Cronyism Didn't Decline. It Got a Statute.

TL;DR

The share of India’s non-financial sector assets held by five conglomerates rose from 10% in 1991 to nearly 18% in 2021, while the next five groups fell from 18% to under 9% — and the sharpest phase of that rise begins in 2015.¹ Over roughly the same window, the dominant political-funding instrument was anonymous by design: the Electoral Bond Scheme, enacted through a Finance Act, deleted three separate disclosure obligations at once, removed the cap on corporate donations, and produced no scandal for six years because nothing about it was illegal.²³ The Supreme Court struck it down in February 2024 as violating the voter’s right to information under Article 19(1)(a).² The pattern is not uniquely Indian: US dark-money spending hit $1.9 billion in the 2024 cycle, at least $4.3 billion since Citizens United.⁴ Scandal volume tracks what is prohibited. It does not track what is happening.

Legal, filed, audited, and completely uninteresting to a news cycle. Photo: Yury Kirillov on Unsplash.⁵

The Inference That Doesn’t Hold

There is a reading of the last decade that goes roughly: the great scandals ended. No 2G, no coal-block allocation, no Commonwealth Games. Whatever else is true, the era of the enormous corruption headline is over, and that must mean something improved.

It does mean something. The question is what.

A scandal requires three things to exist: a transfer of value, a rule prohibiting it, and someone able to document the breach. Remove any one and the scandal disappears while the transfer continues. Remove the rule specifically, and the transfer becomes a compliance event — filed, audited, entirely lawful, and completely uninteresting to a news cycle.

So “fewer scandals” is consistent with less capture and equally consistent with better-structured capture. Distinguishing them requires measuring the thing itself rather than the coverage of it. Two measurements are available.

Measurement One: Concentration Kept Rising

Viral Acharya, formerly Deputy Governor of the Reserve Bank of India, documented this in a 2023 Brookings paper, and the numbers are unusually clean.

The “Big-5” — Reliance, Tata, Aditya Birla, Adani and Bharti Telecom — saw their share of total assets in India’s non-financial sectors rise from 10% in 1991 to nearly 18% in 2021. Over the same period the next five business groups went the other way, from 18% in 1992 to under 9%.¹ Acharya’s own gloss is the important one: the Big-5 grew not just at the expense of the smallest firms but at the expense of the next-largest ones.¹ This is not the ordinary story of scale beating fragmentation. It is the top five pulling away from firms that were themselves very large.

The shape of the growth has two phases. Until 2010 the strategy was breadth — expanding into more sectors, eventually reaching over 40 two-digit industrial categories. From 2015 it became depth: acquiring larger shares within sectors already occupied.¹ Their share of merger and acquisition activity doubled from under 3% in 2015 to 6% in 2021, with no comparable rise for the next five groups.¹

Acharya considers and rejects the obvious alternative explanation — that this reflects Big-5 groups buying up distressed companies through the bankruptcy code — because the acceleration begins in 2015, before the first insolvency cases were resolved in 2018.¹ He attributes it instead in part to a deliberate industrial policy of creating national champions through preferential project allocation, alongside regulators declining to act on predatory pricing, and notes that high tariffs mean these groups largely do not face international competition in their domestic markets.¹

He names the risk directly: political connections and inefficient project allocations, related-party transactions inside opaque group structures, too-big-to-fail leverage, and crowding out of entrants.¹ And he finds a downstream price effect — a 10% rise in Big-5 share within an industry is associated with 2.7 percentage points higher wholesale price inflation the following year.¹

The comparison he draws with Korea is worth keeping. The chaebol model also built national champions, but those firms competed against international peers rather than sheltering behind high tariffs, and were accompanied by serious reforms in land, labour and power.¹ The point isn’t that large conglomerates are inherently a pathology. It’s that the specific Indian configuration — concentration plus protection minus factor-market reform — is the version most exposed to the crony failure mode.

Measurement Two: The Money Channel Was Made Anonymous by Law

The Electoral Bond Scheme is the cleanest available illustration of legality doing the work that secrecy used to do.

It arrived in the Finance Act, 2017, presented on the reasoning that India had not developed a transparent method of funding political parties.² What it actually did was delete transparency in three places at once:²³

• Representation of the People Act: parties no longer had to disclose contributions received through bonds in their contribution reports.
• Income Tax Act: parties were relieved of the obligation to maintain a record of who gave through bonds.
• Companies Act: the requirement that companies disclose political contributions in their profit-and-loss accounts was removed — and separately, the cap limiting corporate donations to 7.5% of three years’ average net profit was abolished entirely.

Read that list again as a design. Not “the donor is anonymous to the public” but: the party need not report it, the party need not record it, and the company need not disclose it, and there is no longer any ceiling on the amount. Loss-making companies and shell entities became eligible donors in unlimited sums, with no trail at any of the three points where a trail normally exists.

The small-donor justification does not survive the denomination data. The Association for Democratic Reforms reports that 94% of contributions came in the ₹1 crore denomination;³ in one 2023 tranche, 95.1% of the bonds sold were ₹1 crore units.⁶ This was not a scheme through which ordinary citizens contributed privately. It was a corporate channel.

