The design of Taiwan's coercive institutions had a profound effect on the patterns of violence that the island experienced during martial law. In particular, the reforms made to the coercive apparatus in the early 1950s, described in Chapter 3, help explain why Taiwan experienced a drop in state violence in the years that followed. Before that, during the late 1940s and early 1950s, the exclusivity of the KMT's coercive institutions isolated coercive agents from Taiwan's society. Social exclusivity had two major effects: first, the security forces had a more difficult time identifying and dealing with popular threats without resort to public, indiscriminate violence, and second, exclusivity created out-group resentment between the Mainlander and native Taiwanese populations that incentivized worse violence. Fragmentation of the coercive apparatus also fostered violence – against civilians as well as between rival security organizations – by hampering the flow of intelligence on popular policing and by providing incentives for inter-agency competition. As a result, state violence was relatively high during the period from 1945 to 1955.

Following the reforms that were documented in Chapter 3, however, the coercive apparatus became unitary and socially inclusive. This new organizational design provided fewer social and material incentives for coercive agents to engage in violence. Because social inclusivity provided better access to information on Taiwan's society, and because reductions in fragmentation improved internal coordination, the coercive apparatus’ intelligence capacity also improved. These changes enabled coercive agents to rely on surveillance and pre-emptive, discriminate repression rather than public, reactive, and indiscriminate violence. As a result, state violence declined around the time that these reforms were completed, in 1955, and remained low for the rest of the martial law period.

The first section of this chapter describes the drop in state violence that occurred in Taiwan in the mid-1950s, with an emphasis on several key aspects of that violence that are either omitted from or inaccurately depicted by existing studies. The second section illustrates how exclusivity and fragmentation worked through the intelligence and incentives pathways to contribute to higher levels of violence in the late 1940s and early 1950s, while the third section analyzes how inclusivity and a unitary coercive structure altered incentives and intelligence capabilities in ways that decreased violence after 1955.