Pakistan's macroeconomy has been pulled back from the edge. Its people have not. This is an account of the gap between the two — and of the 105 million young Pakistanis whose working lives will be decided inside it.
In the year to June 2026, Pakistan did almost everything an economist would ask of a country in crisis. Growth accelerated to 3.70%. The current account was brought to near balance. Reserves roughly doubled from their 2023 lows to $18.5bn. The fiscal deficit narrowed to around 3.2% of GDP with a primary surplus. Remittances hit an all-time record. The IMF signed off on its third review. By the metrics that determine whether a state can pay its bills, Pakistan is meaningfully safer than it was three years ago.
Over almost exactly the same period, the average urban Pakistani became 19% poorer in real terms, and the poorest fifth of urban households lost 23% of their real income. Per-person beef consumption fell 42%. Milk fell 10%. Households cut the share of their budget spent on their children's education from roughly 4% to 2.5%. Measured food insecurity rose from 15.9% to 24.4% of the population.
Both of these are drawn from official Pakistani government data. Neither is spin. The central finding of this report is that they are not in contradiction — they are causally linked. Pakistan stabilised its balance sheet substantially by transferring the cost of adjustment onto household consumption: through the inflation of 2022–24 that permanently repriced the cost of living, through electricity and gas tariffs raised to cost recovery, and through a petroleum levy that now raises more than Rs 1.5tn a year. Stability was purchased. Ordinary households paid for it.
For the young, this arithmetic is harsher still. Pakistan adds roughly 3.5 million people to its labour force every year and creates something closer to 2.5–3.1 million jobs. The shortfall compounds annually. Youth unemployment stands at 12.9% against a national 7.1% — and, in the single most damning statistic in Pakistan's labour data, unemployment rises with education. A Pakistani with no schooling faces 4.4% unemployment. One with a postgraduate degree faces 11.7%. A woman with a university degree faces 23.9%.
Against that, the honest opportunity set is narrower than the national conversation suggests but not empty. IT and software exports have grown at a 16% compound rate for five years to roughly $4.5bn. Freelance earnings will cross $1bn for the first time. Solar installation has become a genuine mass trade almost by accident. Legal overseas employment is running at three-quarters of a million placements a year with new quotas opening in Italy and Qatar. What these have in common is that they reward certified, verifiable, specific skill — and that Pakistan's training system, which reaches perhaps 13% of new labour market entrants, is not built to supply it.
The state has stopped bleeding. It has not started building.
The FY2026 numbers describe a genuine and hard-won recovery. Real GDP grew 3.70%, up from 3.18% in FY2025 and 2.6% in FY2024. Large-scale manufacturing grew 6.1% — its best in four years, with 16 of 22 tracked sub-sectors expanding. Services grew 4.09%. The economy reached Rs 126.9tn, about $452bn, with per-capita income at roughly $1,901.
But the target was 4.2%, and the miss matters. Pakistan needs sustained growth well above 5% simply to absorb its own demographics; at 3.7% growth against 2.55% population growth, real income per head rises barely over one percent a year. This is the arithmetic that explains why a "recovery" year can coincide with rising measured poverty.
Inflation is where the household experience diverges most sharply from the official narrative. Average CPI ran at 23.4% in FY2024, collapsed to 4.5% in FY2025, and averaged roughly 6.7–7.2% across FY2026. That collapse was real, and it was the government's proudest achievement.
It did not last. From a near-zero trough of 0.28% in April 2025, year-on-year inflation climbed to 7.3% by March 2026, 10.9% in April, and 11.7% in May 2026, closing the year at 11.1%. Three forces drove it: the Middle East conflict that erupted in late February 2026 and pushed Pakistan's oil import bill up by roughly $1bn in April alone; the pass-through of those global energy costs into domestic tariffs; and the food-supply damage from the 2025 monsoon floods. Core inflation — 8.7% urban, 7.9% rural — confirms this is not purely an import shock.
