Forecasting Broking Income: Building a Commission Budget for FY27

Most broking budgets take last year's commission line and add a growth percentage that sounds defensible in a board meeting. For FY27 that method fails on the evidence, because the two numbers it depends on have stopped moving together.

In FY2024-25, non-life commission expense across the industry reached roughly INR 47,266 crore, up from about INR 39,601 crore the year before. That is an increase of close to 19 percent. General insurance premium over the same period grew about 8.5 percent. Commission grew at more than twice the rate of the premium it was paid on.

That gap can be read two ways. Optimistically, distribution is capturing a larger share of a growing market and a broker with a good grid should assume yield expansion. Pessimistically, a distribution cost line growing at 19 percent against 8.5 percent premium growth is exactly the pattern a regulator notices. IRDAI has been collecting distributor-level commission data from life insurers to base reform on evidence, which is not the behaviour of a regulator intending to leave the number alone.

A budget assuming the divergence continues and one assuming it closes produce FY27 numbers several crore apart on a mid-sized book. You do not know which, so the budget must be built with that assumption visible, isolated and changeable without rebuilding the model. This post is about how to construct it so that being wrong is survivable and legible.

The first mechanical error in most broking budgets is that the base is a revenue figure. Revenue is an output. The base is a policy list.

Build the expiring register: every policy expiring in FY27, at policy level, carrying insurer, line, client, expiring premium, current commission rate, and commission earned this year. This is a query, not a project. Firms that cannot produce it in a day have discovered something more important than their budget.

Two adjustments matter before this register becomes a base:

• Strip the non-recurring. Project covers, one-off marine declarations under an open policy, a contractor's all-risks policy on a job that finished, single-year transactional placements. These earned commission this year and will not recur, and leaving them in the base guarantees a miss.
• Normalise the current-year distortions. A mid-term cancellation that reduced this year's earned commission, or a large endorsement that inflated it, should not propagate into the FY27 base. Budget from the annualised run-rate of the cover as it stands at year-end, not from what happened to land in the ledger.

What remains is the recurring renewal base: what the firm would earn in FY27 if every account renewed at the same premium and rate. On most established commercial books that is 70 to 85 percent of total income, and the only part of the budget grounded in something that already exists. Getting it wrong is unrecoverable, because every subsequent assumption is applied to it.

Once the base exists, the single assumption that dominates the FY27 number is retention. It also gets stated in the sloppiest way of any input in the model.

"We retain 90 percent" is three different claims:

1. Retention by policy count, which flatters every book, because the accounts a firm loses are rarely its smallest.
2. Retention by premium, which is closer but still misleading, since premium and commission move on different rates by line.
3. Retention by commission, the only version that belongs in a revenue budget, because it weights each lost account by what it actually paid the firm.

The spread between the first and third on a real book is routinely five to ten points. A firm reporting 92 percent count retention may be at 84 percent on commission, and the difference is the budget.

Assume retention at segment level, not as a firm-wide constant. Group health behaves differently from fire, a PSU account differently from an SME package, a two-year-old relationship differently from a twelve-year-old one. Three or four segments with their own rates, each from the firm's trailing three-year experience, beat a blended figure that is right on average and wrong everywhere.

New business is where budgets become fiction, because it is the only line with no anchor in an existing policy. The discipline is to make it a bridge from a pipeline rather than a number chosen to close a gap.

Build it in four steps:

1. Name the pipeline. Identified opportunities with a client name, a line, an expected premium and an expected incept date. Not a market-size estimate. If the named pipeline cannot support the budget, either the budget is wrong or the pipeline work has not been done, and both are worth knowing in February rather than October.
2. Weight by conversion, using your own rate. Firms consistently overstate this. The measurable version is placements won divided by opportunities formally pursued over the last three years, cut by segment, because a competitive PSU tender and a referred SME account do not convert at the same rate.
3. Apply commission by line, not a blended rate. New business mix is rarely the same as the existing book's mix, which is the whole point of pursuing it.
4. Phase by incept date, then apply the earning fraction. An account that incepts in January earns one quarter of its annual commission in FY27. Budgets that credit a full year of commission to a Q4 win overstate the year by a material amount, and it is the commonest single arithmetic error in broking budgets.

That last step deserves emphasis: new business earns, on average, roughly half its annual commission in the year it is won, because wins spread across the year. A firm budgeting INR 4 crore of new-business commission is implicitly targeting INR 8 crore of annualised wins. Stating both keeps the sales target honest and the revenue budget achievable.

Two assumptions get collapsed into one line called "rate" and then argued about at cross purposes. They move in opposite directions often enough that separating them is not pedantry.

