The NDA now controls India's upper house with 141 seats. The Transgender Bill, introduced days before the majority was even sealed, shows what that arithmetic means in practice.

The BJP-led NDA crossed the Rajya Sabha majority mark on March 16, 2026, sweeping 9 of 11 contested seats through cross-voting and the collapse of the opposition. Three days earlier, the government introduced a Transgender Persons Amendment Bill that strips self-identification rights, imposes medical gatekeeping, and operates retroactively. Together, these events mark the end of India's upper house as a functioning institutional constraint on executive power.

I have been thinking a lot about Douglass North lately. North won the Nobel in Economics in 1993 for showing something that seems obvious once stated but is routinely ignored in practice: that institutions are not decorative. They are the rules of the game. They determine who gets what, how disputes are resolved, and what the powerful can and cannot do. When institutions work, they reduce what North called "transaction costs" of collective life. When they are dismantled, those costs do not disappear. They are transferred, usually downward, onto the people with the least capacity to bear them.

I thought of North on March 16, 2026, watching the Rajya Sabha election results come in from Bihar, Odisha, and Haryana. The NDA swept 9 of 11 contested seats. Cross-voting, absenteeism, and what can only be described as opposition self-immolation handed the BJP-led coalition roughly 141 seats in a 245-member house, 18 above the simple majority mark of 123. For the first time in decades, a single ruling coalition commands this comfortable a margin in the upper house.

Three days earlier, on March 13, the government had already introduced the Transgender Persons (Protection of Rights) Amendment Bill, 2026 in the Lok Sabha. The bill strips self-perceived gender identity from law. It replaces a simple self-declaration process with a five-stage medical-bureaucratic gauntlet. It operates retroactively, declaring that self-identified persons were never covered by the 2019 Act. And it creates a sentencing regime where forcing someone into a transgender identity carries life imprisonment, while actual violence against transgender persons carries a maximum of two years.

The timing was not coincidental. The bill was introduced before the majority was sealed because the government already knew the majority was coming. The Rajya Sabha was the last institutional dam. It has now broken. This essay is about what flows through.

What Happened on March 16: The Mechanics of a Majority

This section matters because the how of the majority tells you something the what cannot.

In political economy, George Tsebelis's veto player theory argues that policy stability depends on the number of institutional actors whose agreement is required to change the status quo. The more veto players, the harder it is to pass legislation, but also the harder it is to undo rights already established. The Rajya Sabha was designed to be precisely this: a federal veto player, a "Council of States" where staggered six-year terms, indirect election by state legislatures, and regional diversity would temper the majoritarian impulses of the Lok Sabha. When the Rajya Sabha functions, it forces deliberation, coalition-building, and compromise. When it does not, the executive faces no meaningful legislative constraint.

The NDA's 141-seat majority did not emerge from a single election. It is the accumulated result of the BJP winning 14 state governments and NDA allies governing five more. But the March 16 elections crystallised the shift. Of 37 seats across 10 states, 26 were won unopposed. The remaining 11 went to the polls in Bihar (5), Odisha (4), and Haryana (2). What happened in those three states was less an election than a demonstration of institutional decay.

In Bihar, the Mahagathbandhan needed 41 first-preference votes (the quota) to win a seat. They claimed the numbers. They did not deliver. Four opposition MLAs simply did not show up. RJD's Faisal Rahman stayed in Delhi "for family reasons." Three Congress MLAs switched off their phones and were unreachable. The RJD candidate received 37 votes instead of 41. BJP's fifth candidate, Shivesh Kumar, won with just 30 first-preference votes, thanks to second-preference transfers. NDA swept all five seats. Tejashwi Yadav accused the BJP of "engineering defections." But the Congress MLAs later offered a more revealing explanation: they abstained because RJD had fielded a Bhumihar candidate without consulting them. A caste grievance outweighed coalition discipline.

In Odisha, three Congress MLAs, Sofia Firdous, Ramesh Chandra Jena, and Dasarathi Gomango, voted for a BJP-backed independent instead of the joint BJD-Congress candidate. Eight BJD MLAs also crossed over. Total defections: 11. Congress had preemptively relocated 8 of its 14 Odisha MLAs to Karnataka before polling. Even that was insufficient. Firdous, at 35, the party's young minority face in Odisha, is the daughter of Mohammed Moquim, himself expelled from Congress for writing to Sonia Gandhi about internal dysfunction. The generational alienation speaks for itself.

