Free China Movement

Evidence · Human Rights · Democratic Future

Take Action

English 简体中文 繁體中文 Français Deutsch Español

🔎

Civilian Command From Day One: Subordinating Army, Police and Intelligence to Law

October 2026

Download: PDF · Word (DOCX)

Series: Security-Sector Transition & Rule-of-Law Safeguards — research papers accompanying the Security-Sector Transition & Rule-of-Law Safeguards working blueprint. Paper S1.

Security-Sector Transition & Rule-of-Law Safeguards — Policy Paper S1 — Draft for review

1. The question

In a democratic transition, who lawfully commands the armed forces in the first hours, and how is that command made institutional rather than personal? In the PRC (Mainland), the forces answer to a party, not to a state that changes hands through elections. The PRC Constitution assigns direction of the armed forces to the Central Military Commission (PRC Constitution 2018, Art. 93). The 2020 National Defense Law then says the armed forces are “subject to the leadership of the Communist Party of China” (National Defense Law, Art. 21). Once party authority over the military ends, a command gap opens. Two dangers sit inside that gap. One is that a commander fills it himself. The other is that a single civilian personally takes over the party’s former role. This paper asks how to close the gap on the first day without falling into either danger.

2. The hub position

The Security-Sector Transition & Rule-of-Law Safeguards blueprint (the blueprint) defines civilian control as control by institutions under law, not by a person. It names representative budget oversight, lawfully constituted executives, professional command, independent courts and audit as the parts of that control (the blueprint). The blueprint does not accept a loyalty declaration or a change of name as proof of normalization. Among its indicators of normalization are the removal of party organizations from command chains, registered units and weapons, and transparent senior appointments. It also sets out a sequence: registration, freeze, interim control, vetting, reorganization, and then exit or lawful integration (the blueprint). Its open questions include standards for manifestly unlawful orders and limits on emergency domestic deployment. This paper takes up the first stage of that sequence: lawful command on Day 0.

3. Evidence and comparative options

The ROC constitutional baseline. The 1947 Constitution of the Republic of China already contains the core norms. The President commands the land, sea and air forces (Art. 36). The organization of national defense is to be set by law (Art. 137). The armed forces must stand above personal, regional and party ties, and must be loyal to the state and protect the people (Art. 138). No party or individual may use armed force as a tool in political struggle (Art. 139). Active-duty military personnel may not hold civil office at the same time (Art. 140) (Office of the President, ROC Constitution). These articles grew out of the 1946 settlement. On 25 February 1946 the National Government and the CCP agreed, with General Marshall as adviser, on a basis for military reorganization. Under it, the President would exercise command through a Ministry of National Defense. Communist forces would be integrated into a reduced national army of mixed formations, and secret or independent armed forces were barred (FRUS 1946, Vol. IX, Doc. 100). The agreement was never carried out. The principles remain the most direct constitutional precedent for nationalizing the army.

ROC (Taiwan): civilianizing command through statute. The National Defense Law and the Ministry of National Defense Organization Law were passed in January 2000 and took effect on 1 March 2002. Under them the Chief of the General Staff, who had reported directly to the President, was placed under the Minister of National Defense, and the minister was required to be a civilian (Jamestown Foundation, 2007). Implementation took time. In 2007 every minister appointed under the law had been a retired senior officer, and analysts expected full civilianization to take up to another decade (Jamestown Foundation, 2007). Many officers resisted the larger civilian role. Over 2000–2008 the number of nonmilitary staff in the ministry grew substantially, and joint political-military exercises were used to build defense competence among civilian officials (Princeton ISS case study). The lesson is that a statute can move the chain of command quickly, while building civilian capacity is a separate, slower task.

Spain: constitutional text tested under pressure. The 1978 Spanish Constitution gives the armed forces the job of guarding sovereignty and the constitutional order, and leaves their structure to an organic act (Art. 8). It gives the Government direction of military administration and national defense (Art. 97) (Spanish Constitution). On 23 February 1981, Civil Guards under Lt. Col. Tejero seized the Congress of Deputies, and troops took to the streets in Valencia. At 1:45 a.m. the King went on national television, condemned the coup and ordered that constitutional order be maintained, and the coup failed (Ara, 2021). The episode shows the value of one publicly known lawful authority that every unit can check against. It also shows the weakness of relying on the personal standing of one figure.

Portugal: ending a military guardian body. After the 1974 revolution, officers kept political power through the Council of the Revolution. The 1982 constitutional revision abolished the Council and barred military officers from political activity (Christian Science Monitor, 1982). Portugal therefore spent eight years under an explicit military guardian body before ending it. A transitional design should set the end date of any mixed civil-military body at the outset.

