your invoicing already generate this number for you, or is it on you to request it manually?

If your business collects payment through Cardcom or Sumit — the two providers oToK's Payments feature connects to for charging customers directly — the allocation number isn't something you request by hand on the Tax Authority's own site. Both platforms have a direct API connection to Tax Authority servers and generate the number automatically the moment a qualifying invoice is issued, the same real-time step the law requires.

The catch is that "connected" isn't a one-time setup you can forget about. Cardcom's own support documentation notes the authorization for this service is valid for three months at a time and needs renewing, with a dedicated button to confirm a given invoice actually received its number rather than assuming it did. Sumit's help center describes the same automatic generation when the Tax Authority connection is live, plus a manual fallback for the invoices where it isn't. Neither failure mode is dramatic on its own — but an expired authorization quietly turns every qualifying invoice issued afterward into one your customer can't use to deduct VAT, and nothing about the invoice itself will necessarily tell either of you that happened.

1. 1Confirm which invoicing system actually issues your tax invoices today — your own Cardcom or Sumit account, a separate bookkeeping tool, or a mix of both — since the automatic allocation-number step only fires on the platform that's actually connected to the Tax Authority.
2. 2If you're on Cardcom, check the service's authorization status rather than assuming a setup done months ago is still active, and use the built-in "check allocation number receipt" confirmation on a recent invoice above the current threshold.
3. 3If you're on Sumit, verify the Tax Authority connection under its Israel Invoices settings, and know where the manual allocation-number button is for the rare invoice that doesn't generate one automatically.
4. 4Re-check your threshold math after June 1, 2026 specifically — an invoice amount that safely cleared the ₪10,000 line in the first half of the year can fall under the new ₪5,000 line in the second half without anyone changing how they price anything.

How big is the late-payment problem for small businesses right now?

Set the compliance question aside for a moment: even a perfectly valid invoice with a correct allocation number still has to actually get paid, and the most recent disclosed data says that's getting harder, not easier. Intuit's 2026 Small Business Late Payments Report — built on its ongoing ~5,000-respondent QuickBooks Small Business Insights panel plus a December 2025 survey of 1,305 US business owners running companies with 0–250 employees — found 59% of small businesses now carry invoices overdue by 30 or more days, up sharply from 47% a year earlier, with an average outstanding balance of $17.7K per business.

The knock-on effects are concrete, not abstract: 49% of surveyed owners said standard payment processing creates a critical or moderate cash-flow gap, 39% said a single late payment made it hard to cover payroll or bills, and businesses already carrying overdue invoices were far more likely to lean on expensive workarounds — 38% increased their reliance on credit cards (versus 21% of businesses without overdue invoices), and 26% said they'd personally delayed their own salary because of it.

A separate February 2025 survey by the Kaplan Group — a commercial debt-collection agency, worth flagging given its business interest in the finding — polled 100 financial decision-makers (CFOs, VPs of finance, controllers) at companies of varying sizes and found 93% report losing revenue to late payments, with 11% losing more than 5% of revenue to outright invoice defaults; the pain was worse, not better, at larger companies and in fast-scaling sectors like SaaS, where 22.2% of respondents lost over 10% of revenue this way.

Is "most businesses fail from bad cash flow" a real, measured finding — or another unsourced number?

This is the point where the data above could tip into a much scarier, much more repeated claim: that 82% of small businesses fail specifically because of poor cash-flow management, a line attributed almost everywhere online to "a U.S. Bank study by Jessie Hagen." We went looking for that study the same way we checked the numbers above — a dated report, a sample size, a named methodology — and came up with nothing. The figure shows up across U.S. Chamber of Commerce-adjacent content, SCORE-affiliated guides and small-business blogs going back years, always with the same attribution and never with a link to an actual document.

That doesn't make cash-flow timing a non-issue — the QuickBooks and Kaplan Group data above is real, disclosed and considerably less dramatic on its own merits. It does mean the 82% figure specifically belongs filed next to other precise-sounding, source-free statistics that circulate because they're quotable, not because anyone can point to where they came from.

Does a reminder's wording actually change whether — and how fast — someone pays?

Here the peer-reviewed research is unusually strong, and it comes from an unlikely place: not invoicing at all, but tax collection. Hallsworth, List, Metcalfe and Vlaev published "The Behavioralist as Tax Collector: Using Natural Field Experiments to Enhance Tax Compliance" in the Journal of Public Economics in 2017, reporting on two large natural field experiments run with the UK's HMRC and the Cabinet Office's Behavioural Insights Team across more than 200,000 taxpayers who had already declared income but hadn't yet paid it.

