Page 52 - 35-1

The Financial Statement Comparability of Listed Foreign Firms in Taiwan: The Effect of Board Interlocks and
Financial Statement Fraud

3. Findings

This study finds that overall, the financial statement comparability of KY companies
is worse than that of non-KY companies, and non-KY companies having board interlocks

with KY companies display poorer comparability than those without interlocks. These
findings indicate that the characteristics of KY companies (high organizational complexity
and overseas operations causing audit difficulties) affect their financial statement
comparability, and that non-KY companies having board interlocks with KY companies

also suffer negative impacts on their financial statement comparability through the
interlocks. After the case of Pharmally, the financial statement comparability has improved
for all non-KY companies regardless of board interlocks with KY companies. However,
the financial statement comparability for KY companies does not improve. Instead, it

shows signs of deterioration compared to non-KY companies. One possible reason is
that the fraud case coincided with the COVID-19 pandemic, restricting auditors from
conducting on-site audits of KY companies with overseas operations. This finding suggests
during the pandemic, the location of company operations is a significant factor affecting

financial statement comparability.

4. Research Limitations/Implications

The primary limitation of this study lies in the timing of the Pharmally Case, which
occurred in August 2020. To calculate comparability after the case of Pharmally using
a complete set of 16 quarters of data, it would be necessary to collect data through
September 2024 to obtain the first complete comparability measure based on 16 quarters.

Unfortunately, as of the final revision, we were still unable to obtain the complete data for
Q3 2024. Due to an insufficient data period, we cannot use the 16 quarters of data as used
in Hypotheses 1 and 2 to calculate the financial statement comparability for Hypotheses 3
to 5. We have to calculate the comparability using data from the first two quarters instead.
However, De Franco et al. (2011) notes when results using different periods align with the

main results calculated using 16 quarters, this limitation should not cause bias. Still, due
to the short post-incident period (only five quarters), we do not observe any significant
difference before and after the fraud case. Future studies with more post-incident data may

44