From Opaque Discretion to Structured Variation

Why a Minimum Digital Kernel Matters for a Diverse State

Modern states face a hard institutional problem. They must make decisions at scale, across departments, jurisdictions, and levels of government. They must do so in ways that citizens can understand, administrators can review, and institutions can coordinate around.

In many states, that is precisely what is missing. Public power is exercised every day, but there is no shared way to represent that exercise of power across institutional boundaries. Rules may exist, facts may exist, decisions may exist, but they are often scattered, implicit, or poorly connected. As a result, decisions become difficult to interpret, compare, review, or challenge.

This problem is especially important in India. A country of India’s scale cannot be governed purely through case-by-case discretion. But nor can it be governed through rigid uniformity. It needs both diversity and coherence, both judgment and structure.

That is the context in which the idea of a Minimum Digital Kernel (MDK) becomes relevant. The kernel is not a blueprint for digitizing all of government. It is not a proposal to automate administration end-to-end. It is not a substitute for institutions, politics, or public reasoning.

It is something narrower and more foundational. A Minimum Digital Kernel is the minimal common representational infrastructure, or shared decision format, through which a state can express public decisions using three linked elements:

• rules

• facts

• decisions

A rule provides the legal, policy, or procedural basis for action. Facts provide the evidence, records, or verified inputs relevant to the case. A decision records the outcome and the basis on which it was reached.

This does not guarantee fair rules or just outcomes. But it does provide the minimum structure through which the exercise of public power can become visible, intelligible, and reviewable across the system.

The Deeper Problem Is Not Digitization

Much of the conversation around digital government starts by asking how to digitize services, automate workflows, or reduce friction. Those are useful objectives, but they miss a more basic issue.

The deeper problem is that public decisions are often hard to interpret outside the institution that made them. Rules may remain implicit in practice. Facts may be fragmented across files, departments, and databases. Decisions may be recorded without any clear account of the reasoning behind them.

When that happens, governance becomes opaque in a very specific way. It is not only slow. It is hard to understand. A citizen may not know why an entitlement was denied. A supervisor may not know why an approval was granted. An auditor may not be able to compare similar cases across offices. An appellate body may struggle to distinguish a justified exception from an arbitrary departure.

This is the gap the Minimum Digital Kernel addresses. Its value lies in giving the state a common way to represent decisions so that public action becomes more:

1. intelligible, because decisions can be explained

2. comparable, because similar cases can be examined together

3. reviewable, because the basis of action can be reconstructed

4. contestable, because decisions can be challenged on stated grounds

That is a more precise ambition than digitization, and a more consequential one.

Not the End of Judgment, but a Change in Its Terms

A common objection to rule-based governance is that public administration cannot be reduced to code. That objection is correct. Many public decisions involve interpretation, balancing, and judgment. Laws are often open-textured. Policies contain ambiguities. Real situations create exceptions. No serious state can function without discretion.

But the existence of judgment does not justify opacity. The question is not whether discretion exists. It always will. The question is whether discretion is exercised in a form that leaves a trace. A public authority may grant an exception, but the exception should be visible as an exception. A decision may involve interpretation, but the basis of that interpretation should be recorded. A caseworker may weigh competing considerations, but those considerations should not disappear into an untraceable file.

This matters most in domains where full codification is neither possible nor desirable. In fact, the kernel is often most valuable precisely where judgment cannot be eliminated, because that is where unstructured discretion is most dangerous.

This is why the goal is not to eliminate variation. It is to move from opaque discretion to structured variation.

Opaque discretion is difficult to review, easy to manipulate, and hard to contest. Structured variation allows institutions to adapt to context while still making public reasoning visible. It preserves judgment while making judgment traceable.

That is where MDK belongs. It does not replace human judgment. It requires that judgment be expressed in a form that can be understood, reviewed, and compared.

Why This Matters More in India

This question is especially important in India because India cannot, and should not, follow the historical path of relatively homogeneous nation-states.

India is a federal, multilingual, civilization-scale polity. Authority is distributed. Administrative capacity varies sharply. Social and economic realities differ across states and cities. A model of governance based on uniform administration would be both unrealistic and undesirable.

But diversity does not remove the need for coherence. In a country like India, governance cannot rely only on local discretion and informal coordination. The scale is too large, the stakes too high, and the need for interoperability too great. The challenge is therefore not to replace diversity with uniformity. It is to create a framework in which diversity can operate without collapsing into fragmentation.

That requires common protocols with plural institutional expression. This is the central idea.

Different states, cities, regulators, and departments may adopt different policies, processes, and institutional practices. But they should still be able to represent the exercise of public power in a common grammar. Decisions should still be expressible in terms of rules, facts, and outcomes. Exceptions should still be identifiable as exceptions. Appeals should still be able to reconstruct the chain of reasoning.

This is what I mean by protocolized diversity.

