The structure of business preferences and Eurozone crisis policies

Abstract

What explains business views regarding policy preferences in the Eurozone crisis? Although recent literature examines the impact of the crisis on citizen views, few studies examine business preferences towards adjustment policies. We present unique data from a new representative survey of 500 high-level firm representatives from Spain to test theories about such preferences, in particular views about the euro, fiscal austerity, and wage devaluation, as well as plausible mechanisms for such preferences. We test three broad families of theories to explain such preferences, focusing on the role of structural firm characteristics, economic hardship, and political leanings of firm managers. We find that first, there is a strong conservative position regarding all of these policies. Second, we find that contra conventional approaches to explaining preferences, for the domestic policies (but not for euro views), the political leanings of firms matter much more than baseline structural characteristics. Third, we find that surprisingly economic hardship does not cause firms to demand more left-wing policies, as it might for voters; in fact, firms that have suffered are likely to be more skeptical of such measures. These findings indicate the need to better measure political orientations of firm respondents and suggest that this is a larger division among firms than previously recognized.

Introduction

What are business preferences over different public policies regarding the current Eurozone economic crisis? What factors best explain variation in such preferences? To what extent can structural characteristics of firms account for such preferences, as opposed to the economic suffering of firms or even the political leanings of business leadership? The European recession, the worst in the continent since the Great Depression, has already produced a wealth of research seeking to explain either cross-national variation in policy responses or variation in voter preferences about parties and policies.¹ In this paper we focus on a greatly underemphasized question by examining the policy preferences of key economic actors—firm managers and owners—regarding the crisis. Understanding such business preferences provides an important theoretical and empirical contribution, as business activity and confidence are determinants of future European growth.² Further, measuring such preferences in a time of crisis allows one to test the portability of political economic models that assume “normal” times, to more difficult economic circumstances. Despite the longstanding belief that firms' preferences matter for policy outcomes, we lack high quality data on such preferences and basic tests of their correlates. We address this lacuna with an original survey of a representative sample of firm managers and owners from a country that was especially hard hit by the ongoing crisis: Spain. This dataset allows us to test competing theories of firm preferences for an array of economic crisis response policies, in particular contrasting the views and correlates of fiscal and labor policies versus currency-based adjustments. The paper provides one of the few large-scale systematic analyses of such policy preferences of firms during the height of the crisis, from the perspective of an economically important country that suffered the effects of the crisis for nearly one decade.

We measure preferences over several key policies that have been debated as responses to the crisis, and test different sets of theories aimed at explaining variation in those preferences. We examine support for commitment to the euro as a currency, fiscal austerity, and internal devaluation (nominal price and wage reductions). In terms of competing explanations, we test three broad categories of variables: structural factors of firms, crisis experience (in particular specific concerns about the crisis and the degree of economic suffering), and finally, the political leanings of firm managers. Where possible, we also test relevant mechanisms that relate these variables to policy preferences.

Some of our results contrast with conventional wisdom about the structural determinants of business' policy preferences. First, based on the descriptive statistics, we note that there is broad support among firms for the euro and for Spain staying in the Eurozone, that firm managers generally believe recent austerity policies are necessary to end the crisis, and that they prefer deflationary policies such as wage cuts. This indicates an overall preference for a strong “conservative” policy package in response to the ongoing recession. Second, in contrast to purely structural models of firm preferences that disregard the political preferences of firm managers, we find that the political orientation of the firm representative matters more for contentious domestic policy preferences than other firm-level covariates, but less so regarding the euro. In fact, many baseline structural features of firms such as size or sector do not correlate in anticipated directions with policy preferences.

In terms of the three policies we study, regarding currency preferences, we find that the key structural variable of export orientation matters in more nuanced ways than the most basic models of international political economy predict, as export-oriented firms are more opposed to a potential Eurozone exit. This opposition among export-oriented firms is more driven by prospective concerns about a euro exit as opposed to retrospective evaluations of the euro. The evidence suggests that these internationalized firms support the euro for reasons beyond devaluation concerns, consistent with other, subtler theoretical models of currency preferences. By contrast, crisis incidence and political orientation matter little for euro views.

Regarding austerity and devaluation, we find that the political orientation of firm managers is substantively more important than most relevant structural aspects of the firm. This political orientation also matters for mechanisms for why firms support austerity and devaluation. Further, firms that have suffered substantially during the crisis do not have significantly different preferences from other firms regarding support for the austerity, but do support internal devaluation. These findings pose a challenge to conventional theories of firm preferences that draw only on structural factors, and highlight greater relevance of political leanings of the firm decision-makers themselves.

The paper proceeds as follows. Section one gives an overview of the relevant literature on the crisis in Spain and the motivation for surveying firms. Section two describes the outcomes of interest and theoretical perspectives. Section three describes the dataset. Section four provides descriptive statistics and results. Section five concludes.

Section 1. Overview and relevant literature

We first provide an overview of the economic crisis in Europe and then turn to summaries of extant literatures that motivate our design and hypotheses.

Background on the Eurozone crisis and the Spanish recession

Spain is a particularly relevant case to examine regarding business preferences related to the crisis, because of the depth and duration of the economic recession, as well its economic importance to the viability of the Eurozone.³ The immediate causes and events leading to an asset bubble, Eurozone crisis, and major recession in Spain in late 2008, as well as the grim economic statistics over the last decade, have been well documented. At the peak of a real estate bubble heavily dependent on access to foreign finance, the freezing of international markets in 2007 and 2008 took the form of a “sudden stop” for Spain, and quickly led to a rapid fall of asset and housing prices and corresponding dramatic economic contraction. Between 2007 and 2013, real Spanish GDP per capita fell by more than 10 percent, and the unemployment rate increased from 8 to 27 percent. Long-term unemployment increased in particular: By 2013, 62 percent of the unemployed population had been looking for a job for one year or more. Positive growth rates only resumed in the last quarter of 2013, and the GDP expanded annually by more than 3 percent in 2015 and 2016. Labor market indicators have improved significantly, but unemployment remains at 17 percent in 2017. Despite these positive economic indicators, questions about the sustainability and strength of the future recovery remain, particularly given the vulnerability of the situation of public finances to an eventual deterioration in international financing conditions.

