Politics, pandemics, and support: the role of political actors in Dutch state aid during COVID-19

How do governments distribute resources across economic sectors during a crisis? And why do some sectors receive more than others? The recent COVID-19 pandemic has highlighted the urgency of these questions. In this paper, we explore the extent to which a political economy perspective can help explain the characteristics of sector-specific state aid in the Netherlands, a traditionally corporatist country. While KLM, the biggest player in the Dutch aviation sector, was promised loans worth €3.4 billion, the horeca (hospitality) sector was denied a similar deal. Limited cross-case analysis eliminates purely economic accounts. We employed process-tracing and analyzed hundreds of national media articles to understand the influence of elected leaders, interest groups, and experts. We find that, against the backdrop of economic concerns, vote-seeking behavior by elected leaders as well as the strength and organization of interest groups influenced how much each sector could expect. Meanwhile, policy-seeking behavior helps explain the form that aid took. Our findings highlight the need to consider fiscal support in political economy terms, even during crises, and to explore the composition of state aid, not just its presence or amount. de dichas ayudas se explica a través del alineamiento político de los decisores políticos. Nuestros resultados subrayan la necesidad de estudiar el apoyo gubernamental en términos de economía política, inclusive durante crisis, y de explorar no solo la existencia o la cuantía de dichas ayudas sino también su composición. Palabras clave: COVID-19; ayuda estatal; apoyo fiscal; rastreo de procesos; análisis de influencia sobre decisiones políticas; electoralismo; lobbies.


INTRODUCTION
As the world locks down to contain COVID-19, a sharp economic recession takes hold (Tooze, 2020). Firms around the world struggle with mandated closures or depressed demand. In response, governments have unveiled massive fiscal stimulus packages. In mid-March, the European Commission loosened its restrictions on state aid, defined as an advantage conferred selectively to specific undertakings by national governments (Baczynska & Koranyi, 2020;European Commission, 2019). Yet not all sectors have been affected to the same degree. Under these conditions of extraordinary crisis, how do governments decide whose burdens to alleviate? Which sectors get priority? And how does the type of state aid compare across awarded sectors?
In this paper, we unpack the logic behind sector-specific support during crises. We do this with a case study of the Netherlands. Widely categorized as corporatist (Visser & Hemerijck, 1997), this case promises exceptional insight into the political economy of state aid. In mid-March, announcing a multi-pronged emergency stimulus, the Dutch finance minister Wopke Hoekstra acknowledged, "this is going to hurt... we can't rule out that there will be businesses that won't make it" (Hoekstra, 2020). Here, we investigate Dutch decision-making regarding which businesses would be spared, and how. To this end, we take aviation and hospitality (in Dutch: horeca, for Hotel-Restaurant-Café) sectors as points of departure, and spotlight the complexities of decision-making during the early stages of the crisis. Neither sector made it to the list of essential occupations published in mid-March by the Dutch government (Rijksoverheid, 2020a) 1 . Both contribute substantially to the Dutch economy and were exceptionally hard-hit by COVID-19 measures. However, while the country's largest aviation group, Royal Dutch Airlines (Koninklijke Luchtvaart Maatschappij [KLM]) was promised €3.4 billion in loans and loan guarantees (Hoekstra & van Nieuwenhuizen, 2020b;Rijksoverheid, 2020b), the national association representing the horeca (Koninklijke Horeca Nederland [KHN]) was denied a similar emergency package. Instead, they were left with non-sector specific compensatory packages and relaxed rules around public gatherings. Complicating matters further, right before this article was published, the Cabinet's commitment to KLM wavered as it declared the stimulus package conditions inadequately met and the company's agreement with its unions broke down, jeopardizing the disbursal of subsequent aid installments (Obbink & Lammers, 2020;Duursma, 2020e). This seeming erraticism underscores the urgency of considering the role of politics in economic decision-making.
These decisions cannot be explained with standard economic logic by which governments allocate capital according to need and strategic value (Tenreyro, 2020). We argue that the relative strength of organized interests and vote-seeking elected leaders sensitive to public opinion help explain how much each sector received. Meanwhile longer-term policy aims mattered more for what the package ultimately looked like in terms of conditions and composition. The next section borrows from existing literature to craft a theoretical framework with expectations about how policy actors influence state aid. Next, we justify our methodological decisions. Case analyses and discussion follow.

THE POLITICS OF ECONOMIC POLICY
State aid raises the central economic problem of how scarce resources are distributed. It is unavoidably subject to political wrangling (Hall, 1986;Schito, 2020). In this paper, we draw from a model developed by Baekkeskov (2016) to evaluate the influence of three actors: elected leaders, organized interests, and experts on the character and composition of state aid. We separate these actors to disentangle their influence, but in reality, any given outcome is likely to reflect myriad interests to varying degrees.

