Effects of dual networks on tax strategies: geography and transaction

Ryo Itoh, Zonghui Li

Research output: Contribution to journalArticlepeer-review

Abstract

This study investigates how a revenue-maximizing tax strategy of local and central governments incorporates dual networks, namely, an inter-firm transaction network and an inter-country geography network. We assume a two-stage game in which governments propose discriminatory tax levels for firms, whereas each firm has an incentive to invest in a country near the foreign branch office of its transaction partner. In our model, the centrality index of the Kronecker product of the two networks describes the interplay among the location choices and tax strategies in the equilibrium. A stronger linkage within each network generally increases demand for investment and in turn raises overall tax levels to exploit the high demand. Although more central firms in the inter-firm network are likely to be levied higher taxes because of their high demand for investment, firms in the highest tax bracket differ among countries depending on their geographical location. Finally, we show that a uniform tax in which firms are not discriminated and networks do not matter is the socially optimal tax, which incorporates all inter-country externalities. We also investigate decentralized tax strategies based on the rule of non-discriminatory (uniform) taxation and show, by comparing social welfare under discriminatory and uniform tax regimes, that restricting tax discrimination improves social welfare.

Original languageEnglish
Pages (from-to)97-128
Number of pages32
JournalJapanese Economic Review
Volume72
Issue number1
DOIs
Publication statusPublished - 2021 Jan

Keywords

  • Geography
  • Inter-firm transaction
  • Katz-Bonacich centrality
  • Tax discrimination
  • Tax strategy

ASJC Scopus subject areas

  • Economics and Econometrics

Fingerprint Dive into the research topics of 'Effects of dual networks on tax strategies: geography and transaction'. Together they form a unique fingerprint.

Cite this