Exports vs. Investment:
How Political Discourse Shapes Popular Support for External
Federico Maria Ferrara (a), Jörg Stefan Haas (b), Andrew Peterson (c), Thomas Sattler (d)
(a) London School of Economics and Political Science
(b) Hertie School, Berlin
(c) University of Poitiers
(d) University of Geneva
Forthcoming in the Socio-Economic Review
The economic imbalances that characterize the world economy have
unequally distributed costs and benefits. That raises the question how
countries could run long-term external surpluses and deficits without
significant opposition against the policies that generate them. We show
that political discourse helps to secure public support for these policies
and the resulting economic outcomes. First, a content analysis of 32,000
newspaper articles finds that the dominant interpretations of current
account balances in Australia and Germany concur with very distinct
perspectives: external surpluses are seen as evidence of competitiveness
in Germany, while external deficits are interpreted as evidence of
attractiveness for investments in Australia. Second, survey experiments
in both countries suggest that exposure to these diverging
interpretations has a causal effect on citizens’ support for their country’s
economic strategy. Political discourse, thus, is crucial to provide the
societal foundation of national growth strategies.
Keywords: current account; growth; competitiveness; financial flows; voter attitudes
JEL classifications: F5 International Relations, National Security, and International
Political Economy; F1 Trade; F3 International Finance
The authors acknowledge financial support from the Swiss National Science Foundation,
grant no. 165480.
Countries exploit the opportunities stemming from economic openness in very
different ways (e.g. Baccaro and Pontusson, 2016; Baccaro and Benassi, 2017; Blyth and
Matthijs, 2017). Some countries, such as Germany and Japan, strongly focus on the
opportunities from international trade and aim at generating growth by maximizing exports.
Others, such as the UK or Australia, rely more heavily on international capital inflows to
boost growth by financing domestic consumption and investment. As a result, the external
economic balance has been identified as a critical aspect of a country’s growth strategy in an
open world economy (Baccaro and Pontusson, 2016, esp. pp. 183 and 191-192). More
broadly, the global macroeconomy is important to understand the workings of domestic
macroeconomic regimes (Blyth and Matthijs, 2017). Export-driven economies have run large
external surpluses, while investment-driven economies have run sizable deficits for most of
the post-Bretton Woods period. Together, they have repeatedly created a need for domestic
and international economic adjustment with adverse effects on international cooperation.
The different growth strategies do not only have the potential to cause international
economic conflict; they can also bring disadvantages for the domestic population. Economies
with large external deficits are often vulnerable to sudden stops in financing, as the Euro
Area crisis has shown (e.g. De Grauwe, 2011; Regan, 2017; Walter, Ray and Redeker, 2020),
or experience negative effects on labor markets in areas that house import-competing
industries (e.g. Autor, Dorn and Hanson, 2013; Cerrato, Ferrara and Ruggieri, 2018). In
perennial surplus countries, wages and domestic investment are chronically low, which has a
negative impact on large parts of the population (Jones, 2009).
This paper, therefore, asks how some countries could sustain their growth strategies
and run such persistent external imbalances without major domestic opposition against the
policies that generate them. The existing literature gives a partial answer to this question. It
identifies wage bargaining institutions as a main determinant of the long-term external
balance (Hall and Soskice, 2001; Hancké, 2013; Johnston, Hancké and Pant, 2014; Johnston,
2016; Manger and Sattler, 2020), which points to an important part of the mechanism. But it
does not explain why this institutional setup receives broader societal support even though it
is by no means obvious that a majority of the population benefits from it. In Germany, for
instance, the exemplar of an export-driven surplus economy, only about a quarter of all
employment is linked to exports (Bundesministerium für Wirtschaft und Energie, 2019).
Nonetheless, German voters and interest groups consistently support policies that generate
external surpluses, such as low-deficit policies (Redeker and Walter, 2020; Hübscher, Sattler
and Truchlewski, 2021). It remains unclear why a large majority of the population tolerates a
large external surplus that deprives them of higher wages and consumption opportunities.
After all, there are plenty political levers that could reduce imbalances.
The toleration of a
long-run imbalance, thus, is a political decision that requires a political explanation.
Our analysis shows that political discourse is a crucial determinant of citizens’
support for policies that produce external imbalances. Building on discursive institutionalism
(e.g., Schmidt 2008; Schmidt 2010), we define political discourse as the process by which
policy ideas are constructed and conveyed to the public by political actors. We conceive of
political actors in a broad manner, with our analysis considering a wide range of groups,
encompassing the media, politicians, technocrats and experts.
We focus on the effects of political discourse on citizens’ policy preferences and
argue that imbalances can be interpreted in two ways: either through a trade logic that
highlights competitiveness, or a financial logic that emphasizes investment and savings. The
two perspectives focus on very distinct economic mechanisms and are consistent with
different sets of policies, or ‘growth models.’ These interpretations, then, affect how voters in
a country think about economic policies and the resulting outcomes. Citizens who are
primarily exposed to the competitiveness perspective are more likely to accept ‘belt-
tightening’ and austerity policies that lead to current account surpluses. Citizens who are
exposed to the investment perspective are less likely to support austerity, but favor
‘loosening’ policies that can lead to current account deficits.
The empirical analysis of Germany and Australia, two countries that represent polar
opposites when it comes to their external economic balance, confirms these conjectures. We
proceed in two steps: first, we examine how the dominant interpretations of the current
account in the political discourse vary between the two countries. Then, we test if discourse
has a causal effect on public opinion.
For the first step, we assess citizens’ exposure to different forms of discourse about
the current account using a structural topic model analysis of 32,010 German and Australian
newspaper articles. We find that the interpretations of the current account presented by the
media differ fundamentally across countries. In Australian newspapers, reporting on the
current account is more likely to mention investment and savings than competitiveness and
productivity. The opposite is true for German newspapers. In the Online Appendix, we
validate these results in a variety of ways: first, we present evidence from a qualitative
examination of media reports on the current account; second, we employ a dictionary-based
approach to confirm the results of the topic model; third, we perform a structural topic model
analysis of the speeches of the Reserve Bank of Australia and the Deutsche Bundesbank to
show that newspaper reports in fact proxy the dominant interpretations among key
policymakers in Australia and Germany. In sum, we provide a big picture of the differences
in discourse about macroeconomic imbalances in the two countries.
For the second step, we build on studies about the effects of framing on attitudes
towards economic policy (e.g., Ardanaz, Murillo, and Pinto 2013; Harell, Soroka, and
Iyengar 2016; Barnes and Hicks 2018) and make use of a survey experiment that tests how
people react to a change in the discourse about the current account. In the experiment, we
randomly expose respondents to the trade-competitiveness perspective and the saving-
investment perspective. This ensures that we properly identify the causal effect of discourse
on attitudes, and not the other way around. The results suggest that citizens’ opinions are in
fact responsive to the different narratives. The two perspectives influence citizens’ approval
of a proposed policy package that would reduce the external imbalance. Popular support for
policies undergirding external balances is thus susceptible to the economic ideas that are
conveyed by different forms of discourse around the current account balance.
To our knowledge, this paper is the first to focus on the effect of discourse on public
support for diverging growth strategies and external imbalances. The political economy
literature has long highlighted the influence of ideas and discourse on policymaking and
institutional change (McNamara, 1999; Blyth, 2002; Schmidt, 2002), and their importance in
stabilizing dominant social groups (Amable and Palombarini 2009). Ideas and discourse are
notoriously difficult to measure, but recent research has made several advancements in this
direction and provided increasing evidence of their influence on economic policy (Chwieroth,
2007; Helgadóttir, 2016; Ferrara, 2020). Our study confirms this by showing how political
discourse affects the policy preferences of the mass public. In this way, discourse helps to
secure societal support for institutional arrangements that embody diverging economic
2 Two Perspectives on External Imbalances
We differentiate between two main perspectives on the current account balance: the
trade-competitiveness and the saving-investment perspective. From the first perspective, the
current account position is defined as exports minus imports plus net income from abroad. A
country will run a surplus when it sells more goods and services than it buys, which implies
an important role for international competitiveness. From the second perspective, the position
is defined as the difference between domestic savings and investment. A country will run a
surplus when there is less domestic investment than there are domestic savings available,
which implies an important role for capital flows.
Both perspectives are equally valid. In fact, both measure the same thing and will, per
definition, yield the same result. However, since they emphasize different driving forces
behind current account dynamics, they guide our thinking in different directions. In other
words, “the way we talk about identities and our models can inadvertently shape the
inferences we draw from them" (Borio, 2016, p. 2). The two perspectives yield diverging
‘policy targets’, which are consistent with different ‘growth models’, as comparative political
economy research recently highlighted (Blyth and Matthijs, 2017; Baccaro and Pontusson,
2016). The current account, thus, represents the international dimension of a particular
growth model as it is the entity that links the domestic to the international economy.