A fall in the inflation rate is not a fall in prices. After cumulative inflation of roughly 23% and 30% in FY2023–24, the price level was permanently reset. When inflation "fell to 4.5%," prices were still rising — from a base that had already broken most household budgets. Nothing came back down. This is the single largest source of the gap between government presentations and lived experience.
Pakistan's external position improved dramatically, and it is worth being precise about why. Reserves reached $18.5bn by end-FY2026, from around $14.5bn a year earlier — about 2.5 to 3 months of import cover, still thin. The rupee held near 278/$.
The engine was not exports. Goods exports actually fell 4.9% to $30.8bn, dragged down by roughly $1.1bn less rice and $1.5bn less food exports overall. The goods trade deficit widened to $33.6bn. Foreign direct investment fell 34% to $1.64bn. What held the external account together was remittances at a record $41.6bn, up 8.6%, with May 2026 alone bringing in $4.25bn — the largest single month in the country's history.
This is a country whose external solvency rests on the wages its citizens earn by leaving. The current account still slipped to a -$139m deficit for FY2026 after FY2025's surplus. Meanwhile the real effective exchange rate reached 106.44, a seven-year high — meaning the rupee is now expensive in real terms, actively working against the export growth Pakistan says it wants.
Here is the number that governs everything else in Pakistan's political economy. In the FY2027 budget, markup and interest payments come to Rs 8,045bn — 43% of the entire Rs 18.8tn federal outlay, and roughly 68% of net federal revenue.
Two-thirds of every rupee the federal government keeps goes to servicing debt before a single school is built, a teacher paid, or a hospital supplied. Defence takes Rs 3,000bn. What remains for the entire federal development programme is Rs 1,000bn — about one-eighth of the interest bill.
The revenue side explains why. The Federal Board of Revenue collected Rs 13.0tn in FY2026 — enough to beat a downward-revised target by Rs 21bn, while missing the original IMF-agreed benchmark by roughly Rs 975bn. Both framings circulated in the press; both are true. Tax-to-GDP sits near 10.2%. The Finance Minister himself conceded the ratio "should be in the high teens."
Agriculture is effectively exempt from income tax. Real estate is systematically under-assessed. Three to four million retailers remain outside the net. Instead, the burden falls on salaried employees, importers, and — through a petroleum levy now at up to Rs 120 per litre — every commuter and freight operator in the country. Public debt stands at 70.1% of GDP including IMF obligations, or 73.8% counting government guarantees; the government prefers to cite 68.5% on a narrower definition.
Pakistan is 22 months into a 37-month, $7bn Extended Fund Facility approved in September 2024, with a companion Resilience and Sustainability Facility. The third review completed on 8 May 2026, taking cumulative disbursements across both facilities to roughly $4.8bn. The IMF's language was approving — "strong implementation despite Middle East war" — while warning that "downside risks are high" and that the State Bank should "stand ready to raise interest rates." The policy rate, cut from a peak of 22% in June 2023, sits near 11% and was held at the July 2026 meeting.
Layered on top was the 2025 monsoon. Between June and mid-September, rainfall ran 23% above normal and the Sutlej, Ravi and Chenab flooded simultaneously — Punjab's worst in four decades. 1,037 people died, including more than 275 children. Some 6.5 to 6.9 million were affected and 2.9 million evacuated in Punjab alone. Nearly 230,000 homes were destroyed or damaged, 92% of them in Punjab. Total damage: Rs 822bn, about $3bn, of which Rs 430bn was agricultural. Roughly a third of the cotton crop was lost. More than 200,000 jobs went with it.
Pakistan's crises are not accidents of bad luck; they are the output of a stable structure. Four features reproduce them:
For FY2027, the IMF projects 3.5% growth with average inflation of 8.4%; the government targets 4.0% with an FBR target of Rs 15.3tn, a 17.6% jump after a year in which it missed by nearly a trillion.
The most important economic document Pakistan published this year was not the budget. It was a household survey.
The Household Integrated Economic Survey for 2024–25 contains a finding that ought to have dominated national debate. Average monthly household income nearly doubled in nominal terms between 2018–19 and 2024–25 — from Rs 41,545 to Rs 82,179, a rise of 97.8%.
