Behbood Certificates, IRIS and a Software Fault: Where Pensioners' Refunds Get Stuck

The passbook records a profit entry on a fixed date every month. No form, no application, no officer — just an entry. For a widow holding a Behbood Savings Certificate, that entry is the month's budget; for a retired employee with a Pensioners' Benefit Account, it pays for medicine and the electricity bill; for a family holding a Shuhada Family Welfare Account, it is the money they would rather not have to ask for again.

That same amount has to travel once a year. It moves from National Savings to the FBR, from the FBR to the IRIS portal, and there it becomes something else. On paper it is exempt profit; on screen it becomes a liability. And between the two stands a taxpayer whose only evidence is a passbook.

This is not a new story, nor is it the only one, but this time the matter has moved beyond an individual complaint to an organisation. The Public Interest Litigation Committee of the Lahore Tax Bar Association has asked the Federal Board of Revenue to correct a software fault in IRIS — the way tax is being computed on the profit of holders of these three schemes is, in the committee's reading, not what the law says. Waheed Shahzad Butt's name is attached to that letter.

There are two holders here — a piece of paper and a screen. Look at both, and there are really three things: a law that grants the exemption, a system that forgets it, and a person caught in between. The most striking thing in this whole affair is not the tax amount. It is that the amount is small enough that almost nobody fights it.

The law's intent versus the system's reading

Pakistan's tax architecture rests on one principle that has not changed in years: the bulk of collection happens at source. Money is deducted before it reaches the taxpayer, and the taxpayer must then show that deducted tax in the return — failing which the deduction is wasted. This is the withholding regime, and it applies to profit on debt too.

The Income Tax Ordinance 2026 sets that framework. Profit on debt is a taxable head, but the Second Schedule of the same Ordinance carries exemptions and relief for certain categories. That is why Behbood, Pensioners' Benefit and Shuhada Family Welfare are not merely financial products — they are legally designed for people with no other dependable income. A widow, a retiree past sixty, the family of a martyr. The names of these categories are themselves a declaration of economic vulnerability.

The money moves through the Central Directorate of National Savings (CDNS), which has both a head office and branches, plus reach through post offices. The holder collects a certificate and leaves. In their world, filing is a once-a-year chore, done either personally or handed to a consultant. But the architecture of that chore is unchanged, and it collides with the entire logic of these schemes.

IRIS is the front door of that architecture. It is the FBR portal where returns are filed, withholding statements deposited, notices received, refund status displayed, and first appeals begun. When the profit on a Pensioners' Benefit Account travels from CDNS to the FBR, it appears in IRIS as a pre-populated figure that the filer must re-pot. The question is not whether the figure appears. The question is which category the system has placed it in, and at what rate it is taxing it.

The committee's letter focuses on exactly this. It asks the FBR to correct the software error. One thing must be said plainly: the letter's arithmetic is not an independent investigation. It is an advocacy committee's document, detailed and internally consistent, but not corroborated by revenue officials in public material. Source quality here is medium-low.

That does not mean the problem is imaginary. It means we should understand the structure first, do the arithmetic second, and arrive last at the thing no letter mentions.

Three layers of arithmetic, and the budget is never one number

The first way to read this is as "the software made a mistake." I read it differently. For one account holder it is an error; for thousands it is a structure, and structures have their own arithmetic.

There are three layers. The first is failure to recognise the exemption. If the system treats the profit as ordinary profit on debt and files it under the standard category, the exemption paper means nothing. The second is failure to match tax deducted at source. CDNS deducted, paid the FBR, and that should appear as a credit in the taxpayer's ledger. If the credit is missing or partial, the tax already paid also counts as "short" in the system's eyes. The third and most important is when both happen together.

In that third layer, what hits the taxpayer is a double loss. Once, the money already deducted at source; again, the money demanded through a notice.

Take the arithmetic as illustration only, not as any specific case figure. Suppose a holder's annual profit is five hundred thousand rupees. If the exemption is applied correctly, the applicable rate is very different. If the system computes it under the standard category, the gap will show either way. But that gap does not occur once on five hundred thousand rupees. It occurs every year, on every return, for every holder who can reach IRIS.

Here is something rarely heard in tax debates. Tax arguments are usually about rates, while the real pain sits in categories. Changing a rate is a political decision and attracts noise; changing a category is a technical decision and attracts none. In the Behbood and Pensioners case the rate is generous. The trouble is the system never reaches that rate, because it has filed the income in the wrong slot.

There is another picture people see without drawing a meaning from it. A return has a place to claim an exemption and a place to show tax paid. For a pensioner, both should be filled from a single number. But most annual filers meet both fields on a screen where they do not know what is being asked. The distance between taxpayer literacy and tax complexity is the real economy of this case.

What the letter is least attended to is not legal but administrative. It does not ask the FBR to legislate. It asks that the system compute according to law already in force and is failing to. This is not a policy fight; it is a fight over implementation of a declared policy, and the difference between the two is vast.

When the cost of winning exceeds the loss

Here begins the part I find most interesting, and the part no advocacy letter will contain.