2.1 The Trade-Competitiveness Perspective
The trade-competitiveness perspective played a major role in the history of
international political economy, especially for the mercantilism–liberalism debate of the 17th
and 18th century (Mun, 1986 ; Smith, 2003 ; Viner, 1948).
mercantilists recommended that countries run an external surplus by exporting more than
they import to increase a country’s power. Although the modern version of this perspective
highlights the role of jobs and growth rather than power, its recommendations are remarkably
similar. Research in the ‘neo-mercantilist’ tradition suggests that export-promoting strategies,
such as exchange-rate undervaluation, promote growth (Rodrik, 2009) and secure domestic
jobs (Krugman, 2016).
Versions of the mercantilist view have recently reappeared in interpretations of global
imbalances and the Euro Area crisis suggesting that the export-promoting strategy of surplus
countries exploits deficit countries. Some assert that Germany consciously undercut the
wages of other Euro Area members, hence, robbing them of significant market shares in
regional and global trade (Flassbeck and Lapavitsas, 2013, p. 14). Other claim that a surplus
is desirable and criticize the Euro Area’s deficit countries for their failure to follow the
German example (Sinn, 2014). This interpretive framework points to persistent current
account deficits in peripheral countries as the root cause of the crisis and appears to be
largely inspired by neo-mercantilist ideas underpinning Germany’s growth strategy.
From a less normative point of view, scholars in the field of comparative political
economy argue that trade plays a decisive role in shaping current account imbalances and
stress the role of institutions in managing wage growth and maintaining competitiveness
(Hancké, 2013; Johnston, Hancké and Pant, 2014; Johnston 2016; Manger and Sattler, 2020).
Specifically, coordinated wage bargaining systems in combination with the broader
institutional framework facilitate wage restraint and limit inflationary pressures (Hall and
Franzese, 1998), which helps export-oriented industries to compete internationally. This
leads to a strong tendency towards current account surpluses. More broadly, countries that
follow different growth models can either rely more on domestic consumption or more on
exports, with diverging effects on the current account.
2.2 The Saving-Investment Perspective
In contrast, the saving-investment perspective discounts trade flows and highlights
international financial flows (Obstfeld and Rogoff, 1996). Already in the 18th century, the
mercantilist focus on exports was criticized on the grounds that it was neither desirable nor
possible to run a surplus and accumulate precious metals forever (Hume, 1752). More
recently, scholars have argued that “forward-looking households and firms ... will generate
current-account balances consistent with efficient resource allocation" (Obstfeld, 2012, p. 14)
and that a current account deficit may be the desirable consequence of real capital
movements (Pitchford, 1989, p. 8). Therefore, the perspective provides little rationale for
actively steering the current account, be it directly via government intervention or via
institutions that support surpluses.
Like its counterpart, the saving-investment perspective plays a prominent role in the
interpretation of global imbalances and the Euro Area crisis. Some scholars consider that
policymakers’ overwhelming focus on restoring competitiveness via wage adjustment was
misplaced and priority should have been given to stabilizing the financial system instead
(Jones, 2011, 2015, 2016). In the case of the US, former Federal Reserve chairman Ben
Bernanke claimed that the “trade balance is the tail of the dog; for the most part, it has been
passively determined” (Bernanke, 2005). Such arguments do not only have important policy
implications for governments, but also for the more normative question whether or not an
external surplus or deficit is desirable or a problem in the first place.
Table 1: Two Perspectives on the Current Account
Trade / Competitiveness
Investment / Savings
Trade, dependent on
competitiveness and wages
Financial flows, triggered by
investment and savings decisions
Growth through exports
Growth through investment
This is not to say that one would never worry about long-term external imbalances
from the saving-investment view. However, even those who see imbalances as useful
indicators of potential financial crises point out the risks of both surpluses and deficits
(Obstfeld, 2012). Others go even further by claiming that the importance of the current
account is overstated, and that more attention should be paid to financial flows instead
(Borio, 2016). In the policy debate, the investment perspective is widespread among
international organizations. Despite the differences between their procedures of
macroeconomic surveillance (Moschella, 2014), both the European Commission and the
International Monetary Fund have recently recommended that Germany act against its large
current account surplus by increasing investment (European Commission, 2016; International
Monetary Fund, 2016). Table 1 summarizes the two perspectives and their implications.
3 Building Popular Support for an Economic Strategy: The Role of
How do the different perspectives matter?
Our analysis suggests that they affect
citizens’ attitudes towards economic policies via political discourse. We define political
discourse as the process by which political actors construct policy ideas and convey them to
the public. By employing the term political actors, we do not just refer to politicians, but
virtually any group that communicates to the public and potentially shapes its preferences:
this includes the media, non-elected policymakers and experts.
We adopt a discursive institutionalist approach and see discourse as a more versatile
and overarching concept than ideas: the term “discourse” may be used to indicate the
interactive processes by which ideas are conveyed as well as their substantive ideational
content (Schmidt 2008). Discursive institutionalism conceptualizes discourse along two
different dimensions: ideas are created, elaborated and justified with “coordinative”
discourse, while “communicative” discourse denotes the process of deliberation and
communication of ideas to the public (Schmidt 2010).
Coordinative discourse is performed by political actors involved in the formulation
and construction of policy ideas. These actors interact with each other to shape the common
understanding of policy problems. Coordinative discourse is typically performed by
coalitions of activists, e.g., “advocacy networks” (Keck & Sikkink 1998), or communities of
experts organized on the basis of shared cognitive and normative beliefs, e.g., “epistemic
communities” (Haas 1992). But coordinative discourse equally involves elected officials and
policymakers discussing with each other, often behind closed doors, as well as
representatives of labor unions, employers’ associations and sectoral interests. All these
actors participate in interactive processes aimed at making sense of complex social
phenomena, thereby producing cognitive and normative policy ideas that define the
boundaries of feasible and desirable policy action.
Communicative discourse regards the action of bringing ideas to the forefront and
conveying them to the public for deliberation and legitimation. The actors typically involved
in this process are political leaders attempting to persuade the public to get (re-)elected, as
well as technocrats sharing their views and explaining their decisions to ensure accountability
and gain legitimacy with the public. Yet communicative discourse encompasses other
political actors as well, including the media and experts (Schmidt 2008: 310). The
communicative discourse of this wide range of political actors feeds into the public debate,
which then affects citizen perceptions and evaluations of economic policies.
Thus, political discourse does not only have important implications for the
construction of policy actors’ consensus over economic policy in coordinative arenas (Hay
and Rosamond, 2002). By means of its communicative function, political discourse also
helps to promote a wider societal consensus about the national interest among those who do
not directly benefit from imbalances. In this paper, we focus on this latter dimension of
political discourse. We acknowledge that different ideas have different origins and may be
constructed in different ways, but the analysis of the ideational roots of different discourses
on the current account and the way they are constructed in coordinative arenas go beyond the
scope of this paper. Rather than focusing on the origins of ideas about current account
imbalances, we choose to assess the implications of the ideas and interpretations conveyed by
political discourse in this policy domain. We provide an empirical assessment of how
political discourse about the current account affects the policy preferences of the mass public.
The current account plays an important role in the political discourse because it is
widely accepted as a key indicator of economic performance (Financial Times, 1988; Lee,
2009). At the same time, interpretations of current account deficits and surpluses differ
widely. Like many key economic terms, the meaning of a current account surplus or deficit is
“contingent on the particular cultural frame and social setting” (Matthijs and McNamara,
2015, p. 225). What a current account surplus or deficit says about the state of the economy
and whether it should be a policy target that requires action by the government is open to
interpretation (Blyth and Matthijs, 2017). The competing interpretations of global imbalances
and the Euro Area crisis discussed in the previous section illustrate this point.
The discourse regarding the “optimal” current account balance varies by country.
Political actors rely on policy ideas to define what is in the general interest and to separate
legitimate from illegitimate political demands (Amable and Palombarini, 2009). The
discourse on current accounts, therefore, is instrumental to generate broad societal support for
a particular growth strategy (Baccaro and Pontusson, 2016). For instance, the
competitiveness perspective motivates the reliance on foreign demand to drive long-term
development in an export-led growth model. In this way, discourse may secure support for a
particular growth model among citizens by defining what constitutes a viable policy solution
for their country (McNamara, 1999; Blyth, 2002; Schmidt, 2002; Best, 2004; Matthijs and
Discursive processes are, of course, not the only possible explanations for policies
that maintain imbalances. Material considerations and institutional factors can play an
important role, given that external adjustment has distributive consequences that affect
welfare across groups in the population (e.g. Frieden, 1991; Walter, 2013; Ferrara and Sattler,
2018). For instance, restrained wage growth and internal devaluation strategies ensure the
export competitiveness of manufacturing industries. These macroeconomic strategies are
favored by coordinated wage bargaining systems, as in the case of Germany, where
manufacturing industries typically constitute an important share of the national economy
(Manger and Sattler 2020). Thus, the prevalence of competitiveness narratives in these
countries may be seen as epiphenomenal to the presence of strong sectoral interests that
benefit from them. However, even in an archetypical surplus country like Germany, only
around a quarter of all employment is directly or indirectly linked to exports
(Bundesministerium für Wirtschaft und Energie, 2019; IWD, 2020). For most Germans,
prioritizing international competitiveness over consumption or investment does not yield
immediate material benefits. Nonetheless, German voters and interest groups consistently
support policies that generate external surpluses, such as fiscal restraint and low-deficit
policies (Redeker and Walter, 2020; Hübscher, Sattler and Truchlewski, 2021).