In Haryana, counting was delayed five and a half hours over ballot validity disputes. BJP minister Anil Vij arrived in a wheelchair with both legs fractured; Congress challenged his vote. Five Congress votes were declared invalid. Congress candidate Karamvir Singh Boudh won by 0.33 vote value, one of the closest Rajya Sabha contests ever.

Here is the critical legal point, and it is one I want to emphasise as a lawyer: the Tenth Schedule anti-defection law does not apply to Rajya Sabha elections. The Supreme Court confirmed this in Kuldip Nayar v. Union of India (2006). These elections are not "proceedings of the Legislative Assembly," which means parties cannot issue legally binding whips. Cross-voting carries no disqualification penalty. The open ballot system, introduced in 2003 to prevent horse-trading, creates transparency without consequence. You can be caught defecting, but you keep your assembly seat.

This is a textbook principal-agent problem. The party (principal) nominates a Rajya Sabha candidate. The MLAs (agents) are supposed to vote accordingly. But the agents face no enforceable penalty for defection, while the incentives for defection (money, ministerial berths, protection from ED/CBI cases) are substantial. When the principal cannot discipline the agent, the institution breaks down. This is not a metaphor. This is what happened on March 16 in three states simultaneously.

What 141 Seats Enable, and What They Do Not

The distinction matters, and it is often missed in commentary that treats the majority as a blank cheque.

For ordinary legislation requiring a simple majority of 123, the NDA's 141 seats face no upper-house obstacle. This covers everything from the Transgender Bill to potential amendments to labour codes, environmental regulations, education policy, and criminal law. The opposition holds roughly 75 INDIA bloc seats plus 27 non-aligned seats (AAP 10, YSRCP 7, BJD 6, BRS 3). Even if every non-NDA member voted together, an implausible scenario given their divergent alignments, they would muster roughly 102 votes against 141.

For constitutional amendments, the situation is different. These require a two-thirds supermajority: 164 of 245 seats. NDA at 141 is roughly 23 seats short. This is why the One Nation One Election bill, introduced in December 2024, remains before a Joint Parliamentary Committee rather than proceeding to a vote. It cannot pass without support from non-aligned parties. The same constraint applies to any attempt to constitutionally entrench Hindutva's policy agenda (a national UCC, for instance, which Uttarakhand enacted at the state level in 2024 but which would require a constitutional amendment to impose nationally if it encroaches on personal law under the Concurrent List).

For money bills, the Rajya Sabha has always been largely irrelevant. It can only delay for 14 days and suggest non-binding amendments. This is the route the BJP exploited extensively when it lacked upper house numbers. The Aadhaar Act (2016) was classified as a money bill despite its sweeping biometric surveillance provisions, bypassing the Rajya Sabha entirely. The Finance Act, 2017 amended 26 enactments governing judicial tribunals through the money bill route, matters entirely unrelated to taxation. The Electoral Bonds Scheme, struck down by the Supreme Court in 2024 as unconstitutional, was also introduced through a Finance Act.

A seven-judge Supreme Court bench to definitively resolve whether non-financial legislation can be classified as a money bill was constituted in November 2019. It remains unconstituted over six years later. With NDA now commanding a Rajya Sabha majority, the practical urgency of the money bill bypass has diminished. The constitutional question lingers, unresolved, a piece of unfinished institutional business that grows less urgent with every seat the NDA gains.

I wrote in my piece on the VB-G RAM G Bill about how the Centre uses technical mechanisms to shift fiscal burdens onto states while claiming credit. The 60:40 cost-sharing formula that replaced MGNREGA's fully centrally funded wage structure was an example. The money bill classification is the legislative equivalent: a technical mechanism for bypassing a constitutional safeguard. Now that the safeguard itself has been numerically overwhelmed, the bypass is no longer even necessary.

The Transgender Bill: A Case Study in Institutional Capture

Let me explain what the Transgender Persons (Protection of Rights) Amendment Bill, 2026, actually does, because the commentary has been heavy on outrage and light on legal architecture. And the architecture is where the damage lies.

The 2019 Act defined a transgender person broadly: anyone "whose gender does not match the gender assigned at birth." It explicitly included trans men, trans women, persons with intersex variations, genderqueer individuals, and those with socio-cultural identities like kinner, hijra, aravani, and jogta. Section 4(2) recognised "the right of every transgender person to a self-perceived gender identity." Under this framework, obtaining an identity certificate required a self-affidavit and an Aadhaar card submitted to the District Magistrate. Over 23,000 identity certificates have been issued under this system (some government sources cite over 32,000).

