State of the art

Previous research has shown that multiple crises over the past decades, along with the EU’s responses to them, have fueled political dissensus within the European Union (Hutter et al., Reference Hutter, Grande and Kriesi2016; Zeitlin et al., Reference Zeitlin, Nicoli and Laffan2019). This was particularly evident during the sovereign debt crisis, when the EU-driven enforcement of fiscal austerity exacerbated public and party-based Euroscepticism, especially in the most affected countries (Brack and Startin, Reference Brack and Startin2015; Braun et al., Reference Braun, Popa and Schmitt2019; De Vries, Reference De Vries2018; Foster and Frieden, Reference Foster and Frieden2021; Hobolt and Tilley, Reference Hobolt and Tilley2018; Rohrschneider and Whitefield, Reference Rohrschneider and Whitefield2015; Conti et al., Reference Conti, Marangoni and Verzichelli2020, specifically on Italy).

In response to the COVID-19 pandemic, however, the EU adopted a markedly different strategy. To prevent the repetition of past mistakes, rather than fiscal restraint, the EU prioritized economic solidarity and risk-sharing mechanisms (Buti and Fabbrini, Reference Buti and Fabbrini2022; Capati, Reference Capati2023; Fabbrini, Reference Fabbrini2024; Schramm and Terranova, Reference Schramm and Terranova2024). This learning was driven not only by economic rationale, but also by a recognition of the political costs of failing to support member states in times of crisis. As Armingeon et al. (Reference Armingeon, de la Porte, Heins and Sacchi2022: 160) observe, the new approach was designed as a “reaction to political vulnerabilities that inevitably follow from economic vulnerabilities” (see also Ladi and Wolff, Reference Ladi and Wolff2021; Schelkle, Reference Schelkle2021; Nicoli and Zeitlin, Reference Nicoli and Zeitlin2024). Notably, they provide evidence that countries with strong anti-EU sentiment were allocated particularly high per capita grants from the NGEU fund. In contrast, the distribution of NGEU grants did not appear to be linked to the severity of the health crisis during the first half of 2020. In other words, by offering grants and demonstrating solidarity, the aim was to deprive populists of the ammunition needed to provoke a political backlash against the EU. Moreover, the RRF marked a departure from the negative conditionality that characterized previous interventions. Instead, it took a carrot-and-stick approach (Vanhercke and Verdun, Reference Vanhercke and Verdun2022) combining financial incentives (“carrots”) with requests for reform (“sticks”) to encourage member states’ compliance while fostering national ownership of policy plans (Munta et al., Reference Munta, Pircher and Bekker2023; Miró et al., Reference Miró, Natili and Schelkle2024; Zeitlin et al., Reference Zeitlin, Bokhorst and Eihmanis2025).

With that being said, our understanding of how EU financial assistance affects domestic political debates remains limited. While existing research indicates that EU funding – especially in the form of structural and cohesion funds – can foster support for European integration at both individual and regional levels (Hooghe and Marks, Reference Hooghe and Marks2005; Chalmers and Dellmuth, Reference Chalmers and Dellmuth2015; Radojičić, Reference Radojičić2024), and may reduce backing for Euroskeptic parties (Henceroth and Oganesyan, Reference Henceroth and Oganesyan2019; Rodríguez-Pose and Dijkstra, Reference Rodríguez-Pose and Dijkstra2020; Vergioglou, Reference Vergioglou2023), these findings are nuanced and context-dependent. This trend aligns with cost-benefit analyses suggesting that economic incentives shape attitudes toward European integration (Foster and Frieden, Reference Foster and Frieden2021). According to a rationalist institutionalist framework (Hix, Reference Hix2007), attitudes toward the EU are not fixed, but responsive to evolving incentives and perceived interests, implying that today’s critics could become tomorrow’s supporters if their circumstances improve.

Drawing on this literature, our first hypothesis investigates whether the NGEU achieved its intended objective: reducing the Euroscepticism that had been exacerbated by the EU’s austerity-driven, conditionality-based approach to the sovereign debt crisis. We therefore propose:

H1: Following the adoption of the NGEU, political parties increased their support for and/or reduced their criticism of the EU.