The overall policy response in Spain was a mix of fiscal austerity and attempts to deregulate the labor market to foster internal devaluation, with very littfle discussion of leaving the euro.⁴ Although the first reaction to the crisis was to implement a series of fiscal expansionary measures, most notably through small-scale infrastructure projects, panic in peripheral sovereign debt markets due to foreign-investor perceptions that public debt was becoming unsustainable (the Spanish public deficit rose to 11 percent of GDP in 2009) forced European countries to adopt fiscal consolidation plans. In May 2010, the Spanish government passed a first set of austerity measures, including a cut of public employee salaries, a reduction in infrastructure spending, a pension freeze, and an increase in the value added tax (VAT). Large public protests partially in response to these policies erupted in May 2011 and October 2011 in Madrid and other major cities, with sporadic public protests occurring since. The Socialist government was replaced by the conservative Partido Popular (PP) in the November 2011 elections, which won with 45 percent of the vote and an absolute majority in Parliament. The government passed further austerity measures, including an income tax increase, an additional VAT increase, major spending reductions across many government departments, and a vigorous attempt to curb spending in subnational levels of government.⁵

Labor market reforms since the onset of the crisis have been politically contentious as well, as both governments sought to “flexibilize” the labor market. The main labor unions called for general strikes in response to both the 2010 (socialist-led) and 2012 (conservative-led) labor market reforms, both aimed at promoting the internal flexibility of firms and reduce worker dismissal costs. Overall, the reforms allowed firms to more easily opt out of collective agreements and act more unilaterally regarding changes in working conditions for a wider set of changes in circumstances, and they have been generally credited with contributing to wage moderation (OECD 2013) and the improvement of employment indicators in the current recovery.

Throughout the crisis, business organizations and in particular the peak association Confederación Española de Organizaciones Empresariales (CEOE) have expressed support for greater labor market flexibility, reforms aimed at reducing bureaucratic intervention, privatization of some public services, and fiscal consolidation efforts focused on expenditure reduction rather than on revenue generation, expressing strong opposition to the different rounds of tax increases. In the first phase of the crisis, they were also concerned about business access to sufficient finance, and in the latter years, about the allegedly insufficient investment in infrastructure caused by years of austerity policies.

During the crisis, successive governments have called unions and business associations to try to negotiate a joint proposal of labor market reform, but both the rejection of workers' representatives of liberalizing labor regulations and the business position on the reduction of firing costs and reforms to facilitate firm–level wage restraint made these agreements impossible. When the government broke this gridlock by unilaterally passing the 2012 labor market reform, the CEOE welcomed it as an important move in the right direction.⁶ Overall, despite attempts on the part of the government to further deregulate the labor market, many labor-market structural rigidities persist, such as high severance costs and a segmented labor market.

Firm preferences and the crisis

As the previous summary illustrates, Spain is a particularly useful test case of firm preferences regarding the ongoing European crisis, because of the depth of the recession and political controversy and division surrounding policies that have been implemented. Our study is motivated by the fact that firm preferences are relevant for understanding the policy consequences (and resolution) of the ongoing European economic crisis, but there have been few tests of theories of such preferences. While some literature addresses firm preferences for different policies in normal economic times,⁷ few studies examine such policy preferences in crisis times, in particular during the most recent Eurozone crisis.⁸ This omission is relevant in debtor countries, where political conflicts over the adjustment process persist.⁹

Systematic firm-level data on these outcomes regarding the most recent crisis is important for theoretical and empirical reasons. First, since business preferences matter for the sustainability of the Eurozone, then understanding variation in and the correlates of such preferences is of natural interest (see, for example, the debate over sovereign risk spreads, concerns of foreign investors, and the actions of firms).¹⁰ Second, understanding firms' preferences can assist in constructing broader political economy models of crisis politics, as businesses can influence policies through direct influence such as lobbying or via conveying information or their preferences to the government.¹¹ More generally, they constitute policy-influencing coalitions with other actors, including organized labor, sector-level associations, and citizens. Finally, beyond the immediate context of the current economic crisis, better understanding the correlates of firm preferences can test whether standard political economy models of such views apply to times of crisis as opposed to “normal times.” Previous work has identified the necessity of understanding large scale shifts in sector preferences during economic crises, as new economic policy packages become politically viable.¹² This study adds to the tradition of comparative political economy scholarship to better understand the preferences of capitalists and firm managers and owners (in contrast with the large literature on voters), but by providing more systematic measurement of their attitudes and preferences from a larger, representative sample of them.

Section 2. Outcomes and hypotheses

Overview of theoretical approaches

In theorizing about business policy preferences during the Eurozone crisis, we test three sets or families of competing hypotheses: structural characteristics of the firm, crisis-specific experiences, and the political orientations denoted by firm managers' responses. We draw on relevant insights from the literature on the impact of the economic crisis on voter preferences, and that of the traditional literature on firm preferences on relevant economic policies during non-crisis economic periods. We suggest that firm preferences over policies might be different in a crisis versus a non-crisis context, and argue that the former context should increase dissatisfaction with status quo institutions, such as existing currency arrangements or institutions. We test theories for multiple relevant public policies because governments have pursued a mix of policies in response, and one should test which firm-level characteristics matter more for different policy views.

We focus on the preferences over a Eurozone exit, fiscal policy (austerity or fiscal adjustment packages), and views on wage devaluation policies. These three policy areas are obviously not exhaustive, but they constitute the most widely discussed issues facing firms and the population of the crisis-hit countries, and there is much academic and public division over the impact of such policies on the European economies. For each policy preference, our theorizing first highlights the most relevant structural variable that should matter for that policy, though we acknowledge that for each policy, multiple structural variables could be relevant, and that the structural variable of interest should differ depending on the policy in question. Where possible, we suggest multiple channels by which characteristics of firms would affect policy preferences. In the empirical section, we describe a set of follow-up questions regarding each dependent variable to test competing explanations for why structural, political orientation, or crisis-specific factors would be correlated (or not) with variation in policy preferences.

We now preview the distinct theoretical perspectives used to explain variation in policy preferences. We operationalize each perspective with a number of different variables. The first perspective emphasizes traditional political economy variables that are the structural characteristics of the firm, including the economic sector in which the firm operates, its size, export orientation, whether there is a strong union presence, and other factors specific to the product and workforce of the firm. Again, the relevance of these factors should differ in importance depending on the policy area in question.

A second perspective considers firms' crisis-specific experiences as a correlate of policy preferences. This category of variables takes as a starting point that a major economic recession should affect preferences. We separate this category into two variables. One is the basic firm experience in the crisis captured by the loss in firm profitability since the onset of the crisis, as firm-level suffering should affect perspectives regarding what policies are most effective to end the crisis. The second set of variables consists of firm beliefs about problems that they currently face due to the crisis. While a large number of beliefs could matter, we focus on those most relevant during the ongoing crisis. These are the beliefs about access to credit, insufficient consumer demand, labor costs (wages and related social security costs), and sensitivity to taxes. We categorize and distinguish between the “macroeconomic” factors of access to credit and demand and the “microeconomic” concerns of wages, non-wage related labor costs, and taxes. We suggest that these specific firm-level concerns about micro versus macro issues that firms face should be linked to policy preferences to address them.

