The $779 Million Mistake

P@SHA, the country's only government-recognised IT industry body, has asked Islamabad to amend Section 154A of the Income Tax Ordinance in the FY 2026-27 federal budget. The proposal would split the current flat 0.25% final tax on IT export receipts into two categories. "Genuine" freelancers, those who can prove multi-client engagement, project-based work, and business autonomy, would keep the favourable rate. The much larger group of full-time remote employees of foreign firms, currently treated as freelancers under the same code, would be reclassified and taxed on a graduated slab: 5%, 10%, 15%, and finally 20% for incomes above PKR 30 million a year.¹

The official framing is fairness. Local software houses, P@SHA argues, cannot match the take-home pay of foreign employers who route compensation through the freelance code. The result is a 22 to 44 percent net-pay gap, an "implicit subsidy" to foreign companies, and a structural drain of senior talent away from the formal sector. The proposal is positioned as a course-correction, not a crackdown.

This article argues something narrower and more specific: that the proposal misdiagnoses the problem, misreads the market, and, if enacted, will damage the only segment of Pakistan's IT export economy currently growing fast enough to matter. It is also, on close inspection, a case study in what happens when a 1992 trade body tries to legislate against a 2026 labour market.

Pakistan's IT and IT-enabled services exports reached a record $3.8 billion in fiscal year 2025, an 18 percent increase over the previous year and the highest figure in the country's history. The sector is now the third-largest source of foreign exchange after textiles and rice, and accounts for roughly 45 percent of all services exports.² Government targets are ambitious: $5 billion by FY26, $15 billion by 2030.³

Sitting inside that $3.8 billion, and growing five times faster than the rest of it, is the freelance and remote-work segment. The State Bank attributes $779 million of FY25 export receipts to individual freelancers and remote employees of foreign firms, up from roughly $410 million the year before. That is a 90 percent year on year increase against the headline sector's 18 percent. Pakistan now sits among the top five freelancing nations globally, with somewhere between 2.3 and 3 million active digital exporters by Payoneer's and PAFLA's count.⁴

This is the picture P@SHA's proposal would intervene in. The asymmetry of growth rates matters: a policy change that nudges remote workers into a tax bracket twenty times their current one is being applied to the segment carrying the most weight in the sector's forward trajectory.

The proposal's central illustration is a comparison table. At a gross monthly compensation of PKR 500,000, a remote worker filing under Section 154A pays PKR 1,250 in tax and takes home PKR 498,750. A local software-house employee on identical gross compensation, taxed under the salaried slab, takes home PKR 393,250. The monthly difference is PKR 105,500, or roughly 27 percent. At the top of P@SHA's table, at PKR 1 million in monthly gross, the differential widens to 44 percent.⁵

P@SHA calls this tax arbitrage. The label flatters the proposal. A more honest reading is that it is a market signal. Foreign employers, Stripe, Microsoft, Shopify, a thousand smaller startups, and the platforms that intermediate them, are willing to pay Pakistani engineers something close to global rates. Local software houses, building services-business margins on a billing model that depends on labour cost arbitrage between Pakistan and the West, are not. The 27 to 44 percent net-pay gap isn't a tax distortion. It is the gap between what Pakistani technical labour is worth in the global market, and what local employers have so far been able to pay.

Closing the gap by raising the remote worker's tax doesn't change what foreign companies pay. It changes where, how, and through whom that money moves.

The graduated schedule P@SHA has proposed for the new Category B is the most legible part of the document, and the most consequential.

Several things are worth noticing at once. First, the proposed rate at the lowest bracket, PKR 6 million a year or roughly USD $21,500, is a twentyfold increase on what these earners pay today. Second, the top bracket triggers at PKR 30 million annually, around $108,000, a band populated almost entirely by senior engineers, designers, and product managers whose foreign employers chose them precisely because they were globally competitive. Third, the schedule introduces a re-classification mechanism, five criteria of which three must be met for Category A status, that effectively shifts the burden of proof onto the individual to demonstrate they are not an employee of the foreign firm.

