How Pakistan’s pension bill swallows development space

Pakistan’s federal budget contains many large numbers, but few explain the state’s shrinking room for manoeuvre as neatly as its pension bill.

In the current financial year, the federal government set aside about Rs1.06 trillion for pensions. That is more than the Rs1 trillion federal development programme and far above the Rs716 billion allocated to the Benazir Income Support Programme. Islamabad is spending more on people who have left government service than on the roads, water systems, hospitals and other projects through which it hopes to expand the economy.

This is not because pensioners have suddenly become unusually expensive. It is the result of a promise made over decades without building the pool of assets needed to pay for it.

Under Pakistan’s traditional system, most public servants receive a defined benefit after retirement. The benefit is linked to salary and length of service, while the government bears the risk and guarantees payment. Employees generally did not contribute to a dedicated retirement account during their careers. Once they retired, their pensions were paid from that year’s budget.

This is an unfunded, pay-as-you-go system. Current taxpayers pay current pensioners, and another batch of employees is added when the next budget arrives.

A defined-benefit pension is attractive to an employee because the retirement income is predictable. For the government, however, it creates an open-ended liability. Salaries rise, pension increases are granted, people live longer and more employees retire. Unless revenues and a pension asset pool grow at a similar pace, the gap lands directly on the exchequer.

That gap has widened quickly. The federal pension allocation was about Rs480 billion in 2020-21. By 2024-25 it had reached Rs1.014 trillion, and the current year’s allocation is around Rs1.06 trillion. Pensions now consume roughly 6.4 per cent of federal current expenditure.

The composition matters too. Of the Rs1.055 trillion shown in the current budget documents, Rs742 billion was allocated for military pensions and Rs243 billion for civilian pensions, with the remainder covering pension increases and the federal pension fund. Roughly seven of every ten rupees in the main pension allocation therefore go towards retired armed forces personnel.

And this is only the federal bill. Provincial governments, autonomous bodies, public universities and state-owned enterprises carry pension obligations of their own. Research by the Pakistan Institute of Development Economics has repeatedly warned that the wider system is fragmented and lacks a matching asset base.

The problem is not merely that the figure is large. Pensions sit inside current expenditure, alongside interest payments, defence, salaries, subsidies and the cost of running government. These expenses are difficult to compress quickly. When revenues disappoint, development spending is easier to postpone than a monthly pension payment.

How did the bill become so unwieldy?

Part of the answer is accumulation. Pakistan recruited millions of permanent public employees under arrangements that could require the state to pay a salary for three decades and a pension for another two. The burden remained manageable while fewer employees had retired and salaries were lower. As larger cohorts left service, the cash demand accelerated.

The design also allowed costs to multiply. Pension calculations were traditionally linked to the last salary drawn, creating a higher base after a late-career promotion or pay increase. Family pension rules extended payments to several categories of dependants. Some retirees could receive more than one pension, while re-employed officials could, in certain cases, draw both a government salary and a pension.

Military pensions add another complication. Armed forces personnel often retire earlier than civilian employees, so payments may continue for a longer portion of their lives. Many then take other employment while continuing to receive a pension earned through service. The question is not whether that pension was promised; it was. The fiscal question is how the government finances promises whose full cost was not funded when they were made.

Pakistan has known about this problem for years. Pay and pension commissions were formed and international examples studied. The prescription barely changed: stop adding employees to the unfunded system, tighten existing benefits and build professionally managed funds. Implementation was another matter.

Khyber Pakhtunkhwa moved first among the major governments. It introduced a contributory arrangement for employees hired from July 1, 2022, under which employees contribute 10 per cent and the government 12 per cent. It also narrowed categories of family beneficiaries, raised the minimum age for early retirement, restricted multiple pensions and tightened benefit calculations.

The federal government eventually followed. In 2024, it approved a defined-contribution scheme for new civilian entrants and announced that the system would later cover new armed forces recruits. Under the federal rules, an employee contributes a fixed share of pensionable salary and the government adds its contribution. The money is meant to be invested through authorised pension fund managers, creating an individual retirement pot rather than another undefined claim on future taxpayers.

Under the old system, the benefit is promised and the government must find the money. Under the new one, the contribution is fixed and retirement income depends on accumulated savings and investment returns. Risk moves partly to the employee, but the liability is funded as it is created.

The government has also begun changing the old system around the edges. Reforms notified in 2025 restricted multiple pensions, shifted benefit calculations away from the final salary alone towards an average of recent pay, and stopped retired civilian employees who return to government jobs from collecting both full salary and pension simultaneously. These changes can reduce leakages and slow future growth, although they do not erase pension rights already earned.

This is the awkward part of pension reform: even a sensible reform can take decades to improve the headline number.

Moving a new recruit out of the old system prevents a future liability, but does little to pay today’s pensioner. For a long transition, the government must finance the legacy scheme while contributing to new accounts. It is paying for the past and saving for the future simultaneously.

That is why the upcoming budget should not be judged only by whether the pension allocation rises or falls. It will almost certainly remain close to, or above, Rs1 trillion. Existing pensioners remain entitled to payment, and any increase announced for retirees will add immediately to expenditure. A contributory scheme cannot make that cost disappear by July.

The more consequential development expected in the 2026-27 budget is the extension of the defined-contribution system to armed forces personnel recruited from July 1, 2026. The military switch was originally intended to begin earlier but was delayed. Because military pensions form the largest component of the federal bill, bringing new recruits into a funded arrangement would close the most important remaining door through which fresh unfunded liabilities are entering.

But an announcement alone will not settle the matter. Three questions should be asked when the budget is presented.

First, who exactly is covered? Restricting reform to new recruits protects existing rights but delays the fiscal benefit. Changes affecting serving employees, retirement ages, family pensions or post-retirement work would produce different costs and must be stated clearly.

Second, where is the money? Employee and government contributions should enter identifiable, ring-fenced accounts managed under transparent investment rules, not quietly become cash for routine spending. The federal budget allocated Rs10 billion for the pension fund in 2024-25 and Rs4.3 billion in 2025-26, but the government has yet to provide a full public account of contributions collected, assets accumulated, managers appointed and investment returns.

Third, is the government reducing the legacy burden or merely slowing its growth? Limiting multiple pensions, averaging salaries for calculations, reviewing family benefits and regulating salary-plus-pension arrangements can produce nearer-term savings. Without such measures, the contributory scheme remains a reform for future finance ministers while today’s finance minister continues paying the old bill.

Pension reform should not be confused with an argument against pensioners. A pension is deferred compensation and, once promised under law, cannot be treated as a discretionary favour. Reform is needed so that the promise can be honoured without steadily displacing spending on citizens who never held a government job.

That requires a bargain. Employees need a credible fund and clear ownership of their savings. Government needs predictable contributions. Taxpayers need consistent rules without exemptions for politically connected groups.

Pakistan has finally begun moving from a system built on promises to one supported by assets. That is an important change, but only the start of a long handover.

When the budget arrives, the eye-catching figure will again be the pension allocation. The more useful clues will be buried nearby: whether military recruits are finally brought into the contributory system, how much has been collected, where those funds are invested, and whether the old scheme is being cleaned up rather than simply carried forward.

The pension problem will not be solved in one budget. The immediate task is more modest: stop making it larger.