A final, less developed theoretical perspective, views firm preferences as reflecting beliefs and interests of firm representatives as ideologically-driven political actors, and thus might generate different predictions about policy preferences.¹³ This is related to previous research that characterizes firm representatives' preferences as reflecting other firm-level characteristics, aside from structural features of the firm.¹⁴ For example, as we detail below, political ideology is a well-known factor that differentiates citizen views; we test whether the political orientations revealed by manager's responses matters for some firm preferences over others. Crudely, firm representatives that have more left versus right-wing political orientations politically could prefer corresponding policies for their firms.

Context for predictions of structural factors: internal versus external adjustment

We build on recent work that more clearly specifies firm and sector-level policy response preferences in the event of macroeconomic or financial crises. Walter offers a useful clarifying perspective that we draw on and test with data on the Eurozone crisis (2013).¹⁵ In her framework, government policies in response to crises can be via internal or external adjustment processes. The former includes fiscal policy (typically reducing government spending) or structural reforms (including labor market regulations that reduce nominal wages); the latter involves adjustment of the exchange rate, typically a depreciation or devaluation of the currency.

How do adjustment strategies from Walter's framework apply in the context of the Eurozone crisis? External adjustment in the form of currency devaluation—Eurozone exit—is of course a potentially costly policy. In the case of the Eurozone, simple external adjustment via currency depreciation cannot be easily pursued by national-level governments in the Eurozone as they do not control monetary policy; external adjustment would entail voluntarily leaving the Eurozone. Thus, it is unsurprising that there is more pursuit of but also division over internal adjustment measures (fiscal and labor-market policies). We now turn to the predictions of each type of explanation for each policy preference of interest (Eurozone exit, fiscal austerity, and wage devaluation).

Preferences for staying in the Eurozone

Given the salience in the political and economic debate over the viability of the Eurozone, we consider preferences over this external adjustment strategy (euro exit), even though national-level governments would have to leave the Eurozone to pursue this strategy. What factors should be most relevant in explaining firm support for a possible Eurozone exit? We first turn to arguments that focus on standard structural characteristics from the literature as to why firms might support or oppose staying in the euro, the most relevant of which is the degree of export orientation of firms. However, the ample theoretical literature on firm currency preferences actually gives mixed predictions regarding the role of export orientation; in this sub-section, we clarify the logics of the contrasting predictions of euro views. We classify expectations as following a retrospective versus a prospective logic. We distinguish between these perspectives regarding euro preferences because the Eurozone itself pre-dates the crisis, and firm preferences for the euro can be linked to retrospective evaluation of its effects, prospective concerns about an exit, or both.

The conventional open-economy logic examining preferences of exporters theorizes that because currency depreciation makes exports relatively cheaper in the foreign countries' markets, those companies should prefer net currency depreciation for exporters, a finding confirmed in previous empirical analysis of exporter firm currency preferences.¹⁶ Thus exporters should ceteris paribus prefer a depreciated euro and a cheaper currency. This basic prediction has been built upon in recent frameworks that examine firm preferences during balance of payments crises, and argues that firms can prefer external adjustment strategies as a government response (currency depreciation) versus internal adjustment (such as domestic price deflation or changes in domestic interest rate policy). During such crises, more export-oriented firms are theorized to prefer external adjustment defined as currency depreciation.¹⁷

However, one should cautiously apply this framework of preferences for external versus internal adjustment strategies to the ongoing crisis, as Eurozone economies cannot easily pursue an external strategy in the form of simple currency depreciation. This option entails leaving the Eurozone, a monetary breakup which has many consequences beyond a change in the nominal exchange rate. As a result, studies that focused on previous balance-of-payments crises are not analogous to the current crisis, as governments in the Eurozone cannot easily pursue standard depreciation strategies.

One approach to distinguishing the conflicting preferences of exporters is by distinguishing between retrospective and prospective views towards the common currency. The retrospective logic argues that since euro membership for firms caused increases in prices and unit labor costs above competing countries in the monetary area, the resulting real exchange appreciation damaged the export sector's competitiveness.¹⁸ Therefore, export-oriented firms should attribute their reduced price-based competitiveness to the currency zone and prefer to leave it, and should be more likely to blame the euro for higher internal costs and inability to depreciate. This supposition is supported by other comparative evidence of firms that experience real exchange appreciations under fixed exchange rate regimes as being more critical of such monetary arrangements.¹⁹ Other evidence indicates that export-oriented firms should be more supportive of currency regimes that allow for depreciation.²⁰ This standard perspective inherently includes a prospective logic, as it assumes that export-oriented firms anticipate the outcome of a depreciating currency of leaving the currency zone, and thus support such depreciation.

An alternative set of political economy arguments focuses more on the prospective concerns that relate to negative implications for firms from not just currency depreciation, but for leaving a currency regime. Such concerns relate to specific scenarios or problems that a euro exit would entail, distinct from the lower value of the currency of final goods. First, if firms rely on intermediate inputs from the euro-area and changes in the nominal exchange rate are not automatically translated into final prices (i.e., low levels of “pass-through”), then the gains from a nominal devaluation from an euro exit would be limited.²¹ The salutary effects of a nominal devaluation for exporters are also more likely to be offset if firms hold debt denominated in the currency that would be relatively appreciated, as the real value of the debt burden would increase.²² Second, leaving a currency zone has consequences for the volatility of the exchange rate as well as its level; such an exit is clearly not analogous to devaluing the national currency in a floating regime to regain competitiveness, as a currency breakup entails much uncertainty about the magnitude of the disruption.²³ This makes the external adjustment option in the context of the euro periphery significantly more costly. Export-oriented firms could have prospective concerns that reduce support for a euro exit, and in the empirical tests below, we arbitrate among these multiple concerns.

We now discuss hypotheses regarding the impact of firm-level economic suffering on views towards the euro. Firms that have suffered during the crisis may have distinct views towards the euro for a few reasons. Following a retrospective logic, they could be more inclined to blame the euro for the economic situation. Previous research in the tradition of citizen retrospective policy evaluation demonstrates that individuals are more likely to hold status quo institutions at fault during difficult economic times.²⁴ Blame attribution serves as a key component in models of formation of policy judgment and assessment of the status quo institutions.²⁵ Applying this logic to firm managers, we suggest that all else equal, the experience of greater economic losses should make firms more skeptical of status quo policy arrangements, and thus they should be more willing to abandon the euro. Firms that have suffered more during the crisis could view the current currency regime as a cause of deterioration of the economic situation for several reasons. They could perceive the currency zone as constraining the domestic government from adjusting monetary (or fiscal) policy to domestic conditions, because the Eurozone implies real exchange rate overvaluation, or because membership in the Eurozone implies the imposition of certain fiscal policies that might reduce growth in the domestic economy. More generally, firms may assume that Spain's economic performance has worsened under the Eurozone, and link that negative assessment to opposition to the monetary regime.²⁶ Thus we hypothesize that firms that have been more economically hurt by the crisis should be more favorably disposed to a Eurozone exit.

