Why Pakistan built the power and still cannot deliver it

The question under all the other questions

Muzamil opens this one with a thesis, not a guest. When Pakistanis argue about the economy, he says, they argue about exports and inflation and the price of flour. He thinks they are starting in the wrong place. The thing that decides whether industry can run, whether services can be delivered, whether exports can ever be competitive, is the cost and reliability of energy. “Energy ki cost define karti hai ke aapki exports competitive hongi ya nahi hongi,” he says — and for as long as he can remember the country has treated it as an afterthought.

His evidence is a single comparison he keeps returning to all episode. In Dubai, where everything is expensive and the power is entirely thermal, a unit of electricity costs him around 27 to 29 rupees. In Pakistan, the same unit costs him roughly 55 — and Pakistan has cheap hydro and cheap nuclear in its mix, which Dubai does not. The numbers don’t reconcile. To find out why, he brings in someone who built the machinery itself: Naeem Shafique, one of the first Pakistani engineers to join General Electric, who set up GE’s power practices across the Middle East and Central Asia before running its power business in Pakistan, all the way to his retirement in 2020.

A 400-ton rotor and a fresh graduate at Tarbela

Before the policy, the apprenticeship. Shafique grew up moving around the country — his father was a civil engineer contracted to the armed forces — and did his mechanical engineering at UET Peshawar, graduating in 1981. He had an MS admission lined up in the US but a year to fill, so he went to visit cousins at the Tarbela dam, where a Canadian GE subsidiary was building the power plant. He asked to see it. He spent a day there and talked his way into an internship, telling a sceptical Canadian plant manager he wasn’t there for a job or for money, only to learn.

The manager’s first line stuck with him: “We are here to do a project in a specified time frame and make money. We are not here for a talent hunt.” Then he slid a drawing across the table and asked Shafique to read it, handed him a micrometer, a vernier caliper. Shafique read them all. He got the pass. His first salary was 2,700 rupees a month in 1981 — when the dollar was about 4.5 rupees, that was serious money, several times what his peers were earning elsewhere.

The MS dream quietly died, and he doesn’t regret it, because of what he saw instead. At a hydro plant everything is enormous: the rotor that turned was around 400 tons, arriving in four pieces to be welded and assembled to fine tolerances. His mentor, Lloyd Welcher, deliberately rotated him through the whole cycle — manufacture, assembly, installation, accuracy checks, commissioning — so that a mechanical engineer came out understanding the electrical side, the generator, the control system. By the end of 1982 the young engineer was leading the commissioning of one of the four units himself. That hardware-up education is what gives the rest of the conversation its weight.

Forty years, and a 100% market share

From Tarbela the parent company recruited him into GE proper. He flew to the GE headquarters near Albany for a thirty-four-week marketing development programme — one of four international trainees in a batch of sixteen — and came back a salesman who understood the equipment from the inside. Singapore, Japan, then back to Pakistan. He stood up GE’s office in Iran in 1992 when sanctions briefly eased, then established GE Power’s Lahore office in the mid-1990s as the first wave of independent power producers arrived.

What followed is a claim he states plainly: from 1998 onward, every combined-cycle gas-turbine power project that went up in Pakistan ran on GE equipment. “I did not lose a single gas turbine order in Pakistan.” He credits three things — staying genuinely active in the market, equipment that out-competed Siemens and Mitsubishi, and a strong service backup, because a thermal plant is designed to run thirty years and can run fifty if it’s maintained. He notes, almost in passing, that there are gas turbines in Pakistan installed in 1968 still running today, more than fifty years on, at a fertilizer plant in Engro’s facility. The point isn’t nostalgia. It’s that the hardware lasts; the system around it is what fails.

His proudest project under this run was a set of RLNG plants in the mid-2010s. The Haveli Bahadur Shah plant came in at 62.4% efficiency — at the time, he says, the most efficient power plant in the world — and proved out even higher in operation. The machinery, in other words, was world-class. So why is the electricity so expensive?

The mix, the dollar, and the engine of circular debt

Shafique’s answer starts with the energy mix. Hydro is cheap. The nuclear that’s installed is cheap. Renewables are cheap. But Pakistan leaned hard into thermal generation without asking the obvious question — where’s the fuel? — and the fuel had to be imported. Imported gas, imported coal, all of it priced in dollars. Once your generation depends on imported fuel, every shift in the country’s economics hits the bill directly.

Then the structural trap. Independent power producers, he’s careful to say, are not unusual or sinister — they exist everywhere, and capacity payments are a standard arrangement worldwide. Without them Pakistan probably wouldn’t have the generation it has. The problem is how the contracts were written: the payments were dollar-indexed. The state buys that electricity in dollars but sells it to consumers in rupees. As long as the exchange rate holds, fine. When the rupee falls, a gap opens between what’s owed to the producer and what can be collected from the consumer — and that gap is the famous circular debt.

