Analyzing Accountability Relationships in a Crisis: Lessons From the Fukushima Disaster

Research output: Contribution to journalArticlepeer-review

5 Scopus citations


The Fukushima Daiichi nuclear disaster became the worst nuclear accident since the 1986 Chernobyl disaster. Many studies have examined the so-called “first-order causes” of the Fukushima disaster, such as economic interests and lax regulations. However, studies examining the disaster have paid little attention to how it escalated during the response phase, that is, “second-order causes.” This study examines the unresolved question using an analytic frame of accountability relationships. The results demonstrate how crisis management organizations of government faced cross-pressures within a web of accountability relationships while dealing with the disaster. In particular, these organizations’ responsiveness to hierarchical accountability had a negative effect on the political accountability relationships. It is the contribution of this research to specifically identify multiple and complex relationships between the types of accountability. Previous studies have mostly treated professional accountability as the single dependent variable. In contrast, this research argues that the other accountabilities—hierarchical and political—can be dependent variables as well. More importantly, inconsistent with the findings of previous studies, the economic pressure of political accountability had no effect on professional accountability relationships as a result of the heroic efforts of a nuclear plant manager.

Original languageEnglish
Pages (from-to)743-760
Number of pages18
JournalAmerican Review of Public Administration
Issue number7
StatePublished - 1 Oct 2018


  • accountability relationships
  • centralized decision making
  • crisis management
  • nuclier power village
  • the Fukushima disaster


Dive into the research topics of 'Analyzing Accountability Relationships in a Crisis: Lessons From the Fukushima Disaster'. Together they form a unique fingerprint.

Cite this