The MGA market has entered its proving phase 

For much of the past decade, the MGA story has been one of expansion. 

UK MGA gross written premium reached approximately £7.2 billion in 2025, around a fifth higher than the previous year, according to Howden Re. More than 300 MGAs now place over 10% of the UK general insurance market, based on figures from the Managing General Agents’ Association. MGAA members reported writing £13.2 billion of premium in 2024. 

Delegated authority has also become structurally important to Lloyd’s, accounting for an estimated 40% to 45% of market premium, depending on the measure used. 

But the conditions that supported that growth are changing. 

Commercial insurance rates fell by an average of 1.2% during the first quarter of 2026, according to the Council of Insurance Agents and Brokers. It was the first overall decline in almost nine years, ending a run of 33 consecutive quarters of increases. Lloyd’s reported price movement of minus 3.7% during 2025. 

That tension has run through our recent conversations with MGA founders, brokers, capacity providers and investors. The sector has reached record scale just as the cycle has begun to turn against it. 

A softer market does not undermine the MGA model. It does, however, test it. The next few years will expose which businesses have developed a defensible underwriting proposition and which have relied on favourable pricing, readily available capacity and continued market momentum. 

The MGA market is often discussed as though it were a single category. In practice, the businesses operating within it have been built with very different ambitions. 

Some were designed to establish a book quickly and sell within three to five years. Many launched during the hard market, when capacity was available on relatively generous terms and premium growth could disguise weaknesses elsewhere. 

Others were built to trade through a full insurance cycle. 

The distinction matters now because the balance of power is shifting. For some newer MGAs, this is the first renewal season in which capacity providers have held more leverage than the underwriting business. 

Carriers are asking harder questions about performance, reporting, reserving and exposure management. Their support is also becoming more selective. An individual MGA can lose its backing even while the delegated authority market continues to grow overall. 

Views differ on how severe the consequences will be. 

The optimistic case is that specialist MGAs should be able to outperform in a soft market. They tend to have lower operating costs, greater speed and more narrowly defined areas of expertise than large composite insurers. Those advantages should give the strongest businesses room to adapt. 

Others expect a more difficult correction. Nick Pomeroy of QRG Specialty has warned of a “day of reckoning” for MGAs that have failed to build sufficient premium or produce strong enough results. AM Best, meanwhile, changed its delegated authority outlook from positive to stable for 2026, pointing to tighter renewal economics and increased scrutiny. 

Both assessments may prove correct. 

A soft market should favour MGAs that can demonstrate underwriting discipline, retain several sources of capacity and explain exactly how their book behaves. It will be much less forgiving of businesses that cannot. 

When MGA leaders are asked what will separate the survivors from the rest, the conversation repeatedly returns to data. 

This is also where the distance between the market’s ambitions and its day-to-day reality is most apparent. 

A frequently cited estimate suggests that underwriting and actuarial teams spend close to half their time cleaning, organising and reconciling data rather than analysing it. The figure appears in research from consultancies and technology providers, including hyperexponential and Oliver Wyman, although there is no single definitive market-wide study. 

The broader evidence is difficult to ignore. The London market’s monthly bordereaux process still takes approximately 48 days, with data quality identified as one of the main causes of delay under the Blueprint Two programme. 

Many MGA platforms remain disconnected from claims and accounting systems. Email is still the default means of transferring information between brokers, MGAs, insurers and reinsurers. 

The commercial consequences are immediate. Claims information that reaches the underwriting team three or six months late has limited value for pricing the next risk. If the feedback loop between claims and underwriting never closes, decisions on pricing, reserving and portfolio management are inevitably being made with an incomplete picture. 

Capacity providers increasingly see reliable data as an indication of how easy an MGA will be to manage. Clean, timely reporting reduces the amount of intervention required from the carrier. It also gives the insurer greater confidence that the delegated portfolio is being understood and controlled. 

That can translate into broader authority, better terms and a more durable relationship. 

There is an important qualification. Some specialist MGAs are not struggling with poor-quality data; they are operating in classes where meaningful historical data barely exists. 

An underwriter covering vacant properties, recycling plants or asylum accommodation cannot simply purchase a conventional dataset or plug into a standardised platform. The technology market’s vision of a fully connected, “360-degree MGA” is usually designed around higher-volume books with repeatable data. 

Specialist underwriting still depends on judgement. Better infrastructure can support that judgement, but it cannot manufacture a credible loss history where none exists. 

The MGA market’s discussion of artificial intelligence has become noticeably more practical. 

Much of the early excitement has given way to a focus on submission processing, document extraction, workflow automation and claims administration. KPMG’s latest insurance CEO survey found that 67% of leaders expect to see a return on their AI investment within one to three years, compared with 21% a year earlier. 

The more useful distinction is between AI as a cost tool and AI as an underwriting tool. 

Most MGA implementations still sit in the first category. Automation is being used to remove repetitive work, process submissions more quickly and control operating costs as market conditions soften. 

That work has value. But it is not the same as demonstrating that AI has improved risk selection, sharpened pricing or reduced claims leakage. 

Capacity providers are unlikely to be persuaded by the number of models an MGA has deployed. They will want evidence that those models have changed decisions and produced better results. 

For smaller MGAs, the obstacles remain stubbornly practical. The cost of many AI products is difficult to justify at their scale. They may also lack the data, technical expertise and implementation capacity needed to turn a pilot into a dependable part of the underwriting process. 

Perhaps the more interesting opportunity sits on the other side of the balance sheet: AI as a risk to insure rather than merely a technology to use. 

