The MGA model in one page
An MGA sits between producing brokers and the carriers whose capital backs the risk. Unlike a retail or wholesale broker, which places business for a client, the MGA acts for the carrier: it underwrites, prices, binds and issues within an agreed mandate, as if it were the carrier's own underwriting desk for that programme.
The model thrives in specialty and niche lines — professional liability for a specific profession, coastal property, cyber for small business, equine, cannabis — where deep expertise matters more than balance-sheet scale. The carrier rents the expertise; the MGA rents the paper and the capital.
Binding authority: the contract that defines everything
The delegated authority agreement (in the Lloyd's market, a binding authority agreement with a coverholder) is the constitution of the relationship. It specifies the classes and territories the MGA may write, maximum limits per risk, mandatory exclusions, rates or rating guidelines, risks that must be referred to the carrier before binding, and the overall capacity — the maximum premium the MGA may write in the period.
Everything the MGA binds inside that envelope is the carrier's risk, automatically. That is precisely why carriers audit programmes intensively: a risk bound outside authority is a dispute waiting to happen, and a pattern of them is how programmes die. The operational question for every MGA is whether authority is enforced at the point of binding or discovered at the annual audit.
Bordereaux: the reporting backbone
A bordereau is the periodic detailed statement the MGA delivers to its carrier: the premium bordereau lists every policy bound — insured, coverage, limits, premium, commission — and the claims bordereau lists every claim with its status, reserves and payments. Monthly is the norm; formats are agreed per contract.
Bordereaux are how the carrier books the business, reserves the claims, pays reinsurers and satisfies its own regulators. Late, incomplete or inconsistent bordereaux are the single most common friction in delegated authority — and because many MGAs still assemble them by hand from spreadsheets each month, errors are structural, not incidental.
Where programmes go wrong operationally
The recurring failure modes are mundane: a policy bound above the per-risk limit because the check was manual; capacity tracked in a spreadsheet updated weekly while binding happens daily, so the programme overshoots; referral-required risks bound without referral because the trigger lived in a PDF nobody re-read; bordereaux submitted late because month-end assembly takes a week.
Each failure taxes the relationship. Carriers price opacity into their terms — lower commissions, tighter authority, more exclusions — or exit the programme entirely. In a market where capacity is the scarce resource, operational credibility is commercial leverage.
The modern MGA operating stack
The fix is to make the authority agreement executable. When limits, exclusions, referral triggers and capacity live as configuration in the policy admin system, every bind validates against them in real time; out-of-authority risks are blocked or routed to referral, and capacity utilisation is always current.
Bordereaux then stop being a month-end project: they generate from the same live policy and claims data, validate against the agreed format and tolerances, and submit on schedule. MGAs that operate this way can show carriers a live, accurate picture of the programme at any moment — and that transparency, more than any pitch deck, is what earns bigger pens and better terms at renewal.