Suppose you are the holder. You believe the computation is wrong. You have three routes: seek a correction in the return, file a formal response to the demand, or go to first appeal. All three cost time, paperwork, and often a professional's fee.

Now look at the amount at stake. On a small sum, the cost of fighting usually exceeds the sum. That does not mean people accept the error. It means they never start. And that behaviour creates a stable equilibrium for the system.

Errors on small amounts persist because challenging them costs more than bearing them. This is the least-said sentence in the whole story. No administrative machine reforms until its errors become expensive to it. Here the cost lands on the other side, which is why nobody is in a hurry.

Second, the asymmetry of evidence. The system produces a number; the taxpayer must produce proof against it. Proof means passbooks, certificates, deduction receipts, perhaps a letter from National Savings. A widow who has never seen a tax notice cannot navigate that paperwork and terminology. Reading her own position requires training she does not have.

Third, connecting both sides: the holder's age and the system's speed do not keep time together. This is not sentiment; it is a design problem. If you build a scheme whose beneficiaries are the least technically equipped, then loading the compliance cost onto them is a policy choice — even an unconscious one.

The hardest truth in this collision

I accept that parts of this letter cannot be fully settled without hearing a revenue official. But one thing is publicly observable: much of IRIS runs on defaults, and defaults are where policy hides.

There are two administrative models. In one, the exemption attaches to the income type — the system knows how this kind of profit is to be treated. In the other, the exemption attaches to the person — the taxpayer's profile records the category, and a special regime follows.

The first model's risk is that every scheme change must be re-taught to the system, and somewhere a step will be missed. The second model's risk is that the taxpayer must register their own category, which may exceed their capacity. Behbood, Pensioners and Shuhada are all defined by the identity of the person, not the flavour of the income. The second model is more logical for them. But if the system runs on the first, the question is not where the error happened. The question is whether any mechanism exists to catch it.

This recalls something visible in every large system. A system cannot see its own error, because the error is treated as correct from inside. Until someone outside points, it goes on counting it as right. A widow's passbook cannot register a complaint in that system, because the system speaks profit and the passbook does not. The bank or any other institution has no reach either.

What the available figures show

Some points in the letter are clear. It comes from a public interest committee that is part of the Lahore Tax Bar Association. It is addressed directly to the FBR. It names three schemes — Behbood Savings Certificates, Pensioners' Benefit Account, Shuhada Family Welfare Account. And Waheed Shahzad Butt's name is attached.

What the document does not contain matters as much. There is no formal FBR response, no independent verification, and the letter's date is not available in public material. When you write about a matter where only one side has spoken, the honest route is to state that side and state that it is one side.

The logic I read in the letter's arithmetic is this: if the system's internal steps were applied correctly, the outcome should match the paper. If the system is already producing something for a holder that does not match the paper, it is doing so because of the system, not the taxpayer.

That is why the letter carries three demands — fix the system so the error does not recur; review affected returns so what has gone wrong is corrected; and, failing both, share the data so the number of affected people is known.

The third demand carries the most force. Because once the number is out, the debate moves from "is there an error" to "how large is it" — and that is where reform begins.

The sentence that forces a thought

Now the part routine reporting skips — what "public interest" means here. When an advocacy committee speaks of public interest, it is raising a voice for a class. There is nothing wrong with that, but it must be understood.

I wondered whether another angle was being left out. The letter says the system is wrong and should be fixed. What it does not say is who benefited most from the error. If someone paid extra tax, no fund exists to compensate them. It is simply an error. And an error means money left someone's pocket and went nowhere.

One more thing deserves clarity, because it is the heart of the matter. An error made by the system is not just an amount to the taxpayer; it is a brick in the trust between that taxpayer and the state. When a holder concludes that no matter how correct I am, this system will find me wrong, they stop filing. And when they stop filing, they drop out of existence in some way. This is not drama; it is a process that happens without sound.

The biggest question in this letter is not how many people are affected. It is how many know they are affected. A widow who has never logged into IRIS does not know how tax is being computed on her profit. And for someone who does not know, no remedy exists.

The open question worth leaving open

There is an outcome here that does not yet exist, and a way to watch for it.

If the FBR says only that "affected taxpayers should file their own correction," there will be no remedy — because those who must correct do not have the information correction requires. If it says "we are fixing the system," a question remains: will that fix cover old returns or only new ones.

If only new, what of those who have already overpaid? If old, the machinery must itself identify who paid how much. The second option is less a technical task than a question of intent.

And one thing matters most — this issue needs a place where it can be heard. Tax administration stories usually concern big numbers and big deals. But the small amount credited monthly to a National Savings scheme is nothing against those figures — nearly free. Miscalculating tax on it means one household's monthly food is reduced. And when the same error repeats every year, the total is enormous even if it never appears at once.

The letter's greatest value is that it has posed a question. And the question is this: when the law says this income is exempt, does the system also need to know that, or only the taxpayer? That question awaits an answer, and until it comes, the passbook entry will not, and neither will the money that was rightfully hers.

And when you think about that money, one thing becomes clear. The true test of a tax system is not how much it collects; it is whose pocket it reaches into when it is wrong.