Finally, we consider the role of the political leanings of managers. Previous studies argue that left-wing individuals are more sensitive to the loss of monetary policy autonomy under a fixed exchange regime with a low inflation objective, and hence they should be less supportive of the common currency. Since fixed currency regimes are more likely to constrain inflation, much of the political economy literature theorizes that more right-wing individuals who are more inflation-averse should prefer maintenance of the currency regime.²⁷ Therefore, if political orientation matters for firm-level preferences, more left-leaning firm representatives should all else equal be more sympathetic to a euro exit.²⁸

Preferences for fiscal policy and austerity

We turn now to hypotheses regarding firm-level views on fiscal austerity, a politically contested issue in the Eurozone periphery. Again, as there is surprisingly little theorizing of firm perspectives on government fiscal responses to an economic crisis, we draw on some of the literature that debates the macroeconomic consequences of austerity.²⁹ While it is unclear what structural factors matter most in explaining firm support for austerity, one straightforward baseline hypothesis is that to the extent that such reductions in government spending dampen aggregate demand, smaller firms are more opposed because they would be more sensitive to reductions in such demand, whereas large firms have more flexibility in weathering short-term domestic economic shocks.

Regarding the set of crisis-experience variables, we hypothesize that firms that have suffered the most during the crisis should be more likely to oppose austerity. This prediction comes from previous research on citizen preferences that identifies a large impact of economic suffering on increased support for government spending;³⁰ we suggest that firms under economic strain should prefer public spending on programs to either directly offset their losses or, indirectly, to stimulate domestic demand.³¹ Related, firms that claim larger concerns related to “macroeconomic” issues (characterized as concerns about access to credit and demand for their products) should be more opposed to fiscal austerity policies, as the main economic critique of such austerity is that the fiscal cuts it entails reduce such demand.³²

In terms of political orientation, most straightforwardly, if the conservative orientation of firm managers matter, we hypothesize that more conservative firm managers should be more supportive of austerity, consistent with much of the European-wide individual-level data linking conservative political orientation with the view (correct or not) that state largesse has been responsible for the poor economic outcomes.³³

Preferences for internal devaluation / wage decreases

We now discuss business preferences for internal devaluation policies and outline several hypotheses. While there are many characteristics of firms that may lead firm managers to prefer lower wages, one relevant variable should be whether the firm has a strong union presence. In the institutional context of Spain's labor collective bargaining laws, we expect that firms with greater union presence should be more supportive of wage devaluation. There are two channels through which this variable might make firm managers more likely to support internal devaluation. One is that unions might play a role in facilitating wage declines and moderating potential industrial conflict.³⁴ Alternatively, if union presence is understood as a proxy for workers' structural power within the firm, in the more conflictual bargaining context of Spanish industrial relations,³⁵ a stronger union presence should increase business support for further reducing wages. Higher worker representation, either because it makes nominal wage cuts needed and/or feasible, should be associated with more support of internal devaluation strategies among firm managers.³⁶

In terms of crisis-experience variables, the degree of economic suffering should also correlate with views on internal devaluation. We hypothesize that firms that have suffered more during the crisis would also favor nominal wage reductions, as this would be a more direct way to reduce costs. Regarding the crisis-experience variable of specific concerns facing the firm, we hypothesize that firms that have more “micro” concerns (such as higher wages) relative to “macro” concerns should support deflationary policies.

Finally, as with austerity preferences, the conservative orientation of firm managers could matter as well. We hypothesize that more conservative firm managers are supportive of internal devaluation and lower wages. Previous research points to conservative skepticism of policies that protect wages with evidence from Europe³⁷ and the United States.³⁸ If the political orientation of firm managers matters for this policy preference, then such managers would prefer to reduce worker incomes as a crisis adjustment strategy. We summarize the main predictions from these variables in table 1, for each of the policies.

Table 1: Predictions of policy views from key variables

Section 3. Design and data

To test these theories we fielded a telephone survey to a nationally representative sample of 500 firms operating in Spain with at least ten employees, drawn from the official directory of firms (Directorio Central de Empresas (DIRCE)) kept by the Spanish National Statistics Institute (INE).³⁹ The respondents in the sample were firm owners or high level managers (CEO or CFO equivalents), and the sample was stratified by sector of operation (industrial manufacturing, construction, wholesale/retail, services), region and firm size (fewer and greater than fifty employees) to guarantee a representative sample composition across these variables.

We obtained data on the dependent variables of policy preferences discussed above, a number of firm-level variables including key structural characteristics, firm-level views on the concerns listed above about the crisis, amount of profits lost since crisis onset, and other questions designed to disentangle different reasons for policy preferences. These are all detailed in the next sub-section. We measure the firms' preferences by assessing those of the firm owner or manager, with respect to what the owner (or firm representatives) perceives is optimal for the firm. To maximize our ability to measure precisely the firm's point of view, many questions began with the opening clause, “from the perspective of what is best for your firm's profitability…” Where possible, all questions reminded the respondent to answer the question from the perspective of the firm.⁴⁰

Variable description and coding

Firm characteristics

We measured the following main characteristics of the firm: the broad industrial sector (industrial manufacturing, construction, wholesale/retail, services), the size of the firm in terms of natural log of workforce size, the size of the firm in terms of annual profits, the percentage of revenues derived from exports, whether and how much the firm exports to the EU area and outside the EU area, degree of perceived union presence in the firm, and percent of the firm's workforce that is on a temporary contract. The firm's broad sector is coded as a binary variable (1 indicating classification in one of the four main sectors and 0 otherwise). All independent variables, unless otherwise indicated, are rescaled 0–1 to ease interpretation. The broad industrial sectors are binary variables indicating if the firm is in manufacturing, construction, wholesale/retail, or services. The union presence variable is a four-point scale rescaled 0–1 that asks the firm representative how important unions are to decisions of the firms (response options: “very important, somewhat, a little important, not important at all”); higher values indicate greater importance.

We also coded an open-ended sector question to measure the tradability of the firm's products and services (regardless of whether the firm actually exports), and used this textual data to also measure the skill intensity of the firm. We coded the latter using a simple coding rule; if operation in the firm's sector requires a university degree or specific vocational skill, it was coded as high skill. Both tradability and skilled are coded binary. We also used the open-ended sector data to match the firm to broad ISIC categories at the single-digit level (21 sectors), and use this to generate a variable of sector level unemployment for each firm.

Measurement of crisis experience and concerns about the crisis

Regarding a key crisis experience variable, the amount of profits lost, we ask whether the firm has lost or gained profits since the onset of crisis, and depending on the response, the amount of profits gained or lost as a percentage. Our main indicator of economic suffering is the percentage loss in profits that a firm reports, rescaled 0–1, with higher values indicating greater loss in profits.

