Where Companies Die in India: What 10,708 Struck-off Companies Tell Founders About State, Age, and Vintage

By MakeItLegit Data Desk · Data source: MakeItLegit gov_event feed, snapshot 4 August 2026

10,708 companies died in the MCA's public struck-off lists with enough information to analyse their state, entity class and incorporation year. Almost every founder starting a company in 2025 assumes they will not be one of them. The data suggests that the first question is not whether a company can be incorporated. It is whether its founders have chosen a structure, activity and compliance rhythm that can survive after incorporation.

This is not a list of failed businesses. A struck-off company may be a dormant shell, a company whose founders stopped using it, a vehicle that was closed voluntarily, or a company removed in a Registrar of Companies enforcement sweep. Some struck-off companies can also be revived. But the list is still a useful record of where corporate intent and corporate maintenance have separated.

The pattern is striking. Companies incorporated in the pandemic-era vintages dominate the data. The geographic distribution is lopsided. Private limited companies account for most deaths because they account for most companies, but one-person companies appear more often than their incorporation share would suggest. And a catch-all activity code — “other professional, scientific and technical activities” — is the single largest activity bucket among the records we could interpret.

The practical lesson is less dramatic than the headline: incorporation is an event; compliance is a commitment. The MCA's public data shows what happens when the second part is treated as optional.

What we analysed

MakeItLegit analysed the MCA public disclosure feeds ingested through its gov_event pipeline. The snapshot contains 11,185 struck-off company events and 215,165 director-disqualification event rows. After deduplicating directors, the data contains 145,188 unique disqualified directors. For the state, entity-class and incorporation-year cuts, we used 10,708 struck-off records with parseable CINs.

The state analysis decodes the state segment of each Corporate Identity Number. The entity-class analysis uses the relevant CIN characters, and the vintage analysis uses the incorporation year encoded in the company record. Activity analysis uses NIC-2008 five-digit activity codes.

There are important limits. A struck-off record is not the same thing as a failure, insolvency or liquidation. Voluntary strike-off is common for dormant companies and unused shells. A company can apply to close itself, while the Registrar can also act under Section 248 of the Companies Act, 2013. A strike-off can be challenged or reversed through the National Company Law Tribunal under Section 252. The director data records disqualification events; disqualification is a statutory consequence of non-compliance, not a finding that a person committed a criminal offence.

The dataset is bounded by the MCA public feeds available to MakeItLegit. It is not a nationwide survival study, and it does not provide the denominator required to calculate the probability that a company incorporated in a particular state, year or class will eventually be struck off. The findings below describe the composition of the observed struck-off records, not the failure rate of every company in India.

Finding 1: Death follows the COVID vintage

The largest incorporation vintages in the parseable struck-off records are 2021, 2022 and 2020:

The 2020–2022 group contributes 2,983 records, or 28% of the 10,708 parseable struck-off companies. That concentration is the clearest signal in the dataset.

It does not mean that incorporation during the pandemic mechanically killed a company. It does suggest that pandemic-vintage companies are heavily represented in the strike-off records now visible in the feed. These were companies formed during a period when entrepreneurs rushed to test new ideas, remote services and digital businesses, while uncertainty changed the economics of established businesses. Some never became operational. Others became operational briefly and then stopped maintaining their legal shells.

There is also an administrative explanation. Companies that became inactive after incorporation did not necessarily disappear from the register immediately. The Registrar's strike-off powers under Section 248 allow companies that meet the statutory conditions for removal to be taken off the register, subject to the prescribed process. Enforcement and clean-up waves therefore create a lag between the point at which a company stops being useful and the point at which it appears in a public struck-off list.

The older vintages tell a different story. The records for 2010, 2011, 2012 and 2013 range from 348 to 377 each in this cut. These older companies are the long tail of inactivity that remained on the register until a later filing failure, voluntary closure or administrative sweep brought them into view. A company can be legally alive for years after its commercial purpose has ended. That gap is one reason founders should not think of annual filing as paperwork that can be postponed until they restart the business.