South Korea: removing a faction within the officer corps. In his first year in office, President Kim Young-sam purged Hanahoe, a private society of army officers that had controlled promotions. Senior commanders, including the Army Chief of Staff, were forced out. Later the former presidents Chun Doo-hwan and Roh Tae-woo were tried for the 1979 coup and the 1980 Gwangju crackdown (Korea Times, 2015). Formal command lines mean little if informal networks still decide promotions.

Chile: closing military prerogatives late. Military-era prerogatives lasted fifteen years into the democratic period. The 2005 reforms reduced the military-dominated National Security Council to an advisory role and abolished appointed and life senators (JURIST, 2005). Reserved powers that are left in place at the start are hard to remove afterwards.

International guidance. DCAF guidance holds that armed forces should be subordinate and accountable to elected leadership. On that view, parliaments approve defense funds, and each security institution needs its own legal framework that respects fundamental rights (DCAF, Oversight and Guidance, 2015).

Options. The comparative record suggests three Day-0 models:

  • (a) Personal command: a head of the interim government takes command alone. This is fast, but it repeats the personalization the hub rejects.
  • (b) Military guardian council: officers run command until elections. This is Portugal’s path before 1982 and Chile’s path before 2005, and in both cases the arrangement became entrenched.
  • (c) A statutory civilian bridge leading to a time-limited command body with a civilian majority, and then to an ordinary ministry under a civilian minister. The draft L02 model law already follows this route. It voids party-military orders at activation and requires civilian co-signature and 72-hour review of temporary orders. It caps the bridge at 90 days and sets up a transitional command body with a civilian majority and a 12-month term (L02 draft model law).

This paper recommends option (c).

4. Recommended design

  1. Single lawful source. The activation instrument names one civilian transitional authority and a pre-designated civilian successor. No person or body without a legal basis can acquire command by claiming legitimacy.
  2. Party command ends at activation. Orders issued in the name of party military organs have no legal force. No one is obliged to obey them, and no one may be punished for refusing them.
  3. Professional continuity, civilian authorization. Serving commanders keep day-to-day operational command for a period capped by statute. Any order beyond routine operations needs civilian co-signature and receives independent judicial review within 72 hours.
  4. Collegial, civilian-majority command body. A transitional command body is set up with civilians in the majority and professional officers and independent oversight members alongside them. Sitting party cadres and intelligence chiefs are excluded. Its powers are listed exhaustively, and it has a fixed term that only legislation can extend.
  5. Statutory civilian ministry. A defense organization law follows the ROC constitutional instruction that defense organization be set by law (Art. 137). It places the chief of the general staff under a civilian minister, as ROC (Taiwan) did in 2000–2002, and sets a deadline for dissolving the transitional body into that ministry.
  6. Political neutrality. Party organs, political commissars and party committees are removed from every unit. Military participation in partisan politics is barred, in line with ROC Constitution Arts. 138–139. No serving officer may hold civil office (Art. 140).
  7. Budget and appointments. From Day 30, all defense and security spending goes through a published budget subject to legislative audit. Senior appointments and promotions above a set rank require public nomination and confirmation by the representative body. This addresses the informal promotion networks seen in South Korea.
  8. Registration before recognition. No unit counts as part of the national forces until its personnel, weapons, sites and communications are entered in the public register described in the blueprint.
  9. Duty to refuse. A statutory duty to refuse manifestly unlawful orders applies, backed by a protected reporting channel to an independent oversight body that must be in place within 30 days.
  10. Domestic deployment. Military forces may not take on ordinary policing. Any emergency domestic deployment requires a published civilian order, a geographic and time limit, and legislative ratification.
  11. No reserved prerogatives. The armed forces hold no seat, veto or guardian role in the constitutional order. Every transitional body has a sunset clause, and renaming a body does not reset its clock.

5. Open questions and risks

  • Legitimacy of the bridge authority. Who names the Day-0 civilian authority if several bodies claim it? A recognition test with independent corroboration will be needed.
  • Capacity gap. ROC (Taiwan) shows that civilian defense expertise takes years to build. Until it exists, there is a risk of nominal civilian control with real military autonomy.
  • Commanders who hold out. The law needs a lawful response to units that refuse to register, one that does not provoke armed confrontation.
  • Interaction with vetting. Removing commanders too fast could empty the chain of command. Retaining them too long could preserve party networks. This must be coordinated with Paper S2.
  • Paramilitary and police forces. Command of armed police and stability-maintenance bodies needs its own handover, coordinated with Paper S3.
  • Sources. The comparative record above relies partly on secondary reporting. Final adoption should rest on primary statutory texts, including the official ROC (Taiwan) National Defense Act.

6. Cross-links

In this series: Vetting Without Purge · Demobilizing the Stability-Maintenance Apparatus · Intelligence Oversight.

Working paper v1.0, published October 2026 as part of the Security-Sector Transition & Rule-of-Law Safeguards blueprint program. This paper will be updated as the evidence base develops.