We also measure the crisis-specific concerns that firms have, with the problems being: access to credit, insufficient consumer demand, high wages of workers, labor costs (including social security and related labor costs), and higher non-labor taxes. Unless otherwise noted, for all outcomes, we use a 0–10 scale, as it is the standard and most understood format in Spanish business barometer contexts. We first measure on this scale how much each of the five factors current pose a problem to the firm, with 0 indicating the factor is not at all a problem and 10 indicating it is the principal problem.⁴¹ We also construct a variable to indicate the relative concern of the “macro” factors of credit and demand versus the “micro” concerns of wages, non-wage labor costs, and taxes. We do this by averaging the scores within the micro and macro category of concerns, taking the difference between the two, and again rescaling again 0–1, with higher values indicating a greater weight of macro concerns relative to micro ones.

Measurement of preferences—policy views

We now turn to measurement of the policy preferences and attitudes regarding the euro, austerity, and internal devaluation/wage adjustment. We first measure the respondent's stated projected consequences if Spain were to cease using the euro. The response options were: the company would have to close, the company's situation would worsen, the company's situation would not change much, and the company's situation would improve. The “worsen” and “improve” options had a further branch with options of “worsen/improve a little” or “worsen/improve a lot.” We code both the trichotomous response options of negative, zero, and positive consequences to the firm. We also code all the response options on a six-point scale as a 0–1 scale, with higher values indicating the firm would benefit from a euro exit.⁴²

We measure support for austerity as a binary question indicating if the respondent thought that the fiscal adjustment measures (commonly understood as the budget cuts that the government has taken) have made the crisis worse, or are necessary to end the crisis (1 coded as austerity making the crisis worse, 0 as the measures being necessary to end the crisis). We measure the final policy preference for internal devaluation on a 0–10 scale, with 0 indicating complete disagreement with the view that internal devaluation strategies would help the firm and 10 indicating total agreement.⁴³

Measurement of political leanings

As we could not gather data on the demographic characteristics of the respondent to ensure his/her anonymity (such as personal ideology, vote intention or choice, and education), we use a proxy to measure the political leanings of the respondent that is as distinct as possible from any measure that might be a function of specific policy preferences. To do so we use the same 0–10 scale and ask questions about how much the firm manager blames the policies adopted by the previous Spanish Socialist government of Zapatero versus the policies adopted by the current Spanish conservative government of Rajoy.⁴⁴ On this scale, 0 meant the factor had “not at all contributed to the crisis” and 10 was labeled as “has been primarily responsible for the crisis.” We use the difference between the blame of Rajoy less Zapatero as a proxy for how “left-wing” the respondent is, as higher values of this measure indicate more blame of Rajoy versus Zapatero. This variable is rescaled 0–1 such that higher values represent more blame of Rajoy relative to Zapatero. Given the constraints on asking directly vote intention or partisan orientation, this is a reasonable proxy for personal political orientation of the firm representative that is distinct from firm-level structural characteristics.

Descriptive statistics

Regarding broad industrial category composition, 25 percent of the firms in our sample are in industry (generally including manufacturing goods); 11 percent in construction, 16 percent in wholesale/retail, and 48 percent in services. 82 percent of the firms are small or medium-sized enterprises (SMEs), defined to be having fewer than fifty employees. The median firm size is twenty (with a mean of eighty); among large firms, the median is ninety-four (mean of 343). 31 percent of the firms export, with most (60 percent), exporting mainly to Eurozone countries. Only 27 percent of the firms with fewer than fifty employees export, whereas half of the large ones do. Among exporters, most are unsurprisingly in the manufacturing/retail sector (56 percent of the manufacturers are exporters), with 25–30 percent in the commercial retail/services sectors. About 32 percent of the sample is coded as high-skilled and 33 percent as representing sectors that are tradable.⁴⁵ 39 percent of the firms view unions to have some influence over decision-making within the firm. Many firms have suffered during this period of economic crisis, a further indication of the validity of the sample. 75 percent of firms report having lost profits, with an average profit loss of near 30 percent; 53 percent report having lost workers (with an average workforce loss of 10 percent, though the average loss among firms that have lost any workforce is closer to 30 percent). Unsurprisingly, the construction sector by far has the highest proportion of economic stress, with 71 percent of the firms in that sector reporting both loss in profits and workers. Tables 2a and 2b provide relevant descriptive statistics on structural characteristics and views of firms.

Table 2a: Descriptive statistics—characteristics of firms

Table 2b: Descriptive statistics—views of firms

Basic business views and preferences

What are the patterns in terms of business concerns and the crisis? Recall we asked firms to assess the relevance of problems relating to taxes, high wages, access to credit, and consumer demand (the latter two constituting more macroeconomic-oriented concerns), as a separate indicator of crisis-specific concerns for firms. On the ten-point scale, taxes are the highest concern (with an average of 8.1); the lowest concern of these five is high wages, though the average is still high, at 6.1. In terms of our measure of conservative orientation, respondents blame the Zapatero administration more, as the average blame of the Zapatero administration is 7.5 out of 10, and for Rajoy's administration it is 6.3.

We now turn to descriptive patterns about support for the euro, austerity, and deflation (lower wages). We find on average high support for the euro as a currency and for staying in the Eurozone, as only 12 percent of firms believe their industrial conditions would improve if Spain were to leave the Eurozone. With respect to domestic policy, firm representatives support status quo “conservative” economic policies. 57 percent believe that fiscal adjustment policies have been necessary to end the crisis. Finally, there is modest support for internal devaluation as a policy applied to firms, with a combined 67 percent support it at the midpoint level or more (40 percent support deflation strategy at a level higher than the midpoint level, and 27 percent do at the midpoint level).

Section 4. Results

Estimation approach and controls

We provide a more systematic test for the arguments discussed above with a set of estimations controlling for relevant firm characteristics, testing the hypotheses presented in table 1. Our approach is to control for baseline structural factors, and in our discussion of results we pay special attention to the hypothesized most relevant structural variable for each policy preference. The regressions control for variables that proxy crisis-specific concerns, and finally the variable indicating political orientation.⁴⁶ Unless otherwise noted, for the OLS specifications the dependent and independent variables are scaled 0–1. By recoding both the dependent and independent variables in this manner, we interpret a regression coefficient as representing a 100*β percentage-point increase in the dependent variable associated with moving from the lowest to highest possible value of the independent variable. Each estimation controls for the following battery of basic firm characteristics, as described and measured above: firm size (measured in the natural log of the workforce); the percentage of a firm's revenue deriving from exports to the EU; a binary indicator for each of the broad industrial sectors (with services sector set as baseline), binary indicators for tradable or non-tradable sector and high versus low skilled, union presence (coded as a perception of 0–1 with 1 indicating higher values of union presence), percentage of workers on a temporary contract, the unemployment of the overall sector, and finally, whether the firm has suffered during the crisis, which we calculate as percent of profits lost (scaled 0–1, with higher values indicating more profit loss).⁴⁷ We present the results for preferences for a Eurozone exit, austerity policies, and deflation, respectively.

