Page 104 - 34-1

In-House Provision of Corporate Services: The Case of Property-Casualty Insurers and In-House Actuarial
Loss Reserve Certification

insurers is used with a dummy variable to distinguish between weak and healthy firms
(e.g., Beaver and McNichols, 1998; Nelson, 2000; Beaver et al., 2003). Petroni (1992)
and Gaver and Paterson (2014) recognize that the incentives of weak insurers are
likely to differ from those of healthy insurers. Thus, this research will investigate the

appropriateness of conducting separate analysis on weak versus healthy firms.
Only two prior papers consider the relationship between loss reserve accuracy and the
use of an in-house actuary. Using a sample of Canadian insurers, Kelly, Kleffner, and Li

(2012) do not find a significant relationship between the use of in-house actuaries and the
absolute value of the loss reserve error. Our research differs from Kelly et al. (2012) with
respect to the methodology used, the loss reserve error definition, and sample analyzed,
largely due to the different institutional settings between Canada and the U.S.
Secondly, Kamiya and Milidonis (2018) analyze the loss reserve error for U.S. P-C

insurers, but limit their sample to firms using in-house actuaries. A distinction between
8
insurers using officer and non-officer in-house actuaries is made. Therefore, their approach
cannot be used to detect whether there is an overall effect on managerial bias from using

in-house actuaries compared to external actuaries. On the other hand, this is exactly the
purpose of the present paper, which is the first to analyze loss reserve accuracy of in-
house versus external actuaries using U.S. P-C insurer data. Furthermore, neither Kelly et
al. (2012) nor Kamiya and Milidonis (2018) explicitly contain hypotheses concerning the
relationship between the enactment of SOX and the loss reserve error in the context of in-

house actuaries.
This research proceeds by estimating regressions in which the loss reserve error is
the dependent variable. The loss reserve error indicates the direction of the reserve error

(under- or over-statement of losses). Under-reserving is relevant because it is an important
9
cause of insolvencies. Also, Eckles et al. (2011) indicates that managers can have an

8 It would be interesting to also include an officer-actuary variable in our analysis. However, this would
entail an unreasonable amount of hand collection of data. Our sample period is 12 years, compared to
three years for Kamiya and Milidonis (2018). Further, information on officer versus non-officer actu-
arial certification only became available in 2007, while our sample period starts in 1999.
9 Best’s Insolvency Study indicates that inadequate pricing/deficient loss reserves is the single largest
cause of insolvency for U.S. P-C insurers, accounting for 42 percent of all insolvencies (A. M. Best
Company, 2012b). Milidonis and Stathopoulos (2014) indicate that financially weak firms are more

96