Supranational emergency politics? What executives’ public crisis communication may tell us

ABSTRACT This contribution engages with the empirical analysis of emergency politics in the EU, arguing that executives’ public communication helps to distinguish crisis management from crisis exploitation. An initial, descriptive text analysis of emergency emphasis in more than 19,000 executive speeches suggests that supranational actors, most notably the European Central Bank, do indeed use rather alarmist language over and beyond objective crisis pressures when their competences are contested. Yet, this behaviour does not appear to be a ubiquitous phenomenon, pointing to the need for more specific expectations on when and why EU executives pro-actively embark on the emergency politics script.


Introduction
Kreuder-Sonnen and White (2021) aim to stimulate a more critical reading of crisis responses in and by the European Union. They want us to move beyond purely functionalist analyses of crisis management. They are not primarily interested in whether executive actors have responded efficiently or effectively to mounting challenges. Theirs is rather a story of crisis exploitation. Building on long-standing thoughts on emergency rule in the nation state, Kreuder-Sonnen and White argue that there is a recurring 'script' of 'emergency politics'. In this perspective, extraordinary circumstances are not solely a challenge for executive actors. Such circumstances also present strategic opportunities to limit the choice set of possible political responses. By reference to sheer necessity and inevitability, executives may utilize crisis events to silence controversial debates and circumvent procedural rules constraining them in 'normal' times. Kreuder-Sonnen and White thus want us to focus on whether and how executives in the EU (ab)use exogeneous threats to pursue policy choices that face political or constitutional resistance.
Not least by making Jean Monnet a prime witness, Kreuder-Sonnen and White draw particular attention to supranational executives. In the unfinished EU polity, the distribution of competences among different executives, and especially the delegation of political powers to supranational institutions, as well as their constitutional restraint, are notoriously contested. And the political resistance that particularly supranational executives face in this regard has grown. While resistance came mainly from individual governments in Monnet's day, today's supranational decision-making must be managed in the context of a broader, societal EU politicization (for recent overviews: Rauh, 2021;Schimmelfennig, 2020). Controversial public debates may indeed incentivise crisis exploitation further.
This focus on the darker side of European crisis responses is as appealing as it is frightening. It provides more explicit agency to long-standing normative concerns about integration by stealth, the prevalence of technocracy over democracy, or processes of alleged de-politicization in and through the EU. It also lends credence to a partisan mobilization strategy that has unsettled political competition in various EU states recently. The often successful populist critique of elitist decision-making as out of sync with popular will can be seen, Kreuder-Sonnen and White claim, as the mirror image of executive behaviour following the emergency politics script (Kreuder-Sonnen & White, 2021; see also White, 2019, ch. 6). Especially these far-reaching normative implications warrantdare I say: necessitatecareful scrutiny. Is executive behaviour during recent crisis periods consistent with the proposed emergency-politics script?
This contribution aims to stimulate debate on how supranational emergency politics can be observed empirically. I highlight that the perceptual nature of crises, as well as the partial observational equivalence of crisis management and exploitation, are key epistemological challenges in this regard. Yet a decidedly comparative perspective on the communicative behaviour of executive actors should provide important clues to emergency politics in action.
An initial descriptive analysis of more than 19,000 public speeches delivered between 1989 and 2020 shows that the European Commission and the European Central Bank (ECB) have indeed increasingly employed emergency language especially when supranational competences were strongly contested. Yet, while the ECB continues to use markedly alarmist language, the European Commission scaled down its public emergency emphasis after 2009 and before the onset of COVID-19. During the Eurocrisis, the Commission actually sent less alarmist signals than many national executives. The latter emphasized the emergency to varying degrees, in patterns that do not squarely match objective crisis affectedness or stated preferences on European responses.
These patterns do not exclude that supranational emergency politics is at play. Yet they also suggest that the emergency politics script is not an ubiquitous phenomenon. Rather we need more careful specification and empirical research about when and why specific EU executives embark on a strategy of more pro-active crisis exploitation.