The data also makes a common mistake visible: confusing incorporation with proof of traction. A certificate of incorporation proves that a legal entity exists. It does not prove that the business has customers, revenue, employees or a plan for the next three years. If the company is formed as an experiment, the founders need an explicit decision point for continuing, restructuring or closing it. Otherwise the experiment becomes a compliance liability.

Finding 2: Death is regionally lopsided — but not where founders think

The state distribution is led by Telangana, Tamil Nadu and Himachal Pradesh in the records we could decode. Here are the top state-code groups in the source data:

Telangana requires a coding note. Older CINs use TG, while newer CINs use TS. The source contains 909 TG records and 63 TS records; the combined Telangana total is therefore 972. Punjab is represented here by the older PN code. The available cut does not provide a separate post-change PB count, so the table does not claim a complete Punjab total across both formats.

The headline is not that one state is inherently bad for companies. It is that strike-off records are not distributed evenly across the country, and the largest counts do not simply mirror the popular founder map. Telangana's combined count is higher than Maharashtra's in this dataset, even though Maharashtra is generally associated with a larger overall corporate base. That may reflect differences in incorporation vintage, registry activity, the timing of enforcement or the mix of companies represented in the public feed. The dataset alone cannot isolate those causes.

Himachal Pradesh is especially notable in the count. Its presence near the top should not be read as evidence that companies incorporated there are more likely to fail. Without a reliable state-by-state incorporation denominator, we cannot calculate that. It may reflect a concentration of certain registered-office or shell-company patterns, or simply the timing and coverage of the relevant registry records. The defensible conclusion is narrower: Himachal Pradesh is disproportionately visible in this observed struck-off set relative to the assumptions many founders make about where corporate enforcement activity is concentrated.

The “safe state” theory is just as misleading. Small north-eastern states and most Union Territories have low counts in a dataset like this, but their incorporation volumes are also low. A low observed count is not a survival guarantee. It may mean that fewer companies were incorporated, that fewer records are available in the relevant feed, or that enforcement and reporting patterns differ.

For founders, state selection should therefore be driven by the real registered office, operating footprint and access to professional support — not by a belief that a particular jurisdiction makes compliance disappear. The registered office is a legal obligation, not a branding choice. A state code in a CIN tells you where the company was registered; it does not rescue an inactive company from annual filing duties.

Finding 3: The one-person company signal is hard to ignore

Private limited companies dominate the struck-off records, as expected. But the entity-class mix is more revealing when read against how founders use these forms:

The base-rate warning matters. Private limited companies make up 91% of these deaths because they are the standard incorporated form for most startups and small businesses. That is not evidence that a private limited company is intrinsically fragile.

The OPC result is different. One Person Companies account for 4.6% of the struck-off dataset, while OPCs account for less than 2% of total incorporations according to the comparison used for this analysis. On that directional comparison, OPCs appear in the struck-off records at roughly two to three times their incorporation share.

That is not a controlled hazard model. We do not have a complete incorporation denominator by year and class in this extract, and we do not know how many OPCs were voluntarily closed. But it is an empirical warning against treating the OPC as the automatic answer for a solo founder who wants limited liability and does not want to involve another director.

The explanation may be structural rather than legal. Many OPCs are formed by founders who are still testing whether there is a business at all. The same founder has to originate the idea, fund it, operate it and maintain the company. If the idea stalls, there is no co-founder pushing the entity through the next filing cycle. The form is not broken; the selection logic can be.

Public limited companies appear at 2.4% of the records, a much smaller share than private limited companies. Public companies are generally chosen more deliberately, with a larger expected operating footprint and more formal governance. Again, this is a composition finding, not a survival claim. It tells us what appears in the observed struck-off list, not that every public company is safer.