Given the complexity of external economic relations, it is plausible that citizens’
attitudes in fact are influenced by the political discourse in their country. Most citizens have
only a vague idea about the usefulness of different growth strategies and their effects on the
external economic balance of their country. Few people would doubt that higher wages or
lower unemployment rates are good for them because the effects are immediate and direct.
But alternative policies aimed at achieving economic growth, and their expected effects on
specific actors and on the economy as a whole, are a much more complex issue.
In such a
situation, a predominant discourse helps people define their interest.
To be sure, when arguing that political discourse influences citizens’ policy
preferences, we do not claim that the effect of ideas and discourse is causally independent
from that of structural realities and institutional factors. We cannot exclude that political
discourse affects public opinion in a way that is conditional upon the material interests and
institutional setting of a given society. For instance, it may be that the competitiveness
perspective resonates more with German citizens for reasons that are related to the structure
of the German domestic economy, its sectoral interests and wage bargaining institutions.
However, the assessment of the influence of political discourse on policy preferences, be it
conditional or independent, remains important to understand the foundations of different
countries’ growth models and is currently a blind spot in the literature.
The news media play an important role for the transmission of the substantive
ideational content of political actors’ discourse to the public and constitute the primary
source of exposure for citizens to the policy ideas of politicians, technocrats and experts.
Media reporting significantly influences the economic views of the public (Barnes and Hicks,
2018; Boef and Kellstedt, 2004). Its effect on voters can even be greater than that of actual
macroeconomic data (Kayser and Leininger, 2015). Furthermore, the importance of framing
effects is well-established (Chong and Druckman, 2007). News play an especially important
role for attitude formation when individuals feel an increased need for orientation because an
issue is relevant, yet ambiguous or hard to understand (McCombs and Reynolds, 2009;
Barnes and Hicks, 2018). We can therefore expect that the coverage of the different
perspectives on the current account balance will shape public opinion about optimal
The implication is that, in a country where the competitiveness view dominates the
discourse about the current account, it is easier for political actors to justify “belt-tightening”
policies that can be politically risky (Hübscher et al. 2015; Hübscher 2018; Hübscher et al.
2020), but which are important to achieve competitiveness and higher exports and, hence, an
external surplus (Baccaro and Benassi, 2017; Haffert, 2019). Citizens who are continually
exposed to this perspective are more inclined to accept these policies because they believe
that they are in their own interest as well as that of the country. In contrast, in a country
where the investment perspective dominates, we can expect citizens to tolerate policies that
generate an external deficit because it can be interpreted as an indicator of high investment
levels. Painful government interventions to reduce the deficit are harder to justify in such an
environment because the investment perspective stresses the ability of private actors to
determine the optimal external balance. The empirically observable implications of this
argument are as follows:
HYPOTHESIS 1: (a) In deficit countries, political discourse highlights the investment
perspective more than in surplus countries. (b) In surplus countries, political discourse
highlights the competitiveness perspective more than in deficit countries.
HYPOTHESIS 2: (a) Citizens who are exposed to political discourse highlighting the
competitiveness perspective accept contractionary policies that are aimed at reducing the
current account deficit more than people who are not exposed to it. (b) Citizens who are
exposed to political discourse highlighting the investment perspective accept expansionary
policies that are aimed at reducing the current account surplus more than people who are not
exposed to it.
4 Empirical Analysis
4.1 Case Selection
As Figure Error! Reference source not found.1 shows, several advanced economies
have experienced sizable imbalances over the last 40 years.
In addition, the standard
deviations in Figure 1 indicate that the current accounts for many countries do not cycle
between deficits and surpluses, but remain either in deficit or surplus for most years
For our analysis, we choose Australia and Germany because the two countries
experienced fundamental differences in the long-term external balance as Figure 1 shows.
Australia has run current account deficits of 3 percent of GDP or more for the better part of
the last 50 years but is nonetheless seen as a particularly successful economic model (The
Economist, 2016). Among the notorious deficit countries listed in Figure 1 it is clearly the
most interesting case. Spain’s high average deficit is heavily influenced by the huge deficits
after joining the euro, while it did not strongly lean towards deficits before. The US also runs
a large deficit, but this is often attributed to the US dollar’s status as reserve currency of the
world (Gourinchas and Rey, 2005; Helleiner and Kirshner, 2009).
Figure 1: Average current account balances (% of GDP) of large advanced economies,
Among the surplus countries, Germany is a particularly intriguing case. The country
has always run surpluses since records about current account balances began, except in the
late 1970s after the oil shocks and in the 1990s after German unification. The country’s
response to these shocks underlines Germany’s role as a prototype surplus country. Even the
enormous costs of reunification pushed Germany into (moderate) deficit only for a decade
because German society made massive efforts to move the current account back into surplus.
Since then, Germany has been running ever-larger surpluses. While membership in Europe’s
economic and monetary union arguably contributed to this development, disaggregating the
data by counterparty shows that in most years, the imbalances were primarily accumulated
vis-à-vis the world outside the euro area (Micossi, D’Onofrio, and Peirce, 2018, 1; Kollmann
et al., 2014, 23). Similarly, Germany already ran persistent surpluses already in the late
1960s, which contributed to the destabilization of the Bretton Woods system, in the 1970s
after the breakdown of the Bretton Woods system, and throughout the 1980s, long before the
start of the Euro Area. It is plausible that without unification Germany’s average surplus
measured as a percentage of GDP would exceed the surpluses of Japan or the Netherlands.
More importantly, German surpluses are enormous in absolute terms. As a result, the
imbalance has a considerable impact on other countries and the world economy as a whole.
In nine of the last ten years, Germany had the largest current account surplus among
advanced economies, as measured in current US dollars. In 2019, the surplus stood at 273
billion US dollars, one and a half times larger than that of Japan and three times larger than
that of the Netherlands. German macroeconomic strategy has therefore spurred a significant
amount of debate (e.g., Bonatti and Fracasso, 2013; Armingeon and Baccaro, 2015;
Bernanke, 2015; den Haan et al., 2016; Dieter, 2018). Our analysis connects to these previous
Figure 2: Current accounts of Australia and Germany (percent of GDP)
4.2 Research Strategy
Following our theoretical discussion and hypotheses, our analysis proceeds in two
parts. The first part, which we present in section 5, examines whether the political discourse
on the current account differs between Australia and Germany. We focus on both media
reporting and the public communication of influential policymakers. Hypothesis 1 would
predict that the competitiveness perspective is more prevalent in Germany and the investment
perspective is more common in Australia. The second part, which we present in section 6,
conducts a survey experiment in both countries. We study the effect of the two different
interpretations of the current account on citizens and how their attitudes towards economic
policies vary with diverging exposure to these theoretical perspectives. Hypothesis 2 would
predict that exposure to the competitiveness perspective increases support for policies that
move the current account towards a surplus, while exposure to the investment perspective
increases support for policies that move the current account towards a deficit.
Current account (% GDP)
1960 1970 1980 1990 2000 2010 2020
For our analysis of differences in political discourse, we proceed as follows. We start
from assessing differences in media reporting. For newspaper reports, we choose quality
publications that provide variation across the ideological spectrum and are sold nationwide.
For each country, we include a left-leaning, a conservative, and a business newspaper. For
Australia, we retrieve articles from the Sydney Morning Herald (left-leaning), The Australian
(conservative), and the Australian Financial Review (business). They represent three out of
four traditional Australian quality newspapers and account for about 70% of sales in that
For Germany, we collect articles from the Süddeutsche Zeitung (left-leaning), Die
Welt (conservative), and the Handelsblatt (business). Frankfurter Allgemeine Zeitung, an
important conservative newspaper, could not be included because of cost restrictions.
However, the media in our sample account for two thirds of nationwide quality daily sales
and each of them is among the top five. In all cases, we adopted the same search procedure
and selected all the articles with reference to the respective country and the terms “current
account balance" or “trade balance" in the main text. Data availability differed by newspaper
Table 2 provides more detailed information on the composition of our text corpus.
Table 2: Composition of Australian and German Corpora of Newspaper Articles
N° of Articles
Sydney Morning Herald
Australian Financial Review
The newspaper articles are examined with a structural topic model (STM), which
allows us to identify the key topics related to competitiveness and investment that appear in
newspaper articles mentioning the current account (Roberts, Stewart & Tingley, 2019). To
validate these results, in the Online Appendix, we provide a qualitative analysis of media
reports and employ a dictionary-based approach, for which we identify key terms associated
with the two perspectives on the current account based on our theory (e.g., Fan 1988, p. 44-
50; Burden and Sanberg 2003; Young and Soroka 2012). Furthermore, we test whether the
discourse of influential policymakers is really consistent with media reporting, which
constitutes an underlying assumption of our theoretical framework. In the Online Appendix,
we use STM to analyze portions of speeches about the current account of the Reserve Bank
of Australia and the Deutsche Bundesbank. Jointly, these approaches paint a detailed picture
of the role played by the two perspectives in the Australian and German debates about
current account balances.
