The Federal Board of Revenue (FBR) is undergoing yet another reinvention, with a proposed transformation plan that introduces a three-wing structure to separate audit, assessment, and enforcement functions. The plan promises faceless audits, centralized risk management, algorithmic case selection, re-audit powers, artificial intelligence, and extensive use of third-party data. The stated objective is to modernize tax administration, reduce discretion, and improve compliance. However, this reform may expose a deeper problem: Pakistan is attempting to digitize tax enforcement before digitizing the economy itself.
A History of Unsuccessful Reforms
This is not the first time the FBR has embarked on a supposedly transformative journey. Over the past quarter century, Pakistan has witnessed a succession of reform initiatives, including Large Taxpayer Units, Regional Tax Offices, Inland Revenue integration, the Tax Administration Reform Project, the Universal Self-Assessment Scheme, Track-and-Trace Systems, Point-of-Sale integration, digital filing platforms, and repeated organizational restructuring. Each has been presented as a solution to the country's chronic revenue problem, yet fundamental indicators remain stubbornly unchanged.
Pakistan's tax-to-GDP ratio remains low compared to its potential. The informal economy continues to thrive, tax litigation has multiplied, and trust between taxpayers and tax administrators has deteriorated. The overwhelming reliance on withholding taxes persists, and the number of active income tax return filers remains disproportionately small relative to the size of the economy.
The Core Problem: Economy Structure vs. Administration
The reason for these failures is simple: most reform efforts have focused on tax administration rather than the structure of the economy itself. The current transformation plan appears to repeat this mistake. The dominant assumption behind faceless audits and algorithmic enforcement is that the principal challenge facing the FBR is the efficient analysis of available information. However, the real challenge is that a substantial part of Pakistan's economy generates little or no reliable information in the first place.
An algorithm can analyze data, but it cannot analyze transactions that are never recorded. A faceless audit can scrutinize electronic trails, but it cannot effectively examine economic activity conducted entirely through cash and informal arrangements. Technology magnifies visibility; it does not create visibility where none exists.
The Cash Economy Dilemma
Pakistan remains a predominantly cash-based economy. Large segments of wholesale and retail trade operate outside effective documentation frameworks. Property transactions are frequently under-declared, and records such as property, utility, withholding tax, and commercial information suffer from inaccuracies, duplication, and inconsistent reporting standards. Agricultural income remains largely outside meaningful tax administration, and federal and provincial databases remain fragmented. Digital payment penetration is limited compared to successful tax jurisdictions.
Against this background, the proposal to strengthen centralized audits, risk engines, and algorithmic selection mechanisms raises an obvious question: who will actually be audited? The answer is equally obvious: those already visible to the state. Banks, corporations, registered manufacturers, salaried employees, documented professionals, and compliant taxpayers generate records. The undocumented economy generates very little. Consequently, the new system risks becoming extraordinarily effective at auditing those who are already documented while remaining relatively ineffective against those outside the documented economy.
Tax Justice Concerns
This is not merely a technological issue; it is a question of tax justice. Recent developments reveal the contradiction even more clearly. While the FBR seeks greater powers of centralized audit, re-audit, and algorithmic scrutiny, significant segments of the trading community continue to enjoy exemptions from meaningful documentation and integration requirements. A trader with an annual turnover of up to Rs. 200 million can remain outside comprehensive e-invoicing, full digital integration, and effective audit coverage. At the same time, large corporations, registered manufacturers, and documented businesses are being brought under increasingly intrusive digital scrutiny.
Thus, two parallel tax universes are being created. One consists of documented taxpayers whose every transaction leaves a digital footprint. The other consists of substantial segments of the cash economy that operate with limited visibility. The algorithm will inevitably audit what it can see, but it cannot audit what policy deliberately allows to remain unseen. The more visible taxpayers become increasingly visible, while the less visible remain largely invisible.
International Lessons: India's Experience
International experience provides valuable lessons. India introduced its Faceless Assessment Scheme in 2019 and subsequently expanded it into a nationwide framework. The objective was to reduce corruption, eliminate local influence, and ensure consistency in assessments. Despite India's far superior digital ecosystem, the initiative encountered substantial criticism. Businesses complained of automated notices generated through incomplete information, taxpayers reported communication difficulties, and professional bodies highlighted situations where centralised officers failed to appreciate commercial realities not fully reflected in electronic records.
