“An order book proves that investors will lend at a price. It does not prove that the borrowed dollars will raise Pakistan’s capacity to repay.”
On September 3, 2026, at a ‘High-Level Dialogue on Taxation for Fiscal Sustainability’ jointly organised by the Asian Development Bank (ADB) and the Government of Pakistan, Finance Minister Muhammad Aurangzeb chose to begin with news from the international capital market. “Pakistan as a sovereign has printed a $3 billion bond,” he said. He called it the largest single transaction in the country’s history and an external validation of Pakistan’s economic direction.
The sentence was delivered at precisely the forum where a harder question should have been asked: what does new foreign debt validate when it is being raised to retire old foreign debt? Shahbaz Rana’s excellent report supplied the missing ledger. Pakistan had obtained a three-month, $3 billion Saudi deposit in April to repay a liability to the United Arab Emirates (UAE). After one extension, the new Eurobond proceeds are intended to repay Saudi Arabia. The sequence is UAE liability, Saudi bridge, global bond.
This is not an argument against refinancing. Every responsible sovereign smooths maturities, replaces short obligations with longer ones and diversifies creditors. It is an argument against converting a liability-management operation into a certificate of economic transformation. No new export plant, power saving, technology platform or human-capital asset appears anywhere in this chain. The identity of the creditor and the date of repayment have changed; Pakistan’s capacity to earn the dollars has not automatically changed with them.
“Printed” is familiar market shorthand for issued or priced. The problem is not the jargon. It is the political alchemy performed with it. A bond is not national income. It is a promise to transfer future dollars to present lenders. When a government presents the promise as wealth, debt management begins to masquerade as development.
WHAT THE MARKET ACTUALLY VALIDATED
The transaction has features worth acknowledging. Pakistan sold $1.75 billion of 5.5-year paper with a 7.5 per cent coupon and $1.25 billion of 10-year paper with a 7.9 per cent coupon. Reported yields were 7.75 and 8.25 per cent. Orders approached $6 billion. Demand for the 10-year tranche shows that investors were willing to hold Pakistani risk beyond the present International Monetary Fund (IMF) programme horizon.
Comparison with 2021 also requires care. Coupons are higher today, but benchmark US Treasury yields are far higher. Pakistan’s spread over the risk-free benchmark is consequently narrower than it was in 2021. That indicates an improvement from the period when default fears dominated the country’s pricing. The present $3 billion is also defensibly described as the largest single offering: the $3.5 billion cited for 2021 combined an April issue with a July reopening.
These facts establish a debt-management achievement. They do not establish the much larger claim attached to it. S&P’s B and Fitch’s B-minus ratings remain speculative grade. Fitch stated that the proceeds were for general budgetary and sovereign financing and identified structural improvement in tax revenue generation, lower debt and lighter servicing as conditions for a future upgrade. The rating is thus not a declaration that fiscal reform is complete. It is a price-sensitive judgement that Pakistan can presently service more debt, provided reform and external financing continue.
An order book answers one question: will investors lend at the offered return? It does not answer whether the borrower will use the money productively, whether taxation is fair, whether exports will rise, or whether citizens will receive better services. Nearly $6 billion of orders for $3 billion of bonds may reflect confidence, scarcity value, portfolio diversification and a generous yield at the same time. Investors were not casting votes on Pakistan’s social contract; they were purchasing a claim on its future foreign-exchange earnings.
The distinction matters because the sovereign spread measures perceived credit risk, not the quality of development. Pakistan can become a better borrower without becoming a more productive economy. The first may buy the time required for the second. It cannot substitute for it.
THE ARITHMETIC BEHIND THE CELEBRATION
At face value, the two tranches carry an annual cash coupon of about $230 million: $131.25 million on the 5.5-year bond and $98.75 million on the 10-year bond. If both remain outstanding to maturity, their scheduled nominal coupons add up to roughly $1.71 billion, before fees and without counting the $3 billion principal that must be repaid. Because the reported yields exceed the coupons, the issue price was below par; the government should publish the exact net proceeds and all transaction costs.