The founder decision should be practical: if the company is genuinely a solo operating vehicle, an OPC may fit. If the founder expects investment, a co-founder, an employee option pool or frequent ownership changes, a private limited company may be the cleaner long-term structure. The mistake is choosing an OPC purely to avoid onboarding a second director, then allowing the company to sit idle because the business case was never tested.

Finding 4: The activity codes that die

The activity code is often treated as a form field to complete quickly. In the struck-off records, it is a clue about how precisely founders described what the company would do.

The leading NIC-2008 five-digit activity codes include:

The largest category, 74999, is a catch-all: other professional, scientific and technical activities not elsewhere classified. It may be accurate for a genuine specialist practice that does not fit another category. But it is also an easy answer when a founder has not decided what the company will actually sell.

There are 1,289 records in this bucket. The number does not tell us that companies using 74999 are more likely to be struck off than companies using another code; we do not have the necessary denominator. It does tell us that a broad, low-information activity description is common among companies that later appear in the struck-off data.

The software-related codes show the opposite interpretation problem. Code 72900 contributes 973 records and code 72200 contributes 320. Together they contribute 1,293 records — a high absolute count. But software and technology are also common incorporation purposes in India. A large count can simply reflect a large underlying population. It would be unsafe to turn these figures into a software failure rate.

The practical standard is simple: use the most specific NIC-2008 code that matches the business the company is actually set up to conduct. If the company pivots materially, review whether its constitutional documents, registrations and activity description still make sense. A code cannot make an inactive business active, but a vague code can make the original intent harder to understand when the company applies for a bank account, seeks investment or answers a compliance query.

Finding 5: Director disqualification is the larger warning system

The company list is only one side of the story. The same MCA public feed contains 215,165 director-disqualification event rows. After deduplication, 145,188 unique directors appear on disqualification lists.

That is more than three times the 11,185 struck-off company-event records in the dataset. The difference is not mysterious: one company can have multiple directors, and the statutory consequence can reach all directors associated with a non-compliant company. A director can also appear in more than one disqualification event row, which is why the event-row count is higher than the unique-director count.

In plain English, Section 164(2) of the Companies Act, 2013 creates a serious consequence for repeated filing failure. Where a company has not filed its financial statements or annual returns for a continuous period of three financial years, the directors who were directors during that period can be disqualified. The disqualification can prevent a person from being appointed or re-appointed as a director for five years.

The important point for founders is that this is not a warning that arrives only after a dramatic collapse. A company can be commercially quiet and still generate a director-level consequence if its annual filings are not maintained. “There was no revenue” is not the same as “there was nothing to file”. An inactive company still needs a deliberate compliance decision: maintain it properly, or close it through the appropriate route.

The disqualification data should also be read carefully. It records statutory disqualification events; it does not establish fraud, criminality or personal dishonesty. A founder may have joined a company that later stopped filing, or may have misunderstood which filings remained due after operations ceased. The remedy is not to ignore the data. It is to check the company's status, identify missed filings, understand whether disqualification has occurred and obtain the appropriate professional advice before incorporating or accepting another directorship.

What founders should actually do differently

The first decision is not “OPC or private limited?” in isolation. It is whether the founder has enough conviction to maintain a company for at least the next three years. Choose the entity that matches the real business and the founder's willingness to operate it. The higher observed OPC share is more plausibly downstream of founders choosing the form to avoid a co-director than evidence that OPC as a legal form is defective.

Do not incorporate a private limited company if the business cannot credibly support annual compliance for three years. Section 164(2) is designed around repeated non-filing, and the director consequence is serious. Budget for accounting, annual returns, financial statements, statutory registers and the other obligations that apply to the company. The exact obligation set depends on the entity and facts; “no transactions” is not a compliance strategy by itself.

If the business is not going to continue, voluntary strike-off through the prescribed STK-2 process is usually a more disciplined path than leaving the entity to be swept from the register. Closing a company does not mean skipping the requirements for closure. It means making the decision while records, directors and banking information are still accessible.