Our second hypothesis suggests that the new phase initiated by NGEU will have a more pronounced impact on the rhetoric of Euroskeptic parties, while its effect on pro-EU parties will be minimal. This expectation is grounded in existing literature that explores the reasons behind parties’ shifts toward or away from Euroscepticism. One dominant explanation for party-based Euroscepticism is rooted in party family traditions and ideological cleavages: mainstream parties (social democrats, liberals) generally support integration, while radical left and radical right parties are more skeptical (Marks and Wilson, Reference Marks and Wilson2000; Hooghe et al., Reference Hooghe, Marks and Wilson2002; Halikiopoulou et al., Reference Halikiopoulou, Nanou and Vasilopoulou2012; Vasilopoulou, Reference Vasilopoulou2018).

A related perspective highlights the mainstream-radical divide, suggesting that mainstream parties tend to support European integration, while radical parties – both left and right – are more prone to Euroscepticism (Hooghe et al., Reference Hooghe, Marks and Wilson2002; De Vries and Edwards, Reference De Vries and Edwards2009; Conti and Memoli, Reference Conti and Memoli2012), and so are antiestablishment and populist parties (Pirro and Taggart, Reference Pirro and Taggart2018; Braun et al., Reference Braun, Popa and Schmitt2019; Taggart and Pirro, Reference Taggart and Pirro2021; Szczerbiak and Taggart, Reference Szczerbiak and Taggart2024). These parties often deploy Euroscepticism as a tool to differentiate themselves from the political mainstream, capitalizing on public discontent with EU institutions. A similar logic applies to challenger parties – those with no prior experience in government – who may use Euroscepticism to gain visibility and contrast themselves with incumbent elites (De Vries and Hobolt, Reference De Vries and Hobolt2020). Importantly, this literature also suggests that Euroscepticism among these parties is often strategic rather than purely ideological, meaning that they may adjust their rhetoric in response to changing incentives, including participation in government and funding provided by the EU. From this perspective, rhetorical positions are not fixed but contingent on context and opportunity structures.

At the same time, the literature is divided on whether and why political parties alter their positions on issues related to European integration. Building on reputational theory and cleavage theory, some scholars argue that mainstream parties would incur “credibility” and internal “cohesion” costs from changing their (generally longstanding) attitudes in favor of European integration (Rohrschneider and Whitefield, Reference Rohrschneider and Whitefield2015). As a result, these parties tend to remain rhetorically consistent, even amid rising public Euroscepticism or the electoral gains of Euroskeptic competitors (Rohrschneider and Whitefield, Reference Rohrschneider and Whitefield2015, p. 141). By contrast, challenger parties – particularly those already skeptical of the EU – may be unlikely to revise their attitude because it is a way to differentiate themselves from mainstream parties (Taggart, Reference Taggart1998) and attract disenchanted voters (De Vries and Hobolt, Reference De Vries and Hobolt2020).

Other works, building on the spatial theory of voting, stress the dynamics of competitive party politics and thus examine the strategies adopted by (generally pro-European) mainstream political parties to compete with Euroskeptic challenger parties (Meijers, Reference Meijers2017, p. 413). Since public opinion has become more Euroskeptic, mainstream parties have an incentive to adopt less pro-European, or even Euroskeptic, rhetoric in order to attract anti-EU voters. The greater the electoral success of Euroskeptic actors, the more likely mainstream parties are to shift their discourse to retain or recapture voters. Yet from this vantage point, rhetorical shifts may also respond to material incentives, not just electoral ones.

We hypothesize that Euroskeptic parties are particularly susceptible to external financial incentives, such as those introduced through NGEU funds, because their opposition to the EU is often contingent on political or economic grievances. If the EU is seen as providing tangible economic benefits, the rationale for maintaining a Euroskeptic stance weakens. This aligns with research suggesting that parties may shift their EU position when faced with material gains or changing public sentiment (Foster and Frieden, Reference Foster and Frieden2021; Taggart and Pirro, Reference Taggart and Pirro2021). Accordingly, we expect that the financial assistance and stabilizing effects of the NGEU, combined with the possibility of framing national ownership of recovery plans, will exert a moderating influence on the rhetoric of Euroskeptic parties.

H2: Following the adoption of the NGEU, traditionally Euroskeptic parties will increase their level of EU support and decrease their oppositional attitude toward Europe more than mainstream parties.

