Commission Concentration Risk: When One Insurer Is Half Your Income

A broking firm's risk register is usually a competent document. It carries professional indemnity exposure, cyber, key-person dependency, regulatory breach, client credit. What it almost never carries is the fact that one insurer pays 45 percent of the firm's income.

The omission is not carelessness. It is that the exposure does not present as a risk. It presents as the firm's best relationship: the carrier that gives the fastest quotes, the underwriter who takes the call at 8pm, the grid the firm negotiated up two years ago and now defends. Every incentive inside the firm points toward deepening it. Nobody has ever been thanked for placing an account with a slower insurer at a lower commission in order to reduce a Herfindahl index.

So the concentration accumulates by a thousand sensible decisions and becomes visible only when it is expensive. The trigger is rarely dramatic. An insurer revises its commission structure for a line, withdraws appetite for a class it has decided is unprofitable, or restructures its distribution team and the relationship that made the arrangement work walks out. None is a catastrophe. Each moves a large fraction of a firm's revenue in a quarter.

This is a risk framework rather than a survival plan. What a small firm should do to stay in business under commission reform is a strategic question worked through in the playbook for small and regional brokers. This is the narrower discipline: measuring revenue concentration, testing it against defined scenarios, setting a tolerance the board can hold, and planning for the loss of a counterparty the firm depends on without ever having decided to.

"We're a bit heavy with one carrier" is not a measurement. Two numbers turn it into one.

The concentration ratio (CR1, CR3) is the share of commission income from the largest counterparty, and from the largest three. It is trivial to compute and easy to explain to a board, which is most of its value.

The Herfindahl-Hirschman Index is the sum of the squared percentage shares of every counterparty. Squaring is the point: it weights large positions far more heavily than small ones, so the index rises sharply as the book tilts. Take a firm placing with six insurers at income shares of 45, 20, 15, 10, 6 and 4 percent. The HHI is 2,025 + 400 + 225 + 100 + 36 + 16 = 2,802.

The number becomes intuitive through its reciprocal. Divide 10,000 by the HHI and you get the effective number of counterparties: here, 10,000 / 2,802 = 3.6. The firm has six insurer relationships and the diversification benefit of three and a half. That single sentence does more work in a board meeting than any table.

For a reference scale, competition authorities screening mergers commonly treat an HHI below 1,500 as unconcentrated, 1,500 to 2,500 as moderately concentrated, and above 2,500 as highly concentrated. Those thresholds were built for product markets, not broking revenue, so they are a borrowed yardstick rather than a standard. They are still useful: a firm at 2,802 should know that on any conventional reading, its revenue base is highly concentrated.

Compute it quarterly on trailing twelve-month commission, and compute it on commission income, never on premium placed. Premium share and commission share diverge whenever grids differ, which is always. A firm can look diversified by premium and be dangerously tilted by the thing that actually pays for the office.

Insurer concentration is the axis firms think of first. It is not the only one, and it is not usually the worst.

• By insurer. Counterparty concentration proper: which carriers pay the income.
• By client. A firm where one corporate group is 30 percent of commission has an exposure that does not care how many insurers it uses. Client concentration is often sharper than insurer concentration and is more often invisible, because the account is spread across five policies in three lines and nobody aggregates it to the parent.
• By line of business. A firm that is 60 percent group health is exposed to a single pricing cycle, a single regulatory conversation and a single claims-inflation trend, regardless of how many carriers write it.

Compute the HHI on each axis. The revealing exercise, though, is the one nobody runs: the cross-tabulation. Concentration on one axis is a manageable exposure. Concentration that lines up across axes is a single point of failure wearing three disguises.

The pattern to look for is a cell that holds an outsized share of income: one insurer, for one line, for one client group. A firm at 45 percent with an insurer, 60 percent in group health, and 30 percent with one corporate group may find that a single cell (that insurer's group health book for that client) is 22 percent of total commission and would disappear on any one of three unrelated events. The marginal HHIs looked survivable. The cell does not.

Scenario analysis fails when the scenarios are apocalyptic, because nobody plans against an outcome they consider fanciful. The useful scenarios are the boring ones with meaningful probability inside a planning horizon.

1. The grid moves. The insurer revises its commission structure for a line and the firm's realised yield on that book drops two points. No relationship ends. Nothing is announced. Model it: take the firm's income from that insurer in that line and multiply by the yield delta. On a concentrated book this is often a larger number than the firm's annual profit, arriving with a quarter's notice and no negotiating position, because the firm has nowhere to move the business.
2. The appetite closes. The insurer exits a class or tightens underwriting to the point where it is no longer competitive. The firm's clients must be re-placed at renewal. The revenue may survive, but at a different carrier's grid, on a different servicing model, with a re-broking effort on every affected account inside one renewal cycle. Cost is real; timing is compressed.
3. The relationship leaves. The underwriting or distribution head who made the arrangement work moves. This is the most common trigger and the least modelled, because it sits in nobody's risk taxonomy. It belongs in this one.
4. The counterparty fails. Low probability, high severity, and structurally different from the others because it strands receivables as well as revenue.