Estimation results—euro preferences

What are the correlates of believing a Eurozone exit would be beneficial or detrimental? The core structural variable of interest is export orientation to the European Union. We first note that in terms of differences between EU exporters and other firms, 42 percent of firms that are EU exporters believe that the Eurozone exit would be detrimental, versus only 26 percent of firms that are not exporters (p < .001). Table 3 presents the main estimation results with the hypothesized variables of interest in the top row. Columns 1–2 present OLS estimations (where the outcome is rescaled 0–1 and higher values indicate more positive consequences). Columns 3–4 present ordered logit estimations where the highest category is the belief that the firm would benefit. In both sets of specifications, we estimate the impact of a binary indicator for whether the firm exports to the European Union at all, as well as a continuous EU export orientation variable. Across all specifications, the EU export orientation variable is consistently negatively correlated with the belief that the firm would benefit from a Eurozone exit.

Table 3: Policy preferences—Euro exit would be positive

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

According to column 1, being an exporter decreases the belief that an exit would be beneficial by approximately five percentage points. Moving from minimum to maximum dependence on exports is linked with 11 percentage points in reduced support for a euro exit. These substantive results are shown in figure 1. Whereas among non-exporters the probability of reporting that a euro exit would be detrimental is only 16 percentage points higher than reporting that it would be beneficial, for a firm that exports half of their total sales this difference increases to nearly 30 percentage points. Thus, we find evidence against the basic open-political economy expectation that internationalized firms might be more likely to favor a euro exit because of the expectation of a currency depreciation.⁴⁸ Therefore, either the gains from nominal depreciation are limited, other concerns outweigh pure price competitiveness concerns, or both.

Figure 1: Exporters versus non-exporters views of a Eurozone exit

Table 3 also shows that having suffered more during the crisis does not seem to affect beliefs about the desirability of leaving the euro. Profit losses do not differentiate firm views on the euro exit.⁴⁹ This casts doubt on theories that extrapolate from individual voter results that negative economic circumstances translate into hostility towards status quo institutions, as crisis experience does not differentiate firm views of a euro exit. Finally, as columns 3–4 of table 3 show, political orientation is not correlated with a substantive shift in the categories of euro consequences (from negative to positive) of a euro exit. While left-wing orientation is positively correlated in the continuous measure, this does not translate into substantive changes in categories. We are thus not confident that political orientation is strongly linked to support for a euro exit.⁵⁰

The above evidence indicates that EU export-oriented firms are in fact more likely to think that a euro exit would be harmful, and surprisingly, that firms with higher profit losses are not more likely to have positive views on the consequences of such an exit. This poses a potential challenge to political economy theories that emphasize the importance of relative currency value as the key reason for firm support for various currency regimes. The evidence also challenges theories that indicate that negative economic conditions should increase dissatisfaction with status quo institutions.

However, as discussed in the theoretical section, the fact that exporters do not prefer a euro exit could be because the consequences of a Eurozone exit are difficult to forecast, and relevant counterfactuals that might occur in the case of an exit have to be considered. We build upon these results by probing more why some firms might believe a Eurozone exit might be detrimental or not, testing two logics, derived from the above theoretical discussion: whether firms' views on the euro follow from a retrospective or prospective logic, or both. We test the retrospective logic by examining which covariates of firms might explain firm blame of the euro as a cause of the crisis. This allows us to test the extent that the covariates correlated with euro preferences also correlate with euro blame. The prospective logic refers to firms' beliefs about the different negative consequences of a euro exit.

To test firms' retrospective assessment of the euro, we use two measures. We measure a binary variable indicating whether the respondent believed that the single currency had some responsibility for the economic crisis facing Spain (1 coded as yes, 0 as no), and blame for the euro as a cause for the crisis on a 0–10 scale. A majority of firms (58 percent) disagreed with the statement that the euro has “played some role” in the crisis. The average score of blame for the euro was 4.1, far less than firm blame of the leadership of Zapatero and Rajoy.⁵¹

Retrospective logic and blame

Table 4 shows estimations examining the correlates of blame, using the same control variables as in table 3. The first two columns consider as the dependent variable the binary indicator of whether firms believe the euro as a factor in the crisis, and the last two consider the alternate measure of blame on the continuous 1–10 scale, rescaled 0–1. Columns 1–2 consider the covariates of interest, and then show the results when controlling for structural conditions, respectively. Both columns show that profit loss is not correlated with a belief that the euro was a factor in the crisis. Columns 3–4, following the same format, show that firms that have suffered more are more likely to blame the euro on a 0–1 scale; the magnitude is .25 points on the blame scale (this is nearly one standard deviation). In contrast, we note that across the different measures of blame, export orientation is uncorrelated with blaming the euro more or less. Overall, the fact that crisis suffering is somewhat associated with euro blame (but not, as we observe in table 4, with a belief in the euro being a main crisis factor, nor, as table 3 shows, with views towards euro exit), while export orientation is associated with opposition towards euro exit (but not towards euro blame) is suggestive that the retrospective logic regarding blame of the euro does not drive firm-level euro preferences. Firms that have suffered during the crisis blame the euro somewhat more than other firms, but are not particularly more euro-skeptic, and exporters are clearly more supportive of the euro in spite of not having a more sanguine view of the role of the euro during the crisis. Further, table 4 shows that many structural variables are uncorrelated with variation in blame assignment of the variable.⁵²

Table 4: Retrospective evaluation of euro

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

Prospective logic

Continuing our analysis of why export-oriented firms may not favor a euro exit, we now explore what consequences would a euro exit have. In the survey, we asked about the following potential concerns regarding a possible Eurozone exit, if it were to occur: access to finance, demand for the products of the company, cost of intermediate goods used by the company, the competitive position of the company, and finally, the ability of the company to adjust salaries and costs. Again, these are on a 0–10, and are rescaled 0–1 with higher values indicating more concern.

Table 5 displays similar estimations as those in in Table 4, but now the dependent variables are the different prospective concerns about a euro exit.⁵³ The results show that export-orientation is positively correlated with concerns in the event of a euro exit: reduced demand, prices of intermediate goods, the effect on firm's competitiveness, and adjustment policies. The magnitudes from moving from none to any EU export orientation are quite large, at 11, 11, 17, and 9 percentage points, respectively. The table indicates that firms have many fears related to leaving the euro (which arguably guarantees access to markets and a stable macroeconomic environment), and that these concerns likely outweigh any gains from nominal devaluation associated with an exit. These findings related to prospective concerns help explain why export oriented firms do not prefer a euro exit at this juncture.

