Pakistan's IT exports are capped by a shortage of engineers

The promise every government repeats

Every new government in Pakistan, Muzamil says, announces an IT export target. Ten billion dollars. Then 30 billion. Then 50 billion. For about ten years, he says, the slogan has stayed a slogan. He spent three years interviewing the CEOs and founders of Pakistan’s largest IT companies on Thought Behind Things to understand why.

A simple comparison makes 20 billion dollars sound easy. India exports about 200 to 250 billion dollars in IT services, and its economy is about ten times Pakistan’s. “My thesis is that we cannot,” he says. This episode is his case.

Three inputs, and only one that matters

Muzamil splits the IT industry into three inputs. The first is human capital: engineers, data analysts and other skilled people. He calls them the raw material. The second is power, the electricity that runs laptops and internet. The third is hardware.

Power and hardware, he argues, are small and stable costs. A laptop is a one-time purchase. Electricity is a small share of the total. That leaves human capital as the input that decides how big the industry can get.

The arithmetic of 20 billion dollars

Next he shows how an average software house makes money. By the global standard he describes, a firm bills about 3,000 dollars a month for each person. An American company would pay 10,000 to 15,000 dollars to hire locally, so 3,000 dollars for a remote engineer in Pakistan, India or the Philippines makes sense. Of that 3,000 dollars, on average, about 1,000 goes to the engineer. About 1,000 covers costs like laptops, rent and electricity. The last 1,000 is profit or reinvestment.

So one engineer brings in about 36,000 dollars a year. Divide the targets by that, and Pakistan needs about 280,000 engineers to reach 10 billion dollars and about 560,000 to reach 20 billion.

Then he looks at supply. He cites a 2022 report by the Pakistan Software Houses Association on the gap between industry and academia. By his reading of it, Pakistan produces about 25,000 IT graduates a year, and only about 10 percent are employable straight away. That is about 2,500 people a year who can step into a job billed at 3,000 dollars. The rest do not vanish. Some leave the field, and others build their skills slowly through courses and internships. But at 2,500 a year, he says, building 280,000 engineers would take about 110 years.

“It’s a supply constrained industry,” he says. He compares it to textiles. You can sell to Zara and H&M all you like, but without cotton, you have nothing to ship.

Churn, taxes and the offshore workaround

A shortage has side effects. Muzamil describes firms winning projects and then finding no one to hire. So they poach from each other. A new hire spends a few months at one company, gets trained, and is pulled away by another. That churn, he argues, makes the whole supply chain inefficient. It also explains a puzzle. The rupee fell from about 160 to about 280 to the dollar, which should have made Pakistani IT much more competitive, and IT services can respond to that quickly. He says the expected growth did not show up because supply could not expand.

Then came higher taxes on salaried workers. IT services, he stresses, run on arbitrage: the lower your local cost, the better you can compete with India, the Philippines and Egypt. In his example, if an engineer on 300,000 rupees a month loses about 50,000 to tax, the firm has to pay about 350,000 to keep the same person. He contrasts this with textiles, which get cheaper gas and electricity and zero-rated inputs, while IT’s main input is now taxed.

He sees two more effects. First, freelancers who earn from abroad are paid as remittances, which are not taxed, so the formal industry now competes with scattered freelancers on unequal terms. Second, he says he has heard that small and mid-sized firms are setting up holding companies in Dubai, Singapore or Delaware. They invoice clients from there, keep profits abroad, and now run payroll from abroad too, so staff receive tax-free income. In his example, an employee earning 500,000 rupees a month could save about 120,000 rupees in tax this way.

He doubts the government can stop this. Taxing remittances, he argues, would push the 30 to 35 billion dollars Pakistan receives each year into hawala and hundi. He expects tax collection to suffer and reported IT exports to look lower as money moves through other channels.

Why India’s path is closing

The India comparison, Muzamil says, hides a key difference. India’s early IT growth came largely from business process outsourcing: call centres, data entry and other work that needs a few days of training. That let India put large numbers of people to work quickly.

He believes AI will close that door for Pakistan. It will start at the bottom of the food chain, with basic design and call centre work, and move up into coding skills that many now see as advanced. He points to a recent comment by the CEO of Nvidia that children should stop learning to code because the future is problem solving.

Here, he says, Pakistan’s education system fails. Universities, boot camps and students all lean on rote learning. He recounts asking the CEO of Arbisoft, one of Pakistan’s top software firms, what skill they most look for in graduates. The answer was problem solving, and Muzamil says the CEO told him they struggle to find it even at top universities. He adds that computer science degrees in Pakistan teach a one-track mind. In his view, the coming world needs people who pair technical understanding with domain knowledge in law, medicine or design.

Whose job is talent

When he talks to the government, Muzamil says, it blames the private sector for parking dollars abroad. When he talks to the private sector, it says building talent is the government’s job. He has little patience for the second answer. “If I am a textile baron and my cotton is not coming,” he says, “I will buy farms and start growing cotton.”

His proposal is for firms to form a consortium and fund talent building together. He anticipates the objection that a firm trains 100 people and keeps only 40. The other 60, he argues, protect those 40, because once the industry’s demand is met, people stop being poached. He also criticises founders who stop at a few hundred employees and feel like kings. “You are sitting as the big landlord of a small village,” he says, in a small ecosystem that barely registers globally.

More supply, in his view, would also lower costs. If engineers were available at 200,000 rupees a month, Pakistani firms could sell at 2,500 dollars and undercut India, Egypt and the Philippines.

A platform to educate at scale

Muzamil does not think more universities are the answer. They are expensive and slow, and he says good teachers do not want to move to remote campuses. He describes a visit to FAST, where he studied. He says the school told him its good instructors were leaving for private sector jobs that pay eight to nine times more.

His answer is a scalable, largely automated digital platform that brings university-level education to the mass market. A world-class instructor’s recorded course, he argues, costs little per student when 100,000 people watch it instead of 100. By his own study, which he invites viewers to check, about 3.5 million Pakistanis turn 18 each year, a number heading toward 5 million, while higher education has only about 350,000 seats.

He calls building that platform his personal life goal. He says he has researched it for three years and has a plan, but not yet the money to build it.