AI liability is beginning to resemble the early cyber market. The potential exposures are large, the available loss history is thin and many of the relevant legal responsibilities remain unsettled. 

Armilla launched what it described as the first standalone AI liability policy in April 2025, supported by Lloyd’s syndicates including Chaucer. By early 2026, it had extended available limits beyond $25 million per organisation. Munich Re has offered its aiSure product for several years, while Lloyd’s Lab alumnus Testudo is writing AI risks with capacity from Atrium and QBE. 

Research from Gallagher Re, MIT and Testudo identified more than 700 AI-related lawsuits in the United States between 2020 and 2025. 

This is precisely the kind of emerging, technically difficult risk that specialist MGAs should be equipped to address. But it will require more than repackaging a technology warranty. The firms that establish this market will need to understand how models fail, where liability accumulates and which exposures can be measured with enough confidence to insure. 

The flow of US capital into the London MGA market is increasingly taking the form of acquisition rather than minority investment. 

Acrisure agreed to acquire technology-led MGA Vave from Canopius in December 2025, completing the transaction in April 2026. Vave, which reportedly produces more than 10,000 algorithmic quotes a day, will become part of Acrisure’s US MGA platform. 

Brown & Brown also moved to acquire UK MGA Pardus Underwriting during the same period. 

These transactions form part of a wider consolidation story. In US wholesale broking, each of the three largest groups places more than $16 billion of premium, while the five largest control roughly half of the market. 

The attraction is not difficult to see. MGAs provide access to distribution, underwriting talent, specialist products and fee income without requiring the acquirer to hold all the associated insurance risk. 

This reflects a broader separation between underwriting expertise and risk capital. Fronting carriers, risk exchanges and third-party capital structures are making it easier for underwriting, distribution and balance-sheet capacity to sit in different organisations. 

Accelerant is among the clearest expressions of that model. Following its public listing in July 2025, the business has continued to position itself as a capital-light platform connecting specialist underwriters with risk capital. 

As the separation becomes more established, the MGA is a natural home for underwriting expertise. But the model only works when information can move reliably between the parties. MGAs may be capital-light, but they cannot afford to be data-light. 

The same forces are reshaping the talent market. Senior underwriters, product leaders and distribution executives are leaving large insurers and brokers for more focused MGA businesses offering greater autonomy and a closer relationship between individual performance and reward. 

The job itself is changing too. Calculation and administration can increasingly be supported by technology. Judgement, portfolio management and relationships are becoming more valuable, not less. 

Lloyd’s is sending two messages to the delegated authority market. 

The first is that MGAs are essential to its growth, distribution and access to specialist risks. The second is that the market intends to scrutinise them much more closely. 

Rachel Turk, Lloyd’s chief underwriting officer, has described a “laser focus” on delegated authority oversight, intended to prevent poorly controlled MGAs from damaging market performance. She has also made clear that new entrants face a high bar, including the ability to provide the data required for proper oversight. 

That burden is felt most sharply by smaller MGAs. They are often required to produce reporting and governance comparable with that of much larger insurance organisations, but with a fraction of the people and resources. 

At the same time, Lloyd’s is encouraging new products and more advanced forms of underwriting. 

The market increased the allowance for its innovation ICX risk code from 2% to 5% of premium in 2024. It continues to run the Lloyd’s Product Launchpad for emerging risks, while the Lloyd’s Market Association has identified enhanced underwriting as a strategic priority. 

Algorithmic and smart-follow models are already estimated to account for around 7% of Lloyd’s premium. 

The direction is understandable. Lloyd’s wants the growth, specialist expertise and product development that MGAs can provide, without accepting weak controls or opaque portfolios. 

The unanswered question is whether its oversight framework can distinguish effectively between an MGA taking a poorly understood risk and one creating a genuinely new class of business. 

That distinction will become increasingly important because the biggest opportunity for MGAs does not lie in distributing existing products more efficiently. It lies in insuring risks that the conventional market has not yet found a way to cover. 

Swiss Re estimates that between 80% and 90% of catastrophe losses in emerging economies remain uninsured. The gap may be wider for newer and less tangible exposures. Intangible assets now represent close to 90% of the value of the S&P 500, but only a small proportion of that value is insured. 

Much of what the insurance industry describes as innovation is still distribution improvement: selling a familiar policy through a better channel or with a faster journey. 

Creating a new insurance market is harder. 

There are signs of progress. Carbon-credit insurance has begun to move from experimentation into commercial use, with CFC, Oka and Artio among the first insurers approved to provide CORSIA-compliant cover in late 2025. Products addressing intangible assets and AI exposures are also beginning to emerge. 

These markets remain small. But they show what the MGA model can do at its best: bring together specialist knowledge, responsive capacity and focused distribution to insure risks that larger organisations have struggled to address. 

The MGA sector is not approaching the end of its growth story. It is entering a more demanding part of it. 

Rapid expansion has concealed a widening gap between businesses that can demonstrate the quality of their underwriting and those that cannot. Softer pricing, stricter capacity negotiations and greater regulatory scrutiny will make that difference harder to ignore. 

The strongest MGAs will not necessarily be those with the fastest premium growth. They will be the businesses that can explain their portfolios, connect claims information to underwriting decisions, maintain more than one source of capacity and show that their technology improves outcomes. 

The most ambitious will go further. They will use the MGA model not simply to distribute established products, but to build credible markets around risks that remain largely uninsured. 

Those questions will shape the discussion at InsTech’s full-day MGA event in London this September, bringing together the founders, brokers, capacity providers and investors dealing with them in real time.