Purpose The purpose of this paper is to focus on the interactive effects of intrinsic and extrinsic factors on R&D output by analysing Chinese-listed industrial family firms. It proposes modelling the moderating influence of quality of government (QOG) on the relationship between family firm governance types (family control and family management) and R&D output from the twin agency perspective (Stulz, 2005). Design/methodology/approach The data set is organised as an unbalanced panel. This study exploits random-effects GLS regression, analysing both cross-sectional and time variation, and estimating the mean effects. The GLS model corrects the variance- and sequence-related problems of linear model random items and remains consistent and robust when the error term is heteroscedastic and non-normally distributed. Findings The findings provide several empirical conclusions: in areas with a higher QOG, family firms with greater family control (i.e. voting rights of the board) achieve more R&D output than firms with less family control; and QOG has no significant interactive effects with family management (i.e. the ratio of family managers among top managers) on R&D output. The main contribution of this paper is to show that in areas with a higher QOG, greater R&D output for family firms depends on greater family control rather than family management. These findings give a better understanding of the interactive influence of inside and outside agency problems in family firms in general and their R&D output in particular across different cities, and may help both family firms' leaders and government policy makers to foster innovation by controlling intrinsic and extrinsic agency problems. Research limitations/implications To date, most family firm innovation research has concentrated upon governance and R&D behaviour (Block, 2012; Brinkerink and Bammens, 2018; Chrisman and Patel, 2012; Lee and O'Neill, 2003). Few studies, however, have been performed from the major strategic (control) and operational (management) orientations, into the influence of outside (QOG) and inside (governance) factors upon innovation. This study attempts to fill that gap. It uses patent counts to measure the economic and technological importance of innovation. It argues that different QOG may lead major controllers or executives in family firms to have different motivations, and hence to approach innovation differently from the agency perspective. Practical implications The main contribution of this study is to show that in areas with a higher QOG, higher R&D outputs of family firms depend on higher family control rather than family management, due to the interactive influence of inside and outside agency problems. When family management is high, the direct effect is high, because family management may reduce the principal-agent agency cost (PAAC), but the interactive effect of QOG and family management is not significant. In areas with high QOG, although family management may reduce the PAAC, principal-principal and altruism agency costs may increase. Based on the twin agency theory, differing inside expropriation issues between strategic (family control) and operational (family management) orientations are the main differentiator, one accentuated by the external expropriation issues of QOG. Social implications These results contribute to a better understanding of family firms in general and their R&D output in particular across different cities. The findings also show of interest for government policy makers who should be aware of the significance of FFs' characteristics for innovation and their incentives to conduct R&D projects. Originality/value The research uses Stulz's (2005) Twin agency concept to analyse the interacting effects of state-level agency problems of governments with firm-level agency problems of family firms on R&D output. This paper answers the main question: What are the interactive effects of QOG and family firm governance on R&D output? The main contribution of the paper is to bridging the current gap in the literature.