Adjusted for inflation, the average urban Pakistani was 19% poorer at the end of that period, and the average rural Pakistani 7% poorer. The poorest urban quintile lost 23% of its real income: what was Rs 24,365 in 2018–19 was worth Rs 18,820 by 2024–25. Six years of work, and the money bought a fifth less.
You do not need a price index to read this. The same survey measured per-person monthly consumption across both periods, and the pattern is a household budget being dismantled from the top down — protein first.
Beef consumption did not fall by 42% because tastes changed. Pulses — the cheapest protein available to a poor Pakistani household, the fallback when meat goes — fell by more than a quarter. When the fallback is being cut, there is no lower rung. Only tomatoes rose, by 15.9%.
Measured food insecurity rose from 15.9% to 24.4% of the population across the same period. Households also cut education spending from roughly 4% to 2.5% of their budgets — a decision that converts a present cash shortage into a permanent reduction in a child's earning capacity.
Set the federal minimum wage, raised to Rs 40,700 a month for FY2026–27, against current prices as recorded by the Bureau of Statistics in mid-August 2026.
| Item | Price (Rs) | Unit |
|---|---|---|
| Mutton | 2,384 | per kg |
| Beef, with bone | 1,267 | per kg |
| Vegetable ghee | 568 | per kg |
| Chicken, live broiler | 440 | per kg |
| Hi-speed diesel | 384 | per litre |
| Petrol | 326 | per litre |
| Eggs | 300 | per dozen |
| Fresh milk | 202 | per litre |
| Sugar | 179 | per kg |
| Tomatoes | 102 | per kg |
| Onions | 67 | per kg |
| LPG cylinder | 4,494 | 11.67 kg |
| Source: Pakistan Bureau of Statistics, Sensitive Price Indicator, 17 cities. | ||
A family of five needs perhaps Rs 25,000–35,000 a month in food to eat adequately at those prices. Rent for a two-room portion runs Rs 25,000 in Karachi's Malir, Rs 33,000 for a five-marla lower portion in Lahore, Rs 8,000–18,000 in a smaller city, and Rs 5,000–12,000 for a single room in a katchi abadi — the informal settlements that house perhaps 40–50% of Karachi. Add electricity, gas, transport, and any medicine at all.
The total lands between Rs 50,000 and Rs 70,000 a month against a single minimum-wage income of Rs 40,700 — and that wage is largely unenforced across the 80% of employment that is informal. A daily-wage labourer earns Rs 800–1,200 a day when there is work. A domestic worker earns Rs 8,000–25,000 a month. This is a structural monthly deficit, and it is closed not by budgeting but by borrowing, selling, and going without.
No single expense generates more anger. The residential tariff structure notified in February 2026 splits households at 200 units a month. Stay under it and you are "protected," paying Rs 10.54–13.01 per unit. Cross it in any single month and you pay unprotected rates — Rs 22.44 per unit on the first hundred units, Rs 33.10 from 201 to 300, up to Rs 47.20 above 700.
A household consuming 190 units pays roughly Rs 5,200 a month. A household consuming 210 units — twenty units more, one hot week with a fan running — pays roughly Rs 10,800. Twenty extra units cost Rs 5,600. Families ration light and fans in August heat to stay below a threshold, and a single miscalculation doubles the bill.
On top of the per-unit rate sit fixed charges of Rs 200–675 per kilowatt, 18% GST, a fuel cost adjustment averaging Rs 2.5–3.5 per unit, a Neelum-Jhelum surcharge, electricity duty, and a Rs 35 television licence fee. The base tariff was raised 19% to Rs 35.50 per unit in July 2024 alone. In August 2023, the bills triggered the largest spontaneous protests of the decade, with residents in Karachi, Lahore and Rawalpindi burning them in the street. As the columnist Zahid Hussain wrote at the time, the surcharges amounted to "extortion by a state unable to collect taxes from the politically powerful landed and business classes."