Elected leaders
Classical democratic theory expects elected leaders to cater to their voters (see Romeijn, 2018 for an overview). Because state aid concentrates benefits in specific sectors and diffuses costs across tax-payers, we can expect voters and their representatives to support state aid in accordance with employment structure (vote-seeking) (McGillivray, 2004, p. 8). However, incomplete information (Downs, 1957) and "extra-rational or irrational prejudice" (Schumpeter, 2010, p. 235) means that the public's position can be difficult to predict. For instance, voters may be more sensitive to potential losses than gains (Weaver, 1986). Besides seeking votes, elected leaders may be concerned with maximizing their impact on public policies they deem 'good' (policy-seeking) or advancing their own career by securing a cabinet position (office-seeking) (Degner & Leuffen, 2016;Strøm & Müller, 2010).
The current coalition government (2017)(2018)(2019)(2020)(2021) in the Netherlands is the third led by Mark Rutte of VVD (conservative-liberals). It was formed in 2017 with two Christian-democratic parties and D66 (social-liberals) after protracted negotiations and one failed attempt. The coalition government assumed power with the smallest possible majority in both the Second Chamber (76 of 150 seats) and the First Chamber (38 of 75 seats). Its platform included commitment to climate action and reduction in profit taxes (VVD, CDA, D66 & ChristenUnie, 2017). Elected officials simultaneously answer to their constituents, who are typically geographically dispersed (Andeweg, 2005, p. 497) and with relatively flexible party allegiances (van der Eijk & Niemöller, 1984, p. 193), leaving Rutte III vulnerable to public sentiment with 2021 elections around the corner.
If the interests of elected leaders drove the outcome, we would expect to find that state aid packages: (a) Are viewed favorably by the public. (b) Contribute to the ambitions of the current coalition platform.

Organized interests
While some view interest groups as merely "mobilizers of minorities" and therefore obstructive of the democratic process (Schattschneider, 1942, p. 193), pluralistic accounts depict them as important intermediaries between citizens and policy-makers (see Albareda, 2018;Flöthe, 2019, p. 166). As regards economic policy, key interest groups are employers' associations and trade unions. In corporatist systems, these 'social partners' are formally included in the policy-making process. However, access does not mean influence. Density, or the proportion of potential members of a given sector or constituency that an interest group can claim to represent matters (Andeweg & Irwin, 2014, p. 189). Lobbying success also depends on an interest group's ability to cohesively aggregate and articulate its interests (Braun, 2016, p. 40). The Netherlands has a long history of corporatism (Visser & Hemerijck, 1997, p. 19). The government consults a tripartite council (the Socio-Economic Council, or SER) consisting of employers, unions and government-appointed experts on every important socioeconomic policy change. Although participating interest groups have a central deliberative role, not all interest groups get to participate. 'Outsider' groups may strategically attempt to gain access to formal processes (Braun, 2016, p. 38).
If interest group mobilization affected state aid outcomes, we would expect to find that rewarded interest groups: (a) Had access to decision-makers. (b) Organize a large share of the total sector. (c) Made cohesive demands that are reflected in the outcome.

Experts
Elected leaders need technical knowledge to make decisions (see Flöthe, 2019). Governments listen to a diverse group of experts that includes public or private research units and think tanks (Campbell & Pedersen, 2015;Rosenthal & 't Hart, 1991). Expert influence is enhanced under uncertainty (Haas 1992;Rosenthal & 't Hart, 1991, p. 251) -a key feature of COVID-19. Previous research comparatively assessing national pandemic responses finds that norm-based advice of leading health experts steered countries' vaccination policies (Baekkeskov, 2016). Economists, who hold a privileged place in modern policy-making, have been lent increased discretion particularly in debates concerning how the economy and society should be governed (Christensen, 2018, p. 3), with their expertise used to boost the legitimacy and rational appearance of a policy decision (Meyers & Rowan, 1977). This is especially pronounced when their role is historically institutionalized in state bureaucracies (Christensen, 2017, p. 12).
Economists are typically skeptical of state aid, apart from to correct market failures, at an early stage of an industry cycle or in "extreme" situations (Falck, Gollier & Woessmann, 2011). However, state aid has staged a tentative recovery in mainstream economic discourse, as the 2008 banking crisis made clear the pervasiveness of market failures (Stiglitz, Lin & Monga, 2014, p. 6). There is some discussion of its merits in boosting strategic trade positions (Schito, 2020, p. 6). Despite comparatively low rates of state aid allocation (Schito, 2020, p. 4), the Dutch have deviated at certain points in their history from their fairly liberal, non-interventionist stance (van Zanden, 1999, p. 177). In the Netherlands, key corporatist institutions seat experts, like university professors that specialize in economics and labour relations, next to unions and employers (Andeweg & Irwin, 2014, p. 192).
If experts affected state aid outcomes, we would expect they: (a) Reward sectors in line with their long-term strategic returns to the Dutch economy.
And that experts: (b) Were consulted throughout decision-making stages.

