Karnataka's Gig Workers Welfare Fee and What It Forces Aggregators to Buy: Insurance Structuring After the February 2026 Notification

In February 2026 the Government of Karnataka notified the mandatory Gig Workers Welfare Fee under the Karnataka Platform Based Gig Workers (Social Security and Welfare) Act 2025. The mechanics are specific. Aggregators must self-declare and deposit a welfare fee on each transaction, calculated as a small percentage of commission with per-transaction caps reported in the range of 50 paise, 75 paise and Re 1 depending on category. Ride-hailing, food and grocery delivery, logistics, e-marketplaces and professional services each sit in their own cap band. The welfare-fee rules require aggregators to calculate, self-declare and pay quarterly, and the obligation runs from mid-February 2026.

This arrives a few months after the central Code on Social Security 2020 came into force on 21 November 2025, which separately empowers the government to draw aggregator contributions of one to two percent of annual turnover (capped at five percent of amounts payable to workers) into a national fund. So a Bengaluru-headquartered platform now faces two parallel contribution regimes plus its existing insurance programme.

The question every risk manager is asking is blunt. If the platform is already paying into a welfare fund that promises accident and disability cover, why keep buying group personal accident and motor third-party cover on top? The short answer is that the fund and the policy do different legal jobs. The welfare fee funds a state benefit pool. It does not extinguish the platform's own tort and contractual exposure when a rider injures a pedestrian, damages a third party's car or dies on the job. Treating the fee as an insurance substitute is the single most expensive mistake a platform can make this year, and brokers who can map the overlap precisely will win these accounts at renewal.

Start with what the fund actually is. A welfare fee is a statutory cess routed into a state-administered corpus that pays notified benefits to registered gig workers. The benefit is owed by the fund to the worker, on the fund's terms, subject to registration, eligibility and whatever schemes Karnataka eventually notifies. It is a social-security instrument, not a liability indemnity. Three gaps follow directly.

First, third-party exposure is untouched. When a delivery rider knocks down a pedestrian or hits another vehicle, the injured party sues the rider and, increasingly, the platform on a vicarious or operational-control theory. The welfare fund pays the rider nothing toward that liability. Only motor third-party cover and the platform's own public liability arrangements respond.

Second, benefit adequacy is thin and slow. Even where the fund eventually pays an accident benefit, the quantum is a notified scheme amount, not the worker's actual loss of earnings or medical cost. A group personal accident policy with a defined capital sum and weekly benefit, plus a group mediclaim layer, fills the gap between a modest state benefit and a family's real loss.

Third, timing and certainty differ. Insurance pays on a contractual trigger that a broker can negotiate and a surveyor can assess. Fund benefits depend on registration being complete, the scheme being live and administrative processing. A platform that tells riders "you are covered by the welfare fund" and nothing else is exposed to reputational and contractual claims when that promise proves slow or partial.

For a platform operating riders or driver-partners in Karnataka, the defensible programme has four distinct layers, each answering a different exposure. Brokers should present it exactly this way so the client sees that no two layers are interchangeable.

• Group personal accident (GPA). Covers the rider for death and disablement from accidents, typically with a capital sum, permanent total and partial disablement scale, and an optional weekly benefit for temporary disablement. This is the layer most often confused with the welfare fund. It is not the same thing: GPA is the platform's contractual promise to the rider, payable on a clean accidental-death or disablement trigger, with sums the platform chooses. See our group personal accident structuring guide for gig platforms.
• Group mediclaim (GMC). Hospitalisation cover for riders and, where the platform commits to it, dependants. The welfare fund may notify a health scheme, but GMC gives a known network, cashless access and a sum insured the platform controls.
• Motor third-party and own-damage. On the rider's two-wheeler or the platform's owned fleet. Third-party cover is statutory under the Motor Vehicles Act regardless of any welfare fee. For owned or leased fleets, an own-damage and fleet programme sits here.
• Platform liability. The aggregator's own public and product liability, plus professional indemnity where the platform gives advice or handles money, responding when a third party sues the platform rather than the rider.

None of these four collapses into the welfare fee. The fee sits beside the stack as a fifth, statutory line. A broker who draws this on a single slide, fee plus four insurance layers, gives the CFO the clarity that the Act's drafting deliberately leaves murky.

The renewal conversation this year is a policy-wording exercise, not a price exercise. The Act and the central Code both define "gig worker" and "platform worker" in ways that may not align with how the platform's existing GPA and liability wordings describe the insured population. Misalignment is where claims fail.

Check the definition of "insured person" in the GPA wording. If it reads "employees" or "persons under contract of employment", it may not clearly capture independent driver-partners who are, by design, not employees. Many gig wordings already use "registered partners" or "enrolled riders", but some carried-over corporate wordings do not. Negotiate an explicit endorsement naming the gig-worker population by reference to the platform's registration records.