The subsequent survey experiment then directly builds on the text analysis by
examining how the perspectives on the current account transported by the dominant discourse
affect popular attitudes. A key question is whether these different interpretations in fact have
a meaningful effect on public support for a country’s growth strategy and the associated
economic policies. The survey experiment allows us to examine whether such a causal effect
exists. Our demonstration of framing effects on attitudes towards external imbalances adds to
similar findings in the literature on government expenditure (Jacoby 2000), trade preferences
(Ardanaz, Murillo, and Pinto 2013), redistribution (Harell, Soroka, and Iyengar 2016) and
attitudes towards austerity (Barnes and Hicks 2018).
In the survey, respondents in both countries are reminded how the current account
balance in their respective country has developed in the past years. In other words,
Australians are confronted with a deficit scenario, while Germans are confronted with a
surplus scenario. We then provide respondents with different interpretations of the situation
that their country faces. Each interpretation matches one of the two theoretical perspectives
attested in the newspaper analysis. A German respondent, therefore, would see an expert
statement that interprets the German surplus either through the competitiveness or the
investment perspective. An Australian respondent would see an expert statement that
interprets the Australian deficit either through the competitiveness or the investment
perspective. We also include a control group that does not see any of the two interpretations.
We simultaneously fielded the surveys to ca. 1,000 respondents in each country in August
After confronting respondents with these interpretations, we ask all of them to what
extent they approve of a set of policies that aim at altering the current account balance. In
Australia, this is a set of “belt-tightening” policies that aim at reducing the current account
deficit. In Germany, this is a “loosening” of economic policies that aim at reducing the
surplus. Since respondents are randomly assigned to a particular interpretation or the control
group, the differences in their responses represent the causal effect of the different
interpretations on respondents’ approval of the suggested policies.
5 Text Analysis
We analyze the media coverage of current account balances using three methods, each
with its own advantages and drawbacks. Reassuringly, they reach the same conclusions. First,
we employ a structural topic model (STM) to identify discourses through the tendencies to
employ any of a number of possible words. We make use of this technique to study the two
corpora of newspaper articles. In doing so, we build on previous studies that have effectively
used STM to model the framing of international newspapers (Roberts, Tingley and Airoldi,
2016; Barnes and Hicks 2018).
In the Online Appendix, we validate the results of this approach. First, we conduct a
qualitative examination of relevant articles in our text corpora, providing examples of how
the two perspectives on the current account balance are represented in Australian and
German newspapers. Second, we confirm the results of the STM analysis by applying a
simpler and easier-to-reproduce dictionary-based approach. Third, in line with recent studies
using topic modeling to analyze the speeches given by central bankers (Moschella and Pinto
2018; Diessner and Lisi 2019; Cross and Greene 2019), we use STM to shed light on
Australian and German central bankers’ communication on current account imbalances. The
results show that the Deutsche Bundesbank’s discourse around the current account differs
substantially from that of the Reserve Bank of Australia, with the former adhering to the
competitiveness view and the latter promoting the investment perspective.
The joint application of this set of methods ensures that the different approaches
validate each other in showing that there are significant differences between Australia and
Germany in the political discourse around current account balances.
5.1 Structural Topic Model
In this section, we use a structural topic model (STM) to identify the presence of word
clusters in newspaper articles that are consistent with our theoretical framework and estimate
their relationship to document metadata. Topic models are increasingly used to systematically
investigate and interpret discourse in large collections of texts (Jacobs and Tschötschel
2019). As explained in detail by Roberts, Stewart and Tingley, STM allows researchers to
discover topics in a text corpus and conduct hypothesis testing about the relationship between
topics and document-level covariates (Roberts, Stewart and Tingley, 2019). Here, we focus
on estimating the proportion of text devoted to topics of interest both across newspapers and
over time. STM has the advantage that it can isolate word clusters that are related to the
competitiveness and investment perspectives, and separate them from other, potentially
confounding, topics. This ensures a more comprehensive analysis of our text corpus than
other methods, such as the dictionary-based approach.
First, we create two text corpora, one for Australian newspaper articles and one for
German ones, and convert text into a structured form, using standard text processing
approaches. In particular, we lowercase and stem words, remove stopwords and numbers, and
reduce the size of the document-frequency-matrix by considering only terms that appear in at
least 2% of the documents to improve estimation efficiency (Proksch and Slapin 2009).
Second, for each country, we run models iteratively and choose the number of topics based
on interpretability (Chang et al., 2009). In both cases, a model with 50 topics gives us a fine-
grained view over the issues addressed in the Australian and German media, and yields topics
that are theoretically meaningful. In Section A.1 of the Online Appendix, we justify the
selection of the number of topics in greater detail and show that this number constitutes a
balanced choice in terms of topic exclusivity and semantic coherence.
In both the Australian and the German case, we can identify three topics that are
highly relevant to our research question. In Table 3, we give an overview of such topics.
“Highest probability" is a simple measure that indicates which words are the most likely to
co-occur in a given word cluster. The FREX metric indicates “exclusive" words – namely,
those that are highly likely in one topic and unlikely in other topics (Airoldi and Bischof,
Table 3: Top Words for Australian and German Topics
invest, busi, small, survey, capit, plan, firm
busi, invest, survey, small, featur, firm, plan
market, investor, fund, bond, year, global, equiti
investor, equiti, bond, portfolio, asset, fund
reform, industri, competit, polici, australia, market, product
reform, competit, tariff, protect, effici, micro-econom, structure
unternehmen, industri, produkt, jahren, markt, entwicklung,
investitionen [company, industr, produc, years, market,
wettbewerb, bewertung, standort, bereich, produkt, unternehmen,
schweden [competition, valuation, location, sector, product,
deutschland, frankreich, euro, euro-zon, wettbewerbsfähigkeit,
spanien, mehr [germany, france, euro, euro area, competitiveness,
wettbewerbsfähigkeit, österreich, löhne, spanien, währungsunion,
griechenland, portug [competitiveness, austria, wages, spain,
monetary union, greece, portug]
anleg, fond, aktien, invest, investoren, manag, jahr [invest, fund,
stocks, invest, investors, manag, year]
fond, invest, anleg, manag, hielten, immobilien, market [fund,
invest, invest, manag, held, real estate, market]
In Australia, the first investment topic expresses an industrial perspective, as the
stemmed word “invest" appears in association with the terms “business", “firm", “small"
“capital", and “plan". The second Australian investment topic, instead, suggests a mere
financial perspective, being defined by the words “investor", “fund", “bond", “global",
“equity", “portfolio", and “asset". The third Australian topic of interest is the only topic
produced by the model that is defined by the word “competitiveness" (“competit" in the
stemmed form). This is associated with policy-relevant terms, such as “reform", “polici",
“market", “product", and “structur". Therefore, the first two considered Australian word
clusters can be easily ascribed to the investment perspective on the current account, while the
third one reflects a narrative of the current account that is much more in line with the
competitiveness perspective. This interpretation of the discussed topics is also confirmed by
text excerpts containing high proportions of these topics, presented in section A.2 of the
In Germany, the term “competitiveness" (“wettbewerbsfähigkeit") plays a more
prominent role, as it is one of the defining terms in two different word clusters. The first topic
sees “competit" (“wettbewerb") in association with business terms, such as “firms",
“industry", “development", “market", and “product" (“unternehmen", “industri",
“entwicklung", “markt", “produkt"). The word “investment" (“investitionen") also has high
probability to appear in this topic, but it is not among the words that are most exclusive to it.
We conclude that this is a topic about the competitiveness of domestic firms, and define this
topic as one about industrial competitiveness. The second German topic is characterized by
the term “competitiveness" in a more international perspective, as it is associated with
references to the Euro Area, as well as France, Spain, Greece and Portugal. “Wages"
(“löhne”) is another defining term of this word cluster, which further indicates that this is a
topic about international competitiveness. Finally, we can detect the presence of a topic that
can be unambiguously ascribed to investment – most notably, in financial terms – as it is
defined by “invest" (both with “anleg" and “invest"), “fund" (“fond"), and “stock" (“aktien").
Thus, we conclude that, for the German topic model, two word clusters are consistent with
the competitiveness perspective, while one is more clearly in line with the investment one.
Section A.2 in the Online Appendix shows relevant text excerpts from the STM analysis of
the German text corpus.
The presence of two investment topics and only one competitiveness topic in
Australia, as well as the presence of two competitiveness topics and only one related to
investment in Germany, is consistent with our theoretical framework. In addition, we directly
test this hypothesis by estimating the expected proportions of these topics over time.