RESEARCH DESIGN
Our single case study examines the Netherlands. Following Gerring (2007, p. 20), we understand a case study as the intensive study of one instance where the goal is to shed light on a larger class of cases or phenomena. Here, we are interested in the political economy of state aid decision-making. Given its corporatist traditions and 'polder model' of consensus-based economic and social policymaking (Braun, 2016), the Dutch case offers ample insight into these dynamics. We study state aid during early stages of the COVID-19 crisis. The Dutch fiscal response started in earnest on 17 March, when finance minister Hoekstra issued a three-month, €20 billion emergency financial assistance package for affected sectors. This scheme included NOW (Noodfonds overbrugging werkgelegenheid), which compensated up to 90% of employee wages for employers facing more than 20% turnover loss, and TOGS (Tegemoetkoming ondernemers getroffen sectoren) which reimbursed entrepreneurs in affected sectors with up to €4000 (Netherlands Chamber of Commerce, 2020). On 20 May, a second emergency package was introduced, injecting an additional €13 billion to help companies until August 2020 (Brandsma, 2020c). Here we focus mainly on the time period between these two packages. 2 In practice, sectors did not reap equal benefit. We focus on two sectors: aviation and hospitality. Both quickly requested and were granted eligibility for the NOW scheme (Vollebregt, 2020). However, on 24 April, the Dutch cabinet announced that KLM would receive an additional €2-4 billion in loans. KLM responded with "great gratitude and appreciation for the outspoken and significant support from the Dutch government" (KLM, 2020). On 26 June, the details of this package were announced: €1 billion in loans directly from the state and €2.4 billion in state guarantees for bank loans for a total of €3.4 billion (Rijksoverheid, 2020b). The package is conditional on KLM meeting a number of conditions.
Hospitality received far less. Days before the first emergency package was announced, the sector's employers' association Koninklijke Horeca Nederland (KHN) was denied the €5.1 billion in state loans it sought. It deemed the generic support measures it received inadequate, stating its frustration with policymakers: "politicians and ministers say that the hospitality industry is very important to them, but we have not yet seen this in concrete solutions and support measures. We feel extremely let down. The current support package is insufficient and does not cover all the costs that many hospitality businesses incur" (van der Reijden, 2020).
These differences in capital allocation cannot be explained by differences in sectoral need. On 15 March, hospitality was among the first sectors mandated to close (Bruins, 2020). Low cash reserves on average make the catering industry especially vulnerable in conditions of revenue loss (Bosio & Djankov, 2020). KHN predicted a 50% loss in turnover for 2020 (De Ondernemer, 2020). Air travel also aches. For KLM's CEO Pieter Elbers, difficulties started with progressive travel bans in early March (De Telegraaf, 2020a). COVID-19 eventually grounded 90% of KLM's fleet, with losses of KLM-Air France reported at €25 million a day (Luchtvaart Nieuws, 2020; Nederlandse Omroep Stichting, 2020).
With both sectors hit particularly hard, our case selection strategy holds sector-specific needs constant. We proceed with cursory cross-case analysis, in which we eliminate differences in economic weight as solely responsible for state aid allocation and composition. The remainder of our analysis relies on process-tracing, a methodology that infers cause by looking for within-case evidence of expected mechanisms (George & Bennett, 2005, p. 138). We conduct "focused empirical tests" (Beach & Pedersen, 2013, p. 34) of our expectations. The data for these tests were qualitative and textual, stemming from two main sources. Firstly, we consulted Dutch industry-specific news sites (Luchtvaart Nieuws for aviation and Misset Horeca for hospitality). Secondly, we used Nexis Uni database to consult top national newspapers, accounting for various political leanings: De Telegraaf (right-leaning); NRC (center-right); AD (center); Trouw (center-left); and Volkskrant (left-leaning). Our search terms specified: only printed articles in Dutch that include the following in the headline or first paragraph of the article: 'steun' , 'staatssteun' , 'economische maatregelen' OR 'compensatie' , as well as 'Luchtvaartmaatschappij' , 'luchtvaart' , 'KLM' , 'horeca' , 'hotel' , 'restaurant' OR 'café. ' In total, this amounted to 142 hits for articles relevant to the aviation sector ('luchtvaart'), and 80 hits for hospitality ('horeca'). All documents came from a 12-week timeframe: starting from 3 March 2020 (roughly two weeks prior to the first press conference held by Prime Minister Rutte) until 22 May 2020 (two days after the second emergency package was announced).
Documents were scrutinized for evidence of the expectations we set above. We analyzed them for descriptive information about the timing and content of key events. They allowed us to gauge which players were invited to the table, identify each sector's asks and subsequent reactions. At the same time, the media does not neutrally convey events exactly as they transpired. Data must be evaluated before it can be admitted as evidence on which to base causal inferences (Beach & Pedersen, 2013, p. 120). Hence, we also analyzed these documents as "vehicles of meaning [that] transcend a written record's purely textual content" (Skarpelis, 2020, p. 393). We paid attention to discourses that signal specific sentiments, and we situated specific justifications offered by policy-makers within a broader context.

