Domestic judicial defiance and the authority of international legal regimes

Abstract

Tensions and occasional overt defiance of international courts suggest that compliance with international regimes is not a self-evident choice for domestic judges. I develop a formal theory of domestic judicial defiance in which domestic and supranational judges vie for jurisprudential authority in a non-hierarchical setting. The model emphasises the role of domestic non-compliance costs and power asymmetries in determining the conduct of domestic and international judges. I argue that the EU represents a special case of a particularly effective international regime. Weak domestic courts have little to gain from an escalated conflict with the European court of Justice. But even domestic judicial superpowers like the German Federal Constitutional Court have strong incentives to seek mutual accommodation with European judges. The analysis also yields new insights into concepts, such as “judicial dialogue” and “constitutional pluralism” that have featured prominently in the legal literature, and suggests new hypotheses for empirical research.

Appendices

Appendix 1: Proof of coexistence equilibrium

Proof is based on the one-stage deviation principle for infinite-horizon games (Fudenberg and Tirole 1991, 110). It demonstrates that no player can gain by unilaterally deviating from the actions prescribed by the coexistence equilibrium in any single period t, for all relevant histories h.

1. Beginning with the Domestic Court at the last node of the stage game after the International Court played M, it is easy to see that, to the extent that no court has been caught deviating (denote this history as $h_{1}$), the Domestic Court does not gain by playing D rather than $\sim D$. As long as no history of deviation exists (history $=h_{1}$), playing D does not only precipitate a constitutional crisis but because it itself constitutes deviation from the equilibrium path it triggers punishment from the International Court in subsequent periods. For the Domestic Court, the present value of compliance in this situation is at a minimum:

$$\begin{aligned} PV_{domestic}(compliance|h_{1})&= -J^{h}+\dfrac{\delta }{1-\delta }\left( \dfrac{1}{2}qpJ^{h}-\dfrac{1}{2}qpJ^{h}+q(1-p)J^{h}\right. \\&\left. -(1-q)pJ^l+\dfrac{1}{2}(1-q)(1-p)J^l-\dfrac{1}{2}(1-q)(1-p)J^l\right) \end{aligned}$$

(1)

This simplifies to:

$$\begin{aligned} PV_{domestic}(compliance|h_{1})=-J^{h}+\dfrac{\delta }{1-\delta }\left( q(1-p)J^{h}-(1-q)pJ^{l}\right) \end{aligned}$$

(2)

For the same history, the present value of defection is at most:

$$\begin{aligned} PV_{domestic}(defection|h_{1})=\dfrac{J^h-C}{2}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^{l}\right) \end{aligned}$$

(3)

Therefore, as long as $PV_{domestic}(compliance|h_{1})>PV_{domestic}(defection|h_{1})$ the Domestic Court is better off playing $\sim D$. On the other hand, if the International Court has been caught deviating either in the current period (by playing M out of turn) or at the end of any previous period (denote this history as $h_{2}$), the Domestic Court must punish the International Court as is prescribed by the Coexistence Equilibrium. In that case, the present value of compliance for the Domestic Court is at a minimum:

$$\begin{aligned} PV_{domestic}(compliance|h_{2})=\dfrac{J^{l}-C}{2}+\dfrac{\delta }{1-\delta }\left( qJ^{h}+(1-q)J^{l}\right) \end{aligned}$$

(4)

whereas the Domestic Court’s unwillingness to punish entails that the present value of defection is at best:

$$\begin{aligned} PV_{domestic}(defection|h_{2})=-J^{l}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^{l}\right) \end{aligned}$$

(5)

(unwillingness to punish means playing M is always safe for the International Court). The incentive compatibility constraint, therefore, is:

$$\begin{aligned} \dfrac{J^{l}-C}{2}+\dfrac{\delta }{1-\delta }\left( qJ^{h}+(1-q)J^{l}\right) >-J^{l}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^{l}\right) \end{aligned}$$

(6)

2. We now consider the International Court at the penultimate node of the stage game. It is easy to see that, given history $=h_{1}$, the International Court has no incentive to defect when it is of type = hawk and the Domestic Court signals $\sim T$; the same applies when the Domestic Court was caught deviating in a previous period; ditto, for $h_{1}$, when the International Court is of type = dove, but the last time the courts were of type = dove the Domestic Court had its way. Now, if there was no deviation in previous period but the last time the two courts were of type = hawk the Domestic Court yielded (denote this history as $h_{1a}$), and the Domestic Court signalled T in the current period, the present value of compliance (which involves playing $\sim M$ in the current period) for the hawkish International Court is captured by:

$$\begin{aligned} PV_{international}(compliance|h_{1a},T)=-\alpha ^{h}+\dfrac{\delta }{1-\delta }\left( p(1-q)\alpha ^{h}-(1-p)q\alpha ^{l}\right) \end{aligned}$$