The 2026 Amendment collapses this definition to three narrow categories. First: persons with recognised socio-cultural identities (kinner, hijra, aravani, jogta, eunuch). Second: persons with intersex variations involving specific congenital characteristics. Third: persons forced through mutilation, castration, or chemical procedures to assume a transgender identity. Everyone else is out. Trans men who do not belong to a traditional community structure. Trans women without medical documentation. Non-binary people. Genderqueer individuals. Anyone whose identity rests on their own understanding of who they are.

The bill deletes Section 4(2), the self-identification right, and replaces it with a five-stage bureaucratic process. An applicant must first undergo a medical procedure. Then, they appear before a designated medical board headed by a Chief Medical Officer. The board's recommendation goes to the District Magistrate, who may, if unsatisfied, refer the case to additional "medical experts" of unspecified qualification. The DM then issues or withholds the certificate. Medical institutions must report gender reassignment surgeries directly to the DM.

Now here is the part that a lawyer cannot read without sitting up. The bill contains a retroactive clause stating that the definition "shall not include, nor shall ever have been so included, persons with different sexual orientations and self-perceived sexual identities." This language purports to retroactively deny that self-identified persons were ever covered by the 2019 Act. It is a legal fiction imposed on an existing statute. And the bill contains no transitional provisions for existing certificate holders. Those tens of thousands of certificates? The bill neither validates nor invalidates them. It simply pretends the category never existed.

Then there is the sentencing disparity. For kidnapping and causing grievous hurt to force an adult into transgender identity: 10 years to life imprisonment, minimum fine of ₹2 lakh. For a child: mandatory life imprisonment, ₹5 lakh fine. For actual violence against transgender persons, compelling forced labour, denying passage, causing physical harm: maximum two years. The state considers the creation of transgender identity a graver offence than violence against transgender people who already exist.

A Development Economist Reads the Bill

Most commentary on the Transgender Bill focuses, rightly, on its constitutional dimensions. The bill contradicts NALSA v. Union of India (2014), where the Supreme Court held that gender identity is "an innate perception of one's gender," not dependent on biology or medical procedures. It contradicts K.S. Puttaswamy v. Union of India (2017), the nine-judge privacy bench that affirmed bodily integrity and self-determination as incident to Article 21. It contradicts Navtej Singh Johar v. Union of India (2018), which struck down Section 377. The Bar and Bench analysis by Bhattacharya and Tripathi demonstrates comprehensively that the bill's legal reasoning is "constitutionally suspect."

But I want to read it differently. As a development economist, what I see in this bill is a familiar pattern: the conversion of rights into conditional entitlements, mediated by bureaucratic gatekeeping, that functions as de facto exclusion.

Consider the five-stage certification process. In development economics, we call this "transaction costs of access." When Hernando de Soto wrote The Mystery of Capital about why poor people in developing countries cannot convert their assets into capital, his central finding was that the bureaucratic process required to formalise property was so expensive and time-consuming that it served as an effective form of exclusion. In Peru, obtaining legal authorization to build a house on state-owned land required 207 administrative steps across 52 government offices, taking nearly seven years. Obtaining legal title to that land required a further 728 steps. The process was technically open to everyone. In practice, it was open to nobody who was actually poor.

The Transgender Bill does the same thing. A trans person in, say, rural Madhya Pradesh, must first access and afford gender reassignment surgery (when India has fewer than 20 qualified surgical centres for gender-affirming care, mostly in metros, at costs running to several lakhs). Then appear before a medical board in the district. Then satisfy a District Magistrate. Then potentially satisfy additional unnamed medical experts. Each step has a cost: travel, documentation, time off work, social exposure. Each step has a gatekeeper with discretionary power.

This is what development economists call "exclusion by design." The system is technically universal. In practice, it excludes precisely the people it claims to serve. I saw this pattern repeatedly when I worked on the Livability Index and Swachh Survekshan studies. Metrics were designed to measure what was visible and administratively convenient, not what mattered for people's lives. A city could score high on "sanitation infrastructure" while its residents lacked functioning sewers. The metric served the dashboard. The dashboard served the minister. The minister served the narrative. The residents served nobody.