Comparing aviation and hospitality in economic terms
Although aviation is a high-value industry, its generous aid cannot be explained in purely economic terms. A strictly technical application of economic prudence would channel capital towards sectors that perform well economically and away from "domestic economic laggards" (Degner & Leuffen, 2016, p. 1144Jansen, 2016). However, in stark contrast to the significant growth experienced by the hospitality sector prior to COVID-19, with its total turnover increasing by 6.2% or approximately €8.2 billion in 2018 (Bluiminck, 2019;Centraal Bureau voor de Statistiek, 2020a), KLM has only recently emerged from dire financial straits. In the early 2000s, KLM counted "loss after loss, " being in such poor financial health that it struggled to buy new aircrafts until it merged with Air France, who took over its remaining debt (Hermanides, 2019). This is a common refrain across Europe -Italy injected €3 billion into its flagship carrier Alitalia despite its financial track record showing losses of some €700 thousand a day (Berti, 2020;Flak, 2013).
Industry can, of course, have economic impact outside of profit-making. Air travel has been hailed as a catalyst of economic and social transformation by facilitating mobility of individuals and goods (O'Regan, 2011, p. 26). In the Netherlands, Amsterdam Airport Schiphol is a major European hub, and by its own measure, accounts for 2-5% of Dutch GDP and opens up the economy to international business (Schiphol, 2020). Responsibility for 56% of Schiphol's passengers -80% including its partners -confers legitimacy to KLM (van Ammelrooy, 2020). These considerations may have influenced its perceived importance. However, they do not tell the full story. Though the hospitality industry may not be as highly acclaimed in national imagination, it allegedly accounts for a comparable proportion -4% -of GDP (KHN, 2019a) and also opens up the Dutch economy by catering to tourists, whose total expenditure amounted to €87.5 billion in 2018 (6.4% higher than the previous year) ( KHN, 2019b).
By employment standards, neither sector is a clear winner either. One report, commissioned by an international airport lobby group and including direct and indirect employment (such as suppliers), put employment in the aviation sector at 370 thousand jobs. The cabinet's calculations use an estimation of 114 thousand direct and indirect jobs at Schiphol (van Bokkum & Kooiman, 2020). KLM itself, however, only employs 35.5 thousand people directly (KLM, 2018). The hospitality sector, for its part, employs approximately 459.5 thousand (De Telegraaf, 2020b) and recently experienced a 20% surge in employment (KHN, 2020). Hence, state aid to the aviation sector during COVID-19, which Gössling (2020) estimates at approximating US $100 billion globally by the end of May 2020, is undoubtedly wrapped up in economic concerns. However, as we have seen, other sectors like hospitality also enjoy strategic value. Fiscal support packages in other cases reflect this. German officials, for instance, saw equal value in both sectors, issuing approximately €5 billion in tax relief for restaurants and bars alongside their €9 billion state aid package for the Lufthansa aviation group (Deutsche Welle, 2020a). For a more complete explanation of the Dutch case, we turn to process-tracing.

Hoekstra's first domino?
On 24 April, when the initial announcement was made that KLM would receive backing to the tune of €2-4 billion, Minister Hoekstra emphasized the company's economic weight. His letter to parliament highlighted KLM as a "jobs engine" and the role played by Schiphol airport for Dutch business, stating: "the cross-continental network of destinations at Schiphol is of great importance to the Dutch economy… many companies choose the Netherlands as their headquarters because of its accessibility" (Hoekstra & van Nieuwenhuizen Wijbenga, 2020a). He described KLM as a lynchpin holding together the economy, referencing loaders and unloaders, distribution centers and main offices. Hoekstra invoked the metaphor of "the first domino … if [KLM] falls, more will fall" (Brandsma, 2020a). In the 26 June letter to parliament detailing the rescue package in greater depth, the same justifications were repeated: "the Cabinet considers it necessary to offer this support due to the importance of KLM for the intercontinental network of destinations that Schiphol offers" (Hoekstra & van Nieuwenhuizen-Wijbenga, 2020b; emphasis added).
In the above section, comparative insights cast aspersions on the viability of Hoekstra's justification. More in-depth qualitative analysis into the standpoints of political and corporate contemporaries, as well as the timing of the announcement, lend further urgency to the need to dig beyond purely technical, economic explanations. First, Hoekstra's claims were subject to contestation by experts. An economist at Vrije Universiteit Amsterdam was "not very impressed" with the jobs argument, stating that the employment figure is calculated too liberally (e.g. by including indirect jobs) and that it would matter predominantly if the Netherlands were experiencing a structural job shortage or if it concerned particularly high-value employment, neither of which was true (Lijesen as cited in van Bokkum & Kooiman, 2020). Moreover, evidence that companies select headquarters for airports as opposed to the tax climate, housing prices or education is sparse (Lijesen as cited in van Bokkum & Kooiman, 2020). As far as strategic value is concerned, Schiphol itself recognizes that noise and air pollution concerns limit the growth potential of aviation (Schiphol, 2020). Its future is even less certain in a post-COVID-19 world in which citizens increasingly work from home. On a company level, the cabinet acknowledged KLM's financial difficulties in their 26 June letter to parliament, admitting that they have "for a long time had concerns about Air France-KLM's performance" (Hoekstra & van Nieuwenhuizen-Wijbenga, 2020b, p. 6).
A final nail in the coffin of a purely economic explanation is the timing and chaotic nature of the announcement. Air France-KLM is a joint venture between Dutch and French companies, but each airline remains separate, with Dutch and French governments owning a 14% stake in their respective airline. Each opted to extend state aid only to their national airlines, generating concerns about a possible split in the partnership (Toman, 2020). Still, both governments were ostensibly in 'close touch' to coordinate their responses to COVID-19 (Hofs & Mebius, 2020). However, on 24 April, the French finance minister caught Hoekstra off-guard by calling at 7:10 PM, fifty minutes before he went public with the French state aid package (Duursma, 2020c). The Dutch scrambled to follow his lead, leaving an "improvised impression" of Hoekstra's own announcement (Duursma, 2020c). At that point, the distribution of state loans and loan guarantees had yet to be agreed with the banks, the European Commission's approval was still outstanding, and there was no clarity on whether the package would be two billion or indeed double that (Duursma & Witt Wijnen, 2020;Waterval, 2020). These details would take another two months to hash out. Rather than cool-headed calculation, the initial promise to KLM betrays a hasty decision-making process subject to international political pressures. This introduces the possibility that economic justifications were exaggerated to sell to the public under conditions of uncertainty and haste (Meyers & Rowan, 1977). A more complete understanding of the fiscal support package therefore requires casting a wider net.