It is not uniform administration. It is not central control. It is the ability of diverse institutions to remain diverse while expressing their decisions in a form that is intelligible across the system. For India, this may be the more realistic path to more legible and contestable governance.

A Brief Historical Lesson

Historically, more structured forms of governance did not emerge because states discovered better administrative tools. They emerged when growing economic and political complexity made arbitrary authority too costly, and when states developed ways to standardize, remember, and review decisions across larger systems.

That historical lesson matters because state capacity has always depended partly on administrative memory. What digital systems make newly feasible is a modern form of decision memory. They make it possible to connect rules, facts, interpretations, exceptions, and outcomes across very large and diverse administrative systems in ways that were previously expensive or impractical.

That does not create accountability by itself. But it changes the conditions under which accountability, coordination, supervision, and policy learning can occur.

A Concrete Example: Building Permissions

Consider a building approval in a city.

Today, such a decision may depend on multiple departments, scattered records, discretionary interpretation, and opaque exceptions. An applicant may be told that the file is pending, rejected, or approved with conditions, but may have little clarity on why. A superior officer may find it difficult to reconstruct the decision later. An appeal may become an argument over relationships rather than reasons.

Under a Minimum Digital Kernel, the city would not need to digitize every urban function at once. But it would require that every building approval be represented through a common structure:

• the applicable rules, such as zoning, setbacks, height restrictions, and land-use norms

• the relevant facts, such as ownership records, parcel dimensions, plan submissions, and geospatial verification

• the final decision, including any approval, rejection, deviation, or exception, along with the basis for that outcome

Now the decision becomes interpretable.

A reviewer can see whether the rule was misapplied or whether the facts were incomplete. A senior administrator can identify patterns of unexplained deviation. An applicant can challenge the decision on stated grounds. If an exception is granted, it no longer disappears into the file. It becomes visible, trackable, and reviewable.

And that has institutional consequences. Supervision improves because similar decisions can be compared. Appeals improve because reasons are available for challenge. Audits improve because deviations can be traced systematically. Policy improves because repeated exceptions can reveal that a rule itself is poorly designed, outdated, or inconsistent with lived reality.

The significance here is not just better record-keeping. It is that the state becomes more capable of learning from its own decisions. The same logic applies beyond urban governance, whether in welfare eligibility, local licensing, land administration, or environmental clearances. In each case, the issue is not whether judgment disappears. It is whether judgment can be seen, compared, and reviewed.

Visibility Does Not Solve Politics, but It Changes Its Medium

A fair critique of any digital-governance idea is that information alone does not produce accountability. This is true. Many states already know where problems are. Audit reports exist. Complaints are filed. Violations are visible to insiders. Yet behavior often does not change because institutions lack incentives, authority, or willingness to act.

MDK does not solve that problem on its own. What it does do is change the medium in which politics happens. When rules are implicit and decisions are poorly represented, politics happens in the shadows: in access, in discretion, in informal interpretation, in unrecorded exceptions. When decisions are represented through explicit rules, documented facts, and visible outcomes, politics shifts onto more reviewable terrain.

There will still be bargaining. There will still be conflict. There will still be attempts at capture. But those struggles increasingly occur through explicit rules, recorded exceptions, and contestable decisions rather than through invisible transactions.

That is not the end of politics. It is a change in its institutional medium.

Why the Kernel Matters

The Minimum Digital Kernel should not be understood merely as a technical optimization or a better workflow layer. Its strategic importance lies elsewhere. It gives a diverse state a minimal common way to represent the exercise of public power.

That does not guarantee good governance. But without such a shared representational layer, every institution becomes harder to understand outside itself. Records may exist, but decisions cannot be compared. Policies may exist, but their application cannot be reconstructed. Appeals may exist, but the original basis of action remains unclear. Variation may exist, but it is not governable variation.

With such a structure, institutional diversity can coexist with system-wide intelligibility. This matters especially in India, where transitions toward more structured public reasoning are unlikely to come through one grand bargain or one nationally uniform administrative model. They are more likely to emerge through multiple sectoral and federal bargains, unevenly and incrementally.

Some domains will move faster than others. Some states will adopt more structure sooner. Some sectors will continue to rely heavily on judgment and exceptions. That is normal.

The role of the kernel is not to erase this unevenness. It is to give these diverse institutional arrangements a shared substrate through which public action becomes intelligible, comparable, reviewable, and contestable at scale.

The Real Promise of MDK

The promise of the Minimum Digital Kernel is not that it will eliminate discretion, automate justice, or solve the political economy of the state.

Its promise is more specific. It allows public decisions to be expressed in a form that can travel across institutions. It allows variation without abandoning intelligibility. It allows judgment without abandoning traceability. It allows diversity without abandoning coherence.

In a democracy of India’s scale, the future of governance may depend less on eliminating discretion than on ensuring that discretion can be seen, compared, and contested.