While we emphasize the strategic adaptation of party rhetoric in response to the NGEU, it is important to acknowledge that such rhetorical shifts may emerge through multiple mechanisms. For governing parties, moderation may reflect a direct material interest in managing and implementing EU recovery funds, which creates incentives to project alignment with EU priorities. For opposition parties, however, rhetorical moderation may be driven more by reputational concerns or anticipatory incentives – such as the desire to appear credible governing alternatives, to respond to changing public opinion, or to position themselves for future coalition opportunities. These distinct pathways highlight that rhetorical change is not exclusively a function of incumbency or ideology, but also of the strategic and reputational logic that parties face within a dynamic political and economic context. Our hypotheses reflect this multidimensional understanding, expecting that parties – especially those with Euroskeptic profiles – will moderate their rhetoric following the launch of the NGEU, regardless of their immediate position in government or opposition.

Research design and methodology

This study examines patterns of Euroscepticism in the Italian Chamber of Deputies (the lower chamber) from February 2018 to December 2024, a period spanning two legislative periods, four cabinets, and marked by the launch of the NGEU program in 2020. The analysis relies on a newly collected dataset of 30 plenary debates concerning the “Communications of the Prime Minister” about upcoming European Council meetings (for more information, see section 2 in the Appendix).

Examining Euroscepticism in the context of plenary debates offers several methodological advantages. These meetings are reasonably transparent, as deliberations with the Prime Minister are broadcast, and stenographic minutes from the meetings are made available to the public. This provides us with sufficiently fine-grained data to identify shifts in tone and content in party positions toward the EU. Furthermore, compared to other institutional settings, most notably standing committees, plenary debates are arguably more visible to the public, providing parties with greater incentives to engage in signaling behavior to their constituents (Williams and Spoon, Reference Williams and Spoon2015) and issue competition strategies. As a result, we argue that plenary debates are an ideal setting for advancing our understanding of the forces driving Euroscepticism at the national level.

The units of analysis in our study are the oral contributions made by individual MPs in response to the Prime Minister’s briefings. Overall, our dataset consists of 725 statements hand-coded independently by two researchers. Intercoder reliability was high (Cohen’s κ = 0.89), and disagreements were reconciled through team discussion. We decided to exclude 138 statements made by MPs from the mixed group, as it is a composition of smaller parties with sometimes different ideological leanings, making it challenging to attribute a coherent rhetorical strategy on European affairs to it. For each of the remaining 587 statements, we assessed the presence of support and/or opposition to the EU. Political science literature offers no univocal definitions of these concepts. One possible operationalization concerns the object of (dis)agreement, namely, whether it is directed at the policy, at conduct in politics, or toward the political system (Persson et al., Reference Persson, Karlsson, Lehmann and Mårtensson2024). This categorization somewhat resonates with the widely recognized distinction between “hard” and “soft” Euroscepticism (Szczerbiak and Taggart, Reference Szczerbiak and Taggart2008). Hard Euroscepticism refers to a principled rejection of the European integration project and opposition to the transfer of national sovereignty to supranational institutions, such as the EU. In contrast, soft Euroscepticism encompasses criticism of the EU's current or future direction, particularly regarding the expansion of its competencies. While acknowledging the theoretical relevance of these approaches, this piece focuses on the overall dynamics of party-based Euroscepticism. We, therefore, consider all forms of opposition and support as significant insofar as they are directed toward the EU.

In line with previous studies on parliamentary behavior, we distinguish between two types of opposition: the expression of criticism and the presentation of alternatives (see, for instance, Karlsson and Persson, Reference Karlsson and Persson2022; Persson et al., Reference Persson, Karlsson, Lehmann and Mårtensson2024). Although both involve some extent of disagreement, the distinction between the two is crucial since they usually serve two different purposes. On the one hand, explicit criticism is mostly voiced by actors expressing opposition toward general issues, such as EU governance, treaties, and institutions (Senninger, Reference Senninger2017). On the other hand, actors presenting alternatives address specific policies promoted and implemented by the EU (Garritzmann, Reference Garritzmann2017). Thus, we regard the latter as a more moderate form of opposition. Based on these distinctions, we coded each statement into one or more of the following categories: (1) support, (2) criticism, and (3) alternative (see Appendix, sections 3 and 4).

Table 1 illustrates the distribution of parliamentary statements across the three categories for each political party, providing both absolute values and percentages relative to their total number of statements. First of all, we observe that the distribution of statements varies extensively across parties. These results, in part, stem from the fact that some are not present for the entire study period: Action (Azione, Az), Italy Alive (Italia Viva, IV), and Us Moderates (Noi Moderati, NM) were absent during the Conte I cabinet. In part, it has to do with the fact that some have a small delegation in parliament and are given less speaking time (+Europa and Südtiroler Volkspartei). In consideration of these disparities, the descriptive section will show data on all parties, whereas the multivariate analysis will only focus on the six biggest parties with a continuous presence in parliament during the study period.