Unified organized interests, divided public opinion
Organized interests and public sentiment are important for understanding KLM's state aid. KLM had first-rate access to policy-makers, having had a direct seat at the cabinet table from which they pursued "intensive discussions" with the cabinet for weeks (Waterval, 2020). Although, in theory, KLM employees are splintered across numerous different unions (there are five for ground staff alone), the company and its employees hashed out their disagreements in private and acted in unison publicly. For instance, Hoekstra wanted KLM to cut costs, explicitly mentioning wage moderation (Duursma & Witt Wijnen, 2020). CEO Elbers, citing his company's "tradition of sticking together during times of crisis, " entered into dialogue with the eight unions active at KLM and the works' council (De Telegraaf, 2020a). The chair of the pilots' union, 80% of whose members are from KLM, expressed his solidarity, explaining: "we are dealing with a new reality... if cuts are a part of that, we won't get in the way as long as it fits in the overall restructuring" (Duursma, 2020d). Other unions made similar gestures, for instance by signing a letter of intent to contribute to KLM's recovery on the eve of the parliamentary discussion on 6 May 2020 (Somsen, 2020).
Dutch public sentiment is somewhat ambivalent. On the one hand, the Dutch public views KLM, the oldest operating airline, fondly (Pontzen, 2019). KLM's advertising agent explains its public image: "the Dutch are proud of their export produce: flowers, DJ's. KLM belongs to that... More than other companies, KLM is steeped in emotion. Flying awakens more emotions than producing steel or lightbulbs" (Duursma, 2020b). Rutte's own acknowledgment in March of KLM's status as 'national icon' suggests the political risk to letting KLM suffer (Duursma, 2020b). However, affections are waning. When Dutch infrastructure minister described KLM as "our blue pride" (onze blauwe trots), an MP for the Greens retorted, "you're not speaking on my behalf... I have no need for terms like 'blue pride' when it comes to KLM" (Hofs, 2020). Mixed feelings stem predominantly from its environmental and noise pollution (Duursma, 2020b). Mistrust of pilots' high pay also contributes, with pilots likened to "glorified bus drivers" (veredelde buschauffeurs) in the media ("KLM alleen zonder", 2020). This accelerated when, in mid-April, a report revealed hundreds of KLM pilots living abroad to avoid taxes ("KLM alleen zonder", 2020). Despite these reservations, parliament backed the Cabinet's intention to support KLM (Stellinga, 2020b). On the whole, the public did too; according to one survey, 75% of voters supported state aid and thought VVD should not impose too stringent conditions (De Vente, 2020). However, on the latter point, political factions disagreed. A parliamentary debate on KLM state aid offers insight into how the package took shape.

Parliament demands conditions
The 6 May parliamentary debate featured opposition parties (the Greens, Party for the Animals) and coalition partner D66 demanding clear sustainability conditions (Duursma & Witt Wijnen, 2020). Furthermore, Labour and coalition partner ChristenUnie pushed for wage and bonus cuts (Markus, 2020). D66 demanded a reduction in night flights. In the end, KLM's support came attached to two sets of conditions, mirroring these concerns. The first is about using taxpayer money efficiently, i.e. KLM not paying out any dividends for the duration of the loan (maximum six years), reducing costs by 15% and coming up with a restructuring plan to improve its competitive position. The second relates to the "public interest" (Hoekstra & van Nieuwenhuizen-Wijbenga, 2020b, p. 5), and includes bringing down CO 2 emissions to 2005 levels and blending in renewable fuels. Additionally, KLM commits to helping reduce the number of night flights at Schiphol to 25 thousand (Hoekstra & van Nieuwenhuizen-Wijbenga, 2020b, p. 5). These conditions contradict expert opinion: a law professor at Nijmegen specializing in state aid explained that mandating costly investments hinders the ultimate goal of keeping the business afloat (Duursma, 2020a). Noteworthy is the reduction in night flights, since it is precisely night flights that allow KLM to serve in one of its key economic capacities: as a transport node between the US and Asia (De Telegraaf, 2020c). The conditions are therefore best understood with reference to policy-seeking behavior on behalf of elected leaders, given the prominence of climate action in par. 3.1 of the coalition agreement (VVD, CDA, D66 & ChristenUnie, 2017).
Any role of experts appears principally confined to a later stage of the decision-making process once the range (2-4 billion) was already determined. In the cabinet's 24 April letter to parliament, they explain that the amount was drawn up after a long period of "intensive contact" with KLM, Air France-KLM, and the French state about what would be necessary to keep it afloat (Hoekstra & Wijbenga van Nieuwenhuijzen 2020a, p. 1). The second letter to parliament on 26 June specifies that "the form by which this need can best be met" was determined "in concert with external, independent advisors" (Hoekstra & Wijbenga van Nieuwenhuijzen 2020b, p. 1). Given the prestige awarded to independent economic advisors, especially in countries like the Netherlands where these roles are historically institutionalized (Christensen, 2018), it seems unlikely that elaborate expert consultation would have gone publicly unannounced if present at earlier stages.

