groups that are not themselves subject to the infringement process. The disrupted
government control over labor inspectorates lead to slower resolution of infringement
cases. Figure 2 shows that as the number of fire alarm oversight mechanisms increases,
the likelihood that a case will end with a letter of notice decreases. Conversely, the
likelihood that a case will end with a reasoned opinion or ECJ referral increases. When
the fire alarm score is 0, the predicted probability that a case will end with a letter of
notice is .64. When the fire alarm score is at its highest, this probability drops to .47. At
the same time, when fire alarm score is 0, the probability that a case will end with an ECJ
referral is .12. However, when fire alarm score is at its highest, the probability nearly
doubles to .23. These changes in predicted probabilities are not as dramatic as the change
in probability connected to police patrol oversight. Furthermore, the ordered probit
results show that this relationship is significant only in the models that do not include
ideological measures. There is weak support for the predictions about fire alarm
oversight but the relationship has neither the magnitude nor the significance observed
with regard to police patrol oversight.
INSERT PREDICTED PROBABILITY FIGURE FOR CRISES HERE
Prolonged Political Crises
Prolonged political crises disrupt Member State government function and the chain
of command from the Commission through the government to the Member States’ labor
inspectorate. The predicted probability figure shows that when cases overlap with a
prolonged national crisis their chances of ending with a letter of notice decrease and their
chances of ending with an ECJ referral increase. Without crises, the predicted probability
of a case ending in a letter of notice is .63. Cases that coincide with prolonged crises end
with a letter of notice with a probability of only .28. Conversely, cases that do not
Convention