For each, the output is one number: revenue at risk over twelve months, stated in rupees and as a share of income, with the profit impact after variable costs. Not a heat map. A number a board can compare against the firm's reserves.

The fourth is a different animal, putting the income stream and the balance sheet at risk at once. The receivable side of it, meaning commission already earned and unpaid, is worked through in the piece on insurer insolvency and broker receivables risk. For concentration purposes the point is narrower: a firm at CR1 of 45 percent has bet its continuity on one counterparty's solvency, and should know it has, even if it judges the probability remote.

Two firms can hold identical HHIs and face entirely different risks, because concentration only bites to the extent the position cannot be moved. Severity is a function of substitutability, and substitutability is testable.

Ask, for the largest insurer position, four questions with dated evidence rather than opinions:

1. Is there an alternative carrier with appetite for this business? Not in principle. Named, and quoting on this class within the last twelve months.
2. What is the yield on the alternative? If the incumbent's grid is two points above every alternative, the position is not substitutable at the same economics, and the concentration is real regardless of how many carriers exist.
3. How long does the switch take? Re-broking is bounded by renewal dates, so the practical answer is one full renewal cycle, and longer where the client has a multi-year arrangement or a lender-mandated carrier.
4. Does the client have a view? Corporates with an insurer-security policy, a captive fronting arrangement or a financing covenant naming the carrier cannot simply be moved because the broker's risk register asks for it.

The answers convert the abstraction into a schedule. A 45 percent position genuinely re-placeable at comparable yield inside two renewal cycles is manageable. A 20 percent position no other carrier writes at a workable rate is more severe, and it is the one most firms would not have flagged.

Test it rather than assert it. The cheapest test is to place a small tranche of the concentrated book elsewhere and find out what it costs. A firm that has never placed outside its dominant carrier will discover the price at the worst possible moment.

Most broking firms have no stated position on how much revenue concentration they will accept, which means the answer is whatever last year produced. A usable statement has four parts, and its value is entirely in the fourth.

1. The metric. CR1 and HHI on trailing twelve-month commission, computed on each of the three axes and on the cross-tab, reported quarterly.
2. The tolerance. A number the firm is willing to defend. Something in the region of CR1 at 30 percent and an effective counterparty count of five is a defensible starting position for a mid-market commercial firm, but the figure matters less than the fact that one exists and was chosen deliberately.
3. The escalation trigger. A level at which the position must be reported and answered rather than noted. Breaching tolerance should oblige a named plan with dates, not a paragraph in the quarterly pack.
4. The exception process. The part that determines whether the statement is real. Concentration is built by good decisions, so a firm that refuses profitable business to hold a ratio will simply stop following the statement. What the statement must do instead is make exceeding the tolerance a decision somebody makes, with a name attached and a review date, rather than a drift nobody owns.

For a firm below the INR 25 to 50 crore revenue band where broking economics start to support the fixed cost of compliance, technology and specialist staff, this arithmetic is harsher. A smaller firm has less absolute headroom, so the same concentration ratio carries more existential weight, and the tolerance should be tighter precisely where commercial pressure to concentrate is strongest.

Business continuity planning in broking firms addresses premises, systems and people. It rarely addresses the loss of a counterparty, which is the disruption most likely to actually threaten the firm.

A continuity plan for a concentration event answers four questions before the event:

• Detection. What would tell the firm early? Quote-to-bind ratios falling with a carrier, referral rates rising, an appetite letter, a senior departure, a change in the insurer's own results. These are observable months before revenue moves, and none of them is watched unless somebody is asked to watch them.
• The re-placement queue. If the position had to move, in what order? The queue is built from renewal dates, not from client importance, because renewal dates are the only thing that governs when the business can actually move.
• The capacity check. Does the firm have the placing and servicing capacity to re-broke that volume inside one cycle, on top of business as usual? Usually not, and knowing the shortfall in advance converts a crisis into a hiring decision.
• The cash bridge. Revenue moves faster than the cost base. A firm losing a quarter of its income in a renewal cycle needs to know, before it happens, what its facility headroom and covenant position look like on the way down.

None of this prevents the event. It changes what the event costs, which is the only thing continuity planning has ever done.

Bring three things to the board: the HHI on each axis with the effective counterparty count, the revenue at risk under the four scenarios in rupees against the firm's reserves, and the substitutability evidence for the largest position. Then propose a tolerance and let the board choose. A board that decides to run at CR1 of 45 percent with the number in front of it has made a decision. A board that has never seen the number has made the same bet without knowing, which is the only indefensible version.