India's IT Tale: Part II
Asynchronous Developmentalism and the form of Indian development.
2026-09-10(Thu) 21:00 AST
The State, or not the State
India has had many relatively successful industries globally. India makes generic pharmaceuticals for the world; the Indian space program is remarkable for its budget, but probably the biggest Indian crown jewel is the IT field.
I talk at length about the history of that field in the previous article, and Jon of Asianometry has made his own video on this topic, but my discussion for this article is who gets to claim the credit for the success of Indian software. On one hand, the relative consensus is that the success belongs to the private sector; a view shared by anyone from Indian governmental ministers to random Indians you talk to on the Internet, and in-between the general class of journalists. On the other hand, there is a revisionist strain, promoted by academics like Sharma, Balakrishnan, and Kapur, each with their own view that the state played a relatively major role in the development of the Software sector; Sharma's view of the state as venture capital, Balakrishnan's view of “strategic intent,” and Kapur's view of “accidental industrial policy.”
The private view comes from the fact that, relative to other industries, India did indeed step out of the software field and had a relatively laid back approach at face value. There is a similar tech story to Indian software; the hardware. That had far more state involvement, and yet far more failure: Fairchild Semiconductor's offer to build a chip plant in 1963 was rebuffed as the state wanted the plant to produce less than what Robert Noyce proposed; Homi Bhabha's attempted collaboration with the Control Data Corporation for a memory manufacturing facility died when Bhabha did, with the CDC not having trust in the rest of the Indian government to continue the deal without red tape; Delhi Cloth Mills’ (yes, a textile company) agreement with Sony to produce calculators for the company was rejected by the state in the 70s when it refused to give DCM a licence; the failures can go all the way to even the modern era, with the government's failure to appease AMD's concerns about going ahead with a three billion dollar fab in Andhra Pradesh in the late 2000s. This doesn't inspire confidence in the Indian government being able to coordinate much of anything capital intensive, and the few overt governmental moves in the software field were also obstructive; the delay of Tata Consultancy Services in actually getting a licence at all, for one example.
But to say the Indian state had no role in coordinating the software sector's rise would be just as misleading; both passively and actively it did so. Passively, the usage of English as the language of instruction, entrenched by the 1960s, and the contemporarily-maligned “overinvestment” in technical and engineering education resulted in a large, talented, Anglophone talent pool. More actively, the centre of Indian software (Bangalore) benefitted from agglomeration effects; Bangalore was built in an almost planned manner to host some of Indian science's biggest institutions, such as ISRO, HAL, IISc, and DRDO's Electronics and Radar Development Establishment. As also discussed in detail in my previous article, the state had a rather disciplinary role: companies were required to export software worth twice the value of an imported computer, a form of export discipline most known for being applied in South Korea; the state electronics corporation UPTRON’s equity in Hindustan Computers Limited; Karnataka providing credit to Infosys, and the Department of Electronics’ fight with the Department of Telecommunications, resulting in the former illegally importing French telecom equipment in order to undercut the expensive high speed links offered by the latter's telecom monopoly.
Clearly, it's a mixed bag. I personally lean towards the revisionist theory, but the form and shape of this state involvement is rather…unique, or remarkable, and it extends if you go beyond IT. We'll need to do some theory here…
Asynchronous Developmentalism
Sociologist Peter Evans, in his book Embedded Autonomy, claims that in economic development, a state can play four roles: it can be a custodian, a midwife, a demiurge, or a husbandman. A custodian state is the state as policing; it regulates and mostly leaves the sector alone. A midwife state is the state as assisting in birth; while still leaving the sector alone, it uses subsidies, tariffs, capital or credit to encourage the sector to develop itself. A demiurge state is the state as producer; the state gets involved in the sector with SOEs. A husbandman state is the state as nurturer; often the evolution of a successful midwife state that goes from assisting in creating new industries to the nurturing of and partnering with existing industries, for example in R&D spending.
Evans’ book is, ironically, directly about the informatics/IT industries of a few selected countries, India being one of them. He lambasts India as having an “incoherent” industrial policy in the IT sector; the Indian state attempted to play demiurge and custodian, doing so poorly and absorbing already scarce state capacity, putting it in a poor position for the husbandry it was starting to attempt when the book came out. In his ranking of the likely successes of the studied countries, Evans puts Brazil and India in the same tier. Embedded Autonomy came out in 1995. Since that time, the Indian IT sector exploded in growth, with a golden age in the 2000s and the 2010s, now making up a significant portion of the Indian economy; Brazil’s, on the other hand, is a footnote.