Figures 3a and 3b present time series estimates showing the evolution of these topics
from the late 1980s to 2018 in, respectively, Australia and Germany. Figure 3a clearly shows
that the industrial investment topic always played a minor role in Australia. In contrast, the
estimated proportion of the competitiveness topic in Australia was relatively more prominent
in the late ’80s and early ’90s. However, discussions about financial investment in reference
to the current account have gained increasing importance over time: by the mid-90s they
started receiving more attention than issues of competitiveness, and this has consistently
remained so until nowadays. It may be no coincidence that this shift happened when the so-
called “Pitchford thesis" gained prominence (Belkar, Cockerell and Kent, 2007). This thesis,
put forward by the Australian economist John Pitchford, suggests that a current account
deficit is not necessarily a problem and can be optimal (Pitchford, 1989). This is the case if
rational individuals decide to borrow money from abroad and repay these loans with returns
from their investment.
The picture for Germany is the opposite. Figure 3b exhibits a consistent pattern of
prevalence of the two competitiveness topics vis-à-vis the financial investment one: with the
only exception of the years preceding the Global Financial Crisis, the two topics defined by
the term “wettbewerbsfähigkeit" are estimated to always have greater coverage than financial
investment. Also, it is important to notice how the German media’s view of competitiveness
has shifted over time from a more domestic perspective, expressed by the industrial
competitiveness topic, to a European one, expressed by the cluster of words ascribed to
international competitiveness. Our analysis suggests that the Euro Area crisis has greatly
contributed to shaping German sensitivity to competitiveness issues. This is consistent with
our expectations on the evolution of German political discourse from the Reunification
period to nowadays.
Figure 3a: Time Series Estimates of Relevant Topics for Australia, 1989-2018
Figure 3b: Time Series Estimates of Relevant Topics for Germany, 1987-2018
Overall, these results confirm the presence of two different narratives of the current
account balance in Australia and Germany. In Australia, the discourse around the current
account, here proxied by media reporting, tends to highlight issues of investment more than
in Germany. Importantly, financial investment is devoted most attention, which is consistent
1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Australian Expected Topic Proportion
Topic Competitiveness Financial Investment Industrial Investment
1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
German Expected Topic Proportion
Topic Industrial Competitiveness International Competitiveness Investment
with the investment perspective’s emphasis on international financial flows. Conversely, the
German media tend to highlight issues of competitiveness in relation to the current account
balance more than Australian ones. Hence, the results suggest that the discourse in Germany
tends to promote a neo-mercantilist view on the current account.
5.2 Additional Analyses
In the Online Appendix, we extend the text analysis to validate the results of STM
and provide additional evidence of the divergence of in the political discourse about the
current between Australia and Germany. First, we perform a qualitative examination based
on a close reading of articles in German and Australian newspapers that report about the
current account, covering the period for which all newspapers are available (2000-17).
Second, we use a dictionary-based approach to validate the results of this section and show
that there are significant differences in the interpretations of the current account in the
Australian and German media. Third, we use STM to provide evidence that the discourse of
key policymakers in the two countries is similarly biased towards different interpretations of
the current account. Analyzing the speeches of the Reserve Bank of Australia and the
Deutsche Bundesbank, we provide evidence that Australian central bankers refer extensively
to inflows of foreign capital when discussing the current account balance, while they do not
talk much about issues of competitiveness. By contrast, German central bankers focus much
more on competitiveness and pay little attention to the role of international capital flows.
While the analyses in this section and in the Online Appendix suggest that the
Australian and the German perspectives on the current account differ, they do not clarify
whether the documented divergence in discourse has a causal effect on individuals’
perceptions of macroeconomic issues related to the current account. After all, different types
of discourses could be the consequence rather than the cause of how individuals in different
countries conceive of the functioning of the economy. By employing an experimental
research design, the next section directly addresses this issue.
6 Survey Experiment
We conduct a survey experiment to determine whether the different types of discourse
on the current account have a causal impact on citizens’ policy preferences. A priori, one
could imagine that citizens are simply uninformed about current account imbalances, that
their opinions are determined by their cultural and media context, or that they are the result of
their personal economic situation, such as whether they work in an export-oriented industry.
An experiment allows us to evaluate how exposure to the competitiveness or investment
perspectives influences their opinions and isolate this effect from other spurious correlations
between these opinions and demographic, social, or economic differences.
In employing this experiment as a test of the theory that political discourse drives
external imbalances, we rely on a few key assumptions. First, we presume that the policies
adopted are determined in general by public opinion (democratic responsiveness). Secondly,
we are assuming that the publics to which policy is responsive are like the participants of our
study (external validity), in the sense that they read newspapers or consume other news that
contain the investment and competitiveness frames as we identified above in newspapers.
Finally, since such experimental effects have been shown to diminish over time, we expect
that such exposure is repeated and/or is more common in the lead-up to important policy
decisions (on the basis that journalists write about issues relevant to upcoming decisions).
In the experiment, participants are exposed to one of three scenarios. All three
scenarios present a basic explanation of current account balances, but (1) the “no framing”
condition presents no further interpretation of the imbalance. In the two experimental
treatments, participants read additional text that interprets the current account balances either
recurring to (2) the competitiveness or (3) the investment perspective. For details of the
question wording, see Figure 4, in which for simplicity we present the version for Australia
(the German wording is available in the Online Appendix). As Australia regularly runs
current account deficits, respondents were told that their country faces a deficit and,
following the possible treatment conditions, were asked whether they would support policies
to reduce the deficit. The experiment in Germany, which is a surplus country, presented the
opposite condition, in which the investment treatment frames the surplus as proof that
companies are reluctant to invest in the country, while the competitiveness treatment
interprets it as the results of high competitiveness and low production costs. In the German
survey question, the government proposes a “loosening” of economic policies to reduce the
surplus. To minimize the likelihood that responses are driven more by the nature of the
communicator rather than by the nature of the message, we decided to talk about the opinion
of generic “experts” rather than politicians or other actors in politicized policy domains. This
remains consistent with our theoretical framework. As Schmidt (2008: 310) notices, experts
are also often involved in processes of communicative discourse.
Figure 4: Wording of Vignette and Survey Question (Australia)
While the policy package that respondents evaluated is hypothetical, the subjects were
presented scenarios that reflected the situation of the country in which the survey was
conducted. That is, in Germany, respondents were informed that the country was running a
consistent current account surplus and considered policies that would reduce this surplus,
while in Australia they were informed about a deficit and policies to reduce it. One might
object that the descriptions should be identical in both countries, such as by assigning
participants in both countries to either a deficit or a surplus treatment as well as to the
treatments suggesting how these imbalances should be interpreted. However, such an
approach would simply add an artificial and unrealistic counterfactual, for example asking
Germans to believe their country runs deficits, or imagine a world in which it did so, while
reducing the power available to analyze the experimental condition that is of interest.
The survey experiment was conducted in Germany and Australia in the summer of
2018 with 2,043 respondents. Respondents came from the survey company Respondi’s
standing panels. The surveys took place between August 6 and August 26, 2018.
Respondents were screened to match the sex and age profile of each country based on census
data (for ages 18-65). The survey included questions for other political economy experiments
and the order of appearance of the experiments was randomized. The median respondent
required 18 minutes to respond to the full survey, so here we drop respondents who took less
than five minutes to respond, as it is practically impossible to respond meaningfully to the
questions in such a short period.
6.1 Survey Experiment Results
First we consider the success of the randomization, not because there is any particular
doubt about the survey firms’ computer randomization but because of the small attrition
created when respondents do not complete the survey or are dropped because they completed
the survey in less than five minutes. We present summary statistics and balance tests in the
Online Appendix, which confirm that the randomization was effective as expected with
respect to these covariates.
To examine the causal effect of the treatments on participants’ opinion, we first look
at the approval for the policy package to reduce the imbalance, that is to reduce the deficit in
Australia and to reduce the surplus in Germany. Figure 5 shows the raw distributions of
approval ratings for a package to address the current account balance, by country. We align
the responses for the two countries so that a five on the x-axis means strongly approving of a
policy package, while a one means rejecting it. The proposed policy package reduces the
surplus in the German case, while in Australia it reduces the deficit. First, comparing the
overall trends (ignoring treatments), we see that Germans are generally more reluctant to
approve a policy that would reduce their surplus than Australians are willing to approve a
package that would reduce their deficit. Although a fair amount of Germans responded
“approve” or “strongly approve” (4 or 5), more than 40% Australians did so. Second,
considering now the different treatment groups (presented in different shades of grey), it is
clear from these raw counts that the treatments caused people to move in the expected
directions. In Germany, the investment perspective convinced more people to approve a
policy package to reduce the surplus, while in Australia the competitiveness treatment caused
more people to support a package to reduce their deficit.
