Page 106 - 34-1

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

By way of preview, Chow test results indicate that it is not appropriate to pool weak
and healthy insurers together in one regression analysis, in contrast with many previous
loss reserving studies (e.g., Gaver and Paterson, 2001, 2004; Grace and Leverty, 2010,
2012). Hence weak and healthy insurers are analyzed using separate regressions. Empirical

results indicate that significant under-reserving occurs for insurers using in-house actuaries
to certify reserves; the degree of under-reserving is significantly higher for weak than for
healthy insurers. This result is a clear indication that the loss reserving behavior of weak

versus healthy insurers is different, requiring a statistical analysis which can account for
this systematic difference. In addition, the effect of SOX on weak publicly-traded P-C
insurers using in-house actuaries is insignificant (in the PSM sample). Thus, contrary to
many previous studies of the effect of SOX on financial reporting, this study reveals that
SOX did not necessarily lead to more conservative financial reporting for weak insurers

using in-house actuaries. But it did lead to more conservative reporting for weak insurers
using external actuaries. 13
This research contributes to the literature in several ways. Approximately 60% of

total P-C insurer reserves in the U.S. are certified by in-house actuaries, but the degree
of managerial bias associated with using in-house actuaries using U.S. P-C insurer data
has not been investigated until this study. This research is important, too, because it
14
contributes to the financial literature on the role of managerial discretion in accounting
reports. The use of in-house actuaries may increase managerial bias due to the influence/

pressure of insurer management on in-house actuaries in some cases. This research also
provides evidence that the results and inferences from analysis of pooled loss reserve error

13 We acknowledge that one should also be concerned about the size of the reserve if the reliability of
financial reporting is the focus of SOX (Marlo and Nyce, 2006). Because we are more interested in
the difference in behavior of weak insurers versus healthy insurers in response to SOX, we do not
report the results of the absolute value of loss reserve errors. In untabulated results, we find that the
absolute value of loss reserving errors increased after SOX for healthy, publicly-traded stock insurers,
and this is due to more over-reserving after SOX. Thus, SOX led to more conservative financial
reporting for healthy U.S. publicly-traded P-C insurers. For weak publicly-traded P-C insurers, the
absolute value of the loss reserve error decreased after SOX, indicating that SOX led to less under-
reserving by weak publicly-traded P-C insurers. Results are reported in an earlier version of the paper,
and they are available upon request from the authors.
14 This estimate was obtained from the NAIC database over the period 1999 to 2010.

98