I have spent time in flood-hit villages in Sindh and cramped one-room homes in peri-urban Punjab, watching women describe BISP payment days in precise detail: bus fares to disbursement points, hours in queues, biometric failures requiring return trips and, too often, agents quietly deducting their “share” before handing over what remains. This overlooked leakage is rarely part of the national debate on Pakistan’s social protection system, where attention stays fixed on state spending rather than how much actually reaches beneficiaries.
Over the past five years, Pakistan has channelled more than Rs 2.5 trillion into BISP, its largest social protection commitment, even as poverty rose from about 22 percent to nearly 40 percent, pushing around 20 million more people below the line while the program expanded. I do not say this to dismiss BISP, which has delivered real relief—helping with school supplies, medicines and short-term recovery after shocks—and, for many women, remains the only income that reaches them directly without mediation through male relatives, a reality that should not be minimized.
But two things can be true at once: BISP cushions shocks, yet as currently designed and run it cannot lift households out of poverty—partly because cash transfers were never built for that role and partly because a meaningful share of the money never reaches intended recipients. My work has taken me to a part of this debate often avoided: the claim that “cash transfers can’t beat inflation” is true but incomplete, as it assumes every rupee reaches a beneficiary and is merely insufficient, whereas the more uncomfortable reality is that a significant portion never arrives at all.
The Auditor General’s reports make this clearer than any NGO survey. In 2023, auditors informed the Public Accounts Committee that around Rs 19 billion in BISP funds had been illegally drawn by about 143,000 government employees, including over 2,500 senior officers, who registered spouses and relatives as beneficiaries of a program meant for the poorest households. A 2023–24 audit flagged Rs 141 billion in irregularities, including payments to individuals without verified national identity records and withdrawals from dormant accounts. A later three-year audit found roughly Rs 96 billion paid to ineligible recipients, alongside further embezzlement involving officials across multiple levels of seniority.
Then there is the registry itself. The National Socio-Economic Registry, the database BISP uses to decide who qualifies, still excludes millions of the poorest households even as it claims to cover the vast majority of the population — a gap recently confirmed by independent World Bank analysis. Because the registry is updated only periodically rather than continuously, families that fall into poverty between survey waves simply don’t exist in the system, while others who have since climbed out of poverty continue drawing support because no one removed them.
Put plainly, Pakistan does not only face a “cash versus jobs” problem but a “cash that leaks before it lands” problem—no additional budget fixes a sieve. This is not to excuse the state from investing in livelihoods; rather, it strengthens the case: if a meaningful share of BISP’s Rs 2.5 trillion is lost to fraud, ghost beneficiaries, exclusion errors and rent-seeking along the disbursement chain, then the real comparison is not “Rs 2.5 trillion in cash versus Rs 2.5 trillion in jobs programs,” but how many billions actually reached poor households versus what the same money could have built.
And even the portion that did arrive faced an economy that ate it alive. Food inflation that, at points, exceeded 40 percent. Electricity bills consuming half a laborer’s monthly wage. Real wages falling as jobs disappeared. A quarterly stipend, however reliably delivered, was never going to outrun that.
Labor-intensive infrastructure, SME financing and agriculture value-chain development consistently generate somewhere in the range of 3,000 to 12,000 jobs for every billion rupees invested, depending on the sector and design. Applied even loosely to Rs 2.5 trillion, that is a range running from roughly 7 to 30 million jobs — enough, on conservative assumptions, to support tens of millions of people and meaningfully bend Pakistan’s poverty trajectory. We chose cheques instead and then failed to fully protect even that choice from theft.
From where I sit — inside program design, not above it — three things need to happen together, not in sequence.
First, fix the plumbing before debating the volume. Make the registry dynamic and continuously verified rather than updated once every few years. Tie biometric verification to real penalties for tampering. Give the Auditor General’s findings teeth — recovery, prosecution and public reporting on a fixed timeline, not a parliamentary committee hearing the same numbers again eighteen months later.
Second, narrow cash transfers to their proper purpose: rapid, time-bound relief after floods, earthquakes and displacement and a tightly means-tested floor for households that genuinely cannot work — not a permanent feature of normal economic life for tens of millions of people. Third, redirect the savings — both from tighter targeting and from plugged leakages — into livelihoods: public works that build real assets, SME credit lines with reasonable collateral requirements and agricultural value-chain investment that creates jobs in the rural areas where most BISP beneficiaries actually live.
None of this is a case against social protection. I have seen what it prevents. It’s a case against pretending that writing bigger cheques, through a system we already know leaks, is the same thing as building an economy where fewer people need a cheque at all. Relief should be a bridge. We have built it into a destination — and left the toll booth unguarded along the way. Pakistan doesn’t just need to spend differently. It needs to spend honestly and then spend on growth. Survival first, yes — but survival was never meant to be the whole plan.
—The writer is Ex-Chairman, National Disaster Management Authority.