This is where honesty requires showing the disagreement rather than picking a number.
| Estimate | Rate | Basis |
|---|---|---|
| Official national poverty line (Cost of Basic Needs) | 28.8% | FY2024–25, Poverty Estimation Committee; up from 21.9% in 2018–19 |
| World Bank PERA projection | 25.3% | FY2023–24, national line |
| PIDE projection | 22.9% | FY2025, Cost of Basic Needs method |
| World Bank, $4.20/day (2021 PPP) | 44.7% | Lower-middle-income international line; ~47% cited for 2025 |
| World Bank, $3.00/day (2021 PPP) | 16.5% | Extreme poverty; was 4.9% at the old $2.15 line |
The headline "nearly half of Pakistan is poor" comes from the fourth row, and it is a legitimate figure — but it is not primarily news about 2025. The World Bank revised its international poverty lines to 2021 purchasing-power parity, mechanically raising measured poverty across all lower-middle-income countries. PIDE's assessment is that 82% of the apparent jump is the line revision and only 18% is real deterioration.
That correction matters, and it should not become a comfort. Every method, including the government's own, shows poverty rising — the official rate is up nearly seven percentage points in six years. Deterioration is not in dispute; only its magnitude.
Separately, an IPC assessment covering December 2025 to September 2026 found 7.5 million people in crisis-level acute food insecurity or worse — but note the scope carefully: that analysis covers only 45 vulnerable rural districts of Balochistan, Sindh and Khyber Pakhtunkhwa, home to about 35.6 million people. It is not a national figure, and it is frequently misreported as one.
Aggregates describe the shape of the pressure. Individual accounts describe the mechanism.
Zubaida Bibi, a widow renting in Bilal Colony, Taxila, pawned her thirty-year-old engagement ring with a neighbour and borrowed the rest to pay one month's electricity bill. She hoped to recover the ring when her pension was released.
Reported in Dawn, September 2023
Ahmed Zaman, 40, sole earner for a family of six in Karachi, earns Rs 150,000 a month — comfortably above average. He pays Rs 35,000 in rent, and one electricity bill exceeded Rs 40,000: half his salary on two line items. "This month, I asked my wife for some help and she sold her gold earrings which she got at our marriage."
Reported in Dawn, September 2023
Hussain Shabbir, 25, pools income with his father and brother for a combined Rs 130,000 a month. The family sold its car to clear debts. "I return money to one person and immediately find myself in need to borrow more from another for the next month."
Reported in Dawn, September 2023
Ehsan Ali sells poultry in Taxila. His sales fell by around 60%. Customers who used to buy chicken now buy pota, kaleji and feet — the offcuts. He survives on orders from hotels and marriage halls.
Reported in Dawn, September 2023
The coping mechanisms follow a recognisable sequence: cut meat, then milk, then pulses; run a tab with the shopkeeper; borrow from a committee or a relative; pawn or sell gold; sell the vehicle; move a child from private school to a seminary, where fees are nominal — one Faisal Road seminary administrator reported enrolment doubling in two months as middle-class parents arrived; withdraw the child from school entirely; skip the doctor and the prescription.
At the far end of that sequence the cost stops being financial. Suicide is now the fourth leading cause of death among Pakistani youth, per the Pakistan Institute of Living and Learning. In Tharparkar district, local reporting attributed 277 of 496 suicides between 2022 and 2024 to unemployment. And in the brick kilns around Lahore, The Guardian documented Shafeeq Masih, bonded by a Rs 900,000 debt to the kiln owner that "however hard he worked, just kept growing," selling a kidney for Rs 400,000 to a broker who came to the kiln. It did not clear the debt. "Whatever they put in writing, we can't question that. They see us as slaves."
Pakistan's flagship safety net, the Benazir Income Support Programme, pays Rs 13,500 per quarter under Kafaalat — Rs 14,500 for enhanced categories. That is roughly Rs 4,500 a month, covering perhaps 15–20% of a family's minimum food bill. It reaches around 9 million households on a budget above Rs 700bn, rising to Rs 838bn and 12 million families in FY2027. It is real money that keeps millions from destitution, and it is not remotely calibrated to the cost of living it is meant to offset.