Observing emergency politics through executives' public communication
How can we distinguish a crisis manager from a crisis exploiter? Initially, one is tempted to look at the measures that executives enact in response to an exogenous threat. A functionally motivated crisis manager striving to restore normality should pursue targeted measures designed to eliminate the threat as efficiently as possible. A crisis exploiter interested in using the threat for specific policy goals, in contrast, is likely to overshoot by designing measures going well beyond the immediate demands that a threat lays bare. As others have noted before me, however, assessing such proportionality runs into two fundamental challenges.
First, crises are least partially perceptual in nature. While identifiable events at contained points in time usually serve as a trigger, the existence, magnitude, and persistence of the resulting emergency is not entirely objective (Billings et al., 1980;see also Kreuder-Sonnen, 2019, pp. 63-64, 72). Whether specific events are read as existentially threatening depends on how successfully they can be related to fundamental features of the affected political system. Crises are socially constructed through 'framing contests' among competing political interests (Boin et al., 2009). The measures that executives enact in response therefore do not provide easily accessible information that would distinguish crisis management from crisis exploitation: the left-hand side of the proportionality equation is blurred.
Second, crises involve high levels of uncertainty and urgency by definition. Where low levels of information combine with high time pressure even the most well-meaning crisis manager may not enact the most appropriate and proportional countermeasures. Overshooting, unintended consequences, or inconsistencies with extant policies are not surprising in demanding conditions even if one assumes the best of intentions on part of executive decision-makers (Allison, 1969). Also from this perspective, the measures that executives enact do not readily provide information to distinguish crisis management from crisis exploitation: the right-hand side of the proportionality equation is blurred.
The proportionality of the enacted measures thus provides only limited insight. On the level of outputs, functionally motivated responses will often be observationally equivalent to more strategically driven policies. Exactly the same characteristics that make emergency politics so appealing for constrained executives in the Kreuder-Sonnen and White model also render their empirical assessment tricky. That executives respond drastically to a (widely shared perception of) emergency is in itself insufficient evidence for the existence of the emergency politics script. The normative bite of the argument rather lies in the strategic motivations behind such executive responses. Yet how can we observe these motivations if not in the policies that executives enact?
One major behavioural implication of the emergency politics model, in my view, lies in the public communication of executive actors. Kreuder-Sonnen and White (2021) themselves stress that policies are only one side of emergency politics. Equally constitutive of emergency politics are executives' attempts to cultivate the crisis nature of the circumstances to then rationalize their measures in the light of the proclaimed emergency (see also Scicluna & Auer, 2019). Especially such communicative behaviour should help distinguish functionalist accounts from emergency politics, or crisis management from crisis exploitation.
Consider the ideal-typical, purely functionally minded executive first. Confronted with a widely shared perception of crisis, restoration of normality in the most effective and efficient way possible should be their prime concern. This executive will try everything to avoid the situation getting out of hand further. In this regard, 'communication is the essence of crisis management' (Coombs, 2010, p. 25). To enable rational responses to a threat, the affected audiences need to be addressed appropriately, not least by mitigating the negative emotionsanger, fear, anxiety, etc.that typically accompany the shocks and uncertainty of a crisis (Jin & Pang, 2010). In the 'framing contests' of crisis interpretation, an executive interested in efficiently and effectively responding to an exogeneous threat should thus engage in reassuring and comforting communication signals.
Contrast this with the ideal-typical executive in the emergency politics model. For these actors, a widely shared perception of an exogenous threat is not only a problem to solve but also an opportunity to circumvent critical debates and procedural constraints. Adding such policy motives to the picture means that a quick alleviation of a given threat is not the prime strategic interest of such actors. The emergency politics modelmuch in line with the speech-act theory in the securitization literaturerather predicts an interest in keeping the threat level up. This way executives can dampen critical debates about different courses of action and achieve deference from those governed (Buzan et al., 1997). In the 'framing contests' of crisis interpretation, an executive interested in exploiting the crisis should thus send comparatively more alarmist communication signals.