Table 5: Prospective evaluation of euro exit

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

Table 5 also shows that firms that have suffered during the crisis are more concerned with lack of demand, the prices of intermediate goods, and adjustment costs. The coefficients for concerns about both of these concerns are quite large, affecting the concern scale by 26, 20, and 29 percentage points, respectively. Firms that have suffered, as well as export-oriented firms, are systematically more concerned than other firms with several potential implications of a euro exit. This is again supportive of a general prospective logic, linking economic suffering to concerns about a euro exit, although in the latter case, recall economic suffering does not correlate with actual preferences for leaving the euro. Table 5 also shows that firms that have more macroeconomic oriented worries are more concerned with issues related to financial access and aggregate demand issues, as hypothesized. Our final conclusion from table 5 is that political orientation remains uncorrelated with these specific euro prospective concerns, consistent with the evidence above that finds weak or little evidence that political orientation as measured here matters for euro-related preferences. Overall, we interpret this as stronger support of specific prospective concerns about a euro exit as opposed to the retrospective logic that focuses on firms' blame attribution of the euro.⁵⁴

The summary from this evidence on the euro is that given the complicated and contradictory predictions that political economy models give for a major currency exit, the main structural variable of EU export orientation is linked to prospective concerns about such an exit, and overall against such an exit. The magnitude of this variable is larger than that of economic suffering or political orientation, which seem to matter less for this preference of external adjustment.

Austerity preferences

We now turn to discussion of the results regarding austerity preferences. Because this question was binary in format, we estimate a logistic regression where “1” indicates opposition to austerity, and “0” support. Table 6 examines the dependent variable of opposition to austerity. The first column considers just the firm-level demographic variables and profit losses, and the latter columns introduce macroeconomic concerns and political orientation, respectively. We find that our standard firm structural characteristics do not explain much variation in the opposition to austerity, as few variables, including the hypothesized variable of firm size, matter. Nor are there differences across sector categories or most other characteristics. Further, in terms of crisis-experience variables, economically suffering firms are no more likely to be opposed to austerity.⁵⁵ Thus the connection between economic suffering and demands from the state for spending (as found in surveys of citizens) do not appear in this firm-level data. The specific way in which the crisis experience matters is via specific macro-level concerns; as the latter columns show, firms that have macro-oriented concerns about demand and credit are more likely to oppose austerity (moving from minimal to maximal such concerns increases opposition to austerity by nearly 30 percentage points). The role of the crisis and views towards austerity thus matters in terms of concerns that firms have regarding the crisis, but not necessarily raw losses.

Table 6: Policy preferences—opposition to austerity

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

When we consider our third family of hypotheses, we find that political beliefs matter a great deal. Our proxy of left-wing orientation is correlated with opposition to austerity by a large magnitude, much more so than any structural indicator. A one rank increase difference in blame between the two governments corresponds to a 5 percentage point increase in opposition to austerity. Figure 2 displays the size of this effect, with predicted opposition to austerity plotted against increasing left-wing political orientation. As the figure shows, moving from minimum to maximum left-wing orientation leads to extremely large changes in opposition to austerity (with a peak of 0.8 probability of opposition). This suggests that what matters in terms of relevant variation in firm-level policy preferences regarding fiscal austerity is not the economic characteristics of the firm, but rather, the political orientation of its managers.

Marginal effects based on estimations in column 3 of table 6; 95% confidence intervals.

Figure 2: Left-wing orientation and opposition to austerity

We build on these austerity results to ask, why is economic suffering not correlated with opposition to austerity? To partially address this null finding, we posed a follow-up question that assessed one plausible reason why firms may support austerity, which is the extent to which the firm representative believes that the size of the public sector has made the crisis worse. This variable is also coded 0–10 (with 0 indicating it has nothing to do with the crisis and 10 being that it is the principal factor explaining the crisis), and rescaled 0–1 as with the other variables. We first note overall that 81 percent believe that the public sector's role in the crisis is above the midpoint level, and that 25 percent rank the public sector's role a 10, that it is one of the principal factors explaining the crisis.

Table 7 presents estimations where this belief in the public sector as a problem for economic recovery is the dependent variable, with the same firm-level covariates and format of columns as in table 6. Again, as with the models predicting opposition to austerity, we find little evidence that basic structural indicators such as industrial sector differentiate firms much. As the most complete specification in column 3 shows, controlling for the other covariates, firms that have economically suffered are in fact more likely to believe that the size of the public sector is a problem. Moving from minimum to maximum suffering increases negative views of the public sector by 18 percentage points. This finding indicates that the non-correlation between economic suffering and views on austerity cannot be accounted for by firm views about the public sector, as firms that have suffered actually have higher skepticism about the size of the public sector.⁵⁶ It could be the case that firm skepticism of the public sector is offset by some other concern, unmeasured here, that results in lack of correlation of firm suffering with views on austerity. Table 7 also shows that firms that have more macroeconomic-oriented concerns are, as with their opposition to austerity, less likely to believe that the public sector is a source of the crisis. These findings of crisis-specific variables support the view that even though economic suffering does not really differentiate firms in terms of austerity policy views, specific macroeconomic concerns do. In fact, firms that have suffered believe the narrative that the public sector is a culprit for the crisis, although that does not translate into outright austerity support. Finally, as the third column of the table shows, left-wing oriented firm managers are much less likely to believe that the public-sector size has been a cause of the crisis; the magnitude of this variable is close to that of profit loss in size (the minimum to maximum shift in left-wing orientation is approximately 16 percentage points). This finding further supports the relevance of political orientation for austerity views.

Table 7: Believe public sector size is cause of crisis

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

Support for wage devaluation policies

We turn to our final policy outcome of interest, support for internal devaluation. Table 8 examines the correlates of the belief that policies resulting in nominal price and wage cuts are a good strategy for firms. As with opposition to austerity, the first baseline specification shows that standard covariates do not explain much variation in support for deflationary policies.⁵⁷ But firms with greater union presence are strongly more supportive of this strategy by about 18–20 percentage points (moving from minimal to maximum union presence). We also find a consistent positive relationship between having suffered in the crisis and deflationary policies, although this effect is imprecisely estimated in the first specification (p < .12); the magnitude in columns 2 and 3 of minimum to maximum profit loss is nearly 20 percentage points. Firms that have macro-oriented concerns about demand and credit are also less likely to support deflationary policies (moving from minimal to maximum values on such concerns increases opposition to deflation as a strategy by 17–18 percentage points). Again, as with opposition to austerity, regarding our proxy of political orientation, we also find that this variable is correlated with opposition to deflationary strategies by a large magnitude. Moving from the smallest to largest amount of the political orientation variable leads to a significant 24 percentage point reduction in support for deflationary strategies. Thus, consistent with expectations from literature on ordinary voters, left-wing managers are much more opposed to deflation.