Health and education show the same shortfall. Government health spending is under 3.4% of GDP; households pay roughly 53% of all health expenditure out of pocket, which is how a single hospital admission becomes a family's financial catastrophe. Enrolment in the Sehat Sahulat card reduces out-of-pocket costs by 19.9% against a 60% design target. And 25.37 million children aged 5–16 are out of school — 28% of the age group, at 45% in Balochistan and 39% in Sindh. Of those who are in school, 77% of ten-year-olds cannot read and understand a simple passage.
Pakistan's demographic dividend has a deadline, and the country is spending it.
The 2023 census counted 240.5 million Pakistanis growing at 2.55% a year — the fastest in South Asia. 40.6% are under 15. Some 105.7 million, nearly 44% of the country, are between 5 and 24. On current trajectory the population passes 350 million by 2050.
The labour force reached 85.6 million in the 2024–25 Labour Force Survey, up from 71.8 million four years earlier: about 3.5 million new entrants a year against roughly 2.5 to 3.1 million jobs created. Unemployment rose from 6.3% to 7.1%, adding 1.39 million people to a total of 5.9 million unemployed.
By age, the burden is unambiguous: 12.9% for 15–24 year olds, the highest of any cohort, and 8.3% for 25–34. Women face 10.5% against men's 6.0%, peaking at 14.9% for women aged 25–34. Khyber Pakhtunkhwa carries the worst provincial rate at 9.6%, ahead of Punjab's 7.3%.
Every functioning labour market rewards education with better employment prospects. Pakistan's inverts the relationship.
An illiterate Pakistani has a 4.4% chance of being unemployed. A postgraduate has 11.7%. A woman with a university degree has 23.9% — she is more than five times likelier to be jobless than a man who never went to school. Historical work by PIDE tracked graduate unemployment climbing from 9.2% in 2001–02 to 17.9% by 2014–15, with 2020–21 figures by discipline that read as an indictment of enrolment planning: agriculture 29.4%, engineering 23.5%, computer science 22.6%. Female computer science graduates faced 51.5% unemployment; female engineers 42.6%.
Five mechanisms produce this:
Unemployment statistics only count people looking for work. Pakistan's central labour fact is how many women never enter the count. Female labour force participation is 25–26% under the current standard — around 21% on the older one — among the lowest in the world, and just 11% in urban areas. The World Economic Forum ranked Pakistan 143rd of 146 countries on economic participation and opportunity in 2023.
Nationally, 28.4% of young people are not in education, employment or training: 36.0% in Balochistan, 32.7% in KP, 32.0% in Sindh, 24.8% in Punjab. That population is overwhelmingly female. The barriers are concrete rather than abstract — no safe transport, unpaid care work falling entirely on women, household decisions about employment made by men, and weakly enforced workplace harassment law. This is also the largest single reserve of unused economic capacity Pakistan has: World Bank and IMF modelling suggests closing the participation gap could add 12–30% to GDP over fifteen to twenty years.
Pakistan's survey counts you as employed if you worked one hour in the reference week. A man who spent an afternoon loading trucks and then nothing for six days is employed. So is an unpaid worker in a family shop. With 80.4% of employment informal, a third in agriculture, 39.8% working 50 hours or more a week and officially measured underemployment at an implausible 1.6%, the headline rate measures the wrong thing.
Both the World Bank and PIDE say so explicitly: the emphasis on "at least one hour" understates underemployment and low-quality work. The real question in Pakistan is not who has a job — it is who has a job that pays enough to live on. Average monthly income is Rs 39,042, below the new minimum wage. Even the widely cited gig economy is mostly not what it sounds like: online platform work is 2.9% of employment, and 97.1% of that is physical — riders and delivery — with only 2.9% genuinely digital.