An unsuccessful ask
While a relatively well-organized sector, hospitality failed to secure a sector-specific aid package despite asking for it, due in part to a lack of public salience and an inability to make unified demands. Contrary to aviation, the hospitality sector is heterogeneous, including catering services and hotel industries. Nearly 90% of companies in hospitality employ fewer than 10 persons, with many working as freelancers or on flex-contracts, meaning that Dutch collective labour agreements for the sector come close to legal minimum standards (Grunell, 2012). Core organizations representing the sector include: employers' associations like KHN, MKB-Nederland and VNO-NCW; as well as employees' unions such as Federatie Nederlandse Vakbeweging (FNV), CNV Vakmensen and De Unie-Federation of managerial and staff unions. Organizational density is relatively high, with KHN's members collectively employing over half of hospitality workers. Not surprisingly, KHN organically mobilized and led negotiations for the sector. Unlike KLM who principally represented themselves as a company and on behalf of the aviation sector, hospitality therefore relied on their sector's association, with no core organizations taking a similar lead in negotiations.
On 15 March, hospitality was among the first sectors to be closed by the government when COVID-19 hit, 3 creating immense pressures for the industry early on. KHN was relatively well-organized from the start, lobbying alongside MKB-Nederland and VNO-NCW to request €5.1 billion in a combination of subsidies and loans (NRC Handelsblad, 2020). The sector was not granted this but was eligible for NOW and TOGS schemes as of 6 April (Posthumus, 2020c). Short-term policy asks contained more flexible, reduced working schemes, and postponing income tax payments. Longer-term asks included an emergency fund for the sector, enabling interest-free bridging loans, guarantee funds and special tax advantages (Posthumus, 2020a(Posthumus, , 2020b. However, nearly 2 months after these asks were outlined, the sector remained malcontent with what was available to them (van der Reijden, 2020).
As other sectors faced similar issues (i.e. bankruptcies), elected leaders signaled their support indirectly, speaking to a broader gamut of sectors. For example, alongside catering, State Secretary Keijzer emphasized entrepreneurs in tourism or retail as requiring additional assistance (Algemene Dagblad, 2020). Outcomes included delays in tax payments and interest rate reductions. The NOW scheme was principally designed as more direct compensation. However, satisfaction with this scheme was short-lived. KHN deemed the calculation of reimbursement skewed, given that it was based on turnover for January 2020 -two months before measures were imposed -as well as only on gross salary (not including insurance premiums or holiday bonus). The result was a reimbursement of, in reality, roughly 70% (and not 90%) of wage costs, with seasonal businesses who operated under capacity in January particularly dissatisfied (Bluiminck, 2020c). Additionally, evidence suggests that the hospitality sector lacked the public salience from which KLM benefited, and was therefore considered less urgent to support from a vote-seeking perspective. Using nearly identical search terms, our document search identified 44% more hits in national media outlets for KLM than it did for hospitality, suggesting relative silence in the public sphere.
All said, the heterogeneity of the sector was both a weakness and a strength. A strength, because its diversity made its demands applicable to broader organized interests from which elected leaders could seek aggregate political and electoral support. However, it was also a weakness, obstructing the viability of a sector-specific package.

KHN shifts its focus -with mixed success
By early April, hospitality had the largest number of bankruptcies of all sectors (Simon, 2020b). As its economic forecast grew grimmer, the sector strategically changed their approach. While their first (unsuccessful) lobbying approach centered on receiving state compensation, KHN replaced this with an approach centered on making money from customers. As a result, KHN used its access to policymakers to demand that businesses reopen at first opportunity, instead of prioritizing the reimbursement of lost revenues. Moreover, KHN consciously sought to increase its presence in the media (van den Anker & van Drimmelen, 2020). The result was coalition parties increasingly referencing the sector during parliamentary discussions on the second emergency package (Posthumus, 2020d). KHN simultaneously pursued internal discussions with the cabinet, outside of the public eye, in which they pushed for demands of a different nature (van den Anker & van Drimmelen, 2020). As a result, on 5 May, PM Rutte announced catering services could be reopened 1 June, despite other sectors remaining closed (i.e. sports clubs).
Despite this favorable momentum, obstacles still presented themselves. A lack of sectoral cohesion -between industry partners and between employers' associations and unions -posed issues for the sector down the line. KHN brought together industry partners like brewers, wholesalers, suppliers, and real estate parties, to sign a voluntary pact and collaborate internally. Economic Minister Wiebes and State Secretary Keijzer expressed support for this cooperation: "it cannot be the case that all damage ends up with the catering entrepreneur" (Bluiminck, 2020a). This generated some positive outputs. For instance, Heineken announced free cancellation on orders and events and 50% rent reduction on 130 properties the company owns (Hermanides, 2020;Simon, 2020c). However, tensions surfaced with online booking platforms whose strict cancellation conditions on reservations threatened the survival of countless hoteliers. KHN tried to negotiate a compromise between hotel partners, but eventually demanded regulatory bodies to intervene to assess fairer government support measures for smaller hotels (Bluminck, 2020b;Simon, 2020a).
Additionally, hospitality lacked cohesiveness among its employer association and unions. The lead-up to the second emergency package concerning new conditions of the NOW scheme triggered sectoral disagreement. The scheme originally contained dismissal fees -fines to beneficiary companies who fire staff for economic reasons. The revised scheme would eliminate these fees to provide more flexibility to employers in restructuring their businesses. Despite this aligning with KHN's demands, that would place more money in employers' pockets, creating significant tension with its unions, who called these decisions "a slap in the face to employees" (Pelgrim, 2020b). FNV board members stated, "we understand that companies are thinking about restructuring. But then it has to be done in a social way and in consultation with employee representatives'' (Pelgrim, 2020a).