Table 8: Policy preferences—support deflation

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

In the theoretical section we suggested two reasons why firms with a higher union presence might prefer lower wages, as unions can help structure wage reductions to avoid conflict, or, because firm managers might believe that strong union presence keeps wages artificially too high. While we do not have a direct test that can arbitrate among these variables, we use a follow-up question that asked about firms' views about the economic consequences of unions. The question asked the firm to assess on a 0–10 scale whether unions were beneficial or not, with 0 indicating that the firm thought that unions were completely negative, and 10 indicating completely positive. 44 percent of firms rated unions as 0 on this scale, and only 22 percent of firms rated unions as a score higher than 5. If a negative view of unions is correlated with union presence, then that would indicate that one potential channel of support for deflation is via negative views of unions and their presence in firms.⁵⁸

To test for this, we estimate as before an OLS model where the firms' view of unions is rescaled 0–1, with higher values indicating more positive views of unions, with the same demographic controls and attitudinal variables as before. We present these results in table 9. The key results we note are that the structural variable of union presence in the firm is positively correlated with positive perceptions of unions (not negative, as one might expect), and that the variable of profit loss in these specifications are uncorrelated with union views. Thus, the connection between the union presence variable in table 8 with support for deflationary strategies is not necessarily via union hostility. Table 9 also shows that, as with previous results, left-wing oriented firm representatives are much more positive towards the role of unions; this again by far is the highest magnitude variable in the estimation, as moving from minimum to maximum left orientation increases support for unions by 28 percentage points on the scale, or more than one standard deviation. This finding further supports the notion that on contentious domestic policies, proxies for political orientation are more correlated with policy views than firm-level structural characteristics.⁵⁹

Table 9: Belief that unions are positive

Standard errors in parentheses

* p < 0.10, ** p < 0.05, *** p < 0.01

Section 5. Discussion

In this study we present one of the few systematic analyses of firm-level preferences for policies to respond during times of economic crisis, with a large representative survey of firms from a crisis-hit country. We present a wealth of new data that measures both key policy preferences and a battery of potential mechanisms relating firm characteristics to such preferences. We first summarize the main takeaways from the study, and then discuss implications and extensions.

Our evidence indicates that firms overall support a more conservative status quo policy package, the components of which include staying in the Eurozone, support for fiscal austerity, and support for internal devaluation. Regarding potential reasons for such preferences, we find a majority of firms have a positive disposition towards the euro and are concerned about many possible negative scenarios of a euro exit. On the domestic front, we find that a majority of firms believe that the public sector does have a role to play in the duration of the crisis, and believe that unions are a negative economic force. Overall, these results suggest that to the extent that the general Spanish public, or populations of other crisis-hit countries, are opposed to the current status quo policies of responding to the crisis, firm representatives seem on balance more supportive.

We observe variation in such preferences, and our objective was to test different families of theories to explain such variation, with particular theoretical attention to the roles of structural factors, economic suffering, and political orientation. From our analysis, we draw several main conclusions, some of which challenge existing assumptions and call for refinement of measurement and theories of business policy preferences. Returning to the table of predictions, we first note that, the political leanings of firm representatives matter substantively more for domestic policies than for the euro, which previous theories do not focus on. And in the case of domestic policies, overall, our measure of political orientation is often the highest magnitude variable and strongly correlated in hypothesized directions—left-wing firm representatives are markedly more anti-austerity, anti-deflation, and correspondingly less likely to blame the public sector and think unions are negative. However, this political orientation is not consistently correlated with views about the euro. This indicates that political orientation might be more relevant regarding policy options where the firm-level consequences are either more directly visible, or where the political debate around these issues has been more politically polarized, as is the case with spending and labor-market reforms in Spain and in other Eurozone economies. The strength of this finding suggests that future research should focus on how and under what circumstances these political leanings of business matter for policy-making.

Second, a striking finding from the data is that many basic structural variables such as the sector, size, and tradability of products matter relatively little for understanding variation in firm preferences. Knowledge of these obvious characteristics reveals little about how the firm representative thinks about many of these policies. This is not to say that structural features do not matter—in two of the three policies, we laid out predictions regarding export orientation (for the euro), and union presence (for wage deflation), and found that such characteristics do matter. But it is surprising how many basic features of firms are less relevant regarding most policy preferences.

Third, we find little evidence that economic suffering itself is a major driver of policy preferences for firms. It matters little regarding euro attitudes and for austerity. In fact, regarding the latter, we suggest one reason why firms that have not suffered are not more anti-austerity is that they have the opposite set of attitudes towards the state; those firms believe that the size of the public sector is a persistent factor prolonging the crisis. As economic suffering is correlated with skepticism of the public sector and support for deflation, this is evidence that profit loss does not necessarily firms to the left, as it appears to for voters in crisis-hit countries. Regarding our other measurement of crisis-specific concerns, we do find that firms that have greater concerns about aggregate demand and credit are more opposed to austerity and deflation. Thus, in further theorizing about the role of economic losses, it is important to disentangle raw losses from specific concerns facing firms.

This broad conclusion from the data has theoretical implications. The fact that political orientation of firm managers seems to matter more than economic suffering, and that the latter does not seem to greatly matter for many domestic policies in anticipated left-wing directions, could partially explain muted demand in the business community for austerity and the absence of coalitions with other groups, such as organized labor. Economic losses in fact make policies to reduce wages even more attractive and are linked to skepticism of the public sector. On issues where political orientation might match more conventional rhetoric about partisan support for various policies, the political division among firm managers could matter much more so than previously thought.

The objective of the study was not to definitively arbitrate among these competing families of theories, as certainly all should be relevant in explaining firm preferences for certain policies. Rather, we view one of our contributions as showing the importance of additional variables beyond structural features that have an independent effect on firm preferences, such as firm-level concerns about the crisis and political orientation. Future work might more precisely disentangle additional mechanisms linking both orientation and crisis concerns to other policies. There certainly could be selection effects in terms of certain politically oriented firm managers choosing particular industries or believing policies are beneficial for firm profitability. Another contribution we think should be extended upon is understanding why firms that have suffered more do not appear to be a large constituency for ending austerity, as they have stereotypical conservative views of the role of the state in the crisis.⁶⁰ Certainly, sampling much larger firms would provide a complementary view of how firms of very different sizes might view the crisis differently. Our results indicate the value of measuring more in-depth other aspects of firm-level managers' political orientations, as those are most tightly linked to policy views.

A final important open question is how these preferences translate into business strategies to influence policies. Future research should more rigorously measure the relationship between specific policy preferences and incentives to either convey such preferences to lobbying associations, government authorities, or political parties directly. This agenda would allow for testing which policy preferences are in fact most important to firms.