By Charl Cronje, Partner, Lane Clark & Peacock LLP
1. The ultimate view
The ultimate view is about how much the settlement value of reserves will differ from the best estimate today. This is the view normally used in capital models. This typically involves statistical methods, supplemented by deterministic scenarios. The problem with the ultimate view is that it is a theoretical number, likely to emerge over a long period of time, making it hard to validate and to use as a practical management tool. Long before we reach the ultimate development point, todays reserves will have been overtaken by reserves on new business.
Another weakness is that, although firms pay significant attention to the 1 in 200 probability ultimate reserving risk (because it may drive regulatory capital), the quality of the modelling at lower percentiles can be variable. The best firms manage this by testing lower percentiles against scenarios to ensure that the entire distribution appears plausible. A key part of managing the entire distribution is assigning some “ownership” of the lower percentiles to the reserving team, ensure that results are consistent with the uncertainty modelling done as part of regular reserving work.
2. The one-year view
The one-year view is arguably the most important from a management perspective. Large deteriorations over a financial year can adversely impact a firm’s share price and market reputation, and we’ve seen examples of this leading to the exit of senior management or turning a firm into an acquisition target.
Most firms measure one-year risk by applying a “recognition factor” to ultimate risk. The thinking is that only a certain portion of the ultimate variability in reserves will become apparent over the next year. For short-tail classes of business, the recognition factor may be relatively high eg 90% or more. For longer-tailed classes the recognition factor may be much lower, eg 50-60%. The problem is that, by linking one-year risk to the ultimate risk in this way, the weaknesses of the ultimate approach above similarly affect the one-year view.
The best firms have a library of reserve deterioration scenarios linked to an appropriate set of risk drivers like inflation, economic activity, legal precedents, changes in development patterns, case reserving strength and more. They regularly update their views on the importance and uncertainty of each of the risk drivers and then ensure that the 1-year view of reserve risk appropriately captures these. In my view, this is one of the most powerful tools for staying on top of reserve risk and avoiding surprises.
3. The reasonable range of best estimates
Reserving actuaries are normally asked to produce best estimate reserves, which represent the mean of the distribution of potential outcomes. Because of the uncertainties in claims development and in the reserving process itself, there tends to be a range of best estimates that could be considered reasonable, and this can be surprisingly wide on some types of business like motor large injury. In my experience, the range of reasonable best estimates is an under-explored area. I’ve seen examples of where firms have been surprised by how different their internal best estimate of reserves can be from that produced by an external reserve review, a regulatory review or due diligence from a potential acquirer. I’ve also seen examples of where firms thought they were in the middle of the range and then, a couple of years later, discovered that they had drifted near the bottom of the range, leading to difficult discussions about reserve strengthening.
So, what should firms do? First, seek to quantify the size of the reasonable range of best estimates. This is usually done by considering alternative plausible sets of reserving approaches (business segmentation, development patterns, initial expectations, reserving methods, inflation allowances etc). This can involve a lot of work. Tools like LCP InsurSight can industrialise the process by considering potentially hundreds of combinations of alternative reserving assumptions, together with quality and bias measurements for each, to help articulate the reasonable range more scientifically.
The best firms don’t stop there. They consider whether the range itself is moving up or down, or getting wider or narrower (again by linking to views on risk drivers). And, most importantly, they actively consider what point in the range they want to target, to maintain consistent best estimate reserving over time.
What next?
My key suggested takeaways are as follows:
- Consider whether the lower percentiles of ultimate reserve risk are consistent with the view of uncertainty from the reserving team as well as common-sense scenarios that are recognisable to management.
- Develop your library of one-year reserve deterioration scenarios, linked to appropriate risk drivers which are regularly reviewed. Use this to challenge the one-year view of risk.
- Decide where you want to be in the range of reasonable best estimate reserves and put in place the right measures to enable you to track and report on this over time.
Looking forward to talking about the contribution of claims handling to reserving risk next time!
|