Table 1. Type of parliamentary statements by party

Note: Own data.

In substantive terms, the data reveal distinct party strategies in their engagement with EU affairs and their attitudes toward the EU. Italy Alive (95.2% of statements falling within the “support” category), PD (84.6%), and US Moderates (82.1%) exhibit a strong tendency to support EU policies, projecting an image of broadly pro-European parties. Conversely, the League (75.7% of statements falling within the “criticism” category) and FdI (58.5%) are the most critical parties, which is in line with their historically Euroskeptic positioning (e.g. Caiani and Conti, Reference Caiani and Conti2014; Conti et al., Reference Conti, Marangoni and Verzichelli2020). Further evidence of this critical stance is that, in 19 out of 53 critical statements by the League, the attack is directed against the European polity itself, not just its policies (for FdI, this figure drops to 9 out of 48).

Meanwhile, the M5S and the Left (Sinistra Italiana, SI) display a more balanced approach, distributing their engagement across support, critical, and alternative proposals. Notably, the Italian Left (65.6%) and, once again, the League (64.3%) register the highest share of “alternative” statements, suggesting an active role in shaping policy debates, not merely reacting to existing measures. All in all, the high share of alternative proposals among both ideological camps challenges the conventional Euroskeptic vs. pro-European dichotomy, highlighting that EU policy debates in parliament are more complex than a binary opposition. It suggests that political contestation over EU affairs frequently manifests as argumentative engagement rather than outright rejection.

For our multivariate analysis, we created three models, where the dependent variable is a dummy variable, taking the value of one if a party uses any of the three categories of statements. To test our first hypothesis about the impact of NGEU (H1), we created a time categorical variable taking one if the debate takes place after July 2020, namely, after the European Council meeting when NGEU was adopted.

Whereas party categories are considered a control variable in the first three models, they are included in interaction with NGEU to test our second hypothesis. In line with our reading of the relevant literature, we expect the impact of NGEU to be stronger for traditionally Euroskeptic parties, increasing their level of EU support and decreasing their oppositional attitude toward Europe to a greater extent than for mainstream parties (H2). To classify a party as Euroskeptic, we cross-checked two expert survey sources: populist (Rooduijn et al., Reference Rooduijn, Pirro, Halikiopoulou, Froio, Van Kessel, de Lange, Mudde and Taggart2023) and Chapel Hill Expert (CHES) surveys conducted in 2019 and 2024 (Jolly et al., Reference Jolly, Bakker, Hooghe, Marks, Polk, Rovny, Steenbergen and Vachudova2022, Rovny et al., Reference Rovny, Polk, Bakker, Hooghe, Jolly, Marks, Steenbergen and Vachudova2025). According to these surveys, three parties can be classified as such in the two legislative terms under study: FdI, the League, and the M5S. With that being said, their stances toward European integration varied over our study period. According to CHES data, the League maintained consistent and strong opposition. In contrast, M5S and FdI showed notable moderation in their EU positions, suggesting a shift away from hard Euroscepticism (see Table A1 in the Appendix). Including this ideological shift helps contextualize our subsequent analysis and reinforces our understanding of how evolving parties’ EU positions relate to changes observed in parliamentary rhetoric. If Hypothesis 2 is correct, then as Italy becomes one of the greatest beneficiaries of NGEU funding, these parties should moderate their Euroskeptic tone.

An important moderating factor we should control for is whether the speaker is affiliated with a party in government (1) or opposition (0) at the moment of intervening in the debate. Building on the responsibility-responsiveness dilemma (Mair, Reference Mair and Mair2014), the hypothesized effect is that parties in government moderate their Euroskeptic positions, while parties in opposition have less incentive to do so. Governing parties are expected to act responsibly, abiding by EU commitments, operating within the supranational constraints set by the EU, and collaborating with EU institutions. By contrast, opposition parties can be more responsive to their electorate and the prevailing public opinion (Spoon and Klüver, Reference Spoon and Klüver2014; Spoon and Williams, Reference Spoon and Williams2017). As a result, these parties have more leeway to criticize the EU and the EU-level compromise made by the government (De Vries and Hobolt, Reference De Vries and Hobolt2020). This also chimes well with the literature on oppositional behavior to the EU in parliament. Of particular interest are works that investigate it by ascribing considerable explanatory power to party characteristics, such as the degree of Euroscepticism (which relates to our H2) or whether parties are in government or opposition (Hoerner, Reference Hoerner2017; Senninger, Reference Senninger2017; Tuttnauer, Reference Tuttnauer2018; Karlsson and Persson, Reference Karlsson and Persson2022).