Figure 5: Histogram of Policy Approval Responses by Treatment, Country
We then perform a quantitative assessment of the causal effect of the two treatments
compared to the control group. Figure 6 presents the main results in graphical form. It shows
the ordered logit coefficient estimates of the two treatments for four specifications estimated
separately for respondents in Germany and Australia. The top (darkest) points of each quartet
are estimates from models with no controls; upper middle points are estimates from models
controlling for gender and age; lower middle points are estimates from models including
gender, age and a measure of the reliance on exports of the industry in which the respondent
; and bottom (lightest) points are estimates from models including gender, age,
reliance on exports, sophistication and political values
. 95% confidence intervals are
indicated with thin bars, whereas thick bars indicate 90% intervals. Higher values correspond
to greater approval of the policy package aimed at reducing the external imbalance (i.e.,
“loosening” policy package to reduce the surplus in Germany; “belt-tightening” policy
Approval of Policy Package
Approval of Policy Package
package to reduce the deficit in Australia). The complete results from the estimation of all
four models are presented in Table 5A in the Online Appendix.
Figure 6: Treatment Effects in Australia and Germany (Ordered Logistic Regressions)
Compared to the “Control” group, who was exposed to no framing, respondents in the
“Investment” treatment express greater approval towards policy measures aimed at reducing
the current account surplus in Germany. This effect is statistically significant at the 10% level
in three out of four specifications. By contrast, the magnitude of the “Competitiveness”
treatment is very small and never significantly different from zero in Germany. The opposite
is true for Australia. Here, compared to the “Control” group, respondents in the
“Competitiveness” treatment are significantly more likely to be in favor of policy measures
aimed at reducing the current account imbalance in three out of four specifications. The
causal effect of the “Investment” treatment on policy approval is small in magnitude and
never statistically significant. The point estimates are stable across specifications, consistent
with the randomization procedure working well across observables. Moreover, the difference
between the “Competitiveness” group and the “Investment” group is statistically significant
at the 5% level in all the specifications
In the Online Appendix, we verify the sensitivity of these results to a different model
specification. In Figure 3A and Table 6A, we replicate the results presented, respectively, in
Figure 6 and Table 5A, using OLS instead of ordered logistic regressions. Using a different
model specification does not substantially alter the main picture of results. The main
-0.2 0.0 0.2 0.4 -0.2 0.0 0.2 0.4
difference between OLS and ordered logit is that the “Investment” treatment in Australia is
estimated to have slightly lower statistical significance and does not attain the 10%
confidence level. However, the difference between the “Competitiveness” group and the
“Investment” group remains statistically significant at the 5% level in all the specifications in
both Germany and Australia, thereby pointing to a significant divergence in the causal effect
of the two treatments on respondents’ approval of policies aimed at reducing the current
Figure 7: Expected Values of Approval of Policy Package to Reduce Current Account
To better quantify the difference between the three groups, Figure 7 illustrates the
expected probabilities of each of the five response categories in Germany and Australia,
based on the estimates obtained from the first of the four specifications presented in Figure 6.
In Germany, the expected probabilities of the five categories are 5%, 21%, 50%, 23% and
2% for the “Competitiveness” group, very similar to the “Control” group. However, exposure
to discourse informed by the investment perspective makes respondents more supportive
towards the reduction of the German current account surplus: the two highest categories of
support for a “loosening” policy package receive, respectively, an expected 28% and 3% of
responses from subjects in the “Investment” group. The opposite is true in Australia, where
respondents in the “Competitiveness” group are estimated to “approve” and “strongly
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
approve” a “belt-tightening” policy package with, respectively, a 34% and 9% probability,
while the corresponding probabilities for those in the “Investment” group are 30% and 7%.
These results provide support for Hypothesis 2, suggesting that the discourse around
the current account balance can influence opinion on policies that help determine that
balance. Although the exposure to different interpretations is limited in our experiment, the
findings provide evidence that citizens’ opinions are not fixed but respond to dominant
interpretations of the current account. In particular, individuals respond more strongly to the
framing that is not dominant in the political discourse in their country, namely the
competitiveness view in Australia and the investment perspective in Germany. This result
can be seen as a consequence of the systematic divergence of discourse between the two
countries that was uncovered in the previous section. Since citizens are consistently and
permanently exposed to the competitiveness view in Germany and the investment perspective
in Australia, it should come as no surprise that they do not significantly update their attitudes
towards the reduction of their country’s current account imbalance compared to the
respondents in the control group. At the same time, our analysis suggests counterfactually
that, if citizens were widely exposed to different interpretations of the current account
balance, their opinions on policies to adjust these imbalances would also change. Thus, the
findings of this paper challenge the idea that societal support towards macroeconomic
imbalances in these two countries is to be taken for granted and could not be reversed if
different strategies of communicative discourse around the current account became dominant
in countries with external imbalances.
This paper sheds light on popular support for external imbalances and the policies that
generate them by examining the policy ideas conveyed to citizens via political discourse. Our
comparative analysis of Australia, a notorious deficit country, and Germany, a notorious
surplus country, reveals important differences. Australian newspapers and policy actors tend
to discuss the current account balance in the context of capital flows and view the deficit as
proof that their country is a highly popular investment destination. By contrast, German
newspapers and policy actors discuss the current account predominantly in the context of
trade. They consider their country’s surplus to be an expression of superior competitiveness
and successful economic policy.
These differences are compatible with two distinct theoretical perspectives on
external balances, which represent the international dimension of different national growth
models. The investment perspective stresses the importance of capital flows that are the result
of rationally acting firms and households. The competitiveness perspective stresses the role
of wages and competitiveness. Our paper presents a systematic analysis of how different
forms of discourse about the current account generate popular support for external
imbalances and the underlying economic strategies that generate these imbalances. Political
debates, e.g. about the imbalances in the Euro Area, have repeatedly pointed to the presence
of such distinct theoretical lenses in different countries (Brunnermeier, James and Landau,
2016; Jones, 2016). Our analysis allows us to explore this claim in a larger context, beyond
the politicized debates surrounding the Euro Area crisis.
The results show how political discourse helps secure support for diverging national
growth strategies and the domestic economic institutions that back these strategies (Hall and
Soskice, 2001; Baccaro and Pontusson, 2016; Manger and Sattler, 2020). Although
institutional complementarities are important determinants of economic policies, these
arrangements must be supported by a broad societal coalition in order to be durable. We
show that the presence of distinct types of discourse about the current account balance helps
generate societal support for the policies and outcomes that follow from domestic economic
institutions. Discourse, therefore, has a stabilizing effect on a country’s macroeconomic
regime, or growth model. At the same time, this also suggests that growth models are by no
means unchangeable. Popular support for a growth model can change over time if the
discourse of influential political actors changes, together with underlying policy ideas.
Our findings indicate that resolving international economic imbalances might be as
much about communication as it is about economics, and points to the need to better
understand the sources and determinants of different types of discourse on the current
The experimental results suggest that viewing economic imbalances simply as the
result of different policies falls short of a satisfying explanation. Even assuming that such
policies are welfare maximizing and optimal given countries’ different factor endowments, it
is important to understand how they are maintained, given that, as our results show, citizens’
support for them depends in part on the interpretation they are offered. While beyond the
scope of this paper, further research is needed on the question what or who is driving such
differences in discourse upstream. One promising approach would be to check whether
imbalanced economies with structures that differ from those of Australia and Germany
display similar political discourses. In the absence of such studies, our paper makes no claim
to generalizability, but rather serves as a starting point that underlines the importance of
seeking out more information about these processes.
Airoldi, E. M. and Bischof, J. M. (2016) ‘Improving and Evaluating Topic Models and Other
Models of Text’, Journal of the American Statistical Association, 111, 1381-1403.
Amable, B. and Palombarini, S. (2009) ‘A Neorealist Approach to Institutional Change and
the Diversity of Capitalism’, Socio-Economic Review, 7, 123-143.
Ardanaz, M., Murillo, M. V. and Pinto, P. M. (2013) ‘Sensitivity to Issue Framing on Trade
Policy Preferences: Evidence from a Survey Experiment’, International Organization,
Armingeon, K., and Baccaro, L. (2015) “The Crisis and Germany. The Trading State
Unleashed”, in Complex Democracy, edited by Volker Schneider and Burkard Eberlein,
165–183, Springer International Publishing.
Autor, D. H., Dorn, D., and Hanson, G. H. (2013) ‘The China Syndrome: Local Labor
Market Effects of Import Competition in the United States’, American Economic
Review, 103, 2121-2168.
Baccaro, L. and Benassi, C. (2017). ‘Throwing Out the Ballast: Growth Models and the
Liberalization of German Industrial Relations’, Socio-Economic Review, 15, 85–115.
Baccaro, L. and Pontusson, J. (2016) ‘Rethinking Comparative Political Economy: The
Growth Model Perspective’, Politics & Society, 44, 75–207.
Baldwin, R. and Giavazzi, F. (2015) The Euro Area Crisis. A Consensus View of the Causes
and a Few Possible Remedies, London, Centre for Economic Policy Research.
Barnes, L. and Hicks, T. (2018) ‘Making Austerity Popular: The Media and Mass Attitudes
Towards Fiscal Policy’, American Journal of Political Science, 62, 340–354.
Belkar, R., Cockerell, L. and Kent, C. (2007) Current Account Deficits: The Australian
Debate. Reserve Bank of Australia Research Discussion Paper 2007/02, accessed at
http://www.rba.gov.au/publications/rdp/2007/pdf/rdp2007-02.pdf on January 10, 2019.
