Reinsurance and Capital Structure: Evidence From the United Kingdom Non-Life Insurance Industry

Abstract

ABSTRACT Using a data set consisting of statutory returns of U.K. non-life insurers from 1985 to 2002, I find that insurers with higher leverage tend to purchase more reinsurance, and insurers with higher reinsurance dependence tend to have a higher level of debt. My results are consistent with the expected bankruptcy costs argument, agency costs theory, risk-bearing hypothesis, and renting capital hypothesis. I also find that the impact of leverage on reinsurance will be weaker for insurers that use more derivatives than those that use less. Moreover, high levels of derivative use increase the leverage gains attributable to reinsurance. INTRODUCTION According to Modigliani and Miller (1958), reinsurance and financing decisions are irrelevant in a world of perfect capital markets. Why then would insurers buy reinsurance? The primary reason is that in reality the markets are imperfect. Insurers usually take out reinsurance cover to provide protection against catastrophic losses, mitigate policyholders' concerns about insurer insolvency, enhance the insurer's ability to bear risk, and reduce expected tax liability. In a similar line of argument, based on incentives to make value-maximizing decisions, the demand for capital structure change arises due to capital market frictions. The capital structure literature (e.g., Titman and Wessels, 1988) has suggested that firms actually have a target debt ratio, which is influenced by several factors such as size, and profitability. MacMinn (1987) and Plantin (2006) have argued that reinsurance and capital structure might be jointly determined. To my knowledge, the current study is the first research using panel data on a sample of 350 U.K. non-life insurers to simultaneously examine the impact of leverage on reinsurance and the reverse causation from reinsurance to capital structure. Specifically, I construct a two-equation structural model and employ a two-stage least squares (2SLS) regression to estimate it. Prior studies (e.g., Graham and Rogers, 2002; Dionne and Triki, 2004; Aunon-Nerin and Ehling, 2008; Zou and Adams, 2008; Bartram, Brown, and Fehle, 2009) use the simultaneous equations to examine the relation between derivative/insurance hedging and the debt ratio. Unlike their study, I investigate the relation between reinsurance hedging and leverage. Moreover, I estimate the model using the fixed effect vector decomposition (FEVD) approach proposed by Plumper and Troeger (2007) in order to address the problems of estimating time-invariant/rarely changing variables. The motivation for this research is twofold. First, since research on the reverse causality from reinsurance to leverage has never been conducted, this current study aims to fill in the gap in the literature. The second motivation is that insurers, unlike other ordinary business firms, not only use derivatives to hedge their investment risk but also use reinsurance to hedge underwriting risk. Most of previous studies (e.g., Graham and Rogers, 2002; Aunon-Nerin and Ehling, 2008; Bartram, Brown, and Fehle, 2009) find that high leverage increases derivative/insurance hedging and that hedging has a significant positive effect on leverage. Dionne and Triki (2004), however, find that leverage has a positive effect on hedging, but firms do not necessarily hedge to increase their debt capacity. Zou and Adams (2008) show that property insurance expands firms' debt capacity, whereas leverage alone does not result in the purchase of more property insurance. The empirical evidence on the relation between hedging and leverage varies. In my research, I wish to examine whether similar relations exist between reinsurance hedging and capital structure. Consistent with the expected bankruptcy costs argument, the agency costs theory, and the risk-bearing hypothesis, I find that leverage exerts a positive influence on reinsurance purchases. I also find evidence to support the renting capital hypothesis, in that higher leverage is associated with more reinsurance purchases. …