Premium rate movement is what the insurer charges the client. It flows through mechanically: a 6 percent rate rise on a renewed account at an unchanged commission rate produces 6 percent more commission for no additional work. It is the line that moves the budget most for the least effort, and the one the firm controls least. Assume it per line, since property, motor and group health are not in the same rate cycle in any year, and state it as a range rather than a point.

Commission yield is the firm's own rate on that premium. It moves for different reasons: renegotiation with an insurer, a change in the insurer's own commission policy, or an account moving from one carrier to another with a different grid.

Mix shift is the third thing, and it moves the blended yield without either of the first two changing. If large-risk commercial business realises 8 to 10 percent and mid-market package and SME business realises 12 to 15 percent, then a firm that wins one large corporate account and loses three SME accounts can grow its premium handled while its commission income falls. Its blended yield will drop and nobody will have cut a rate.

Model these as three separate rows, each with its own owner. Combining them produces a budget nobody can explain in October when it misses, because the miss cannot be attributed to anything.

An annual number divided by twelve is not a budget. It will breach every monthly review for two quarters before recovering in a fourth-quarter rush the firm then treats as a triumph.

Broking income in India is phased by facts the firm already knows:

• April is heavy. Indian commercial renewals cluster at the financial-year boundary, and a firm with a PSU or large-corporate book may take a fifth of its annual commission in that month.
• The public-sector tender calendar is not negotiable. Where a broker is appointed through a tender, both the placement and the appointment run on the client's cycle, not the firm's plan.
• Group health follows employer cycles, which cluster but do not align with April.
• Insurer statement lags shift recognition even when the placement was on time. Phase the budget on the same basis as the ledger it is compared against, or the variance report measures the accounting policy rather than the business.

Phase from the expiring register itself, which already carries every renewal date on the book. The base phases itself. Only new business needs a judgement, and the honest one is that it lands late.

The payoff is that a monthly variance becomes information. A firm that phases evenly learns nothing from being 30 percent behind in May. A firm that phased on its own renewal calendar and is 30 percent behind in May has lost a specific account, and knows it in May rather than in December.

Every FY27 broking budget is being built against an unfinished rulebook, and the temptation is to pick the outcome that suits the plan and move on. The discipline is to state precisely what is known, what is expected, and what is neither.

What is in force. The IRDAI (Payment of Commission) Regulations, 2023 removed product-wise caps, so commission is set by each insurer's own policy. The IRDAI (Expenses of Management, including Commission, of Insurers) Regulations, 2024, effective 1 April 2024, caps aggregate insurer expense at roughly 30 percent of gross written premium for general insurers and 35 percent for standalone health insurers. The Sabka Bima Sabki Raksha (Amendment of Insurance Laws) Act, 2025, in force since 5 February 2026, restored IRDAI's power to cap distributor commissions, imposes no cap itself, and none has been made under it.

What is expected but not settled. As of the date of this post the commission overhaul consultation paper had not been published. IRDAI Chairperson Ajay Seth indicated in early July 2026 that it is expected by end-July. Everything reported as being under consideration (staggered or trail commission across the policy life, remuneration weighted toward advisory and servicing work, product-wise caps differentiated by complexity and tenure, tighter disclosure of remuneration) is a proposal. None of it is a rule, and the paper's actual contents are not knowable from the reporting. Separately, the FY26 and FY27 revisions to insurer expense limits are anticipated, not notified.

A budget's value is realised in the variance report, and a variance report that says "commission income 8 percent below plan" has told nobody anything actionable. If the model was built in the layers above, the variance decomposes into causes that each belong to a person.

Decompose every month into five components:

1. Retention variance. Accounts in the expiring register that did not renew, valued at budgeted commission. Owned by the servicing team, diagnosable by account name.
2. Rate variance. Renewed accounts where the premium moved differently than assumed. Largely market, and useful mainly as a signal for the next assumption.
3. Yield variance. Renewed accounts at the same premium but a different commission rate. Owned by whoever negotiates with insurers.
4. New-business variance, split between volume (fewer wins than planned) and timing (the wins happened but incepted later). These have entirely different remedies, and conflating them produces a sales conversation that fixes nothing.
5. Timing variance. The placement happened on plan and the statement has not arrived. Not a business problem, and it should be visible as such rather than absorbed into the other four.

Run this monthly and reforecast formally twice: in Q2, when enough of the April renewal cluster has landed to test the retention assumption, and in Q3, when the reform position should be clearer and the pipeline is either real or is not.

The point is not accuracy. On a book of any size the FY27 number will be wrong. The point is that when it is wrong, the firm can say which of five assumptions was wrong, by how much, and who is doing something about it. That is a budget. The alternative is a target with a decimal point.