On 15 February 2024, a five-judge Constitution Bench struck the whole structure down — the scheme and the enabling amendments to the Companies Act, Income Tax Act and Representation of the People Act — holding it violated the voter’s right to information under Article 19(1)(a).² That is the same clause discussed elsewhere on this site in a different context, and it is doing real work here: the Court’s reasoning is that you cannot exercise the franchise meaningfully without knowing who is funding the people asking for it.

When the disclosures finally came, the distribution was heavily skewed. Since April 2019 the BJP received ₹6,060.51 crore, the Trinamool Congress ₹1,609.53 crore and the Congress ₹1,421.86 crore, with regional parties collectively taking roughly ₹5,221 crore.⁶ SBI missed the Court’s 6 March deadline; the Election Commission published the data on 15 and 17 March 2024.⁶

That skew is real and worth stating plainly. So is the second fact: parties across the spectrum, in states run by most major formations, used the instrument. Which explains something otherwise puzzling — how a scheme this consequential survived six years without a serious cross-party effort to repeal it. The challenge came from civil-society petitioners and the courts, not from the opposition benches, because there weren’t many benches with a clean interest in ending it.

This Is Not an Indian Peculiarity

The same mechanism runs in the country most often held up as the alternative.

Since the Citizens United decision in 2010, dark-money groups — nonprofits and shell companies not legally required to disclose their donors — have spent at least $4.3 billion on US federal elections, reaching $1.9 billion in the 2024 cycle alone, nearly double the previous record of $1 billion in 2020.⁴ Of the 2024 figure, $1.3 billion went as contributions to super PACs, with the rest split across online, television and directly reported spending.⁴ The Brennan Center’s own caveat is that this substantially underestimates the true scale, since streaming ads, influencer payments and some physical advertising cannot be reliably tracked.⁴

None of that is a scandal. All of it is legal. It generated a constitutional doctrine rather than an arrest.

Naming the parallel matters for the argument. If legalised opacity were a story about one country’s institutions, you’d fix it with better institutions. That it appears in a mature common-law democracy with strong disclosure traditions, arriving by judicial decision rather than legislative stealth, suggests something more structural: wherever money must reach politics and disclosure is costly to donors, the durable equilibrium is a legal channel that does not require disclosure. The route differs. The destination doesn’t.

What the Argument Actually Claims

Precision matters here, because this is the kind of piece that invites overreach.

It does not claim that electoral bonds caused the rise in industrial concentration. Nobody has demonstrated that, the two datasets do not speak to each other, and the concentration trend begins well before the scheme existed. Anyone asserting a causal chain from a specific donation to a specific allocation needs specific evidence, and it usually isn’t available — which is, of course, the design.

What it claims is narrower and, I think, hard to dispute: the frequency of corruption scandals is a poor proxy for the level of state–business capture, and we should expect the two to diverge exactly when capture is formalised. The licence raj generated scandals because discretionary permits were traded illegally, so every transaction was potential evidence. Remove the illegality and the evidence stops being produced. The quiet is a property of the channel, not of the flow.

Which gives a practical test for the next time someone points at a clean stretch of headlines. Ask what happened to the measurements: concentration ratios, markups, the share of political funding whose source is knowable, the number of sectors where a single group holds a dominant position. If those moved the wrong way while the scandals stopped, you are not looking at reform. You are looking at a channel that was upgraded.

Where I’d Hold This Loosely

Three limits, and the first is significant.

I have not established that scandal volume actually fell. I’ve treated it as a premise because it is the common perception and the starting point of the argument I’m addressing, but “were there fewer major corruption scandals in 2015–2025 than in 2005–2015” is an empirical question I have not measured, and perception of scandal frequency is heavily shaped by media economics and news-cycle memory. If the premise is wrong, the piece is answering a question nobody should have asked. I’d defend the conditional — if scandals fell, here is why that isn’t reassuring — more confidently than the antecedent.

Second, industrial concentration is not the same thing as cronyism, and Acharya is careful about this even where I have compressed him. Groups can grow through genuine efficiency, better capital access, or competent management. His evidence for the crony reading is circumstantial — the timing, the tariff shelter, the markup differential, the M&A pattern — rather than a demonstrated transfer. It is a strong circumstantial case. It is still circumstantial.

Third, and cutting against my own framing: legalising a flow is not automatically worse than criminalising it. There is a real argument that bribery driven underground is more distorting than donation conducted above ground, and that the honest fix is disclosure rather than prohibition. I find that argument reasonable in principle — and note that the electoral bond scheme was the exact opposite of it. It legalised the flow and removed the disclosure, which is the one combination the argument doesn’t support. That is what the Court said too.

---

Footnotes

1. https://www.brookings.edu/wp-content/uploads/2023/02/BPEA_Spring2023_EM-Panel_Acharya_unembargoed_updated.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

2. https://www.scobserver.in/cases/association-for-democratic-reforms-electoral-bonds-case-background/ ↩ ↩² ↩³ ↩⁴ ↩⁵

3. https://adrindia.org/content/supreme-court-declaring-electoral-bonds-unconstitutional-monumental-defence-democracy ↩ ↩² ↩³

4. https://www.brennancenter.org/our-work/research-reports/dark-money-hit-record-high-19-billion-2024-federal-races ↩ ↩² ↩³ ↩⁴

5. https://unsplash.com/photos/looking-up-at-modern-glass-skyscrapers-against-a-bright-sky-6eqpNnKLbUI ↩

6. https://en.wikipedia.org/wiki/Electoral_bond ↩ ↩² ↩³