When domestic opportunity closes, Pakistanis leave. The Bureau of Emigration's registered departures for work tell the story precisely:
Three-quarters of a million people a year, registered — and the real number is higher, because these figures exclude anyone who left on a visit visa and found work, anyone undocumented, students, and most professionals who emigrated to Western countries outside the recruitment-agent system. More than 14.2 million have gone since 1971.
In 2025, 69.5% went to Saudi Arabia, followed by Qatar, the UAE and Bahrain. The skill composition is the part worth sitting with: 61.1% unskilled, 29.1% skilled, and only 2.4% classified as highly educated. Within that small slice: 3,795 doctors, 5,946 engineers, 5,659 accountants and 1,640 nurses in a single year. The largest single occupation was drivers — 163,718 of them.
Pakistan is simultaneously exporting its cheapest labour and its scarcest professionals, and importing back $41.6bn that keeps the currency standing.
For those without the qualifications or the agent fee, there is the irregular route, and it kills. Roughly 300 Pakistanis were among the more than 700 people who drowned when the Adriana went down off Pylos, Greece in June 2023. In January 2025, on a newly opened route from Mauritania toward the Canaries, 43 Pakistanis died of 86 aboard. Smugglers charge Rs 3.5–4.5 million per person, typically raised by selling land or livestock, and the traffic concentrates in Gujrat, Sialkot and Mandi Bahauddin. Official recognition of the risk is quiet but telling: the state's death grant for emigrant workers was raised from Rs 800,000 to Rs 1 million in 2025, and 71 recruiter licences were cancelled.
The sentiment data resists both despair and triumphalism. An Ipsos survey for VOA in February 2024 found 74% of young Pakistanis want to stay in Pakistan, with 23% considering leaving — down from a 32% peak in October 2022. Emigration intent rises with income: it is disproportionately the better-off and better-educated who plan to go, which is precisely the composition that damages the country most. Sixty percent said political leaders do not understand their concerns; half thought elections would not change their daily lives.
Then a genuine turn. Gallup's World Poll for 2025, published February 2026, found 25% of Pakistanis "thriving" against 19% "suffering" — the first time thriving has exceeded suffering since 2018. In 2024 only 11% were thriving, a figure that tied the global record low alongside Lebanon and Afghanistan. The share saying living standards were improving more than doubled to 31%. Separately, 51% expected 2026 to be better than 2025, against a 37% global average.
Stabilisation registered. People noticed the currency holding and the shops staying open. That is real, and it is fragile: 71% still see corruption as widespread, 61% disapprove of national leadership, and the improvement is measured from a floor that was among the lowest ever recorded anywhere.
Pakistan's youth are marketed a great deal of hope. Some of it is a business model.
The honest framing is this: no single sector will absorb 3.5 million entrants a year. But several will absorb hundreds of thousands, and they share one property — they reward specific, certified, verifiable skill rather than a general credential. Below, each opportunity is assessed on realistic scale, realistic earnings, and real barriers.
ICT export earnings grew from $2.11bn in FY2021 to $3.81bn in FY2025, and were running 19.7% ahead through the first nine months of FY2026 — a full year near $4.5bn. A 16% compound rate. Some 34,420 IT and ITeS firms are registered; in FY2024 new IT companies were 15% of all new business registrations in Pakistan.
Earnings: junior developer Rs 60,000–120,000/month; QA Rs 45,000–80,000; IT support Rs 35,000–60,000.
Honest scale: perhaps 150,000–250,000 export-generating workers. Real, well-paid, and not a mass solution.
Barriers: internet shutdowns cost the economy $892m–$1.62bn in 2024 across 9,735 hours of disruption; X has been blocked since February 2024; PayPal still does not process outbound payments; and the government's $15bn export target has no credible roadmap — at the current 16% growth rate it arrives around 2034.
Pakistan ranks fourth globally in freelancer supply. Some 2.37 million are registered. Freelance export earnings hit $856m in nine months of FY2026, up 51%, and will cross $1bn — one in every four ICT export dollars. Tax treatment is genuinely favourable: 0.25% on remitted foreign income, with up to 50% retainable in a dollar account.