In a nutshell, the communication of a crisis manager should be geared to calming things down, while the communication of a crisis exploiter should aim at sustaining the level of crisis perception instead. One must be clear, though, that I am speaking about differences in degree here. Even the wellmeaning manager often cannot avoid highlighting the crisis nature of given events. But the crisis exploiter will frame matters as an emergency in a much more pronounced and, relative to the original crisis-triggering events, much longer lasting manner (i.e., also during what Vivien Schmidt, 2021, calls the 'slow burning' phase of a crisis). Based on these considerations I derive an empirically observable implication from the Kreuder-Sonnen and White model: If the emergency politics script is taking hold in European governance, we should see higher levels and prolonged emphasis on the state of emergency in the public communication of those European executives whose competences and policies are politically contested and/or constitutionally constrained.

Emergency emphasis in the public communication of European executives
To get an initial empirical impression, this section examines emergency language in the public communication of the European Commission and European Central Bank. Both hold executive powers delegated from the national level through competences specified in the European Treaties. Both actors, however, have been repeatedly accused of attempting to extend their executive reach over and beyond these delimited competences. Both actors have also seen their competences heavily contested in intergovernmental and public debates. And for both actors there is initial evidence that their public communication responds to controversial debates (Moschella et al., 2020;. I have scraped all 10,948 English-language speeches of European Commissioners since 1989 from the Commission's online press archive, 1 as well as all 2,225 speeches from the ECB president and directors since 1997 from the Bank's speech data (European Central Bank, 2020). The EUSpeech collection (Schumacher et al., 2016), furthermore, allows us to compare this supranational communication to 6,127 speeches by heads of state or government from nine EU member states as well as 236 speeches from the Council president during the high period of the Eurocrisis (2009)(2010)(2011)(2012)(2013)(2014)(2015).
Measuring emergency emphasis in these speeches needs to optimize three things. First, given the corpus size, the measure must provide reliable comparisons at scale. Second, it should be sensitive to the possible rhetorical creativity that political actors may employ when invoking or alluding to a state of emergency. Third, the measure should be agnostic as to the policy content of specific crises or language changes over recent decades.
To construct such a measure, I resort to what the recent text-as-data literature has denoted as latent semantic scaling (Rheault et al., 2016;Wantanbe, 2020). It starts from just a few selected words that are intuitively associated with the polarity of interest. Here I want to capture the emphasis on emergency ('crisis ', 'danger', 'peril', 'hazard', 'threat', 'risk', 'disaster', 'uncertainty', 'uncertain') over normality ('normal', 'safety', 'stability', 'regularity', 'routine', 'calm', 'usual', 'certainty', 'certain'). With unsupervised machine-learning one then gathers a longer list of words that are semantically similar to either side in the language domain of interest.
As a substantially broad and also time-invariant representation of political English, I use 1,856,597 speeches given in the House of Commons between 1985 and 2019 (drawn from ParlSpeech, Rauh & Schwalbach, 2020) and 'learn' words associated with the latent normality-emergency scale with the help of a word-embedding model (Spirling & Rodríguez, 2020). Such algorithms formalize the idea of distributional semantics in linguistics according to which 'you shall know a word by the company that it keeps' (Firth, 1957). In this view the meaning of a word is embodied by the words in its immediate contextwords that repeatedly and consistently occur in the same context are semantically similar. Accordingly, the algorithm uses a moving context window to traverse through the reference corpus, builds a word-to-word co-occurrence matrix, and then reduces the number of dimensions by optimizing the prediction of present words given the presence of other words and vice versa. In result, each word is described by an n-dimensional vector along which semantic similarity can be measured. 2 I accordingly calculated the cosine similarity for each word spoken in the House of Commons to the average vector of the normality and emergency seed terms, respectively. Subtracting one from the other then tells us where each word lies on the latent semantic dimension between both concepts. The 250 terms closest to each side of this dimension are then used to scale each of the 19,541 speeches of supranational executives along the weighted frequency of these terms. Figure 1 illustrates the measure along 100 exemplary terms drawing a speech to either the normality or the emergency pole. Figure 2 plots the annual averages of the normality-emergency scale for the European Commission and ECB speeches and compares them to the average scale value in a random sample of 100,000 House of Commons speeches during the 1985-2019 period. For preliminarily assessing the claim that supranational emergency politics has taken hold in the EU, these descriptive data contain three key messages.