The longer maturities remove a near-term repayment cliff, which is valuable. They replace rollover risk with duration, interest and exchange-rate risk. Every depreciation of the rupee raises the domestic resources required to buy the dollars for coupons and principal.
Foreign-currency borrowing is safest when it finances an activity that earns or saves foreign exchange. Borrowing dollars for a rupee budget deficit, or to roll a bridge loan that financed another repayment, leaves the exchange-rate mismatch with the public.
The institutional setting deepens the concern. Shahbaz Rana reported that the Debt Management Office has been without a permanent director general since January 2020. A transaction proclaimed as the sovereign’s largest single international bond issue should be accompanied by a permanently led debt office, a published maturity-risk strategy and a transaction statement specifying pricing, allocation, fees, net proceeds, use of funds and the liability retired.
The wider balance sheet explains why this is more than a semantic dispute. State Bank data put total debt and liabilities at Rs. 99.59 trillion at end-June 2026, including Rs. 97.88 trillion of debt. External debt and liabilities stood at $138.85 billion. Pakistan serviced $21.59 billion externally during fiscal year (FY)2026; $10.14 billion fell in the final quarter alone, 2.63 times the preceding quarter as principal repayments surged.
There has been real relief on the domestic side. Total debt and liabilities servicing declined from Rs. 13.16 trillion in FY2025 to Rs. 11.97 trillion in FY2026, principally because lower interest rates cut domestic mark-up. That improvement should be preserved. It should not be confused with a reduction in the debt stock or with an escape from external refinancing dependence.
A LOW DEFICIT, BUT WHAT KIND OF CONSOLIDATION?
The minister also pointed to a fiscal deficit of 2.6 per cent of GDP, a third consecutive primary surplus, Federal Board of Revenue (FBR) collection approaching Rs. 13 trillion and a rise in the tax-to-GDP ratio from 8.8 to 10.3 per cent. The FY2026 deficit is indeed the lowest in 22 years. Official fiscal operations show a deficit of Rs. 3.313 trillion and a primary surplus of Rs. 3.634 trillion.
Composition again changes the verdict. The improvement was supported by nearly Rs. 1.97 trillion in domestic debt-servicing savings, a combined provincial surplus of Rs. 1.449 trillion and Rs. 1.567 trillion from the petroleum levy. FBR still missed the revenue target agreed with the IMF. Lower interest costs are welcome, provincial discipline is necessary and the levy raises cash. None by itself proves that the federal tax base has become equitable or durable.
The petroleum levy is classified as non-tax revenue and remains outside the divisible pool. Much of what is reported as direct taxation consists of withholding, advance, minimum or presumptive imposts collected on transactions rather than on accurately assessed net income. Sales tax and import-stage taxation load the price of necessities and inputs.
A person below the taxable threshold may still pay advance income tax through a mobile phone, electricity bill or banking transaction. FBR will keep on increasing collection while deepening the burden on people already documented and taxed.
Nor has privilege disappeared. FBR’s Tax Expenditure Report 2025 estimated concessions and exemptions for FY2023-24 at Rs. 2.435 trillion: about Rs. 1.237 trillion in sales tax, Rs. 545 billion in income tax and Rs. 652 billion in customs duty. Even this exercise is constrained by methodology and data. A fiscal system cannot credibly lecture citizens about non-compliance while Parliament renews large preferences without transparent beneficiaries, distributional analysis or expiry dates.
The correct measure of fiscal consolidation is not a single deficit ratio. It is whether recurrent obligations are structurally contained; whether taxes shift from captive transactions to real income, wealth, rents and unearned gains; whether refunds are paid; whether development and social spending survive; and whether the economy’s future revenue and foreign-exchange capacity expands. Consolidation produced by cheaper interest, federal levies, provincial cash surpluses and compressed demand can be useful breathing space. It is not a new fiscal order.