Bernanke, B. S. (2005) ‘The Global Saving Glut and the U.S. Current Account Deficit’,
January 11, 2019.
Bernanke, B. (2015) “Germany’s Trade Surplus is a Problem”, accessed at
problem/ on 27 March 2020.
Best, J. (2004) ‘Hollowing Out Keynesian Norms: How the Search for a Technical Fix
Undermined the Bretton Woods Regime’, Review of International Studies, 30, 383–404.
Blyth, M. (2002) Great Transformations: Economic Ideas and Institutional Change in the
Twentieth Century, Cambridge, Cambridge University Press.
Blyth, M. and Matthijs, M. (2017) ‘Black Swans, Lame Ducks, and the Mystery of IPE’s
Missing Macroeconomy’, Review of International Political Economy, 24, 203–231.
Bonatti, L. and Fracasso A. (2013) ‘The German Model and the European Crisis’, Journal of
Common Market Studies, 51, 1023–39.
Borio, C. (2016) ‘On the Centrality of the Current Account in International Economics’,
Journal of International Money and Finance, 68, 266–274.
Brender, A. and Pisani, F. (2010) Global Imbalances and the Collapse of Globalised
Finance, Brussels, Centre for European Policy Studies (CEPS) Paperbacks.
Brunnermeier, M. K., Harold, J. and Landau, J-P. (2016) The Euro and the Battle of Ideas,
Princeton, NJ, Princeton University Press.
Bundesministerium für Wirtschaft und Energie (2019) Fakten zum deutschen Außenhandel,
deuschen-aussenhandel.html on February 10, 2020.
Burden, B. and oanberg, J. (2003) ‘Budget rhetoric in presidential campaigns from 1952 to
2000’, Political Behavior, 25, 97–118.
Cerrato, A., Ferrara, F. M., and Ruggieri, F. (2018) ‘Why Does Import Competition Favor
Republicans?’, paper presented at the 2018 APSA Annual Conference in Boston,
available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3147169.
Chang, J., Gerrish, S., Wang, C., Boyd-Graber, J. L. and Blei, D. M. (2009) ‘Reading Tea
Leaves: How Humans Interpret Topic Models’, Advances in Neural Information
Processing Systems, 22, 288–296.
Chong, D. and Druckman, J. N. (2007) ‘Framing Theory’, Annual Review of Political
Science, 10, 103–126.
Chwieroth, J. M. (2007) ‘Testing and Measuring the Role of Ideas: The Case of
Neoliberalism in the International Monetary Fund’, International Studies Quarterly, 51,
Cross, J. and Greene, D. (2019) ‘Talk Is Not Cheap: Policy Agendas, Information Processing,
and the Unusually Proportional Nature of European Central Bank Communications
Policy Responses.’ Governance, 33, 425-444.
De Boef, S. and Kellstedt, P.M. (2004) ‘The Political (and Economic) Origins of Consumer
Confidence’, American Journal of Political Science, 48, 633–649.
De Grauwe, P. (2011) The Governance of a Fragile Euro Area, CEPS Working Paper No.
den Haan, W., Ellison, M., Ilzetzki, E., McMahon, M. and Reis, R. (2016) ‘The Danger of
Germany's Current Account Surpluses: Results of the CFM and CEPR Survey’, accessed
cepr-survey on 27 March 2020.
Diessner, S. and Lisi, G. (2019) ‘Masters of the ‘Masters of the Universe’? Monetary, Fiscal
and Financial Dominance in the Euro Area.’ Socio-Economic Review, 18, 315-335.
Dieter, H. (2018) Stubbornly Germany First. Options for Reducing the World’s Largest
Current Account Surplus, Berlin, SWP Comment No. 48.
European Commission (2016) Country Report Germany 2016 - Including an In-Depth
Review on the Prevention and Correction of Macroeconomic Imbalances, Commission
Staff Working Document, accessed at
https://ec.europa.eu/info/sites/info/files/cr2016_germany_en.pdf on January 10, 2019.
Fan, D. P. (1988) Predictions of Public Opinion from the Mass Media: Computer Content
Analysis and Mathematical Modeling, Westport, Greenwood Press.
Ferrara, F. M. (2020) ‘The Battle of Ideas on the Euro Crisis: Evidence from ECB Inter-
Meeting Speeches’, Journal of European Public Policy, 27, 1463-1486.
Ferrara, F. M. and Sattler, T. (2018) ‘The Political Economy of Financial Markets.’ Oxford
Research Encyclopedia of Politics, July. DOI:
Financial Times (1988) ‘Fifty Years of the D-Mark: Currency Has Become a Symbol of
National Pride’, London, p. 7 (June 24).
Flassbeck, H. and Lapavitsas, C. (2013) The Systemic Crisis of the Euro – True Causes and
Effective Therapies, Rosa Luxemburg Stiftung Studien, accessed at
and-effective-therapies/ on January 08, 2019.
Frieden, J. A. (1991) ‘Invested Interests: The Politics of National Economic Policies in a
World of Global Finance’, International Organization, 45, 425–451.
Gourinchas, P-O. and Rey, H. (2005) From World Banker to World Venture Capitalist: US
External Adjustment and the Exorbitant Privilege, NBER Working Paper 11563,
accessed at http://www.nber.org/papers/w11563 on January 16, 2019.
Haffert, L. (2019) ‘Permanent Budget Surpluses as a Fiscal Regime’, Socio-Economic
Review, 17, 1043–1063.
Hall, P. A. and Soskice, D. W. (2001) Varieties of Capitalism: The Institutional Foundations
of Comparative Advantage, Oxford, Oxford University Press.
Hall, P. A. and Franzese, R. J. (1998) ‘Mixed Signals: Central Bank Independence,
Coordinated Wage Bargaining, and European Monetary Union’, International
Organization, 52, 505–535.
Hancké, B. (2013) Unions, Central Banks, and EMU: Labour Market Institutions and
Monetary Integration in Europe, Oxford, Oxford University Press.
Haas P. M. (1992) ‘Introduction: epistemic communities and international policy
coordination’, International Organization, 46, 1-35.
Harell, A., Soroka, S. N. and Iyengar, S. (2016) ‘Race, prejudice and attitudes toward
redistribution: A comparative experimental approach’, European Journal of Political
Research, 55, 723–744.
Hay, C. and Rosamond, B. (2002) ‘Globalization, European Integration and the Discursive
Construction of Economic Imperatives’, Journal of European Public Policy, 9, 147–167.
Helgadóttir, O. (2016) ‘The Bocconi boys go to Brussels: Italian economic ideas,
professional networks and European austerity’, Journal of European Public Policy, 23,
Helleiner, E. and Kirshner, J. (2009) The Future of the Dollar, Ithaca, Cornell University
Hübscher, E. (2018). The Clientelistic Turn in Welfare State Policy-Making: Party Politics in
Times of Austerity. ECPR Press.
Hübscher, E., Kemmerling, A. and Sattler, T. (2015) ‘Austerity for the Win? The Effect of
Fiscal Consolidation on Political Support for the Government’, paper presented at the
Annual Meeting of the European Political Science Association, June 25-27, Vienna.
Hübscher, E., Sattler, T. and Truchlewski, Z. (2021) ‘Voter Attitudes towards Fiscal Trade-
Offs: Evidence from Three European Countries”, paper presented at the Conference on
“The Politics of Macroeconomic Policies,’ Max Planck Institute for the Study of
Societies (MPIfG), January 14-15, available at: 10.13140/RG.2.2.31085.20962.
Hübscher, E., Sattler, T. and Wagner, M. (2020) ‘Voter Responses to Fiscal Austerity’,
British Journal of Political Science, online first, DOI:
Hume, D. (1752) Political Discourses, Edinburgh, Printed for A. Kincaid, and A. Donaldson.
International Monetary Fund (2016) Germany Article IV Consultation, IMF Country Report
16/202, accessed at https://www.imf.org/external/pubs/ft/scr/2016/cr16202.pdf on
January 18, 2019.
IWD (2020) Viele Jobs hängen am Export, accessed at https://www.iwd.de/artikel/viele-jobs-
haengen-am-export-459120/ on March 12, 2020.
Jacobs, T. and Tschötschel, R. (2019) ‘Topic Models Meet Discourse Analysis: A
Quantitative Tool for a Qualitative Approach.’ International Journal of Social Research
Methodology, 22, 469–485.
Jacoby, W. G. (2000) ‘Issue Framing and Public Opinion on Government Spending’,
American Journal of Political Science, 44, 750–767.
Johnston, A. (2016) From Convergence to Crisis: Labour Markets and the Instability of the
Euro, Ithaca, NY, Cornell University Press.
Johnston, A., Hancké, B. and Pant, S. (2014) ‘Comparative Institutional Advantage in the
European Sovereign Debt Crisis’, Comparative Political Studies, 47, 1771–1800.
Jones, E. (2009) ‘Shifting the Focus: The New Political Economy of Global Macroeconomic
Imbalances’, SAIS Review, 29, 61–73.
