Vinay Kumar’s essay, “Is the IAS India’s Deep State?”, raises an important question about an institution that has exercised enormous influence over the functioning of the Indian state.
My perspective is shaped by a rather long association with the system. I spent 26 years in the IAS as an insider, working with political executives, colleagues in the civil services, specialists and public-sector organisations. After taking voluntary retirement, I spent more than 12 years in the private sector, where I continued to interface with the government and with individual IAS officers, but this time from the other side of the table.
I have therefore seen the system from both within and without. I have seen the IAS at its best and at its worst. I have experienced first-hand both the extraordinary influence of the service and its capacity to resist change.
So I find myself agreeing with much of Vinay’s criticism.
But I do not agree that the IAS is India’s “deep state”.
It is deeply entrenched. That, however, is not quite the same thing.
A “deep state” generally refers to unelected organisations or networks that are said to operate behind the scenes, protect particular interests and exercise power without being elected.
A permanent and influential bureaucracy is not necessarily a deep state.
The IAS exercises influence. It provides continuity. It can shape how decisions are made and implemented.
But it does not constitute an alternative centre of political authority.
That distinction matters.
The criticism is not new
Many of the deficiencies identified by Vinay are real and have been discussed for decades.
The administrative system can be excessively procedural. It can reward caution over initiative, emphasise compliance over outcomes and diffuse responsibility through layers of hierarchy, consultation and approval.
I have seen enough of government to have little difficulty accepting these criticisms.
There is also an obvious paradox in attempts to deregulate government. The machinery asked to simplify procedures and reduce permissions is often the same machinery that has evolved around those procedures and permissions.
I therefore have no hesitation in agreeing that a bureaucracy finds it inherently difficult to reform or deregulate itself.
Where I differ is in treating this as evidence of a “deep state”.
What I saw from inside
One reason I am reluctant to generalise about the IAS is that my own experience was never of a monolithic institution.
I have worked with officers who were extraordinarily capable, courageous and innovative. I have also encountered officers who were excessively cautious, defensive or simply unwilling to take responsibility.
I have seen officers use their position to push through difficult reforms. I have also seen how easily procedure can become a substitute for decision-making.
The same service can therefore produce both the reformer and the status-quo defender.
Nor is the relationship between the political executive and the bureaucracy simply one of politicians wanting change and bureaucrats resisting it. In practice, it is much more nuanced.
Also Read: We Are About to Give AI the Keys. Who Is Guarding the Door?
Political leaders need administrators for institutional memory, policy advice, coordination and implementation. Administrators, meanwhile, work under political leadership whose priorities can change with governments and ministers.
Having occupied senior positions in government, I never regarded the IAS as an autonomous power centre standing above the political executive.
The bureaucracy can influence government. It does not constitute an alternative source of democratic political authority.
An observation about politicians
There is another perspective that I find particularly revealing.
Senior political leaders across the country, belonging to very different political parties, have dealt with IAS officers for decades. As elected representatives, ministers and chief ministers, many have worked with them at the district, state and Union levels.
They have undoubtedly encountered officers they considered arrogant, obstructive or excessively cautious. They have also worked with officers who were responsive, pragmatic and willing to take difficult decisions.
Yet, in my own experience, I cannot recall a seasoned political leader from a major political party arguing that the IAS, as a system, should simply be discarded because there is a demonstrably better alternative.
That is worth thinking about.
It certainly does not mean that the present system is satisfactory. Nor does it invalidate the considerable criticism directed at it.
But those who have actually had to run governments appear to recognise something that is sometimes missing from the debate: an institution can have serious shortcomings without being dispensable.
Which brings us to the more difficult question:
If not the IAS, then what?
The real paradox
For me, this is more interesting than the “deep state” argument.
Some of the sharpest criticism of the IAS has come from IAS officers themselves.
Retired officers in particular have been remarkably candid about bureaucratic complacency, excessive hierarchy, generalist dominance, weak accountability and the reluctance to take decisions. I have myself heard and participated in many such conversations over the years.
Yet after all these years of criticism, there is still no widely accepted blueprint for fundamentally restructuring the system.
That is the paradox.
We seem to have become much better at diagnosing the IAS than at designing its replacement.
The difficult part is institutional design
Take the criticism that the IAS is too generalist.
There is a reasonable case for giving domain specialists a much larger role in areas such as health, education, infrastructure, energy and technology. But someone still has to coordinate across these domains when problems do not respect departmental boundaries.
If specialist cadres replace generalist administrators, who provides that coordination?
If outsiders are brought into senior positions, how do we ensure that expertise is accompanied by the ability to manage large public institutions and reconcile competing interests?
And if we create several specialist bureaucracies, how do we ensure that they do not develop the same institutional rigidity and entrenchment that we currently criticise in the IAS?
These are not arguments against reform. They are reminders that replacing one institutional structure with several others does not necessarily solve the underlying problem.
India needs expertise, but it also needs coordination. It needs continuity, but also accountability. It needs administrative authority, but that authority must be matched by responsibility for outcomes.
Finding that balance is the real challenge.
Lateral entry: an experiment, not a blueprint
Lateral entry was one attempt to address part of this problem by bringing people with specialised expertise and experience from outside the traditional civil-service structure into senior positions. The policy was introduced at the Centre from 2018, and the government told Parliament in March 2026 that 63 appointments had been made at Joint Secretary, Director and Deputy Secretary levels through lateral entry since then.
There was merit in the idea.
But lateral entry was never likely, by itself, to transform the administrative system. An outsider entering a powerful institution inevitably encounters its procedures, hierarchies, informal networks and institutional culture.
The experience has therefore been mixed.
That does not necessarily mean lateral entry was a mistake. It demonstrates something more fundamental:
Changing the occupants of a few positions is very different from changing the architecture of the institution.
And that is where the debate needs to go.
From criticism to design
The IAS has survived governments of very different political persuasions and successive attempts at administrative reform. Its resilience can be frustrating. But it is also one reason successive governments continue to rely upon it.
Having experienced the system both as an insider and, for more than 12 years, from outside government, I do not think the useful question is whether the IAS should be defended or condemned.
It should be neither romanticised nor demonised.
The more useful question is what an improved administrative system should look like.
Which functions genuinely require generalist administrators? Where should domain specialists have greater authority? How should senior appointments be made? How should performance be measured? How should authority and responsibility be aligned? And how do we make an administrative system capable of continuous reform without allowing the existing system to determine the limits of that reform?
These are difficult questions. But they are more consequential than the label “deep state”.
The IAS undoubtedly has serious shortcomings.
It is also deeply entrenched.
But neither proposition, by itself, makes it a deep state.
The real test of the argument is not whether we can produce a compelling critique of the IAS. We clearly can. The test is whether we can design something better—and explain convincingly how we get from the system we have to the system we want.
We have had decades of diagnosis.
Perhaps it is time for a serious conversation about design.
