Page 92 - 臺大管理論叢第32卷第1期

The Effects of Halved Imputation Tax Credits and Wealthy Tax on the Dividend Policies of Listed Companies:
A Comparative Study of Family and Non-Family Firms in Taiwan

agency problem existed between the controlling shareholders and the minority ones
thereby prompting them to avoid an abrupt shift in dividend payouts to alleviate that
agency problem. (Gugler, 2003; Pindado, Requejo, and de la Torre, 2011). Therefore, we
conjecture that family firms may alter their dividend payout ratios in the years immediately

preceding and after the 2015 tax reform but with no predetermined predicted directions.
Additionally, we conduct several Ordinary Least Squares (OLS) regression models
and supplemental tests to examine our research questions. Our sample firms are selected

from Taiwanese firms listed on the Taiwan Stock Exchange and on over-the-counter
markets. The final sample consists of 8,194 firm-year observations spanning from 2010 to
2016.

3. Findings

Our empirical results show that firms with high imputation credit ratios and high
shareholdings of individual directors and supervisors have relatively higher dividend

payout ratios in the year preceding the 2015 tax reform, and that the increase in dividend
payouts is only significant for family firms. Further, we also find that after the 2015 tax
reform, firms with high imputation credit ratios and high shareholdings of individual
directors and supervisors do not pay relatively lower dividends. The results are the same
for both family and non-family firms. Finally, our additional analyses indicate that firms

which are likely affected by the tax reform pay higher dividends in the year preceding
the 2015 tax reform and lower dividends in the year after the reform, consistent with
the tax clientele hypothesis. However, this result is only notable for non-family firms,
suggesting that family firms are more concerned about non-tax costs in deciding whether

to opportunistically take the tax benefit of dividend for individual investors by changing
their dividend policies in response to the 2015 tax reform.
To examine the robustness of our results, we conduct sensitive tests, including
replacing our regression models with Difference-in-differences (DID) models, and

excluding the observations in 2015 because it is the year that the tax reform takes effect.
The results of these sensitive tests show that our findings remain robust to different
regression models and sample periods. In addition, we conduct several additional analyses.
We find that firms which are likely to be affected by the tax reform do not use return

84