First, the public communication of both supranational executive institutions tends to emphasize the emergency situation more than the average speech of elected British MPs. For the majority of years, we see significantly more alarmist language in the public speeches of EU Commissioners and ECB officials than in typical parliamentary debates. This does not readily prove, but is consistent with the claim that the striving for supranational executive discretion is an underlying motivation here.
Second, we also see that supranational executives have indeed increasingly and markedly emphasized emergency conditions in the last 15 years especially. This initially supports the view that the emergency politics script is taking hold in EU decision-making, but caution is warranted. As noted, even the most well-meaning crisis manager will need to speak about crises if objectively undeniable trigger events occur. And during the period of heightened emergency emphasis that we can observe in Figure 2, the EU indeed experienced a number of such events ranging from the Lehmann collapse in 2008 and the Eurocrisis triggered in 2009, to the sudden surge of migrants seeking refuge in the EU during summer 2015, to the spread of COVID-19 in early 2020. As such, we cannot readily infer that the heightened emergency emphasis is driven by the desire to keep the threat level up, over and above objectively challenging conditions. Third, however, the data also shows that such 'objective' crisis affectedness cannot be the sole driver of executives' communicative behaviour. There is notable variation over time, and especially across the two executive institutions under observation. Even when both face similar exogenous conditions, their inclination to emphasize the emergency in their public communication differs markedly at different points in time. Let us briefly consider each institution in comparison.
For the Commission (dots and solid line in Figure 2), the first local maximum of emergency emphasis can be observed around the negotiation, conclusion, and thorny ratification phase of the Maastricht Treaty 1991-1993. This is no definitive proof but it is consistent with the view that executives stress emergencies especially when their discretion is threatened. The Maastricht negotiations revealed that the Commission would not benefit from further authority transfers as much as in the prior Single European Act. National governments moved away markedly from rolling out Monnet's 'community method' further. Moreover, the protests against the Treaty, the highly contentious debate in the UK, and the extremely close referenda in Denmark and France signalled waning public support for further supranational empowerment. Seen from the Kreuder-Sonnen and White model, the Commission had incentives to adopt alarmist language. But its public emergency-talk fell again afterwards. It rose only in 2005, coinciding with another highly politicized and consequential period for supranational discretion around the failure of the Constitutional Treaty. Emergency emphasis in the Commission's public communication then grows especially with the onset of the financial and then Eurocrisis. Surprisingly, though, it then reverts to markedly lower levels from 2010 onwards. This does not easily fit the 'objective' evolution of the Eurocrisis at the time. And it challenges the emergency politics argument, as the Commission's extant and future discretion was contested during this period. The surge of Eurosceptic parties, and especially the debates about the Commission's involvement in the Troika and its role in fiscal supervision, should have incentivised this institution to play along with the emergency script. Yet, at least according to the present data, it did not.
The trajectory of emergency emphasis in the public communication of the ECB (triangles and dashed line in Figure 2) looks rather different. In the earliest available speeches of the ECB president and directors, the emergency emphasis is low. In fact, it is much lower than in the average Commissioner or MP speech. The ECB officials' public emphasis on emergency, however, rises in parallel to the global financial crisis in the years 2007-2008. With the subsequent onset of the Eurocrisis, and the growing spread of interest rates on government bonds of EMU states during 2009-2011, the ECB's emergency language takes a particularly marked leap. Given that the economic challenges during these crises touched on the ECB's competences rather directly, this pattern alone does not necessarily indicate crisis exploitation, though.
More interesting from the perspective of the emergency politics argument is the observation that the alarmism in the ECB's public communication did not revert to its prior levels even as the immediate market pressures waned, notably after the famous 'whatever it takes' statement by Mario Draghi in July 2012. This speechwith a value of .69 an extreme outlier on the normality-emergency scalecalmed the markets rather effectively as can be seen in the dramatically decreasing spread in yields on 10-year bonds of Eurozone countries in its immediate aftermath. But from a political perspective it arguably did so by signalling and paving the way for the ECB's outright self-empowerment. Draghi indicated a potentially limitless intervention by the Bank in sovereign bond markets, an idea that was shortly thereafter institutionalized in the OMT programme which added the idea of conditionality to the bond-based bailouts the ECB had been pursuing since 2010. Together with its role as an austerity enforcer in the Troika and its increasing involvement in banking supervision, the ECB thereby moved clearly beyond its Treaty mandate, confined to monetary policy and price stability, and extended its discretion to fiscal and economic policy (Kreuder-Sonnen, 2019, ch. 5). These moves were and are strongly contested, not just within the affected societies but also among integration-wary circles in domestic politics, as indicated by the repeated complaints raised by the German Constitutional Court between 2015 and 2020 (Scicluna & Auer, 2019). Given this combination of executive self-empowerment and growing political contestation, an executive actor following the emergency politics script should uphold its public emergency emphasis even if the immediate crisis pressures wanewhich is consistent with the aggregated communication patterns that we see here.
