As discussed in the Democratic Decay section, in recent years we have seen the spread of democratically-elected governments that act to fundamentally undermine or even dismantle the democratic system e.g. in the USA, Poland, Hungary, Brazil, India and Indonesia.

These governments cause a lot of damage to the democratic system. Here, we can draw a broad distinction between two forms of damage:

(i) ‘constitutional damage’: this results from governmental attacks on core state institutions, which

• impair meaningful constraints on political power (e.g. through independent courts);

• enable centralisation of state power (which reduces citizens’ protections from arbitrary power); and

• distort meaningful representation for citizens (especially by curtailing the political opposition in the legislature and changing how elections work).

(ii) broader ‘democratic damage’: this encompasses a wider range of trends including:

• the rise of often ambiguously populist-authoritarian leaders and parties;

• negative changes in the behavior of political actors;

• curbs on civil society actors;

• majoritarian backlash against hard-won rights to equality;

• declining public trust in democratic institutions and fellow citizens;

• the decline of rationality and belief in objectivity;

• the disruption of any sense of shared facts, anxieties about deepening anomie;

• and ‘sharp power’ attacks by authoritarian regimes on democratic rivals, including election interference and the disinformation campaigns.

It is fair to argue that, after an anti-democratic government is voted out of power, we should focus on the broader democratic damage they have caused, including often badly damaged citizen trust in government and heightened political and social polarisation.

However, in many contexts it is very difficult to achieve anything without first repairing the damage to core state institutions, such as the courts, prosecution bodies, and the public service. We call this damage 'constitutional damage' and the process of fixing it 'constitutional repair'.

For instance, if the top court is no longer independent, it will be an obstacle to law reform. If the prosecution service is aligned to the previous government, it will not act in a non-partisan manner.

It is important to emphasise that constitutional repair is merely a first step. It focuses on bringing the democratic system back to a baseline of functionality before any bigger issues across the democratic society can be addressed.

Contexts of constitutional repair are distinctive because they are different to two types of constitutional transition that are more familiar to us:

(i) transitions to democracy from full-on authoritarianism (e.g. Communism or military dictatorship in Central and Eastern Europe, Brazil).

These contexts allow for fundamental and far-reaching change and tend to share 4 features:

• the existing authoritarian regime is unable to continue and there is a broad understanding among political forces that the transition is toward a new form of democratic political regime;

• the transition usually involves the adoption of a new democratic constitution to found the post-authoritarian democracy;

• this, in turn, allows for institutional innovation to address the perceived inability of authoritarian-era institutions to implement the new democratic constitution (e.g. the establishment of a constitutional court); and

• the context can require (and justify) extreme measures that would not be justifiable in a stable democracy e.g. purging the courts, or adopting ordinary law to bridge the gap between the authoritarian and democratic constitution.

(ii) major constitutional reform in stable democratic contexts: These contexts involve states that have been democratic for decades (e.g. Ireland, Canada, Costa Rica, Japan) and include major alterations to institutions, power structures, constitutional values or conceptions of the state, which can be deemed a form of constitutional transition.

These contexts share 3 features:

• wholesale constitutional replacement is rare but change can include e.g. adoption of a bill of rights, establishment of a new constitutional court, or changes to territorial organisation (e.g. devolution in the UK)

• unlike the first paradigm above, there is no political regime transition.

• changes are ordinarily achieved in compliance with the procedures laid down in the Constitution or ordinary law.

How is constitutional repair different?

In contrast to the two paradigms above, constitutional repair contexts are much more ambiguous:

• there may be extensive constitutional damage but no rupture with democratic rule, with little space to adopt a new constitution or new institutions as a reparative measure.

• Extreme measures might be required (e.g. court 'unpacking') but often there will be an expectation that democratic and rule-of-law norms appropriate to a stable democracy should be respected.

• Anti-democratic forces remain powerful and capable of winning the next elections.

A variety of factors will affect how constitutional repair is viewed and approached. These are some of the factors at play:

• how we frame the challenge: repair might be viewed as simply as restoration, i.e. an attempt to return the system to how it functioned before the anti-democratic government came to power. However, this might be impossible because of the nature of the constitutional damage caused by that government (e.g. it may be impossible to recreate the Constitutional Court as it previously existed). It might be more fruitful to think in functional terms e.g. how do we achieve an independent top court, even if it looks different to the previous court?

• affective and strategic considerations: there may be a strong political pressure to recreate the prior system as much as possible due to attachment to that system or possibly due to pressure from external actors (e.g. foreign governments, international organisations);

• our own analytical biases: Scholars and experts focused on constitutional design might characterise constitutional repair as requiring constitutional amendment or even replacement even where this is not possible, which can lead us to overlook more practical solutions (e.g. dealing with judges whose appointments are questionable through court presidents' case-management techniques).

It is certainly true that constitutional repair needs to be handled very carefully to avoid its operating as simple political revenge, or what Samuel Issacharoff calls “legal revanchism”.

For instance:

• how can a new government remove judges questionably appointed under the previous anti-democratic government without this seeming to replicate the measures taken by the anti-democratic government to undermine the democratic system? (see e.g. USA, Poland, Hungary)

• how can a new government remove the leadership of key bodies (e.g. prosecution service, media regulators) without doing further damage to the principle that every government should respect the independence of these institutions? (see e.g. Brazil, Poland)

The difficult truth is that there is no easy answer. A new government will often be faced with a difficult trade-off between:

(a) the continuing damage and dysfunction caused by leaving such questionable appointees within the system; and

(b) the potential damage that removing them may cause, to the endurance of democratic and rule-of-law norms, and the potential normalisation of such extreme measures.

These trade-offs have to be considered in light of the common context that anti-democratic political forces remain strong and have the capacity to win future elections.

It seems uncontroversial to say that extreme measures such as purging the courts, vetting judges, or removing leaders from state institutions before their mandates expire should be a measure of last resort.

Factors to consider

A number of factors might be considered when thinking through whether norm-breaking reparative measures can be justified:

democratic reform context: is the damage caused to the system, and the prospect of anti-democratic political forces returning to power, so pressing that extreme measures need to be considered? This will not be determinative as a lone criterion.

reform options: are there any other options that might achieve the same or similar results e.g. if constitutional damage includes a less independent or legitimate top court:

• could introducing term-limits or changing bench selection achieve the same results as court-packing, by depoliticising the court?

• could letting judicial terms expire naturally (in fixed-term systems) address the problem adequately?

articulated reform purpose: how can the new government go to greater lengths than its predecessor to fully explain and justify an extreme measure. This may partly involve what Tamir calls "self-negating" statements, e.g. "Court-packing is bad in principle but necessary for the following reasons...". This alone would be insufficient.

reform process: a new government might be able to differentiate its actions from the previous anti-democratic government by involving a wider array of actors, especially non-partisan actors (including possibly civil society, citizens, and international actors), in its reform process.

repetition risk: a careful mix of rhetorical, processual, and design approaches to constitutional repair, followed by measures to strengthen institutions against future interference, may help to address the risk of a future government taking extreme measures to again undermine the system.