Seeking broader political gains
At this juncture, the employers of the hospitality industry found strange bedfellows in policy-seeking elected leaders. Discussions concerning dismissal fees aligned with clauses set out in Rutte III's 2017 coalition agreement, which target the de-regulation of hiring practices for employers (i.e. para. 2.1, dismissal) (VVD, CDA, D66 & ChristenUnie, 2017, p. 25). This explains why political contestation in parliament mirrored the schism between unions and employers that was emerging within the sector, with center-right coalition parties advocating on behalf of employers, and (left) opposition parties siding with unions. The result was that "[the] new emergency packages expose political choices... division in times of crisis is a risk for the cabinet" (Brandsma, 2020b). Under these political conditions, Rutte's coalition could enjoy a two bird-one stone approach: making good on broader electoral promises by granting favors to small businesses, while solidifying specific COVID-19-related schemes. Nevertheless, likely as a result of cross-party pressure, some compromise ensued in the conditions of the support. Dismissal fees were removed on the condition that employers avoid firing employees in large numbers and equip employees with retraining (Bluiminck, 2020d).
Altogether, limited public salience and the relative ease with which hospitality could be grouped in with other sectors translated into a denial of targeted sector-specific support. Organized interests then changed tack, strategically lobbying to reopen businesses. Further lobbying success was complicated by heated disagreements between different industry partners, as well as between employers and unions. These conflicts mirrored political dynamics as the policy-seeking center-right coalition sought to reward small businesses as promised in their coalition agreement. The ultimate outcome -with dismissal fines remaining valid only under certain conditions -reflects cross-party compromise. Expert input directly concerning state aid for the hospitality sector was otherwise not publicly disclosed (i.e. in the media); merely subsequent commentary on longer-term economic and societal shifts stemming from this crisis. For instance, Dutch economists remained divided on use of wage subsidies and whether it aided the economy long term, and how to best protect rising numbers of unemployed citizens sitting outside the economy (Stellinga, 2020a).