Figure 3 zooms in to 2009-2015 period when arguably all European executives were focussing strongly on the Eurocrisis and compares the average emergency emphasis of the Commission and the ECB to that in public speeches of the Council president and national executives. This cross-sectional perspective shows that we cannot attest a decidedly supranational emergency politics script. While the ECB's outlier role is reaffirmed, the Commission figures among those actors with the least alarmist public communication, whereas the Council president falls somewhere in between. As a group, supranational executives did not systematically emphasize the emergency more than the group of national executives.
The emergency emphasis adopted by national heads of state or government also tells us something about multilateral emergency politics (Kreuder-Sonnen & White, 2021). During the high period of the Eurocrisis, national executives did not send a unified emergency message but differed in the communicated alarmism. Strikingly, the pattern does not easily sit with explanations rooted in crisis affectedness, domestic political resistance, or political preferences regarding the countermeasures discussed at the time. Neither do national budget deficits (and the related risks of sovereign default), the degree to which austerity politics stirred domestic conflict, nor the distinction between creditor and debtor states (or promoters of austerity and redistribution) align with the degree to which the respective national executives communicatively played up the state of emergency during the Eurocrisis.

Conclusions
Kreuder-Sonnen and White initiated this debate with the expectation that the crises the EU has faced and still faces provide strategic opportunities to supranational actors willing to overcome constitutional and political constraints. This claim has far-reaching normative implications and should compel political scientists to scrutinize the proposed emergency politics argument empirically. To stimulate a debate on the systematically observable implications, I argue that executives' communication provides some leverage with which to discriminate between mere crisis managers and more forthright crisis exploiters pursuing political goals beyond immediate crisis alleviation: while the former should try to calm the affected audiences down, the latter have an interest in comparatively higher and more sustained public crisis perceptions beyond the triggering events.
An initially descriptive, large-n text analysis of emergency language in public speeches by the European Commission and ECB provides mixed evidence for the expectation that supranational executives increasingly and consistently play along the emergency politics script. Certainly, crisis affectedness alone does not seem to explain when and how strongly supranational executives emphasize emergency conditions in public communication. And indeed, supranational actors tend to publicly play up the emergency when their attempts to expand their discretion meet resistance in public and intergovernmental debates. Especially the ECB stands out in this regard. Yet we also find patterns that deviate from the script. Supranational executives, on average, are no more prone to employ emergency language than national executives. For the European Commission we partially observe restraint in employing emergency language, even in periods when its discretion was questioned in public and intergovernmental debates. And, at least at face value, the emergency emphasis of national executives does not readily align either with the factual and political constraints they encountered during the Eurocrisis.
Clearly, these initial descriptions are a starting point only. My aggregate description has been rather agnostic regarding the actual contents and varying policy portfolios of the speaking actors. Future research should identify the specific topics in the speeches, match topics across speakers (see Roberts et al., 2020 for respective tools), and then link the patterns of emergency emphasis within topics to more objective crisis indicatorse.g., regarding the influx of migrants during and after 2015 or the spread of Covid infections in more recent periods.
Yet and still, the initial patterns shown here are consistent with the view that pro-actively engaging in emergency politics is part of the strategic toolkit of European executive actors. But this strategy is not ubiquitous. We need more specification of the conditions incentivising executives to embark on pro-active emergency politics. More finegrained hypotheses could focus the specific political goals that executives pursue, the specific constraints they face, and the relationship of these factors to specific crisis-triggering events. I hope that the arguments and tools presented here prove useful in developing and testing such arguments in future.