Page 161 - 36-1

In Taiwanese political practice, legislators are directly elected representatives of
the people who hold legislative power. Because the legal basis for government
policies must be authorized by the legislature, legislators are motivated to serve their
constituents’ needs and to pass relevant laws to secure re-election (Sheng, 1997).
Therefore, this study argues that firms that make political donations during elections
employ a more direct means of engaging in political activism in exchange for
potential political resources and benefits. Furthermore, following the implementation
of the Political Donations Act in Taiwan, political contribution data from
profit-seeking enterprises must be submitted to and publicly reported by the Control
Yuan, providing a channel to identify firms that engage in political activism and
approach politicians. This study hypothesizes that when firms engage in political
activism and seek to establish political connections with legislators, it may indirectly
influence the policy guidelines of administrative agencies. This could enable firms to
gain early access to political information, respond to future changes in the political
and economic environment, and strategize for innovative policies. However, corporate
political activism may also divert funds from innovation budgets and increase
political costs, making firms more vulnerable to political interference and creating a
“political resources curse.” Therefore, as previous research has not yet clearly
NTU Management Review Vol. 36 No. 1 Apr. 2026
outlined the advantages and disadvantages of corporate political activism on corporate
innovation, this study focuses on corporate political activism through political
donations in legislative elections to investigate its impact on innovation capabilities.
2. Research Design
2. Research Design

This paper examines the impact of political activism on corporate innovation
This paper examines the impact of political activism on corporate innovation
capabilities. Since corporate political activism is a self-selection behavior, firms can
capabilities. Since corporate political activism is a self-selection behavior, firms can
choose whether to donate to legislators through political contributions. To minimize
choose whether to donate to legislators through political contributions. To minimize
estimation bias, this study employs Heckman’s (1979) two-stage model. The first stage
estimation bias, this study employs Heckman’s (1979) two-stage model. The first
uses a probit model to identify the factors influencing firms’ decisions to engage in
stage uses a probit model to identify the factors influencing firms’ decisions to engage
political activism while controlling for election-year and industry fixed effects to calculate
in political activism while controlling for election-year and industry fixed effects to
calculate the inverse Mills ratio (IMR). The probit regression model is as follows:
the inverse Mills ratio (IMR). The probit regression model is as follows:

)= + + + +
0 1 2 � � � 4
+ + + +
� � � � � � � � � � � � �
�
+ � � + +
��
� �
��
� �
�
+ + +
��
��
��
� �
� �
� �
(1)
+ ∑ + ∑ . (1)

In the second stage, this study analyzes the impact of corporate political activism
In the second stage, this study analyzes the impact of corporate political activism
on innovation capabilities by incorporating the IMR estimated in the first stage into the
23
on innovation capabilities by incorporating the IMR estimated in the first stage into the
ordinary least squares regression to correct for self-selection bias. Political activism is
ordinary least squares regression to correct for self-selection bias. Political activism is
proxied by whether firms engage in political donations (DONATE ) or by the amount
i,k
proxied by whether firms engage in political donations ( ) or by the amount
of political contributions to legislators (DOS_F ). We measure corporate innovation
���
i,k
of political contributions to legislators ( ). We measure corporate innovation
���
capabilities (the dependent variable) by using the adjusted number of patents (ADJ_
capabilities (the dependent variable) by using the adjusted number of patents
PATENT ) and patent citations (ADJ_CITATION ) to eliminate time heterogeneity, based
(ADJ_ ) and patent citations
i,t (ADJ_ ) to eliminate time
i,t
���
���
heterogeneity, based on patents ( ) and patent citations ( ) over a single
on patents (PA ) and patent citations (CITA ) over a single period or subsequent years.
i,t
i,t
���
���
period or subsequent years. Additionally, this study further controls for election year
Additionally, this study further controls for election year and industry fixed effects and
and industry fixed effects and establishes the following regression model:
establishes the following regression model:

( ) = + ��� ( )
�
���
���
�
���
+ + + + ��� +
���
�
�
���
�
���
���
�
�
+ + + + + ( ) �
� ��� � ��� � ��� �� ��� �� ���
+ + ∑ + ∑ + . (2)
(2)
���
��

This study further examines the moderating effect of differences between
This study further examines the moderating effect of differences between electronics
electronics and non-electronics industries on the relationship between corporate
and non-electronics industries on the relationship between corporate political activism
political activism and innovation capabilities. We assign a dummy variable equal to 1
if a firm belongs to the electronics industry and 0 otherwise. We analyze Innovation
153
capabilities of electronics across through interaction terms ( × or
���
���
× ), and the regression model is as follows:
���
���

( ) = + ( )+ ×
��� ��� � � ��� ��� � ��� ���
( ��� × ) + + +
�
���
���
�
���
�
���
+ + ��� + + +
��
�
���
���
�
���
�
���
�
+ + + ( ) � +
�� ��� �� ��� �� ��� ��
+ ∑ + ∑ + . (3)
���

3. Findings

This study analyzes listed firms in Taiwan from 2008 to 2019 to examine how
corporate political activism impacts innovation capabilities. We manually collect data
on political donations made by profit-seeking enterprises to legislators from Control
Yuan, using them as a proxy for political activism. We also compile patent and patent
citation data from the M-Trends Patent search and analysis system, and utilize the
number of patents held and cited by Taiwanese firms at the United States Patent and
Trademark Office (USPTO) as proxies for corporate innovation capabilities
(Hagedoorn and Cloodt, 2003). The results reveal a significant negative relationship
24