Leverage and reinsurance effects on loss reserves in the United Kingdom’s property–casualty insurance industry

Abstract

We examine the relation between loss reserving errors, leverage and reinsurance in the UK’s property–casualty insurance industry. We find that financially weak insurers under-estimate reserves to reduce leverage, and so pre-empt costly regulatory scrutiny. However, at very high leverage, insurers over-reserve, suggesting a non-linear relation between leverage and reserving policy. We also investigate whether monitoring by reinsurers reduces reserving errors, and find that highly reinsured insurers are less likely to make loss reserve errors. However, the use of proportional reinsurance does not affect loss reserve accuracy.

Keywords:

• leverage
• reinsurance
• loss reserves
• insurance
• UK

JEL Classification:

• G22

Acknowledgements

The suggestions of Iain Clacher, Joy Jia, Clive Lennox, Ola Oyekan, Steve Pavelin, Joan Schmit, Vineet Upreti, two anonymous referees, and Mark Clatworthy (Editor) are very much appreciated. The paper also benefited from the comments of seminar participants at the University of Lancaster and the University of Auckland, as well as conference attendees at the 2013 American Risk and Insurance Conference, Washington, DC. However, the usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Purnanandam (2008) suggests that the risk to investors of financial distress/bankruptcy increases when corporate debt-to-asset ratios reach about 90% – i.e. ‘at very high’ leverage levels. Our analysis also indicated that across all lines of insurance business examined, over-reserving tended to increase when claims-to-asset ratios reached the 90% threshold.

2 The insurance underwriting cycle begins after periods of large losses when premium rates rise thereby increasing profits, and attracting inflows of capital into the property–casualty insurance sector. Thereafter, excess capital induces price cutting in competitive insurance markets (Cummins and Danzon 1997).

3 From 1 April 2013 the statutory supervision and regulation of UK insurers has been conducted by the Prudential Regulation Authority (PRA), while matters of insurance market operations are regulated by the Financial Conduct Authority (FCA). The PRA is part of the Bank of England, and the FCA is an independent regulatory body which is accountable to HM Treasury.

4 Unlike US insurers, which report using generally accepted insurance accounting principles (US GAAP), major UK insurance firms have since January 2005 reported their business activities in line with International Financial Reporting Standard 4 (IFRS 2004), which changed in the accounting and financial reporting rules for UK and other European publicly listed insurers, including disclosure of reserve movements. However, IFRS 4 was only applicable towards the tail-end of our analysis period (1991–2005), and unlikely to significantly affect our analysis. Also, since 1 January 2015 non-publicly listed UK (and Irish) insurers report under accounting standard FRS 103 – a standard that is largely consistent with IFRS 4.

5 Public media sources are replete with reports that the UK insurance regulator has concerns about the release of prior years’ reserves to meet short-term profit and dividend targets. For example, see the Financial Times article of 14 November 2014, ‘Regulators to probe insurers over drawing down reserves’. Such regulatory unease often results in UK insurers being compelled to make reserve additions in order to meet statutory solvency targets and ease regulatory concerns about their future financial condition.

6 The under-investment incentive is an agency cost of debt problem that relates to the risk, particularly in highly leveraged states, that shareholders may not reinstate productive assets following a severe loss event as the gains from reinstatement accrue to debtholders rather than themselves. In such a situation, shareholders may exercise their ‘default’ put option under limited liability and voluntarily liquidate the firm. However, Garven and MacMinn (1993) note that the under-investment problem can be mitigated by (re)insurance contracts in that the proceeds from (re)insurance claims can be used to reinstate impaired assets after unexpectedly severe losses, thereby, minimising the risks (and costs) of financial distress and/or bankruptcy.

7 Although proportional and non-proportional treaties are used across property and liability lines of business, non-proportional treaties are commonly used in catastrophe and liability lines of insurance (Winton 1995). However, we control for the possibility that choice of reinsurance treaty could be driven by line of business in our regression analyses.

8 To further test for multicollinearity we compute variance inflation factors (VIFs) for our independent variables. All VIF values are well below the benchmark value of 10, with the largest being 2.46 (e.g. see Kennedy 2003). Therefore, bias due to multicollinearity is unlikely to be problematic in this study.

9 Ai and Norton (2003) note that interpreting coefficient estimates on interactions can be problematic, especially in non-linear (e.g. logistic) specifications, unless the mediating effect is estimated using consistent cross-differences that account for all model covariates. We thus used the standard cross-partial derivatives estimator to deal with this issue.

10 The five main lines of business written in the UK’s property–casualty insurance market are: personal accident & health, motor, property, legal liability, and miscellaneous & pecuniary insurance.

11 This percentage is derived from multiplying the standard deviations for the KFS and Weiss under-reserving errors in Table 1 with the relevant coefficient estimates in Tables 3 and 4.

Additional information

Funding

The funding and logistical support provided to Elena Veprauskaite by the UK Government Actuary's Department is gratefully acknowledged.