Jones, E. (2011) ‘Macroeconomic Imbalances and the Sovereign Debt Crisis’. In Hubner, K.
(ed) Europe, Canada, and the Comprehensive Economic and Trade Agreement, London,
Jones, E. (2015) ‘Getting the Story Right: How You Should Choose between Different
Interpretations of the European Crisis (And Why You Should Care)’, Journal of
European Integration, 37, 817–832.
Jones, E. (2016) ‘Competitiveness and the European Financial Crisis’. In Caporaso, J. A. and
Rhodes, M. (eds) The Political and Economic Dynamics of the Euro Area Crisis,
Oxford, New York, Oxford University Press.
Kayser, M. A. and Leininger, A. (2015) ‘Vintage Errors: Do Real-Time Economic Data
Improve Election Forecasts?’, Research & Politics, 2, 1–11.
Keck M.E. and Sikkink K. (1998) Activists Beyond Borders: Advocacy Networks in
International Politics. Ithaca, NY, Cornell University Press.
Kollmann, R., Ratto, M., Roeger, W., in’t Veld, J. and Vogel, L. (2014) What drives the
German current account? And how does it affect other EU member states?, DG ECFIN
Economic Papers No. 516.
Krugman, P. (2016) ‘Trade Deficits: These Times Are Different’, Paul Krugman Blog,
accessed at https://krugman.blogs.nytimes.com/2016/03/28/trade-deficits-these-times-
are-different/ on January 25, 2019.
Lee, O. (2009) ‘Devaluation: Ready, Steady, Devalue’, accessed at
on January 25, 2019.
Manger, M. and Sattler, T. (2020) ‘The Origins of Persistent Current Account Imbalances in
the post-Bretton Woods Era’, Comparative Political Studies, 53, 631-664
Matthijs, M. and McNamara, K. (2015) ‘The Euro Crisis’ Theory Effect: Northern Saints,
Southern Sinners, and the Demise of the Eurobond’, Journal of European Integration,
McCombs, M. and Reynolds, A. (2009) ‘How the News Shapes Our Civic Agenda’. In
Jennings B., Zillmann, D. and Oliver, M. B. (eds) Media Effects, London, Routledge,
McNamara, K. (1999) ‘Consensus and Constraint: Ideas and Capital Mobility in European
Monetary Integration’, Journal of Common Market Studies, 37, 455–476.
Micossi, S., D’Onofrio, A. and Peirce, F. (2018) On German External Imbalances, CEPS
Policy Insights No. 2018/13.
Moschella, M. (2014) ‘Monitoring Macroeconomic Imbalances: Is EU Surveillance More
Effective than IMF Surveillance?’, Journal of Common Market Studies, 52, 1273–1289.
Moschella, M., and Pinto, L. (2018) ‘Central Banks’ Communication as Reputation
Management: How the Fed Talks Under Uncertainty’ Public Administration, 97, 513-
Mun, T. (1986 ) England’s Treasure by Forraign Trade, Reprints of economic classics
Fairfield, NJ, A.M. Kelley.
Obstfeld, M. (2012) ‘Does the Current Account Still Matter?’, American Economic Review,
Obstfeld, M. and Rogoff, K. S. (1996) Foundations of International Macroeconomics,
Cambridge, MA, MIT Press.
Petty, R., Priester, J.R. and Briñol, P. (2009) ‘Mass Media Attitude Change: Implications on
the Elaboration Likelihood Model of Persuasion’. In Bryant, J., Zillmann, D. and Oliver,
M. B. (eds) Media Effects. Advances in Theory and Research, London, Routledge,
Pitchford, J. D. 1989. ‘A Sceptical View of Australia’s Current Account and Debt Problem’,
Australian Economic Review, 22, 5–14.
Proksch, S.-O. and Slapin, J. B. (2009) ‘How to Avoid Pitfalls in Statistical Analysis of
Political Texts: the Case of Germany’, German Politics, 18 , 323-344.
Redeker, N. and Walter, S. (2020) ‘We’d Rather Pay than Change the Politics of German
Non-Adjustment in the Euro Area Crisis’, The Review of International Organizations,
Regan, A. (2017) ‘The Imbalance of Capitalisms in the Euro Area: Can the North and South
of Europe Converge?’, Comparative European Politics, 15, 969-990.
Rho, S. and Tomz, S. (2017) ‘Why Don’t Trade Preferences Reflect Economic Self-Interest?’
International Organization, 71, 85–108.
Roberts, M. E., Stewart, B. M. and Tingley, D. (2019) ‘Stm: R Package for Structural Topic
Models’, Journal of Statistical Software, 91, 1-40.
Roberts, M. E., Tingley, D. and Airoldi, E. M. (2016) ‘A Model of Text for Experimentation
in the Social Sciences’ Journal of the American Statistical Association, 111, 988–1003.
Rodrik, D. (2009) ‘The Real Exchange Rate and Economic Growth’, Brookings Papers on
Economic Activity, 2008, 365–412.
Schmidt, V. A. (2002) ‘Does Discourse Matter in the Politics of Welfare State Adjustment?’,
Comparative Political Studies, 35, 168–193.
Schmidt, V. A. (2008) ‘Discoursive Institutionalism: The Explanatory Power of Ideas and
Discourse’, Annual Review of Political Science, 11, 303–326.
Schmidt, V. A. (2010) ‘Taking Ideas and Discourse Seriously: Explaining Change through
Discursive Institutionalism as the Fourth ‘New Institutionalism’’. European Political
Science Review, 2, 1-25.
Sinn, H-W. (2014) ‘Austerity, Growth and Inflation: Remarks on the Euro Area’s Unresolved
Competitiveness Problem’, The World Economy, 37, 1–13.
Smith, A. (2003 ) An Inquiry into the Nature and Causes of the Wealth of Nations,
New York, Bantam Classics.
The Economist (2016, October 27) ‘Aussie Rules: What Australia Can Teach the World’,
Viner, J. (1948) ‘Power versus Plenty as Objectives of Foreign Policy in the Seventeenth and
Eighteenth Centuries’, World Politics, 1, 1–29.
Walter, S. (2013) Financial Crises and the Politics of Macroeconomic Adjustments,
Cambridge, Cambridge University Press.
Walter, S., Ray, A., and Redeker, N. (2020) The Politics of Bad Options: Why the Euro
Area's Problems Have Been So Hard to Resolve, Oxford, Oxford University Press.
Young, L., and Soroka, S. (2012) ‘Affective news: The automated coding of sentiment in
political texts’, Political Communication, 29, 205-231.
Some identify global imbalances as an important cause of the Global Financial Crisis of
2007-08 (Brender and Pisani, 2010). Others suggest that imbalances within the Euro
Area have been a crucial trigger of the Euro Area crisis (Baldwin and Giavazzi, 2015).
US dissatisfaction over the large current account surpluses in other countries has been
steadily growing during the past years and even turned into a major political conflict
after the 2016 US presidential election.
These, among others, include fiscal policy, changes to depreciation rules, the value-
added tax, or the government’s ability to influence wages via the legal framework.
Mercantilists recommend that countries run an external surplus by exporting more than
they import. This strategy leads to an accumulation of foreign assets, which is seen as an
effective strategy to increase a country’s wealth.
For better readability, we will refer to the two perspectives simply as ‘competitiveness
perspective’ and ‘investment perspective’ in the remainder of the text.
As the example of trade policy shows, citizens find it difficult to assess the trade-offs
that are associated with international economic flows (Rho and Tomz, 2017).
Intervening variables such as an individual’s attitude and personal environment can
mitigate the impact of media reporting on opinions (Petty, Priester and Briñol, 2009).
People may choose to consume news that confirm their pre-existing beliefs and reject
information that does not fit into their worldview (for an overview of the debate, see
Barnes and Hicks, 2018). However, at least in the long term, media reporting can be
expected to have a long-term effect on how the current account balance is interpreted.
The current account data comes from the IMF Balance of Payments Statistics.
The Sydney Morning Herald has a regional focus, but in order to ensure ideological
variation, we have decided to include it nevertheless. The Age, the fourth quality
newspaper and only alternative, is also regional.
For Die Welt, The Australian and the Sydney Morning Herald, we used LexisNexis. For
the Süddeutsche Zeitung and the Australian Financial Review we resorted to Factiva.
Finally, for the Handelsblatt, we retrieved data directly from the official website.
“% exported” looks at the industry in which the respondent works, and reports what
share of that industry is based on exports (based on SITC codes).
Sophistication is measured by the main dimension of a principal component analysis of
five sophistication questions. The left-right scale is a composite score of four questions
on social policies reduced to one dimension using principal component analysis. The
questions ask about support for (1) redistribution of wealth (2) state-ownership of public
services and industries, (3) whether government should take responsibility to provide for
individuals, and (4) whether people can only get rich at the expense of others.
The only exception is the first specification (i.e. the one without controls) of the ordered
logistic model in Australia, where the difference between the estimated coefficients of
the two treatments is statistically significant at the 10% level.
E.g., the relative power of social coalitions. See, e.g., Haffert (2019).