Empirical analysis

Table 2 shows the estimates of six regression models testing H1 (models 1–3) and H2 (models 4–6). We calculate standard errors clustered for each European Council object of the Prime Minister’s communication to account for external events and political dynamics that dominated the European summit on that occasion, and that might have affected all parties’ positions (for instance, summits focused on the Russian attack against Ukraine in February 2022). In the following, we report both average predicted effects (APE) and average marginal effects (AME) to provide a comprehensive picture of how party discourse has evolved. APEs illustrate overall trends in party rhetoric before and after NGEU adoption (Figures 1–4), while AMEs isolate the specific impact of our explanatory factors on party rhetoric. This allows us to distinguish between broad temporal shifts and the magnitude of the effect.

Table 2. Logistic regression results

Note: Coefficients from logistic regression models predicting the likelihood of different types of political statements (Support, Criticism, and Alternative), with and without interaction effects. Standard errors are clustered around the European Council to which the debate is dedicated and are reported in parentheses. Statistically significant coefficients are indicated with

+ p < 0.10 (+),

* p < 0.05 (*),

** p < 0.01 (**),

*** and p < 0.001 (***).

The empirical results indicate that transitioning from the pre-NGEU to the post-NGEU period is associated with an increased probability of parties expressing supportive statements toward the EU, along with a reduced likelihood of critical and alternative proposals. Figure 1 illustrates these effects using APEs. Although these results align with our theoretical expectations, only the increase in supportive statements approaches conventional levels of statistical significance. Specifically, AMEs reveal that the adoption of the NGEU correlates with approximately a 10 percentage-point increase in supportive rhetoric. At the same time, the estimated reductions for critical and alternative statements are smaller and statistically indistinguishable from zero. Thus, our first hypothesis receives partial support: the adoption of the NGEU is primarily associated with increased pro-EU rhetoric, but evidence for a corresponding decline in critical discourse remains inconclusive.

Figure 1. Average predicted probability of making a statement about the EU before and after NGEU.

Concerning our other control variable, government status emerges as a powerful explanatory factor for variations in party discourse toward the EU, significantly increasing support and alternative statements (AMEs: 33% and 11%, respectively) and substantially decreasing critical statements (AME: −22%). This finding suggests that the moderation of Euroscepticism cannot be attributed solely to the economic incentives provided by the NGEU but also reflects institutional and strategic pressures associated with government responsibilities.

But were traditionally Euroskeptic parties more affected by this shift than their mainstream counterparts? The analysis of supportive statements reveals a significant increase in the likelihood of making pro-EU statements for two of these parties: FdI, whose APE of supportive statements rose dramatically from 0% in the pre-NGEU period to 50% post-NGEU, and the League, whose probability increased from 7% to 32% (Figure 2). In contrast, M5S showed only a slight change. When examining AMEs, the transition to the NGEU period increased the probability of supportive rhetoric by 45% for FdI and 20% for the League. Our analysis of APE for critical statements similarly highlights this moderation effect: both right-wing parties emerge as the only ones noticeably softening their critical positions (Figure 3). In substantive terms, the introduction of NGEU contributed to reducing the likelihood of FdI and the League issuing critical statements by 30% and 19%, respectively. Finally, the analysis of alternative statements reveals limited overall change in APEs (Figure 4). An important exception is M5S: moving from pre-NGEU to post-NGEU decreases their probability of making an alternative statement by 29 percentage points.

Figure 2. Average predicted probability of making a support statement about the EU before and after NGEU by party.

Figure 3. Average predicted probability of making a critical statement about the EU before and after NGEU by party.

Figure 4. Average predicted probability of making an alternative statement about the EU before and after NGEU by party.