DISCUSSION AND CONCLUSION
In this paper we explored the dynamics of state aid under crisis conditions. Our starting point was the different fiscal support received by two non-essential but heavily hit sectors in the Netherlands. While KLM, the foremost player in Dutch aviation, was promised state aid valued at €3.4 billion, the hospitality sector asked for a similar amount and did not receive it. We found this particularly puzzling given that both KLM and hospitality make sizable contributions to the Dutch economy: in fact, hospitality's growth and employment figures are stronger. Suspecting political dynamics, we developed a set of expectations using existing literature on public health crisis responses (Baekkeskov, 2016). These pointed to the ways that policy outcomes can reflect the interests of elected leaders, organized interests, and experts. We then used process-tracing and limited cross-case analysis to make inferences on the same.
While economic reasons are often used to justify fiscal support, for both sectors, our findings highlight the presence of political dynamics in the character and composition of these packages. In particular, we find evidence of vote-seeking behavior and organized interests during decision-making surrounding how much each sector received, while policy-seeking behavior mattered for how they received it, i.e. the nature and composition of their support.
Elected leaders exhibited vote-seeking behavior by responding to public sentiment. For KLM, public opinion about the company was increasingly ambivalent given environmental costs and high pilot salaries, yet the political clout of the century-old source of 'our blue pride' was substantial (Duursma, 2020b). Conversely, hospitality's initial lack of 'national icon' status helps explain why they were left empty handed after requesting a similar package as KLM. Compared to hospitality, KLM articulated their demands much more harmoniously, as unions rallied behind the company. This was not equally matched by hospitality due to the heterogeneity of the sector, leaving industry groups and organized interests at loggerheads and success evasive even when they modified their demands. Despite access and density being a common denominator between the two sectors (Andeweg & Irwin, 2015), it was the cohesiveness of aviation that appeared more important for their recognition by elected leaders (Braun, 2016). In hindsight, recent developments lend credence to the importance of unified organized interests. After the Cabinet rejected the terms of KLM's collective agreement on 30 October 2020, conflict with unions over the duration of wage cuts left the entire deal hanging in the balance until the pilots' union VNV acquiesced on 3 November (Duursma, 2020e).
In fact, though outside the scope of our paper, the Cabinet's bravado in October 2020 casts greater aspersions on purely technical arguments for state aid, since KLM's economic value remained constant while its promise of aid did not. Our research examining earlier stages offers insight into relevant political dynamics. In line with Degner and Leuffen (2016), we identify a prevalent role for policy-seeking behavior. However, we find that longer-term policy aims of the coalition primarily show up in later stages of decision-making, dictating the composition of state aid. Supporting each sector became contingent on select conditions that aligned with broader policy agendas. This was manifest in tensions along party lines, reflected in the 6 May parliamentary debate on KLM aid, as well as the 'win' for hospitality's employers (and a loss for their unions) around the elimination of dismissal fees that promises to deregulate hiring and firing practices. Both reflect climate and labor policy goals outlined in the 2017 coalition agreement. Additionally, organized interests for hospitality successfully secured non-financial support in the form of an early right to re-open.
Lastly, we found minimal (aviation) or no (hospitality) evidence of expert input in the documents we consulted. This is a peculiar finding, as previous research underscores the prominence of experts in health crisis decision-making (Baekkeskov, 2016). Though experts typically hold privileged discretion in allocating tax-payer's money (Christensen, 2018), they were rarely mentioned in public documents except to criticize the package and otherwise only appeared significant at the end of the decision-making process: for instance, to determine the exact amount to be allocated to KLM within a predetermined range, or the percentage that would be direct loans versus state guarantees. This leaves room to speculate the extent to which experts were involved directly in initial discussions, as expert input was not explicitly mentioned in national media reports-an otherwise increasingly reported feature in media (Albaek, 2011, p. 337).
All things considered, dynamics identified may have been amplified under crisis and high-pressure conditions. Additionally, heavy reliance on media sources gave us limited insight into exact conversations behind closed doors. The case is ongoing, and new dynamics came to light during research and revisions. Future research would benefit from interviews with policy-makers to clarify their decisions -explicitly cross-examining to what extent and at what point experts were consulted throughout this process, as well as interviewing interest groups about their strategies. Moreover, survey data about public opinion on public spending, specifically concerning aviation and hospitality sectors could be valuable. More in-depth, comparative cases might help to shed light on whether similar dynamics play out in different political contexts. For example, how did French support to their half of the Air France-KLM partnership differ from the Dutch?
Limitations notwithstanding, our findings underline the importance of examining the composition of state aid, and not just its presence or amount. Any map of the distribution of benefits and costs during crises must include the conditions, eligibility criteria and form of these benefits. Costly conditions attached to KLM's loans are noteworthy, and the hospitality sector found out quickly that technical details of reimbursement under NOW, and later revisions that threatened job security, hampered their ability to reap these benefits. Moreover, COVID-19 shows that modifications to business regulations, such as rules for reopening, crowd management and travel bans, can constitute support.
More broadly, the Dutch decision to support KLM so generously and the hospitality sector haphazardly points to the potential gains from viewing emergency state aid as industrial policy. As mentioned, we demonstrated that purely rational economic justifications on efficient capital allocation did not fully explain the patterns of aid allocation. Falck et al. (2011) describe how the 2007-8 financial crisis accelerated government support for national champions, including aviation. They show how these policies "seem much more readily understandable in political-economy terms than in economic-efficiency terms" (Falck et al., 2011, p. 8). Our findings suggest that COVID-19, too, has amplified commitment to national champions, whether as a result of powerful organizing strategies or their historic importance. This happening in a traditionally liberal setting is noteworthy (van Zanden, 1999, p. 177). However, the conditions given to KLM also show tentative steps taken to use industrial policy to facilitate green growth, as anticipated by Rodrik (2014) -which otherwise may have been absent without the policy-seeking behavior of parties in a fairly contested and chaotic political landscape.
COVID-19 thus represents a critical juncture. By this is meant, following Collier and Collier (1991, p. 27), a transition in which certain directions of change are established and others are foreclosed, producing distinct historical legacies. Not only is COVID-19 likely to permanently change consumer behavior, but if the '1.5-meter economy' stays with us for long, it will also mandate business practice adjustments that many small companies may not be able to meet (Stellinga, 2020a). Dutch hospitality is dominated by small and medium-sized businesses, and remains the hardest hit with approximately 60% expecting bankruptcy of their company if the crisis lasts longer than twelve months, and businesses over 50 persons projecting lower risks (Centraal Bureau voor de Statistiek, 2020b). Moreover, Business Survey Netherlands claims roughly half of the private sector may not survive COVID-19 (Centraal Bureau voor de Statistiek, 2020b). In this context, targeted support for larger multinational companies, especially those under financial duress, during crisis has the potential to substantially alter the structure of the entrepreneurial landscape, nationally and internationally. Critical junctures are rarely automatic responses to external shocks; instead, there is room for political responses to these events (Collier & Collier, 1991, p. 31). And as much as there is scope for intervention, there is scope for criticism, by scholars, political actors and, in this case, fellow corporations. For instance, Germany's substantial aid to aviation giant Lufthansa triggered competitors such as RyanAir to appeal to the European Court of Justice, claiming Lufthansa merely needing state aid "as a war chest to fight off competition, [not] to survive" (Deutsche Welle, 2020b).
Companies like KLM can embrace the window of opportunity to drastically adjust their business model in accordance with political priorities (e.g. meet emission targets) (Phillips, 2020). Some have also thrown their own lifeline to smaller companies: see, for instance, Amazon's Small Business Relief Fund for companies of under 50 employees (Amazon, 2020). Nonetheless, certain Dutch social media users have wondered why KLM does not receive applause instead of €4 billion, on the premise that "it worked for health care workers. " Behind the sarcasm is a genuine concern that essential sectors are inadequately compensated and non-essential sectors too powerful. One commentator noted that the hospitality sector "makes it seem like it's an essential sector, which is of course nonsense" (van den Anker & van Drimmelen, 2020). Regardless of public sentiment, policy actors will continue to advocate for their own interests. If state aid -emergency fiscal stimulus included -is going to be wielded as a political instrument even in the heartlands of liberal Europe, it may be worth considering more conscious assessment tools -in other words, those that weigh fairness and sustainability just as heavily as national pride and lobbying strength.