The Transgender Bill's Statement of Objects and Reasons reveals the same logic. The government's justification is that rights-based legislation cannot allow individuals to "acquire" a qualifying status on their own terms. It claims the Act was always intended to protect only those who are transgender "for no fault of their own and no choice of their own," i.e. biological conditions or entrenched socio-cultural identities. People whose identity rests on self-perception fall outside this frame.

As the Bar and Bench analysis points out, this reasoning is incoherent. Most rights-conferring statutes work exactly this way. The Special Marriage Act, 1954, confers legal rights on people who choose to marry under it. The Citizenship Act, 1955, grants rights to people who choose to obtain citizenship through its procedures. There is nothing unusual about a legal framework that allows individuals to access rights through a deliberate act. The "genuine beneficiary" argument, that welfare must be restricted to prevent "misuse," is the same argument used against every universal programme from MGNREGA to PDS to education. It is the ideological scaffolding of conditionality, and it has the same effect everywhere it is deployed: the poorest, most marginalised, most administratively invisible people are the ones excluded.

Grace Banu, a prominent Tamil Nadu activist, said it plainly: "This is a complete Hindutva bill. Hijra and Kinnar appear in their texts, so they get listed. Our identities do not fit their imagination, so we are erased." CPI(ML) Liberation's Central Committee compared the bill's provisions to the 1871 Criminal Tribes Act. Kalki Subramaniam, Southern Region Representative on the National Council for Transgender Persons, publicly threatened to resign, calling the bill a violation of "constitutional guarantees and the Supreme Court's NALSA judgment."

The Pattern: When Veto Players Disappear, Rights Follow

The Transgender Bill is not an outlier. It belongs to a legislative pattern that Daron Acemoglu and James Robinson would recognise: the incremental conversion of inclusive institutions into extractive ones. Each individual step is presented as a reform. Cumulatively, they reconstitute the relationship between state and citizen.

The Citizenship Amendment Act (2019) introduced religion as an explicit criterion for citizenship for the first time in Indian history. The Farm Laws (2020) were pushed through the Rajya Sabha by voice vote, with the Deputy Chairman refusing to call a division despite opposition demands. They were later repealed after a year of farmer protests, but not before exposing how thin the procedural constraints had become. Article 370's abrogation was effected through a Presidential Order while J&K was under President's Rule, bypassing the requirement for consent from a constituent assembly that no longer existed. UAPA amendments (2019) gave the government power to designate individuals as terrorists without judicial process; the conviction rate under the law has consistently remained in the low single digits.

A telling statistic captures the broader trend. Under the BJP government, only 25% of bills were referred to parliamentary committees in the 16th Lok Sabha, compared to 71% and 60% in the 15th and 14th Lok Sabhas, respectively. By the 17th Lok Sabha, the figure had fallen further to 16%. The Farm Laws were not referred to any committee. The Aadhaar Act was classified as a money bill to bypass the Rajya Sabha entirely, sidestepping the scrutiny a previous version had received from a parliamentary committee. The VB-G RAM G Bill that replaced MGNREGA, converting a rights-based employment guarantee into a centrally controlled scheme with 60:40 state cost-sharing, was rushed through with minimal deliberation. Speed replaces scrutiny. Arithmetic replaces argument.

At the state level, the Maharashtra Freedom of Religion Bill 2026, passed in the state assembly the same week as the Rajya Sabha elections, extends this pattern. Maharashtra becomes the 13th Indian state with anti-conversion legislation, all but two enacted or strengthened under BJP governments.

Meanwhile, in the same Rajya Sabha elections that sealed the NDA's majority, Menaka Guruswamy was elected on a TMC ticket from West Bengal, becoming India's first openly queer Member of Parliament. Guruswamy is the Senior Advocate who led the legal team that struck down Section 377 in the Navtej Singh Johar judgment of 2018. She enters a Parliament where the government is introducing legislation that contradicts the constitutional principles she helped establish. The irony writes itself. But irony is not a strategy.

What Can Be Done

I do not want to end with the usual gesture of despair. Things can be done. Some are legal, some institutional, some political.

The Transgender Bill must be challenged on constitutional grounds. The bill has been referred to a Parliamentary Standing Committee. This is a procedural opening. Community organisations should use the committee stage to build a public record of testimony. If the bill passes, it will face constitutional challenges grounded in NALSA, Puttaswamy, and Navtej Singh Johar. The principle of non-retrogression, that the state cannot roll back rights already recognised, would be central to such a challenge. The retroactive clause ("nor shall ever have been so included") is particularly vulnerable because it purports to rewrite legislative history, a power Parliament does not possess without a constitutional amendment.