Overall, these results suggest that the post-NGEU phase is associated with a softening of hard Euroscepticism, as parties such as the League and FdI move away from outright rejection toward more pragmatic positions on EU matters. Conversely, pro-European parties show minimal variation. Collectively, these findings partly confirm our second hypothesis. They highlight how the demonstration of EU financial solidarity in response to the pandemic crisis may have contributed to fostering a political atmosphere marked by more tempered Euroscepticism. Overall, these findings suggest that while the NGEU may have helped reshape parties’ critical rhetoric toward the EU, its impact on their tendency to put forward constructive criticism (i.e. alternative statements) was more limited. This highlights the need for a future examination of the strategic repositioning by specific parties, particularly the M5S, to make sense of these results.

While our results illustrate significant temporal patterns, it is crucial to remember that they do not establish a direct causal relationship between NGEU and changes in party discourse. Moreover, our analysis cannot confirm whether shifts in rhetoric resulted primarily from a utilitarian logic based on parties’ cost-benefit calculations or from broader perceptions of increased EU legitimacy due to its responsiveness to recent crises. Additionally, we cannot yet determine whether party rhetoric leads or follows public opinion (Hooghe and Marks, Reference Hooghe and Marks2005; Hooghe, Reference Hooghe2007; Williams and Spoon, Reference Williams and Spoon2015). To partially address this last limitation, we complemented our analysis with public opinion data from the Eurobarometer, which provides biannual measurements of citizen support for the EU (Table A9 in the Appendix). Our analysis reveals a strong correlation between shifts in public support for the EU and our NGEU-based periodization, suggesting that the adoption of the NGEU coincided with a significant increase in EU support (on this, see also Papageorgiou and Immonen, Reference Papageorgiou and Immonen2023; Unger et al., Reference Unger, Sirsch, Stockemer and Niemann2023). On average, our index rose from 56 before NGEU to 65 after, potentially reflecting a positive public response to the EU’s financial intervention and crisis management efforts. When we replace the binary NGEU variable with our continuous measure of public EU support, we find that for every 10 percentage-point increase in our index, the likelihood that parties express supportive statements rises significantly by approximately 14%. Conversely, the probability of making critical or alternative statements EU is negatively associated with the likelihood of making critical or alternative political statements, though the strength and certainty of this relationship vary across models. Thus, while public opinion appears to be a clear driver of increased supportive rhetoric (a trend broadly aligned with the patterns observed in Table 2 for the NGEU dummy), we find limited evidence that it independently reduces critical or alternative forms of contestation toward the EU.

Another issue to discuss is the difficulty of fully disentangling the effects of NGEU and incumbency due to data limitations in our multivariate model. To address this, we conducted additional robustness checks based on subsets of the data in which party status remained constant (e.g. FdI in opposition; League and M5S in government), which are suggestive of rhetorical moderation independent of role-switching (Tables A10–A13), though the evidence is not uniform across parties or categories of statements. An analysis focused exclusively on the Conte II government, where party roles were fixed, does not show clear signs of rhetorical moderation among Euroskeptic parties following the NGEU's adoption in July 2020. We attribute this null result to the short postadoption period before the government’s collapse in January 2021, which left little time for parties to adjust their rhetorical strategies.

To further assess whether the rhetorical impact of NGEU differed between government and opposition parties, we estimated two alternative model specifications. First, Table A14 reports estimates from a binary logistic regression with party fixed effects, including an interaction term between in-government status and the post-NGEU period indicator. Second, Table A15 presents results from a multilevel (mixed-effects) logistic regression with a random intercept for party–cabinet combinations, which removes the collinearity between party fixed effects and the interaction term while allowing intercepts to vary by party–cabinet unit. In both specifications, the estimated interaction term is small in magnitude and not statistically significant for any of the three dependent variables, indicating that government and opposition parties exhibited broadly similar changes in discourse following NGEU. The consistency of this result across fixed-effects and random-intercept specifications suggests that the observed moderation cannot be attributed solely to incumbency-related incentives.

Finally, we also conducted two additional robustness checks. First, we re-estimated our models excluding all statements made between May and December 2020, the period immediately following the announcement of NGEU, to assess whether this salient debate window influenced our results. As shown in Table A16 in the Appendix, the direction and magnitude of the coefficients remain substantively unchanged, though confidence intervals widen slightly due to the reduced sample size. Second, we used an alternative temporal cutoff – August 2021 – corresponding to the date Italy received €24.9 billion in prefinancing from the European Commission (13 August 2021), to test whether the receipt of the first tranche might constitute a more meaningful watershed (Table A17). Re-running our logistic regression models with this alternative periodization likewise did not substantively alter our results.