THREE LENDERS, ONE REFORM LOOP
The setting of the minister’s announcement brings the multilateral role into focus. The IMF, World Bank and ADB have not ignored Pakistan’s tax weakness. They have diagnosed it repeatedly, financed reform projects, supplied consultants, supported information systems and conditioned lending on revenue and administrative measures. Many individual prescriptions—fewer exemptions, wider bases, digital records, stronger audit, better cash management and improved federal-provincial coordination—are sensible. The puzzle is why so much diagnosis and financing have produced so little structural change.
The World Bank’s record offers a controlled experiment. Its independent evaluation concluded that the 2005 Tax Administration Reform Project failed to raise revenue and identified familiar defects: complexity, a narrow base, low compliance, inefficient administration and weak provincial collection.
The subsequent Pakistan Raises Revenue project began with $400 million. In 2025 the Bank added $70 million, taking its financing to $470 million, and cited 1.5 million new taxpayers, fewer withholding lines, upgraded IT and better tax-expenditure reporting as gains.
Those outputs are not worthless. Registration is not the same as an active, accurately assessed tax base; a new filer may submit a nil return or file only to avoid punitive non-filer withholding. A portal can reduce friction, or it can digitise a bad law.
Data analytics can identify evasion, or become another instrument of arbitrary notices if audit selection, privacy, reasons and appeal rights remain weak. The decisive indicators are the tax gap, voluntary payment on assessed income, refund time, audit yield sustained after appeal, compliance cost and the distribution of the burden. These remain secondary in most reform scorecards.
ADB occupies a similar double role. FBR recorded that $300 million was disbursed in December 2023 after the first subprogramme of the domestic resource mobilisation programme. In June 2025, ADB approved an $800 million public-finance package: a $300 million policy loan and a guarantee of up to $500 million expected to mobilise as much as $1 billion from commercial banks.
It was designed to improve tax policy, administration, compliance, expenditure and cash management. Now ADB is co-convening a dialogue on taxation at which Pakistan’s ability to issue another $3 billion of debt is offered as the opening proof of success!
There is no legal conflict in a development bank financing reform and helping a member access capital. There is, however, a dangerous feedback loop when reform success is inferred from renewed borrowing capacity. A programme for fiscal sustainability can itself enlarge liabilities; a guarantee can mobilise commercial debt; prior actions can unlock disbursement; and the next programme can be justified because the tax structure remains unreformed. Unless outcome measures break this loop, the lender evaluates reform partly by the borrower’s ability to keep borrowing.
The IMF supplies the macro framework. Its own Independent Evaluation Office described Pakistan as one of the most prolonged users of Fund resources, under IMF-supported programmes almost continuously since the late 1980s. The present 37-month, $7 billion Extended Fund Facility again promises broader taxation, stronger public finances, restored energy viability and sustainable growth. By May 2026, disbursements under the EFF and the climate facility together had reached about $4.8 billion. Stabilisation, reserve rebuilding and lower near-term default risk are genuine achievements.
The Fund is a crisis lender, however, and its programme horizon favours measurable near-term balances. When revenue underperforms, governments reach for the instruments that produce cash fastest: higher withholding, sales and import taxes, petroleum levies, administered energy-price adjustments and mini-budgets.
These satisfy a quantitative path more readily than taxing politically powerful assets, rebuilding assessment capacity, resolving litigation or renegotiating the relationship between the federation and provinces. The social and growth costs arrive outside the review calendar.
Pakistan’s rulers cannot outsource responsibility for this failure. No multilateral institution passes the Finance Act, issues any statutory regulatory order (SRO), grants an amnesty or refuses to tax a favoured constituency. Domestic elites select and implement the measures that protect them. The lenders’ failure lies elsewhere: in repeatedly accepting proxies for reform, privileging aggregate collection over incidence and institutional legitimacy, and continuing finance after short-lived compliance with prior actions. The result is joint production—domestic political capture supported by an external architecture that rewards short-term bankability.
