Sponsored content: Stephen Kennedy, managing director at Defaqto, says there is an important distinction between the consistency of a data collection process and the extent to which observed premiums can be considered representative of the prices that would be returned in a ‘real world’ situation
UK car insurance premiums are moving upwards. Some sources are quoting year-on-year movements of up to 10% on average and almost 20% for some segments. If you operate in this market, you may well be scratching your head wondering why you have not observed significant improvements in competitiveness and conversion to reflect this.
According to Defaqto’s market price tracking, competitive quoted prices on price comparison websites are now 1.6% higher than last year. From conversations with our broad client base, this is closer to reality. So why the discrepancy between sources?
A robust and consistent methodology is fundamental to producing useful insurance market intelligence. Collecting quotations across multiple distribution channels at comparable points in time, applying systematic quality assurance and investigating unexpected movements can provide a strong basis for identifying trends and changes in competitive positioning.
There is, however, an important distinction between the consistency of a data collection process and the extent to which observed premiums can be considered representative of the prices that would be returned in a ‘real world’ situation.
UK general insurance pricing has become increasingly sophisticated. The premium returned may reflect not only the information submitted within an individual quotation, but potentially a wider range of contextual and behavioural factors. Some of these factors may be observable externally, while others form part of insurers’ proprietary pricing, underwriting, or fraud-prevention processes.
Quote history
Quote history provides a useful illustration. Repeated quotation activity over a relatively short period may contribute to an insurer’s assessment of risk or potential fraud. The effect may not be straightforward – frequency, timing, changes in submitted information and other characteristics may interact in ways that are difficult to identify through external observation alone.
The interval between quote and policy inception date can introduce a further variable. Lead time may influence pricing and may also interact with other rating or behavioural factors. As a result, apparently similar quotation exercises may not always be directly equivalent from the perspective of the insurer’s pricing systems.
This presents a broader methodological challenge for external pricing analysis. Quality assurance can establish that data has been collected correctly, consistently and in accordance with a defined methodology. It can also identify observations that depart from historical patterns. What it cannot necessarily determine is the precise mechanism responsible for the price returned.
Without visibility of an insurer’s underlying pricing and decision rules, it can be difficult to distinguish between a change in general market pricing and an outcome influenced by the circumstances in which quotes have been generated.
None of this removes the value of systematically collected pricing data. Such datasets can provide important evidence of market direction, relative competitiveness and changes over time. It does, however, encourage appropriate care in interpretation.
Consistency and rigorous validation can provide confidence in the quality of the observation. Establishing that an observed premium is comparable with the price that would have been offered to an unaffected customer is a separate question, requiring knowledge that may not be available from external observation alone.