The mechanics are familiar to anyone who has watched lobbying play out in dense regulatory text. Reclassification rules with multi-criterion tests, applied to a labour market the regulator does not currently audit at the individual level, almost always produce three outcomes: a small group of clearly-genuine actors retain the favourable status; a large middle group quietly stops bringing money into the formal banking system; a third group moves.

The Pakistan Software Houses Association was incorporated in 1992 as a trade body for software exporters. It is, today, the only government-recognised representative organisation for Pakistan's IT and ITeS sector. Its membership, depending on whose number you trust, runs from 600 to 1,800 companies, with the higher figure including associate and affiliate members.⁶ It holds a seat at the Special Investment Facilitation Council, coordinates directly with the Ministry of IT & Telecommunication and PSEB, and is consulted by the Federal Board of Revenue on tax matters affecting the sector. When the federal budget is drafted, P@SHA is in the room.

Typically these office bearers, direct a Pakistan-headquartered firm whose primary product is the time of its salaried Pakistani employees, sold to overseas clients on a project or outsourcing basis. These companies compete, directly, with foreign firms offering remote employment to the same talent pool. If a senior engineer at one of their member companies receives a $4,000/month offer from a US fintech, those companies lose. If that engineer is then taxed at 10 to 15 percent on the foreign income instead of 0.25 percent, the net offer shrinks, the local counter-offer becomes more competitive, and the member company retains the engineer.

The asymmetry of representation is the part of this story that is most easily overlooked. P@SHA's 1,800 member companies have a permanent seat at the policy table. The 2.3 to 3 million individual digital exporters whose receipts are the subject of the proposal have no such seat. The Pakistan Freelancers Association (PAFLA), exists and has done useful work on payments infrastructure and training. But it is not a tax-policy lobby with budget-cycle access. The voices of the people most affected by this proposal are, structurally, on X, which is why X is where the response landed.

On 23 May 2026 the Pakistan Software Houses Association's Federal Budget 2026-27 white paper became public knowledge. It proposes to reclassify and tax up to one third of Pakistan's IT export economy at rates between twenty and eighty times the current floor. The white paper carries one institution's name. Twelve people direct that institution. Six hundred and eighty-eight people voted them in.

P@SHA maintains, as required by its constitution, a provisional voter list before each Annual General Meeting. The 2024 list, published at pasha.org.pk in the run-up to the September election that produced the current Central Executive Committee, contains exactly 688 names. The list is split into two categories: corporate members (companies operating in Pakistan's IT and ITeS sector that pay annual dues) and associate members (a smaller set with reduced voting rights).

These 688 voters constitute the entire electorate of Pakistan's only government-recognised IT industry body. They elected the CEC that, eighteen months later, drafted the white paper now under examination.

The population the white paper proposes to reclassify and tax is, on the most conservative count, 2.3 million people. On the standard count used by Payoneer and PAFLA, it is closer to three million. The ratio of P@SHA voters to affected Pakistani digital exporters is approximately 1 to 4,360.

This is not a representation problem at the margin. It is a representation problem at the architectural level. P@SHA's policy access is structured as if the IT sector still resembled its 1992 composition: a few hundred software houses employing the country's small population of formally-trained developers. In 2026 that picture is upside down. The individuals working remotely or freelancing for foreign clients now generate the growth that the headlines describe. The overwhelming majority of freelancers have no practical representation inside P@SHA's voting structure.

The white paper's stated authorship widens this somewhat. Lead authors include one CEC member and one P@SHA staff lead. Contributors and reviewers named on the document's first page include a venture capital firm, a Big Four advisory practice, a research organisation, and several P@SHA member-company principals beyond the CEC.⁹ The wider input is worth noting because it complicates a pure "twelve companies wrote this" reading. It does not, however, resolve the architectural problem. None of the named contributors represent the individual freelancer or remote-worker population whose income the proposal would reclassify. The widening goes upward into adjacent professional services, not outward into the affected workforce.