Three wording traps specific to this exposure

• Scope of cover during work. Confirm whether GPA responds only "while on duty", "during a logged-in session", or twenty-four hours. Riders dispute the boundary constantly. A 24-hour cover removes the on-duty argument at claim stage but costs more; an on-duty cover needs a clean definition of "on duty" tied to the app's session data.
• Aggregation and per-event limits. A single accident involving several riders, or a fund-versus-policy double-recovery question, can trigger aggregation clauses. Read the contribution position so the platform is not arguing with its insurer about whether the welfare benefit reduces the policy payout.
• Liability triggers. Public-liability wordings often exclude liability "assumed under contract". If the platform's rider agreement promises riders certain benefits, ensure that contractual assumption is carved back in or covered separately.

The welfare fee is calculated per transaction, which means the platform now generates, by statutory necessity, a clean quarterly count of transactions by category. That data is gold for an underwriter pricing GPA and motor exposure, and brokers should use it.

Until now, gig GPA pricing leaned on rider headcount and a broad assumption about exposure hours. The welfare-fee filing forces the platform to produce transaction volumes split by ride-hailing, food, grocery, logistics and so on. Each category carries a different accident frequency. A two-wheeler food rider in dense Bengaluru traffic is a different risk from a four-wheeler ride-hailing partner or a slow-moving grocery picker. Presenting category-split volumes lets the underwriter price each cohort rather than blending to a conservative average. For a platform with a large low-risk component, that split usually reduces premium.

For motor, the same logic applies. Commercial motor pricing has been moving with detariffing, and exposure granularity helps. Telematics-derived kilometres and category mix sharpen the rate far better than a flat per-rider charge.

The fee also changes the affordability conversation. A CFO who has just absorbed a new quarterly statutory cost will resist a GPA premium increase. The broker's job is to show that the fee and the insurance are not competing for the same budget line because they cover different losses, and that a thin fund benefit is precisely why the GPA sum insured should hold or rise. Frame the GPA premium as the marginal cost of converting a vague state promise into a contractual, surveyor-assessed benefit the platform can show riders.

Expect insurers to ask for the platform's welfare-fee registration status as part of underwriting. A platform that is compliant on the fee signals a mature risk culture, which supports better terms.

The hard operational problem is what happens at claim stage when a rider is injured or killed and both the welfare fund and the platform's GPA could pay. Get the sequencing right in advance, because grieving families and injured riders will not wait for the platform to work it out.

First, treat the GPA and GMC as the primary, fast-paying layer. The platform controls these, the claim trigger is contractual, and a surveyor or the insurer's claims team can assess and pay on a defined timeline. Build a single intake so a rider or family reports once and the platform routes the claim to the insurer and the fund in parallel, rather than making the family chase two bodies.

Second, document the welfare-fund position clearly. If the fund eventually pays a notified accident benefit, confirm with the GPA insurer in writing (ideally before any loss) that the fund payment is not a collateral source that reduces the policy sum insured. Without that confirmation, an insurer may argue contribution or set-off and shrink the rider's recovery.

Third, on third-party motor claims, the fund is irrelevant and must not be allowed to confuse the file. The injured third party's recovery runs entirely through motor third-party cover and, where the platform is joined, its liability cover. Keep that lane completely separate from the rider-welfare lane.

A practical claims playbook

1. Single reporting channel; one form, parallel routing to insurer and fund.
2. GPA and GMC assessed and paid on the policy timeline, not gated on fund processing.
3. Written collateral-source carve-out so fund benefits do not erode policy sums.
4. Separate, dedicated handling for third-party motor and platform-liability claims.
5. Quarterly reconciliation of fund contributions against claims experience for the next renewal.

This sequencing protects the rider, protects the platform's reputation, and keeps the insurer's file clean.

This is a closing-window opportunity. Most Karnataka aggregators have absorbed the welfare fee operationally but have not yet revisited their insurance programme against it. The broker who opens that conversation first sets the structure.

Build a one-page exposure map for each platform client: the welfare fee as a statutory line, then GPA, GMC, motor third-party and own-damage, and platform liability as four insurance layers, with a plain statement of what each pays and what it does not. Use it to show the CFO that the fee has not made any policy redundant.

Then pressure-test the wordings. Confirm the insured-person definition captures registered gig partners, fix the on-duty versus 24-hour scope, secure the collateral-source carve-out, and check the contractual-liability position on the platform's own liability cover. These are endorsements, not new policies, so the cost is modest and the protection is large.

Finally, harvest the new data. Ask the client for category-split transaction volumes from their welfare-fee filings and take them to market. The granularity will usually improve GPA and motor terms, which funds part of the fee in real money.

Karnataka moved first, but Rajasthan, Telangana and others are advancing similar platform-worker legislation. A platform operating across states will soon manage several welfare-fund regimes alongside one national insurance programme. The broker who builds a clean fund-versus-policy framework for the Karnataka account now has a template ready for every state that follows.

The core message for the client is simple and worth repeating in every meeting. The welfare fee is a tax on doing business with gig workers. Insurance is how the platform protects itself and its riders from what those riders do and what happens to them. The two are complementary costs, not substitutes, and any advice that conflates them is advice the platform should not act on.