WHY THE PRESCRIPTIONS DO NOT TRAVEL
Tax reform fails first because it is treated as an administrative engineering project rather than a constitutional political settlement. Who pays, how much, under which legislature’s authority and in return for what services are questions of representation.
In Pakistan, FBR—the collecting agency—continues to dominate policy and draft annual legislation. Parliament receives dense Finance Bills with little time for public scrutiny. Executive concessions and frequent amendments make predictability impossible. Technology cannot cure a legitimacy deficit designed into the law.
It fails because the tax-to-GDP ratio is treated as both diagnosis and destination. The same ratio can be produced by very different systems. One may tax net income, immovable property, economic rents and consumption through a broad, low-rate value-added tax with refunds. Another may collect at import, withhold on gross transactions, impose minimum tax on turnover and use fuel prices as a revenue machine. The first can support investment and consent. The second can raise cash while suppressing formalisation.
It fails because expenditure remains outside the tax conversation. Citizens see debt servicing, losses of state-owned enterprises, circular debt, pensions without funding, elite privileges and poor municipal services. Asking them to become compliant without demonstrating economy, accountability and benefit is not a compact; it is extraction. The state’s own conduct determines whether digital documentation becomes trust or surveillance.
It also fails because Pakistan is one economy administered by multiple, poorly coordinated tax bodies. The World Bank’s latest fiscal-federalism report correctly notes higher compliance costs and weak provincial revenue effort, while also recognising that reform must operate within the constitutional framework.
Harmonisation does not mean erasing provincial competence. It means common definitions, interoperable data, one registration and return interface, full input-credit recognition, predictable transfers and an agreed collection arrangement under democratic federal consent.
AN ALTERNATIVE ALREADY ON THE TABLE
Our 2020 PIDE volume, Tax Reforms in Pakistan: Historic & Critical View, challenged the donor-backed story that Pakistanis simply refuse to pay. Its concise premise was that “taxes are by-products of economic growth”.
It documented how transaction-based withholding turned businesses, banks, schools and utilities into collection agents; how the Income Tax Ordinance, 2001 was imposed without the parliamentary deliberation required for durable law; and how fragmented administration increased costs while citizens received few enforceable rights.
The 2024 third edition of Towards Broad, Flat, Low-rate and Predictable Taxes, published by PRIME, reached the same conclusion after updating the evidence: Pakistan needs a “lower, simpler, and predictable tax system that is pragmatic, growth-oriented, broad-based and above all, citizen-friendly”.
It proposed a federalised National Tax Agency for collection on behalf of all levels of government, a harmonised sales-tax framework, a Taxpayers’ Bill of Rights and a National Tax Court. These are not devices for weakening enforcement. They are means of making enforcement lawful, inexpensive and credible.
A workable transition should begin with Parliament reclaiming tax policy. Every exemption must appear in a consolidated tax-expenditure schedule naming its legal basis, purpose, estimated cost and beneficiaries; each should expire unless renewed after a public distributional review.
The executive’s power to alter incidence through subordinate instruments must be confined within Articles 77 and 162 of the Constitution. Stable multi-year rate paths should replace annual surprises and mid-year revenue raids.
Income tax must return to income. Minimum and presumptive regimes should be phased out, while withholding is retained only where it is a genuine, fully creditable collection mechanism—principally wages, interest and payments where reliable third-party information exists. Broad, lower and predictable rates should be paired with effective taxation of agricultural income by provinces, under-taxed urban immovable property, economic rents, unearned gains and high-value assets. Compliance should be judged by tax paid on real capacity, not by appearance on an Active Taxpayers’ List.