Section II established that the market gap between local and foreign employers is real. This section shows why the twelve CEC member companies, specifically, need that gap to close, and how the proposed tax is the precise instrument that closes it without requiring them to raise pay.

It is worth stating the economics carefully, because the analysis here is not a moral charge. It is a mechanical one.

A senior Pakistani engineer hired directly by a US technology company through a platform like Deel, Remote, or Multiplier currently earns, on average, somewhere between $30 and $50 per hour, with all of it routed through Section 154A and taxed at 0.25 percent. The same engineer's services, sold to a Western client through a Pakistani services firm of the kind on the P@SHA CEC, bill at approximately $28 per hour (the rate one CEC-member firm publicly lists on its own About page) of which the engineer receives a fraction.

For the engineer, the choice is straightforward in either direction. Direct hire pays multiples more in take-home terms. The CEC-member firm cannot compete on pay, because its own client billing rate caps what it could ever offer. The firm's only economic lever is the engineer's outside option.

At 20 percent withholding, the direct foreign offer's relative attractiveness narrows by roughly forty-five percent. The local firm does not have to raise pay. It just faces less competitive pressure.

This is the mechanism. It is not a conspiracy. It is a coherent collective interest, achievable through one specific tax-policy change, with the entire economics of the move arriving as net gain for the twelve member companies and net loss for the three million unrepresented workers. The fact that every CEC member's business model points in the same direction on this question is not a coincidence. It is the body's defining characteristic.

In the 48 hours after ProPakistani's report of the proposal was posted to X on 23 May, the original post drew more than 183,000 views, 80 quote-tweets, and 46 direct replies. A non-random sample of approximately 40 of those high-engagement posts (top-ranked and most-recent quote-tweets and replies) was reviewed for sentiment, theme, and engagement.

The sentiment distribution was striking. Of the sampled posts, an estimated 95 to 98 percent opposed the proposal, with the heaviest concentration around two themes: that P@SHA was acting as a cartel protecting under-paying members, and that the policy would accelerate either brain drain or off-channel evasion. Engagement on critical content was high, with multiple top-quoted posts clearing 300 to 900+ likes within 48 hours and the single most-engaged post (Saad, @SaadInCyber) reaching 912 likes and 175 reposts.

A representative selection of the most-engaged public responses:

The dissenting and qualified voices are worth recording too, because they exist:

A serious argument has to engage with the strongest version of the position it disagrees with. P@SHA's proposal has three legitimate concerns, and they deserve to be stated cleanly.

First, the classification problem is real. A meaningful share of "freelancers" filing under Section 154A are not independent contractors in any economically meaningful sense. They have one client. They work fixed hours. They receive equity. They report to a manager. They are, functionally, employees of a foreign company, and the use of the freelance code is, at the level of doctrine, a fiscal characterisation that doesn't match the underlying labour relationship. Tax codes that don't match underlying relationships create their own distortions over time.

Second, the cost asymmetry on local employers is also real. A Pakistani software house pays EOBI, SESSI, gratuity, and the full apparatus of formal payroll. It runs HR systems, compliance, audit. Its salaried developers pay tax on the salaried slab, which can run to 35 percent at senior levels. Its remote-employed competitor pays 0.25 percent. Treating these two situations identically for tax purposes does, in fact, create a structural disadvantage for the formal employer.

Third, the IP and value-capture concern has merit. Code written by a Pakistani engineer under direct foreign employment becomes the IP of the foreign company. Code written at a Pakistani software house, even when sold to a foreign client, at least has a Pakistani company in the value chain. There is a coherent national-economic argument that the formal sector creates more durable value than direct labour export.

If the proposal is adopted in substantially its current form in the 2026-27 budget, the damage cascades across three time horizons.

The 688 voters elected an executive committee that is uniformly aligned, by business model, against the policy interest of three million Pakistani digital exporters. This is not because the CEC members are bad actors. It is because, by P@SHA's constitutional design, only people running companies of this kind can stand for election to it.