These concerns emerged despite significantly higher levels of digitization, banking penetration, and information integration than currently exist in Pakistan. The lesson from India is not that faceless assessments should be abandoned, but that technology does not eliminate discretion; it merely relocates discretion. Traditional tax systems concentrate discretion in individual officers, while algorithmic systems concentrate discretion in software architecture, risk parameters, data design, and centralised governance structures. This creates an entirely new category of accountability questions.
Accountability in Algorithmic Governance
Who designs the algorithm? Who determines risk indicators? Who verifies data quality? Who corrects errors? Who supervises automated decision-making? Most importantly, who protects taxpayers against incorrect conclusions generated by defective or incomplete data? These questions acquire even greater significance in Pakistan, where data reliability remains a serious challenge. Taxpayers frequently encounter discrepancies between actual transactions and third-party information reflected in official databases. Poor data inevitably produces poor outcomes, and an inaccurate database combined with automated decision-making can transform isolated errors into systemic injustice.
The debate, therefore, should not be framed as technology versus resistance to technology. The real issue is sequencing. Successful tax jurisdictions generally followed a predictable path: first, widespread financial inclusion; then electronic payments; then mandatory invoicing systems; then integrated databases; then comprehensive taxpayer identification; and only thereafter, advanced risk management and artificial intelligence systems. Pakistan appears determined to reverse the sequence: instead of digitizing the economy first, it seeks to digitize enforcement first.
Consequences of Reversed Sequencing
The consequences may prove costly. More notices may be issued, more audits conducted, and more litigation generated, yet the underlying compliance gap may remain largely intact. This is because Pakistan's principal tax problem is not an audit gap but an identification gap. The state already possesses substantial information about the documented economy—it knows who receives salaries, operates bank accounts, imports goods, is registered, and files returns. The real challenge lies in identifying economic activity that remains beyond effective visibility. No audit strategy can solve that problem; only comprehensive documentation can.
A deeper flaw lies in Pakistan's fragmented tax architecture itself. Income tax is administered by the FBR, sales tax on services by provincial revenue authorities, and property-related information is dispersed among provincial agencies, development authorities, and land registries. Excise, vehicle registration, business licensing, and utility databases operate under separate legal and administrative regimes. The result is duplication, inconsistency, and information silos. No algorithm can compensate for fragmented governance.
Proposed Solution: A National Tax Agency
Before investing billions in artificial intelligence and centralized audit systems, Pakistan must first create a unified information architecture. The logical solution is a National Tax Agency operating through a common technological platform while respecting the constitutional distribution of taxing powers after the Eighteenth Amendment. Such a body would not centralize tax policy but would integrate information, administration, and compliance systems. Provinces would retain their constitutional taxing powers while benefiting from shared databases, common taxpayer identification, and unified enforcement protocols. Without the integration of federal and provincial information systems, algorithmic governance will remain fundamentally handicapped.
What Should Be Done?
The starting point should be universal economic digitization rather than universal audit. Every commercial establishment should gradually be brought within a national digital invoicing framework. Electronic payment incentives should replace excessive reliance on coercive enforcement. Property registries should be fully digitized and integrated across provinces. Federal and provincial tax databases should be connected through common technological standards. Agricultural income reporting should be standardized nationwide. Cash transaction thresholds should be progressively reduced. All business entities above objectively determined turnover thresholds should be brought within mandatory digital reporting systems without political exemptions.
Most importantly, taxpayer rights should evolve alongside enforcement powers. A digital tax administration requires a digital bill of rights. Transparency regarding algorithmic selection, access to information, effective review mechanisms, and independent oversight are not optional safeguards; they are prerequisites for legitimacy. Parliament must also reclaim its role. The architecture of digital taxation cannot be left exclusively to bureaucrats, consultants, and technology vendors. Any transformation affecting millions of taxpayers should be debated publicly through parliamentary committees, professional bodies, chambers of commerce, and civil society.
The purpose of reform should not be to create a more powerful tax bureaucracy but to create a more transparent economy. Without that foundation, Pakistan risks constructing a digital leviathan operating within an analogue economy. Such a system may generate more enforcement activity and even increase short-term collections, but it will not necessarily produce fairness, trust, or sustainable compliance.
The future of taxation does not lie in teaching algorithms how to chase the already documented citizen. It lies in building an economy where economic activity itself becomes visible, verifiable, and accountable. Only then can technology become an instrument of reform rather than merely a more sophisticated instrument of extraction. Pakistan does not need smarter audits before it achieves smarter documentation. It needs a national strategy for digitizing the economy itself. Until that objective is achieved, every new technological layer added to tax administration will merely widen the divide between the documented and undocumented sectors while leaving the country's most fundamental compliance challenge unresolved.
