Understanding HMRC's Risk-Based Approach to Selecting Enquiries

Connect draws on an unusually wide range of data sources for a government system — bank interest reports, Land Registry property transactions, DVLA vehicle records, online marketplace seller data, and overseas exchange-of-information reports from other tax authorities. The system's core function is comparing all of this against the income and gains actually declared, flagging cases where the two pictures don't obviously match.

This doesn't mean every enquiry is triggered by an actual discrepancy — a genuinely unusual but entirely correct return can still be selected simply because it looks different from the norm for its category. HMRC also retains a smaller element of random selection specifically to maintain overall compliance pressure across the system, rather than relying purely on risk indicators that could, in theory, be anticipated and worked around.

HMRC holds data on typical profit margins, expense ratios and turnover patterns for very granular categories of business, from specific types of takeaway food outlet to particular trades within construction. A return that sits well outside the normal range for its specific category, in either direction, becomes statistically more interesting to the risk-scoring system, even before any specific discrepancy has been identified.

Aspect enquiries are usually less onerous to deal with, since the scope is narrower and more clearly defined from the outset, though a poorly handled aspect enquiry can sometimes widen into a fuller review if the initial answers raise further questions.

How far back HMRC can look also depends on risk classification rather than being fixed for every case. The standard enquiry window is a fixed period after a return is filed, but where HMRC believes income has been under-declared carelessly or deliberately, separate and considerably longer time limits apply through the discovery assessment rules.

This is one reason an enquiry that starts narrow can occasionally end up covering several additional tax years once it's underway, particularly if the initial questions uncover a pattern rather than an isolated error.

A meaningful proportion of enquiries HMRC opens conclude with no changes required at all — the risk-based system is designed to flag statistical outliers, not confirmed problems, so being selected is a long way from being found to owe anything.

That said, the process of being enquired into is rarely enjoyable regardless of the outcome, which is exactly why well-organised, contemporaneous records matter: they're what turns a stressful, drawn-out enquiry into a straightforward one that closes quickly with no adjustment needed.

None of the following guarantees you'll never be selected — some element of random and campaign-driven selection is unavoidable — but each one reduces the number of genuine statistical outliers your own return is likely to produce.