Sales taxation requires one interoperable system for goods and services, a broad base, lower and harmonised rates, full input adjustment and time-bound automated refunds. Constitutional assignments can remain intact while a common agency or platform collects and settles shares transparently. A specialised National Tax Court should resolve disputes to a statutory timetable. Digitalisation should deliver pre-filled information, a single account and reasons for every automated action—not more portals, blocked accounts and coercive notices.
The expenditure side must carry equal weight. Loss-making state enterprises, untargeted subsidies, unfunded pension promises, procurement leakage, circular debt and official privileges cannot remain protected while households are taxed through fuel and electricity.
Federal consolidation must not be manufactured by forcing provinces to park surpluses while schools, clinics and local governments remain underfunded. Revenue legitimacy is created where citizens encounter the state: water, sanitation, transport, education, health, policing and justice.
CHANGE THE TESTS—AND THE USE OF DEBT
Future IMF, World Bank and ADB support should be tied to outcomes that citizens and Parliament can verify, not only to laws enacted, databases purchased, registrations added or money disbursed. The scorecard should disclose the share of income-tax collection arising from assessed net income rather than final or minimum withholding; the number of active taxpayers with material liability; tax-gap estimates by sector; effective rates by income group; refund age; audit results surviving appeal; litigation time; compliance cost; and additional spending on essential services. Independent evaluations, consultancy contracts, procurement and all programme conditions should be public.
Debt governance needs an equally concrete rule. Within seven days of every sovereign bond, the Ministry of Finance should publish a transaction report covering the gross and net proceeds, price, yield, coupon, fees, investor allocation, currency risk, maturity profile, use of funds and liability replaced.
Parliament should require that new non-concessional foreign-currency debt normally finance assets or reforms with a measurable capacity to earn or save foreign exchange. Refinancing for crisis management can be allowed, but it must be labelled as refinancing and placed on a declining medium-term ceiling.
The Debt Management Office needs permanent professional leadership and statutory independence from celebratory messaging. A medium-term debt strategy should be stress-tested for exchange-rate depreciation, interest shocks, commodity prices and lost rollovers. Each budget should show not only debt-to-GDP but gross financing needs, foreign-currency maturities, guarantees, contingent liabilities, the domestic banking system’s sovereign exposure and the opportunity cost of crowding out private credit.
These tests would alter the incentives of government and lenders alike. A successful tax project would be one that makes voluntary compliance cheaper, incidence fairer and revenue more resilient during slower growth. A successful bond would be one whose proceeds reduce future financing needs or create the foreign exchange to service it. Market access would remain useful, but it would cease to be the measure of national progress.
THE REAL VALIDATION
Pakistan’s return to the Eurobond market deserves neither reflexive condemnation nor official euphoria. Replacing a three-month bilateral bridge with longer paper can reduce immediate rollover risk. Securing a narrower spread in a higher global-rate environment shows recovery from a dangerous period. The $3 billion transaction may have bought valuable time.
Time is not reform. If the breathing space is used to preserve exemptions, expand transactional taxation, postpone expenditure restructuring and finance the next repayment, the country will have moved further into debtocracy: a system in which fiscal policy, taxation, banking and diplomacy are organised around the ability to borrow and repay. Citizens become sources of cash flow for promises made in their name, without acquiring the productive assets or public services from which repayment capacity should come.
The finance minister was speaking at a dialogue on taxation for fiscal sustainability. The appropriate opening announcement would have been a parliamentary tax compact: which privileges will end, which coercive withholding taxes will disappear, how provinces will participate, when refunds will be paid, how spending will be made accountable and how new revenue will finance human and productive capacity. A bond cannot answer any of those questions.
International investors have validated Pakistan’s present ability to offer a return they consider adequate for the risk. Fiscal sustainability will be validated by something harder: a tax system that citizens regard as fair, an economy that earns more dollars than it must repeatedly borrow, and a state that turns revenue into visible public value. A country escapes debtocracy not when creditors line up to buy its paper, but when it no longer needs to sell new paper to repay the last bridge loan.