The opposition must learn from its own failure. The cross-voting in Bihar, Odisha, and Haryana is not a BJP success story alone. It is an opposition governance crisis. If a party cannot hold its own MLAs in line during an open ballot where the vote is visible, it cannot claim to function as a political organisation. Congress expelled the three Odisha MLAs after the fact. That is the consequence after the damage is done. The structural problem is that INDIA bloc parties have no mechanism for collective discipline, no shared organisational infrastructure, and as a TMC Rajya Sabha MP told The Week, "not one call has been made from the Congress to its allies to discuss how it wants to proceed." A coalition that does not meet is not a coalition.

State governments must use their remaining constitutional powers. The Rajya Sabha majority affects central legislation, but it does not erase federalism entirely. States retain concurrent list powers, the ability to pass their own welfare legislation, and fiscal authority (constrained though it increasingly is by cesses and surcharges that bypass the Finance Commission's divisible pool). States that object to the Transgender Bill's implementation can use their administrative machinery to delay or obstruct enforcement, as states did with the CAA's National Population Register. This is not ideal governance. But when the centre dismantles institutional constraints, federalism becomes the backup firewall.

Civil society must document the costs. The tens of thousands of identity certificates at risk are not abstractions. Each represents a person who gained access to government schemes, employment protections, and anti-discrimination safeguards. If the bill passes, the immediate task is tracking what happens to these individuals. Do they lose access to ration cards, bank accounts, and voter IDs? Are their educational reservations rescinded? This is where development economists and field researchers have a role: building the evidentiary base that courts, committees, and future governments will need.

The constitutional amendment threshold must be defended. The NDA is 23 seats short of the two-thirds majority required for constitutional amendments. This is the last structural constraint. Parties holding those marginal seats, BJD, YSRCP, AAP, and smaller formations, should understand that their bargaining power exists only as long as this gap exists. If they trade their votes for short-term tactical gains, they erase the last check on a constitutional majority that could reshape India's basic structure. Ambedkar designed the supermajority requirement precisely for this scenario. It holds. For now.

The Cost of Institutional Decay

North argued that the difference between prosperous and poor societies is not natural resources or even technology. It is institutional quality. Societies with strong institutions, where the rules constrain the powerful and protect the vulnerable, generate broadly shared prosperity. Societies where institutions are captured, where the rules serve the rulers, generate extraction.

India's Rajya Sabha was, for decades, a functioning institutional constraint. It blocked the Land Acquisition Amendment Bill in 2015. It forced the Prevention of Terrorism Act to a Joint Session in 2002. It compelled governments to negotiate, to dilute their worst impulses, to build coalitions that required listening to regional, linguistic, and minority voices. This is not nostalgia. This is an institutional function.

That function has now been overwhelmed by arithmetic. The Transgender Bill is the first clear product of the new dispensation: a bill that contradicts three Supreme Court judgments, imposes retroactive erasure of rights, and creates a state surveillance apparatus over gender identity. It advances not because it commands a broad consensus, but because the institution that would have forced consensus no longer operates as a constraint.

Development economists know that institutional decay does not announce itself with a single catastrophe. It is cumulative. Each erosion makes the next one easier. Each norm violated becomes the new baseline. The ORF's assessment is direct: "With the re-emergence of one-party dominance, the Rajya Sabha is once again facing increasing challenge as a house of review."

On March 16, 2026, I watched India's upper house lose its capacity to function as what the framers intended: a brake, a mirror, a space for second thoughts. On March 13, I saw what comes when the brake is gone. The Transgender Bill is not the last piece of legislation this majority will produce. It is the first.

The question, as always with institutional collapse, is not whether the costs will be borne. They will. The question is by whom.

Varna is a development economist and writes at policygrounds.press

Further reading:

1. The Transgender Persons (Protection of Rights) Amendment Bill, 2026 (PRS Legislative Research)

2. Transgender Bill, 2026: Narrower definitions, wider constitutional questions (Bar and Bench)

3. 'This Bill is nothing but erasure' (The News Minute)

4. Rajya Sabha polls expose chinks in Opposition (The Week)

5. Re-thinking the Role of the Rajya Sabha in India's Federal Democracy (ORF)

6. Douglass North, Institutions, Institutional Change and Economic Performance (Cambridge, 1990)