Several message variants went into otherwise-identical reminder letters: some emphasized penalties, some named the public services the tax money funds, and some used a plain descriptive social norm — what most people in the same position actually do. That last category won clearly. The best-performing single line tested was: "Nine out of 10 people in the UK pay their tax on time. You are currently in the very small minority of people who have not paid us yet." According to Warwick Business School's account of the results, that one sentence accelerated an estimated £4.9 million in payments within the 23-day sample window, versus more than £9 million accelerated across all the tested messages combined in the same window — and messages that simply made a penalty more salient produced a separate, smaller 3.2–3.9 percentage-point lift in payment rates on their own.

Once a payment actually fails — a declined card, a lapsed standing order — what recovers it?

This matters specifically for recurring billing and installment plans, which is where a payment doesn't just arrive late — it fails outright and needs a second attempt. Subscription-billing platform Recurly, reporting from its own network of subscription businesses (data current to July 2026), puts the median annual involuntary-churn rate — customers lost purely because a charge failed, not because they chose to cancel — at 1.25%, ranging from 0.18% among businesses with the highest per-customer revenue up to 1.69% in higher-risk sectors like education; across its tracked SaaS customers alone it reports recovering over $155 million in payments that would otherwise have been lost.

A second platform, Baremetrics, discloses its own product data (from roughly 148 of its customer accounts as of December 2024) on what actually recovers a failed charge: emailing a customer within 24 hours of a decline gets a 41.29% open rate and a 13.25% recovery rate, a "your card is about to expire" email sent 30 days ahead of that gets a 47.41% open rate and 14.37% recovery — and the same email sent more than 30 days after a failure, once the moment has passed, drops to a 26.83% open rate and just 4.2% recovery.

So what should an Israeli business actually do about invoicing and collections in 2026?

1. 1Treat every invoice near ₪5,000–₪10,000 as needing a valid allocation number from June 1, 2026 onward, not just the large ones — the threshold drop means transactions that never touched this rule before now do.
2. 2Verify your Cardcom or Sumit Tax-Authority connection is actually live rather than assuming a past setup still is — Cardcom specifically requires renewing the authorization roughly every three months.
3. 3Don't split a transaction into smaller invoices to dodge the threshold — it's the exact pattern the Tax Authority's own research says it's watching for, and the accelerated schedule exists specifically to close that gap.
4. 4Write payment reminders that state what most customers actually do ("most clients pay within 14 days of this invoice") instead of only escalating tone or urgency — it's the specific mechanism a large field experiment found measurably outperforms alternatives, at zero added cost.
5. 5Send the first reminder inside the first 24–48 hours after a due date or a failed charge, not weeks later — every dataset in this piece points the same direction: attention (and recovery rates) fall off fast once the moment has passed.
6. 6For recurring or installment charges, don't rely on a single retry — combine an automatic retry with a card-expiry reminder sent in advance, since the two catch different failure causes.

Frequently asked questions

What is Israel's invoice allocation number, and when do I need one in 2026?

It's a 9-digit code the Israel Tax Authority issues in real time before you finalize a qualifying tax invoice — one above the current threshold, with VAT, issued to a registered dealer. The threshold is ₪10,000 (before VAT) from January 1, 2026, and drops to ₪5,000 from June 1, 2026. Without a valid number, the invoice's recipient can't deduct the VAT on it.

Does Cardcom or Sumit generate the allocation number automatically?

Both connect directly to Israel Tax Authority servers and generate it automatically the moment a qualifying invoice is issued through their system — but the connection needs to stay active. Cardcom's authorization for the service specifically needs renewing roughly every three months, and both platforms offer a way to confirm a given invoice actually received its number rather than assuming it did.

Is it true that 82% of small businesses fail because of poor cash-flow management?

That specific figure, usually attributed to "a U.S. Bank study by Jessie Hagen," doesn't trace to any dated report, sample size or methodology we could locate — treat it with real skepticism. The genuinely sourced data is less dramatic but still serious: Intuit's 2026 report found 59% of small businesses carrying invoices overdue 30+ days, and 49% report a resulting cash-flow gap.

Does how a payment reminder is worded actually change whether someone pays faster?

Yes, according to the strongest evidence available on the question: a 2017 peer-reviewed field experiment run with the UK's tax authority across more than 200,000 taxpayers found a reminder using a social-norm line ("nine out of 10 people pay on time") accelerated an estimated £4.9 million in payments within 23 days, outperforming versions that emphasized penalties or urgency instead.

What actually recovers a failed recurring payment — a retry, an email, or both?

Timing matters more than the specific method, based on disclosed platform data from Recurly and Baremetrics: a reminder sent within 24 hours of a failed charge recovers meaningfully more than the same reminder sent weeks later, and a pre-emptive card-expiry email sent about 30 days ahead performs best of all in Baremetrics' own figures. Combining an automatic retry with a timed reminder catches more failure causes than either alone.