Where the Lower-Tail Response Goes

The Italian agreements raise a question that the wage-rate estimates alone cannot settle: why does the lower tail respond weakly when the contractual schedule is broad enough to cover several rungs? This chapter follows two filters between a bargained entitlement and an observed outcome. The first is implementation—whether the contractual floor is reflected in reported pay. The second is labor demand—whether the job and the worker remain in the observed employment pool. Keeping those margins separate makes it possible to distinguish a measured mechanism from a plausible explanation.

A contractual entitlement can be weakened before it reaches pay

The study measures non-compliance in administrative data by classifying a worker as underpaid when their observed full-time-equivalent daily wage is below 95% of the lowest contractual minimum in the agreement that covers them. The threshold is deliberately conservative. It uses the lowest floor in the agreement, not the grade-specific floor that should apply to the worker. If a worker is paid below a higher occupation- or seniority-specific minimum but remains above the agreement's lowest minimum, this measure will not detect the shortfall. The reported rate is therefore a lower bound on true non-compliance, not a complete count of violations.

That definition is useful because it connects the contractual schedule to an actual wage record without pretending that the administrative data reveal every detail of the wage grid. It also gives the analysis a clear interpretation. A worker can be covered by an agreement on paper while receiving less than even its lowest contractual benchmark in the observed daily-wage measure. Formal coverage and effective pass-through are related, but they are not the same thing.

The measured level is not trivial. Non-compliance averages about 5.8% and remains broadly stable, roughly 5% to 6%, throughout 2006–2017. Province-level rates range from about 4% to around 10%, with higher incidence in parts of the South and Islands than in much of the North. This geographic spread is consistent with differences in local enforcement and labor-market conditions, although the study does not treat geography alone as a causal explanation.

The strongest pattern is across agreements. Non-compliance is more common in low-wage agreements and declines as the agreement's median FTE daily wage rises; the reported correlation is about -0.47. The workers classified as underpaid also have lower mean log daily wages, are younger on average, are more likely to be foreign-born and part-time, and are less likely to be male or on an open-ended contract. These characteristics do not prove why underpayment occurs. They show that the measured implementation gap is concentrated in groups already closer to the lower end of the wage structure.

This concentration offers a way to interpret the muted 10th-percentile response. Suppose two agreements receive comparable scheduled updates, but the lower-wage agreement has more workers whose actual pay sits below the contractual benchmark. An increase in the benchmark need not produce the same increase in observed pay if the firms and jobs in that segment were already not fully implementing the floor.