I would now like to introduce something to complete Evans’ theory and his view of Indian industrial policy’s incoherence: I believe the development of the Indian IT sector was a matter of state, but it was a matter of a state that rather unawarely implemented Asynchronous Developmentalism.
An Asynchronous Developmentalist state is a state that performs all four roles of a developmentalist state, but it does so via single “nodes” that are unaware of each other, each role done separately over a span of decades, with no aggregating or synchronising measure; if you look at each one it seems like nothing, but looking at the wider picture it becomes somewhat, unintentionally, coherent. I borrow the term “node” from Indian bureaucratic jargon; India officially designates certain agencies, departments, ministries etc as the “nodal agency” for an entire sector or field; the Department of Electronics was the nodal agency for the electronics sector, which included IT.
The fact that IT development could have happened in a single node, and not require much coordination, made it rather successful. The DoE's war with the DoT's monopoly could not have happened in a state with a more “coherent” industrial policy; were India to have Korean rigidity, it's likely the DoT and VSNL would've kept choking Indian firms with high telecom prices, and that could've significantly delayed the rise of the sector. The problem is that in a field of technology like hardware instead of software, which requires far more coordination and capital intensity, this Asynchronous Developmentalism fails. The Centre (the Indian English term for the Government of India) was rather weak in coordinating capability, and the States were fiscally empty; Karnataka's role with Infosys and Uttar Pradesh's indirect role with HCL would read as fiscal prudence to a Chinese province willing to buy a failed auto plant to turn it into an electric battery one, and those examples were seen as unusually involved by the standards of the States already.
Extension
When you have this theory, you will realise that it explains not just IT but the entire modus operandi of Indian development, and why India is world class in some sectors and rather deficient in others. Every sector in which India is good at required only a single “node” or equivalent which can make do with scarce state capacity, and often things that can be done with scarce state capacity and only a single node are things that played to the strength of India's sheer labour; in other words, labour intensive industries.
IT is labour intensive rather than capital intensive. But so are some other fields: generic pharmaceuticals in an Indian context are labour intensive, you don't really need capital when you have a massive pool of chemistry PhDs to reverse engineer the synthesis processes of drugs, which is exactly what the field began doing in 1970 when the Patents Act made only process patents, and not product patents, legal. Space in an Indian context? Also labour intensive; coordination could be done through a single node, ISRO, that rather than having to go through Indian bureaucracy only had to report to the PM, the way the DoE was also structured. This allowed ISRO to use, again, the massive labour pool of talent available to become a major and global space program despite a shoestring budget. These major, pre-2010s successes all shared the same few traits.
From Labour to Capital
But to say that this same model applies to modern, and especially post-2014 India, wouldn't be an accurate thing to say. I think one of the more underappreciated moves has been the centralisation around the, well, Centre.
Ever since the election of the BJP the Government has found itself with more coordinating capacity; interagency fights like that of the DoE vs the DoT don't happen anymore, and there has been fiscal centralisation thanks to the GST in 2017, which replaced the previous thirty odd tax structures. You also have the creation of the Insolvency and Bankruptcy Code, which played a rather important role in steel development, and PM Gati Shakti, which serves as, quite literally, a “coordination mechanism.”
This effort is starting to show in more coordination-heavy sectors that often have more capital intensity. India now has three semiconductor plants operating, with another one being built; all supported by the Semiconductor Mission. The IBC recycled distressed plants that were on bad loans into the hands of Tata, JSW, and ArcelorMittal-Nippon, as well as steel having consistent demand from the state in the form of constant infrastructure investment, ultimately resulting in a doubling of steel production between 2014 and now. Infrastructure itself being another capital intensive sector, with rail electrification at scale, freight corridors, and a doubling of the pace of highway construction. And then you have the rise of private/commercial spaceflight, green hydrogen, batteries, solar modules; entirely greenfield sectors that have had some heavy husbandry and midwifery from the state.
The record of this hasn't been all positive. Production Linked Incentives, or PLIs, have been the organising factor of plenty of this, and it has had some failures in the form of thin disbursements and outright lapses. The previously mentioned Semiconductor Mission’s successes came after Vedanta-Foxconn's spectacular collapse. Not to mention the consequences of a more centralised state; again, a state with Korean rigidity wouldn't have allowed the DoE to wage its Satcom war. But, overall, India's industrial policy has been remarkably improved as of late. We'll only know for sure whether this more synchronous developmentalist approach will work in decades’ time, but it has turned out better than the previous approach overall.
This article is part II of a two-part series on India’s IT development and asynchronous developmentalism. This will likely be my last article for a while. Read Part I on my website here.