Page 110 - 34-1

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

this applies mainly for weak, publicly-traded insurers using external actuaries. Similar to
independent external auditors, external actuaries have more reputation risk and increased
legal liability after SOX. We expect that weak, publicly-traded insurers using external
actuaries became less under-reserved after SOX. This leads to the following hypothesis:

Hypothesis 3a: Weak publicly-traded P-C insurers using external actuaries became less
under-reserved after SOX than weak publicly-traded P-C insurers pre-
SOX.

In contrast, in-house actuaries working for weak, publicly-traded insurers may face
pressure from firm management to conceal firm weakness even after SOX. Thus, we have
no priors for the effect of SOX on weak, publicly-traded insurers using in-house actuaries
to certify loss reserves.

Conversely, healthy insurers are expected to become more conservative in income
reporting and reserve estimation post-SOX. Thus, we expect that healthy publicly traded
insurers after SOX would bolster reserves. This would reduce the probability of adverse

regulatory scrutiny after SOX and prevent lawsuits against the directors and officers that
could be associated with under-reserving. This hypothesis applies to all healthy publicly-
traded insurers.
Hypothesis 3b: Healthy publicly-traded P-C insurers became relatively over-reserved
after SOX than healthy publicly-traded insurers pre-SOX, regardless of

whether in-house actuaries or external actuaries are used.

3. Methodology

We regress our measures of the loss reserve error on a set of variables designed to
test our hypotheses. Additional variables are included in the model to test hypotheses
appearing in the prior literature concerning income smoothing, taxes, and rate regulation.
Finally, we include control variables representing firm and economic characteristics in the

model as well. The remainder of this section describes the model, the estimation strategy
and variables more fully.

102