Honest distribution: the median active freelancer earns roughly $200–400 a month. Perhaps 100,000–200,000 clear $500/month. Divide $856m by 2.37 million registrations and the average is trivial — earnings are heavily concentrated among skilled specialists, mostly developers and designers.
What works: a real, deep technical skill plus English plus a client-facing portfolio. What does not: a certificate.
The most aggressively marketed "opportunity" in Pakistan is also the least likely to pay. Amazon FBA requires $2,000–10,000 in upfront inventory and a US business entity most Pakistanis cannot obtain cleanly. Shopify dropshipping has sub-5% success rates globally. Yet courses sell for Rs 15,000 to Rs 300,000, and no private seller publishes audited placement or income data.
Documented fraud: a fake entity calling itself "NITSEP" impersonated the IT Ministry and collected Rs 5,400 annual fees from thousands. In October 2025 the NCCIA broke up a training-scam network in Multan, arresting eight (FIR 252/2025 under PECA). In July 2025 the national CERT warned of "honey trap" freelancing scams extorting Rs 1–1.5 million from young people in Punjab.
Rule of thumb: if the pitch leads with the income and not the skill, it is selling the course, not the outcome.
The largest functioning channel by volume: 762,499 placements in 2025. New capacity is opening — Italy has granted a 10,500-worker three-year quota for nursing, caregiving and agriculture; Qatar resumed Pakistani visas after a 19-year gap; South Korea's EPS and Japan's schemes are active; Germany remains under negotiation, not operational.
The bottleneck is not skill. It is language. Germany wants B1/B2 German, Japan JLPT N4, Korea TOPIK 2, Italy Italian. Almost no Pakistani institution offers a combined trade-plus-language pathway. The Philippines built exactly that and captured the global nursing and caregiving market.
Costs and cautions: agent fees of Rs 150,000–400,000, usually financed by family debt, taking 18–36 months of Gulf wages to recover. Verify any recruiter against the BEOE licence list — 71 licences were cancelled in 2025 alone.
The most underrated opportunity in Pakistan, and it happened by accident. As grid tariffs more than doubled and Chinese panel prices collapsed over 40%, Pakistanis went off-grid en masse: solar imports tripled toward $2.1bn in 2024, and by 2025 solar was supplying about 25% of Pakistan's electricity — its single largest source.
Somebody has to mount, wire, and maintain all of it. Earnings: Rs 50,000–150,000/month for an experienced installer. Entry: an electrical apprenticeship plus one to two weeks of solar-specific training. No degree, no English, no visa.
Risk: net-metering and time-of-use policy is unsettled, and every household that exits the grid leaves fewer consumers to carry capacity charges.
Still Pakistan's export backbone. EU imports from Pakistan reached €8.3bn in 2024, of which €7.5bn entered under GSP+ at a 95.1% utilisation rate — €732m in tariff savings in one year, making Pakistan the scheme's largest beneficiary.
The risk is not commercial, it is political. GSP+ rules are revised for 2027, and the EU's July 2026 assessment flagged enforced disappearances, restrictions on expression, and weak enforcement against forced and child labour. A safeguard has already suspended GSP+ treatment for non-fuel ethanol through June 2027. Full suspension is a tail risk, but it would hit millions of textile jobs directly.
The headline numbers are large. The PM Youth Business and Agriculture Loan scheme lends at 0% up to Rs 500,000, 5% to Rs 1.5m, 7% to Rs 7.5m, and disbursed Rs 209bn to 31,700 entrepreneurs in eight months to March 2025. DigiSkills reports 5.14 million "trainings" in nine months of FY2026.
Read the fine print. A National Bank audit found that Rs 215m disbursed in early 2024 was largely willfully defaulted, with no creditworthiness assessment performed; management claimed 80% recovery and produced no evidence when auditors asked twice. "Trainings" are not trainees — the figure counts repeat enrolments. DigiSkills' claim that its trainees earned $1.65bn is self-reported and unaudited. The e-Rozgaar network has 80-plus centres and about 4,600 active users nationwide.
