India exports more software than clothes. How did that happen, and does it matter?

India earns close to $200 billion a year from software and back-office work, several times what it earns selling the world clothes. India also has hundreds of millions of people looking for work that needs no degree, that is the wrong way round.

The two trades need different things from the state. A garment exporter needs a port that loads on time, a customs desk that clears his cloth, land he can build a shed on, and labour rules that let him hire for a season. A software exporter needs a telephone line and a contract. After 1991, India’s telephone and data networks improved beyond recognition, but its roads, ports, and customs desks did not. Its market for skilled service workers stayed loose, while regulation and unions kept formal factory labour bound tight. India specialised where the friction was lowest. Chile, Malaysia, Mexico, and Thailand built better ports and customs desks, so their advantage formed in goods instead.

That outcome has not served the large number of relatively unskilled workers well. When forecasters projected India’s software industry would reach two million jobs, China’s export factories already employed over fifty million people. Software and back-office work today employ about five and a half million Indians directly. That is a genuine success, and it is a rounding error against a workforce of several hundred million.

The 2025-26 Economic Survey presses a harder version of the charge. Services exports, it argues, are “firm-selective rather than system-forcing.” They “do not require ports to function efficiently, customs to clear on time, industrial land to be assembled at scale, or municipal services to work reliably.” Software firms therefore succeed inside a weak state without ever forcing it to improve, since they “adapt around institutional failure instead of confronting it,” and the foreign exchange they earn eases the pressure for reform. The Survey names its own best evidence for this claim: India.

India did not choose software over shirts. Its ports, its customs desks, and its land records chose for it. Remove the factors that shackled industry, and India would not have to pick between backing services and backing factories.