Use a real NIC-2008 activity code. The 74999 catch-all is not automatically wrong, but it is a poor substitute for understanding the business. A specific activity code helps keep incorporation records, tax registrations, banking conversations and future filings aligned.

Finally, treat the registered office as an operating control. Notices must reach someone. Books and records must be retrievable. Filing reminders must have an owner. A company is not maintained by the person who was enthusiastic on incorporation day; it is maintained by a process that still works when the founder is busy, the revenue is low and the original idea has changed.

Data availability

MakeItLegit will maintain a quarterly raw export for journalists, researchers and founders who want to reuse this analysis: download the MCA struck-off and director-disqualification CSV export. Please attribute the data to MakeItLegit's gov_event feed and cite the relevant snapshot date. The export is intended for research and reporting, not for labelling a company or director as fraudulent or failed.

Frequently asked questions

How did MakeItLegit get this data?

The analysis uses MCA public disclosure feeds ingested through MakeItLegit's gov_event pipeline. The 4 August 2026 snapshot contains 11,185 struck-off company events and 215,165 director-disqualification event rows. After deduplication, 145,188 unique directors appear in the disqualification data. State, class and incorporation-year analysis uses the 10,708 struck-off records with parseable CINs.

Is “struck off” the same as “failed”?

No. “Struck off” describes a company being removed from the register under a statutory or voluntary process. It does not by itself describe what happened commercially. A company may have had no operations, may have been created as a dormant shell, may have been closed voluntarily, or may have been removed after non-compliance. A company that traded successfully can also be struck off if its legal maintenance was neglected. Conversely, a company can be commercially unsuccessful without appearing in this specific dataset. The public record is therefore an administrative signal, not a complete business-outcome measure.

Can a struck-off company be revived?

Yes, in appropriate circumstances. Section 252 of the Companies Act, 2013 provides a route to apply to the National Company Law Tribunal for restoration of a company's name to the register. Restoration is not automatic. The applicant must meet the legal requirements and address the facts that led to the strike-off. Anyone considering restoration should verify the company's status and take advice on the application, pending filings, liabilities and operational records.

How is director disqualification triggered?

Section 164(2) is triggered when a company fails to file its financial statements or annual returns for a continuous period of three financial years. Directors covered by the provision can be disqualified from appointment or re-appointment for five years. The feed records the resulting disqualification events. It does not, by itself, say that a director committed a criminal offence.

Does a low strike-off count mean a state is safer?

No. The dataset does not provide incorporation denominators for every state, and low-count states also tend to have fewer incorporations. The state table describes the observed composition of the struck-off records, not a state-level probability of failure or strike-off. State choice should follow the real registered office and business rather than a presumed enforcement loophole.

Does the OPC result prove that OPCs fail more often?

It is a warning signal, not a complete failure-rate calculation. OPCs represent 4.6% of the struck-off dataset, while the comparison provided for this analysis places OPCs below 2% of total incorporations. That points to an observed share roughly two to three times the incorporation share. A proper survival rate would require complete incorporation and closure denominators, cohort tracking and controls for voluntary strike-off and vintage. The safer conclusion is that solo founders should choose an OPC for a real structural reason, not simply to avoid a second director.

What should a founder do if the company is dormant?

First, confirm the company's current MCA status and list the filings and obligations that remain due. Then decide whether the company has a credible operating purpose. If it does, maintain it properly and assign ownership for every recurring filing. If it does not, explore the prescribed voluntary closure route, including STK-2 where applicable, instead of allowing years of non-filing to accumulate. The right route depends on the company's facts, liabilities and filing position.

Where can I compare an OPC with a private limited company?

See MakeItLegit's OPC vs private limited comparison, then use the private limited registration guide for the relevant city. After incorporation, the After-Incorporation tool provides a free checklist for the recurring work that keeps a company in good standing.

The best company is not the one that is fastest to register. It is the one whose founders can explain what it does, maintain it when the first idea fails, and close it properly when the business is over.