The Creamy Layer Was Never the Real Problem. Caste-as-Proxy Is.

Fixing reservation's income thresholds treats a symptom. The real question India keeps avoiding is whether caste, alone, is still the right way to identify who needs help — 75 years after the Constitution first asked that question.

Start With What Everyone Already Agrees On

Nobody serious denies that reservation, as currently designed, is leaking. The creamy layer principle — meant to filter out families who have already climbed out of disadvantage — doesn't apply to SC/ST categories at all, and even where it exists for OBCs, its ₹8 lakh threshold hasn't moved since 2017 while inflation and urban cost-of-living have moved plenty. The result is a system that keeps handing the same benefit to the same relatively well-off families within a reserved category, generation after generation, while the poorest members of that same category — the ones without the information, coaching access, or paperwork to even claim their status effectively — get none of it.

The usual fix on offer is technical: raise the threshold, extend the creamy layer to SC/ST, review it every few years. I think this is exactly the kind of reasonable-sounding reform that lets India avoid the actual question, which isn't "where should the income line sit" — it's whether caste is still, on its own, a reliable enough signal of disadvantage to keep organizing an entire national policy around it.

Why Caste Made Sense in 1950, and Why That's Not the Same as 2026

I want to be precise about this, because it's easy to sound like I'm dismissing history that shouldn't be dismissed. In 1950, caste was not just correlated with deprivation — for the vast majority of Dalit and Adivasi families, it *was* the mechanism of deprivation: land ownership, access to education, entry into professions, even physical presence in certain public spaces were directly gated by caste in ways that had almost no exceptions. Using caste as the targeting variable wasn't a proxy choice back then. It was closer to a direct measurement.

That direct relationship has weakened — not vanished, weakened — over 75 years, unevenly, and reservation itself deserves real credit for weakening it. A meaningful, if still small, professional and political class now exists within SC/ST communities: IAS officers, judges, MPs, business owners, second- and third-generation government employees whose own children grow up with resources, networks, and educational access that most upper-caste rural Indians still don't have. This is not a talking point invented to attack reservation — it's the system working exactly as intended, which is precisely why continuing to treat every member of a reserved category as equally disadvantaged, regardless of that family's actual circumstances, has stopped being accurate.

The Uncomfortable Data Point Nobody Wants to Sit With

Parliamentary committee findings and NSSO consumption data have both pointed to the same pattern: a disproportionate share of reservation benefits in competitive exams and elite institutions goes to applicants from already-mobile families within reserved categories, while first-generation, low-income applicants from those same categories continue to lag on outcomes. If the goal is helping the most disadvantaged people in the country, and the mechanism keeps missing a large share of exactly those people while reliably reaching families who are no longer, by any reasonable measure, the most disadvantaged — that's not a rounding error in an otherwise sound design. That's the design failing at its one core job.

Why "Just Fix the Filter" Isn't Enough

Extending and updating the creamy layer would help — I'm not arguing against it. But it treats caste-based reservation's core assumption as sound and only its execution as flawed. I think the assumption itself needs re-examining: caste, used alone, conflates two different things that used to travel together almost perfectly and increasingly don't — historical group identity and present-day individual deprivation. A Dalit family with two generations of government service, urban schooling, and stable income, and a Dalit family still working as daily-wage agricultural labourers in a drought-hit district, are both currently eligible for the same reservation benefit on the same terms, because the policy is keyed entirely to group membership rather than actual circumstance.

**My honest position: India should move toward a primarily economic and circumstantial basis for reservation — parental education, family income, rural/urban location, first-generation status — layered with continued, targeted support for communities facing documented, ongoing forms of *non-economic* discrimination that pure income-based criteria genuinely can't capture, like caste-based social exclusion, discrimination in housing or marriage, or violence.** That second piece matters and I don't want to wave it away — a wealthy Dalit professional can still face real discrimination that an equally wealthy upper-caste person doesn't. But that specific, narrower harm calls for specific, narrower remedies — anti-discrimination enforcement, hate-crime law, workplace protections — not a blanket admissions and employment quota sized to the entire historical population of a caste group regardless of individual family circumstance.

The Political Reason This Never Happens

Here's the part that actually bothers me most: I don't think this reform is stuck because the underlying arguments are unresolved. I think it's stuck because reservation, as currently designed, has become one of the most reliable vote-delivery mechanisms in Indian politics, and no party — including ones that privately acknowledge the creamy layer problem — wants to be the one seen touching it. Every attempt at even modest technical reform gets read, almost instantly, as an attack on the underlying constitutional commitment to social justice, which makes honest recalibration nearly impossible to attempt without triggering a defensive reflex that shuts down the conversation before it starts. That reflex was arguably justified decades ago, when reform proposals genuinely were often bad-faith attempts to gut protections that were still urgently needed. I don't think it's serving anyone well today, including the communities reservation was built to help.

What I'd Actually Want to See

Not abolition — that's a different, more radical claim than the one I'm making, and I don't think India's non-economic discrimination problem has disappeared enough to justify it. What I want is a shift in the *primary* sorting variable, from group identity to individual circumstance, with caste-linked support retained specifically where it addresses harms that income data can't see. A first-generation SC/ST/OBC applicant from a landless rural family should get more support than they currently do, not less. A third-generation IAS family that happens to belong to a reserved category should get less than it currently does, not the same. That's not a betrayal of Ambedkar's project — I'd argue it's closer to it than the status quo is, since the entire point of the original design was targeting need, and need is what the current system, keyed almost entirely to birth, has gradually stopped measuring well.

Where I'm less certain, and where reasonable people disagree with me: Caste-based social discrimination — in marriage, housing, everyday social interaction — remains real and documented even among economically secure members of reserved communities, and critics of my position argue that income-based criteria will systematically under-protect against exactly this kind of harm, since it has nothing to do with a family's bank balance. There's also a legitimate practical objection: India still lacks reliable, verifiable, non-caste data (an updated caste census barely exists, let alone comprehensive multi-generational income and asset data) to actually implement a circumstance-based system fairly, which means a shift like the one I'm describing could, in practice, be captured or gamed just as easily as the current system is — a risk worth taking seriously rather than assuming away.