A second driver sits underneath it: a demand curve that swings violently. Summer peak has touched 27,000 to 28,000 MW, driven mostly by air-conditioning — what he calls the cooling load. Winter demand drops to 11,000 or 12,000. That leaves something like 15,000 MW of capacity that has to be maintained and paid for year-round while generating no revenue for half the year. Capacity payments come due whether the plant runs or not. The structural mismatch between the country’s projected GDP growth — on which all the generation was planned — and the slower growth that actually materialised means the country built for demand that never fully arrived, and is now paying to keep it idle.

The captive-power argument, and the road not taken

Much of the middle of the conversation is Muzamil pressing a contrarian case, and Shafique engaging it honestly rather than deflecting. Muzamil’s frustration centres on captive power — the roughly six and a half thousand megawatts of generation that industry built for itself in the 1990s, when the grid couldn’t supply reliable, constant power and a process plant going down was unacceptable. Pharmaceutical, textile, fertilizer, any export-oriented process industry put in its own plants.

Muzamil’s challenge is sharp: by 2014, with the economy grown and cheap capital available, did Pakistan really still need to offer dollar-denominated sovereign guarantees to attract foreign investors? Couldn’t it have empowered local players instead — the kind who’d take payment in rupees and keep the money in the country? He points to Thar coal and Engro’s coal operations as the recent success story he most admires: local investors, local fuel, producing units at a fraction of the cost of furnace-oil plants and pulling real industrial growth behind them.

Shafique grants a lot of this — the 2004 policy did bring in mostly Pakistani developers, and he agrees local empowerment matters — but he won’t let one fact slide. The equipment for large power plants isn’t manufactured locally; it comes from abroad, so the capital expenditure is dollarized regardless of who the investor is. And a capital-intensive, cash-hungry industry needs security; a foreign investor weighing Pakistan against better opportunities elsewhere will want a guarantee. Where he agrees fully is the deeper failure: short-term vision. Pakistan missed Thar coal — identified in the 1960s, developed only after 2010 — the way India built out the same Rajasthan-belt coal in the sixties and seventies and now runs nearly half its generation on local coal. And it missed hydro, where every major dam site was mapped in mid-1960s water-management studies. “We missed the hydro first and then we missed the thermal coal.”

The bottleneck nobody funds

If there’s one idea Shafique wants listeners to leave with, it’s this. Pakistan put almost all its money into generation and almost none into transmission and distribution. The result is a country that now has surplus generation it cannot physically deliver.

He points to the nationwide blackout of January 2023. Pick whatever immediate trigger you like — a plant tripping, a fault — the bottom line is that the transmission and distribution system couldn’t handle the load it was asked to carry. It overloaded and collapsed. The grid in large parts is sixty years old and was never upgraded at the pace of the generation bolted onto it. The CPEC-era promise to overhaul distribution after the plants went in simply never happened, and Shafique’s diagnosis is blunt: “incompetency, incompetency, wrong decisions.”

This is also why he’s cautious about the solar enthusiasm sweeping the country. Solar and wind are variable — their output swings with the weather — and a weak grid can’t absorb variable generation without destabilising. Pakistan already curtails the renewable output it can’t manage. So the sequence matters: fix and adapt the grid first, then add renewables alongside it, hand in hand. “Everything goes hand to hand. You can’t leave one thing and do the other” — which, he says, is exactly the mistake the country has already made and is now paying for.

Privatise the last mile, and stop firefighting

On structure, Shafique is unambiguous. The distribution companies should have been privatised long ago — in the nineties or two-thousands. Theft, pilferage, line losses running as high as 40% at some discos: none of it gets solved while the last mile stays in public hands. “Discos should, must be privatised ASAP.” Generation belongs in the private sector too, because a private owner maintains the asset to protect his own revenue. Transmission can stay public, but distribution cannot.

He’s clear-eyed about K-Electric as the cautionary example — a privatisation that began well and was then, in his reading, undone by becoming far too politicised. His fix is to unbundle it: separate generation from transmission and distribution, run the generation as IPPs selling into the network. The lesson generalises. He doubts local investors alone can buy and run the loss-making discos; a global conglomerate is likely needed, and Shanghai Electric’s earlier withdrawal shows how hard that sell already is.

Muzamil closes by asking what radically lowering the bill would actually take. Shafique’s answer refuses the single silver bullet. Some costs you genuinely can’t control — dollar-priced fuel, dollar-denominated payments. But cut the losses, fix the generation mix, end the payment delinquencies and strengthen the grid, and the effect on the consumer is real and large, not one or two rupees. The condition is that these have to happen together. “We have been firefighting. We fix one thing and drop another.” On 2050 he allows himself optimism — the youth, the talent, the mineral reserves, a public that is finally waking up — while doubting the political class has learned anything at all. The awareness, he thinks, is coming from below, and one day it will force the decisions that forty years of building world-class turbines never could.