Asking this body to write tax policy on the freelance segment is structurally equivalent to asking a textile manufacturers' association to advise on garment-import tariffs. The output will be coherent, internally consistent, and self-serving. There is nothing wrong with the body advising. There is something wrong with treating its advice as the IT sector's view.

The Ministry of IT & Telecommunication, the Federal Board of Revenue, and the Special Investment Facilitation Council have a clear procedural correction available before this proposal moves into the 2026-27 budget. Any consultation on Section 154A and the freelance segment must include, as voting participants:

Until this representation is in the room, the P@SHA white paper should be treated as one stakeholder input among several, not as the consensus position of an industry it does not, by any honest count, numerically represent.

If the Ministry of Finance is going to evaluate this proposal seriously in the run-up to the 2026-27 budget, three further tests are worth applying alongside the procedural correction above.

One: revenue realism. Model the policy with realistic compliance assumptions, including the share of current remote workers who would shift to non-Pakistani banking channels at each proposed rate. The historical literature on tax wedges in mobile-labour categories is unambiguous on this: declared income falls sharply, and the elasticity rises with the digital portability of the receipts. A revenue model that assumes static compliance is not a model.

Two: counterfactual stability. The 0.25% Final Tax Regime under Section 154A was meant to provide a multi-year window of predictability for IT exporters. P@SHA's own chairman has repeatedly argued that the sector's biggest unmet need is policy stability. Walking back the regime for the fastest-growing cohort two years into the window cuts directly against that argument.

Three: the right problem. If the actual problem is that local software houses cannot compete with foreign remote-work offers on take-home pay, the correct policy response is to reduce the cost wedge on the local employer, not to inflate the tax wedge on the worker. Cut the salaried slab for IT-sector employees. Extend the 0.25% FTR to local companies on the salary cost of export-bound work. Introduce an R&D credit. The toolkit is large. Reclassifying remote workers is the worst instrument in it.

Pakistan has, in the last three years, accidentally built one of the most interesting digital labour stories in the emerging world. Three million people, mostly under 35, exporting expertise directly into the global economy and bringing the receipts home through formal channels. The growth curve is unmistakable. The geographic reach is unmistakable. The cost to the exchequer of letting it continue under the current 0.25% regime is essentially zero. The alternative outcome, in which these earners are taxed elsewhere or not at all, is the realistic counterfactual.

P@SHA's proposal asks the government to interrupt this story to solve a problem (the take-home pay gap) that does not require an interruption to solve. The gap exists because Pakistani technical labour is being correctly priced by a global market for the first time in the country's history. The right response is to let the formal sector adjust to that price, or to lower its cost of competing for it. Not to tax the people earning it.

The proposal will likely move into budget deliberations regardless of what is said here. The case for keeping it out, or radically narrowing its scope, is in the numbers above. The case for rethinking the institution that wrote it is in the same place.

It is worth saying clearly what this analysis is not. The companies that form the CEC are not villains. Several have built genuine, durable Pakistani businesses. These are accomplishments worth acknowledging. None of them are diminished by the analysis above.

This piece is also not about the individuals on the CEC. The structural conflict identified is institutional, not personal. The same business logic would produce the same proposal regardless of which specific founders and CEOs happened to be in those twelve seats this term.

What this analysis is about is the institutional architecture by which a coherent collective business interest gets translated, through P@SHA's recognised policy access, into national tax policy affecting three million people whose interests run in the opposite direction. That architecture predates the freelance economy. It was designed for a different sector. When the sector changed, the architecture did not.

Fixing this does not require dissolving P@SHA, weakening it as a trade body, or attacking its members. It requires the recognition, by the policymakers who currently treat its voice as singular, that the body's policy advice is one input among many. It is not the view of a sector that long ago outgrew its founding constituency.

Six hundred and eighty-eight voters cannot speak for three million people who never voted for them. Until that arithmetic is acknowledged in the budget process, the proposal under discussion should not advance.