The MGA’s scarce resource is not capacity. It is credible attention.

The MGA’s scarce resource is credible attention

The growth decision that looks easiest on a spreadsheet can be the hardest to absorb in practice. A managing general agent may have capacity for another programme, access to distribution and enough shared infrastructure to process the business. Yet the addition can still weaken the platform if senior underwriters and executives cannot interpret emerging evidence and intervene quickly when results depart from plan.

That is the tension facing expanding MGA platforms. Their apparent constraint is often capacity: whether carriers will provide enough of it, on acceptable terms, to support growth. But once a platform has multiple programmes and partners, another constraint becomes more consequential: the supply of informed management attention. This is not time in calendars or the volume of management reporting. It is the ability of people with relevant judgement and authority to identify an important deviation, decide what it means and act before the consequences compound.

The available evidence supports a narrow but important claim. AM Best’s process for assessing delegated underwriting authority enterprises places governance, talent and underwriting capability within the foundation on which such businesses are assessed. That does not prove that management attention is the binding constraint at every MGA, nor that larger programme portfolios necessarily perform worse. No operating data supplied here link programme count, executive workload and underwriting outcomes. The distinction matters. The case for treating attention as scarce rests not on a universal statistical relationship, but on the operating logic of delegated underwriting: governance and capability have limited value if they cannot be brought to bear when evidence demands a decision.

Capacity can be contracted. Credible attention must be built.

The difference between information and intervention

Most platforms can produce more information more cheaply than they can produce better decisions. A new dashboard can place loss emergence, submission flow, rate movement or bordereaux quality in front of a wider audience. It cannot determine whether a variance is noise, an early sign of drift or evidence that the original underwriting thesis was wrong. Nor can it reconcile conflicting interpretations among programme leaders, actuarial functions, claims teams and capacity providers.

This is where “attention” requires qualification. Management attention is not inherently useful. An executive reading a report without sufficient context may add scrutiny but not judgement. A committee discussing a programme without the authority to change its course may consume time while preserving the status quo. Credible attention combines three elements: access to relevant evidence, the competence to interpret it and the authority to intervene.

That definition has concrete consequences for growth. When an MGA approves another programme, it is not merely adding premium, policies and processing demand. It is increasing the number of underwriting hypotheses that may need to be tested, exceptions that may need to be judged and relationships that may need to be managed. These demands will not arrive evenly. Several programmes can require intervention at once, particularly when they share exposure to the same market shift, distribution behaviour or operational dependency.

The resulting constraint is therefore episodic rather than constant. A platform may appear comfortably staffed during routine months and become attention-constrained when adverse evidence arrives across several portfolios. Average workload will conceal the risk.

Attention debt

A more useful operating concept is “attention debt”: unresolved decisions multiplied by the time they remain unresolved and the portfolio consequence if the concern proves valid.

This is not a proposed accounting measure. It is a management framework for distinguishing a crowded agenda from a dangerous one. Ten minor reporting questions may matter less than one unresolved concern about selection, claims development or delegated authority. The purpose is not to convert judgement into a spurious numerical score. It is to force three questions into the same discussion: What remains undecided? How long has it remained so? What can happen while the organisation waits?

Time is especially important because delay is not neutral. While management debates a response, a programme may continue binding business, renewing accounts or operating under assumptions that are no longer trusted. The eventual decision may be correct but late. In that case, the failure lies less in analysis than in the platform’s ability to mobilise intervention capacity.

Attention debt also exposes a misleading feature of conventional escalation processes. A programme can comply with reporting requirements while accumulating unresolved judgement calls. Conversely, a programme with several escalations may be well controlled if those escalations produce timely decisions. The count of issues says less than their age, consequence and disposition.

Growth should therefore be approved against intervention capacity, not only processing capacity. Before adding a programme, management should ask who would investigate an adverse signal, who could challenge the programme’s leadership, who would decide on remediation and what existing work would be displaced. If those answers depend on the same few executives already serving the rest of the portfolio, the platform has identified a concentration risk even if every operational service-level measure is green.

