Building Commission MIS That Reconciles to IRDAI Returns and the Proposed Disclosure Schedule: Data Model, Monthly Close, and Audit Readiness

Most Indian broking firms run commission reporting as a management exercise: good enough to pay producers and track insurer receivables, tolerant of a few percent of unexplained variance. Two developments have ended that tolerance. First, the draft IRDAI (Insurance Intermediaries) (Amendment) Regulations, 2026, released for comment in June 2026 and still a draft, would require intermediaries to disclose intermediation revenue and other income from insurers in a separate schedule to their financial statements, file audited financials with IRDAI by 30 September each year, and publish them on the intermediary's own website, with stricter disclosure above INR 10 crore of commission income. A schedule that is audited, filed, and public must tie out to the rupee against the ledger, the insurer confirmations, and the annual returns already filed with IRDAI under the broker regulations.

Second, the commission environment itself has become harder to track. The IRDAI (Payment of Commission) Regulations, 2023 replaced product-wise caps with insurer board-approved commission policies, so rates now vary by insurer, product, channel arrangement, and sometimes by campaign period. The IRDAI (Expenses of Management, including Commission, of Insurers) Regulations, 2024 cap insurer spending at roughly 30 percent of gross written premium for general insurers (35 percent for standalone health insurers), and insurers manage that envelope actively, revising broker terms mid-year. A rate table that was static for a decade is now a moving target, and MIS that assumes one rate per product per insurer will drift from reality within two quarters.

Add the regulator's stated interest in remuneration that follows demonstrated servicing effort, and the direction is clear: commission income will need to be tracked per policy, per period, per component, and evidenced on demand. Firms that build that capability now do it on their own schedule; firms that wait will build it during an audit.

The foundational design decision is granularity. Commission MIS fails when it stores summaries; it works when it stores events. The core table should carry one row per commission event, where an event is any occurrence that creates, adjusts, or reverses expected brokerage: new business booking, renewal, endorsement (premium-bearing), cancellation, and insurer adjustment.

Minimum fields per row:

• Policy identifiers: policy number, endorsement number where applicable, client, and the firm's internal placement reference that links back to the placement file.
• Insurer and product dimensions: insurer, line of business mapped to the classification used in the firm's IRDAI annual returns (fire, marine cargo, marine hull, motor, health, miscellaneous, and so on), and product code. If the MIS line taxonomy differs from the returns taxonomy, reconciliation dies at the mapping layer, so maintain a single mastered mapping table.
• Premium basis: gross premium, the premium component on which commission is calculated (which differs where terrorism pool or government-scheme premium carries different terms), and policy period.
• Commission dimensions: commission type (base brokerage, reward or additional remuneration, any fee billed to the client recorded as a distinct type), expected rate, expected amount, GST treatment, and TDS expected under Section 194D.
• Lifecycle fields: booking date, insurer statement reference once the item appears on a bandhan or commission statement, receipt date and amount, and a status flag (expected, confirmed, received, disputed, written off).

Two modelling rules save the most pain later. First, record reward and portfolio-level remuneration as separate rows against a pseudo-policy or insurer-level key rather than smearing them across policies pro rata; the draft disclosure schedule distinguishes income categories, and auditors will too. Second, never overwrite: an endorsement that reduces premium generates a negative commission event, preserving the arithmetic trail from booking to final settlement.

A commission ledger reconciles annually only if it reconciles monthly. The close cadence that mid-size firms have converged on runs in four steps inside the first ten working days of each month.

1. Booking completeness check. Every policy bound in the month must have a commission expectation row. The control is a match of the placement register (or policy administration system) against the commission table; unmatched placements are the first leakage indicator. Target: zero policies older than 30 days without an expectation row.
2. Insurer statement matching. Load each insurer's commission statement and match line items to expectation rows on policy number, endorsement number, and amount. Auto-match typically clears 85 to 95 percent of items in firms with clean policy-number capture; the rest queue for manual matching with a 15-day ageing limit.
3. Variance triage. For matched items where the statement amount differs from expectation beyond a tolerance (many firms use INR 500 or 1 percent, whichever is higher), classify the cause: rate difference, premium-basis difference, GST or TDS treatment, or timing. Rate differences feed back into the rate master; unexplained differences become disputes with the insurer, logged with an owner and a follow-up date.
4. Ledger tie-out. The commission table's confirmed and received totals for the month must agree with the revenue and receivable balances in the general ledger, and TDS deducted per statements must agree with what will later appear in Form 26AS. Sign-off by the finance lead, with the reconciliation pack archived.

The output worth watching monthly is a simple bridge: expected commission for the month, confirmed by insurers, received, disputed, and unexplained. In firms doing this well, unexplained sits below 0.5 percent of monthly commission income. Above 2 percent, the year-end audit will find it, and the draft regulations' 30 September filing deadline leaves little room for a quarter of archaeology after the fiscal year closes.