(7)

Because it involves making an immediate concession on an issue relatively dearer to its judges, Eq. (7) is when the present value of compliance is smallest for the International Court. By playing M out of turn, though, the International Court immediately reveals deviation, which triggers undelayed punishment by the Domestic Court:

$$\begin{aligned} PV_{international}(defection|h_{1a},T)=\dfrac{\alpha ^{h}-\beta }{2}+\dfrac{\delta }{1-\delta }\left( -p\alpha ^{h}-(1-p)\alpha ^{l}\right) \end{aligned}$$

(8)

(Immediate punishment means the International Court incurs the cost associated with a constitutional crisis in the present period while facing the prospect of perpetual punishment in subsequent periods.) Therefore, the International Court is most tempted to deviate from the equilibrium path when it is type $=dove$, given one of the two following situations: the Domestic Court played T in the current period (thus signalling hawkishness) and the last time the two courts were of type $=hawk$ the International Court yielded (denote this history as $h_{1b}$); the Domestic Court played $\sim T$ in the current period and the previous time both courts were of type $=dove$ the International Court had its way (denote this history as $h_{1c}$). The present value of compliance for the International Court in both scenarios is thus captured by:

$$\begin{aligned} PV_{international}(compliance|h_{1b},T)&= PV_{ecj}(compliance|h_{1c},\sim T)=-\alpha ^l+ \\&\dfrac{\delta }{1-\delta }\left( p(1-q)\alpha ^{h}-(1-p)q\alpha ^{l}\right) \end{aligned}$$

(9)

whereas the present value of defection is:

$$\begin{aligned} PV_{international}(defection|h_{1b},T)& = PV_{ecj}(defection|h_{1c},\sim T)=\alpha ^l+ \\&\dfrac{\delta }{1-\delta }\left( p(1-q)\alpha ^{h}-(1-p)q\alpha ^{l}\right) \end{aligned}$$

(10)

Therefore, the International Court is better off complying insofar as:

$$\begin{aligned} -\alpha ^l+\dfrac{\delta }{1-\delta }\left( p(1-q)\alpha ^{h}-(1-p)q\alpha ^{l}\right) >\alpha ^{l}+\dfrac{\delta }{1-\delta }\left( -p\alpha ^{h}-(1-p)\alpha ^l\right) \end{aligned}$$

(11)

3. Finally, we consider the Domestic Court at the communication stage. The only situation in which defection can make the Domestic Court better off is when it is type $=dove$ and the previous time both courts were type $=hawk$ it yielded while the last time the courts were both type $=dove$ the International Court yielded (denote this history as $h_{1d}$).²⁹ In that case, the present value of compliance for the Domestic Court is captured by:

$$\begin{aligned} PV_{domestic}(compliance|h_{1d})=-J^l+\dfrac{\delta }{1-\delta }\left( q(1-p)J^{h}-(1-q)pJ^{l}\right) \end{aligned}$$

(12)

The present value of defection for the same circumstances is:

$$\begin{aligned} PV_{domestic}(defection|h_{1d})=J^{l}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^l\right) \end{aligned}$$

(13)

This reflects the situation where the doveish Domestic Court plays T in the current period, which in turn induces the International Court to infer that its true type $=hawk$. As a result, the International Court plays $\sim M$ and the Domestic accrues payoff $J^{l}$. However, at the end of the period, the International Court observes the Domestic Court’s true type and punishment ensues. So the Domestic Court is better of telling the truth as long as:

$$\begin{aligned} -J^l+\dfrac{\delta }{1-\delta }\left( q(1-p)J^{h}-(1-q)pJ^{l}\right) >J^{l}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^l\right) \end{aligned}$$

(14)

Appendix 2: Proof of international law expansion equilibrium

Proof of the International Law Expansion Equilibrium is straightforward. At the last node of the stage game, the Domestic Court cannot improve its lot by revolting against the International Court, as this brings about a constitutional crisis in the current period without stopping the court of Justice’s activism in subsequent periods. Regardless of history, the present value of compliance for the Domestic Court is at a minimum:

$$\begin{aligned} PV_{domestic}(compliance)=-J^{h}+\dfrac{\delta }{1-\delta }\left( -qJ^{h}-(1-q)J^l\right) \end{aligned}$$

(15)

Positing that the Domestic Court prefers a jurisdictional loss over a constitutional crisis, this implicates that it is strictly better off complying. Knowing this, the International Court has no reason to deviate from M at the preceding node. The lack of credible domestic deterrence entails, in turn, that the Domestic Court’s signal has no effect.