The strongest countercase

The strongest objection is that attention scarcity is largely a process-design problem. Shared services can standardise reporting, compliance and controls. Automation can triage exceptions and direct human review towards the cases most likely to matter. Clear authorities can push decisions closer to programme teams. On this view, oversight should scale with premium if management designs the system properly.

This counterposition deserves more than a ritual acknowledgement. Better process can remove avoidable demands on senior people. Standard definitions can reduce arguments about data. Automated controls can identify exceptions sooner. Delegated thresholds can prevent routine decisions from travelling unnecessarily upward. A platform that treats every variance as an executive matter has created its own bottleneck.

The argument becomes weaker, however, when it assumes that all consequential decisions can be standardised in advance. The difficult cases are difficult precisely because the evidence is incomplete, the relevant pattern is changing or the available remedies impose competing costs. Automation may identify an outlier; it does not establish whether the outlier invalidates the underwriting thesis. Shared services may improve consistency; they do not eliminate the need for accountable judgement.

The practical conclusion is not that technology cannot scale oversight. It is that technology scales observation and routine control more readily than it scales interpretation under uncertainty. Process design can conserve credible attention, but it cannot make the resource infinite.

Forums that alter decisions

This distinction should change how MGA leaders assess management forums. The relevant question is not whether a programme appears on a committee agenda, but whether the forum changes decisions.

A meeting that exchanges information without assigning action, authority or a deadline may create the appearance of oversight while adding to attention debt. Repeated requests for further analysis can become a socially acceptable form of deferral. The incentives are understandable. Programme leaders are rewarded for growth and continuity. Central functions may prefer additional evidence before endorsing a disruptive intervention. Executives may hesitate to challenge a programme that carries strategic importance or supports a valuable capacity relationship.

These incentives can produce second-order effects. As reporting expands, programme teams may optimise for presenting compliance rather than surfacing ambiguity. Executives may become overexposed to frequent but low-consequence exceptions, making it harder to distinguish a genuinely material signal. The most trusted individuals may receive still more escalations, increasing dependence on them and making the platform less resilient.

A useful review of governance would therefore trace decisions rather than meetings: which signals changed appetite, terms, distribution, claims handling or the pace of growth; how long that took; and which forum had the authority to make it happen. This does not presume that frequent intervention is desirable. Good oversight may confirm that no change is needed. But that conclusion should be explicit, attributable and timely.

What would prove the thesis wrong

The attention thesis should be testable. If shared services and automation allow oversight to scale without loss of decision quality, expanding platforms should be able to add programmes while keeping the age of material unresolved decisions stable or falling. They should not become more dependent on a small number of senior decision-makers. Escalations should be resolved within defined periods even when several programmes require review simultaneously.

The opposite pattern would support the thesis. As programme count rises, material decisions would remain open longer, the same executives would appear in a growing share of escalations and management forums would request repeated analysis without changing course. Adverse interventions would increasingly occur after, rather than shortly after, the first credible warning.

These are predictions, not findings from the evidence supplied. They require portfolio-level operating data that are not available here. An MGA could test them internally by examining decision histories, escalation age and concentration of authority. The result may show that its constraint is weak process rather than insufficient attention. That would still be useful: the framework is meant to identify the bottleneck, not predetermine it.

Questions before the next programme

The next growth discussion should go beyond whether capacity and infrastructure are available. If two programmes deteriorate at once, who has the time, knowledge and authority to decide what changes? Which current decisions would wait? How much material attention debt is already outstanding, and where is it concentrated? Which governance forums have changed underwriting or operating decisions in the past year, and which have merely circulated information? What evidence would show that automation is improving judgement rather than producing more alerts? And before approving another programme, can the platform name the intervention capacity reserved for it?

An MGA can purchase systems, organise shared services and negotiate additional capacity. Its harder task is to ensure that credible attention reaches the right programme before drift becomes damage.