Brokers file periodic and annual returns with IRDAI under the IRDAI (Insurance Brokers) Regulations, 2018, covering business placed and remuneration received, broken by insurer and line of business. These returns are the regulator's baseline picture of the firm, and the proposed audited disclosure schedule would sit alongside them. Three documents therefore have to tell the same story: the returns, the audited financial statements, and the commission MIS. Any pair that disagrees is an examiner's opening question.

The reconciliation is best built as a standing three-way bridge, refreshed quarterly:

• MIS to ledger: confirmed commission income per the events table against revenue recognised in the books, with reconciling items limited to defined categories (accrual timing, foreign-currency reinsurance brokerage where relevant, write-offs approved in the period).
• Ledger to returns: revenue in the trial balance mapped to the return's insurer-wise and line-wise grids. The most common break here is classification, not amount: health business booked as miscellaneous, a standalone health insurer's figures merged into a group entity, or fee income reported inside brokerage.
• Returns to insurer confirmations: annual balance and commission confirmations obtained from the top insurers by volume, compared to the return figures for the same insurer.

Run the FY2025-26 figures through this bridge as a dry run even though the 2026 disclosure schedule remains a draft. The dry run surfaces the two structural defects that cannot be fixed in an audit window: line-taxonomy mismatches between systems, and reward income that was booked as a lump sum with no insurer-wise attribution. Both take a quarter or more to remediate properly.

Reconciliation breaks cluster into predictable families. Knowing them in advance turns triage from investigation into classification.

Rate and basis mismatches. The insurer revises its board-approved commission terms mid-year under EOM pressure and the broker's rate master is not updated, so every policy after the revision date books a wrong expectation. Variant: commission calculated on premium including a component that carries different terms (terrorism pool premium on fire policies, long-term motor components, government-scheme business), producing small but systematic differences on every affected policy.

Endorsement and cancellation gaps. Premium-bearing endorsements processed by the insurer without a corresponding commission event in the broker's MIS, and pro rata commission clawbacks on cancellations that hit the insurer statement with no matching negative row. These are the largest single source of unexplained variance in most firms because endorsement volume is high and individual amounts are small.

Timing differences. Policy booked in March, commission on the insurer's April statement; instalment-premium policies where commission follows each instalment under Section 64VB discipline; and 64VB-linked holds where cover and therefore commission awaits premium realisation. Timing items are benign but must be aged and flagged, or they accumulate into a permanent unreconciled tail.

Tax-layer differences. Comparing commission gross of 18 percent GST on one side with net on the other, and TDS under Section 194D deducted at source so cash received is 5 or 2 percent short of the confirmed amount depending on the applicable rate for the period. The fix is structural: store gross, GST, TDS, and net as separate fields, and never reconcile across layers.

Identifier failures. Policy numbers keyed with spaces or prefixes that defeat auto-matching, endorsements reported under the master policy number, and client names used as match keys. Firms that enforce insurer policy-number capture at booking clear auto-match rates above 90 percent; firms that do not sit near 60 percent and drown in manual matching.

If the draft regulations are finalised, the intermediation-revenue schedule will be audited, filed by 30 September, and published. Preparing for that audit is mostly a matter of having six artefacts ready before the auditor asks:

1. The commission events table with full fiscal-year history, exportable, with the status of every row at year-end (received, confirmed-unpaid, disputed, written off).
2. Twelve monthly close packs, each showing the four-step reconciliation and sign-off. The existence of the cadence is itself audit evidence of control.
3. Insurer balance confirmations for insurers covering at least 80 percent of commission income, reconciled to the receivables ledger, with disputes listed and aged.
4. The rate master with change history: which insurer paid what rate on what product from when, referenced to the insurer's communicated terms. Under the 2023 commission framework this is the substantiation for every expectation the MIS ever booked.
5. The income-category mapping: base brokerage, reward and additional remuneration, client fees, and any other income from insurers, mapped to ledger heads and to the proposed schedule's categories, insurer-wise. Firms above the draft's INR 10 crore threshold should assume the auditor tests attribution item by item.
6. Write-off and dispute governance: approvals for amounts abandoned, with authority limits, because unexplained shrinkage between expected and received commission is both a leakage problem and a fraud-control red flag.

A useful internal benchmark before the first external pass: the firm can produce, within one working day, commission income for any month, split by insurer, line, and income category, agreeing to the ledger within 0.5 percent. Firms at that standard describe the annual return preparation as a two-day report run. Firms below it describe it as a three-week negotiation between finance, operations, and memory, and that is the version that does not survive a public, audited schedule.