Free training is worth taking. Treat the announced outcomes as marketing until independently audited.
Anyone advising a young Pakistani to "raise a round" is four years out of date. Funding went from roughly $350m in 2021 and $333m in 2022 to $75.8m across 39 deals in 2023, then $22.5m across just 15 deals in 2024. Unique investors fell from 209 to 97. Airlift raised about $85m and shut down in August 2022 — the cautionary case for dollar-cost, rupee-revenue business models.
The deeper problem is domestic credit. SMEs receive only 5–7% of bank advances, because banks earn risk-free double-digit returns lending to the government instead. The debt trap in Part One is also a small-business financing crisis. Microfinance — Akhuwat, Kashf, Khushhali — fills part of the gap at small ticket sizes.
The gap is arithmetic. Pakistan's technical and vocational system recorded 449,112 enrolments and 241,142 graduates in 2024 — against roughly 3.5 million annual labour market entrants. That is about 13% coverage at enrolment and a 54% completion rate. NAVTTC trained over 100,000 in FY2025 and conducted 366,157 skill assessments for overseas employment. Meanwhile broadband reached 161 million subscribers at 64.2% penetration, and Raast has made instant digital payments universal infrastructure. The pipes exist. The training does not fill them.
Across rigorous evaluations in Pakistan and comparable economies, the pattern is consistent. What works: job-matching and placement services, which outperform training-only programmes; apprenticeships embedded with real employers for six to eighteen months; and language certification combined with a certified trade for overseas placement — the highest-return pathway available to a young Pakistani today.
What reliably does not work: generic one-to-three-month skills courses with no employer linkage, and untargeted credit handed to young people without business development support. Both describe Pakistan's flagship schemes. Notably, no published randomised evaluation of a Pakistani flagship youth employment programme exists — the ILO was engaged to design one and it remained at preparatory stage.
The measures economists and institutions converge on are unglamorous and well known:
Pakistan in August 2026 is a country that has successfully avoided the disaster it was heading toward in 2023, and has not yet begun to solve the problem underneath it. The currency is stable, reserves are rebuilt, the IMF is satisfied, and manufacturing is growing again. These are real achievements and they were not inevitable.
But the mechanism of that stabilisation was a transfer. Household consumption absorbed the adjustment: through the permanent repricing of the cost of living, through electricity tariffs at cost recovery, through a fuel levy that taxes every journey. The result is a country where GDP grows 3.7% while the average urban household is 19% poorer than it was six years ago, where beef consumption falls 42%, and where a woman with a university degree is five times likelier to be unemployed than a man who never attended school.
The binding constraint is now visible and it is not mysterious. Two-thirds of federal revenue services debt because the tax base excludes agricultural income, under-taxes property and cannot document retail. That leaves nothing for the schools that would fix the 77% of ten-year-olds who cannot read, the training that would reach more than 13% of labour market entrants, or the industrial electricity price that would let manufacturers hire. Meanwhile 3.5 million young people enter the labour market each year into an economy that creates fewer jobs than that, and three-quarters of a million leave — the unskilled to Gulf construction sites, and, in smaller but more costly numbers, the doctors, engineers and nurses the country trained at public expense.
What genuine opportunity exists rewards the same thing everywhere it appears: a specific, certifiable, demonstrable skill. Not a degree, not a certificate, not a course bought from a WhatsApp advertisement. A software developer who can actually build, an installer who can wire a rooftop array, a nurse who has passed an Italian language exam — each of these has a real path to a real income. The tragedy of Pakistan's youth policy is that this is knowable, cheap to act on, and mostly not being done.
The Gallup finding — more Pakistanis thriving than suffering for the first time since 2018 — is the most hopeful data point in this report, and the most precarious. It is what stabilisation feels like from inside a household: the shops stay open, the currency holds, the panic subsides. Whether it becomes a trend depends on something Pakistan has not yet attempted at scale, which is taxing those with the ability to pay in order to educate and employ those who currently cannot.