In many Germanic languages, the word for ‘debt’ also means ‘guilt’. This column explores whether the linguistic overlap leads to greater debt aversion in Germany, the Netherlands, Sweden, and Switzerland. Using a guilt-laden term reduces willingness to borrow, lowers approval for government debt, and influences firms’ financing plans and politicians’ framing of fiscal policy. While support for public debt goes down with use of the laden term, the effect disappears when the purpose of the debt is clear. Linguistic framing may help guide borrowing toward socially optimal levels.
In many Germanic languages, the commonly used word for “debt” also means “guilt”, which carries a strong moral connotation. Commentators have long suspected that this linguistic overlap may lead to greater debt aversion. Former French finance minister Bruno Le Maire invoked this idea to explain Germany’s fiscal conservatism: “In Germany, debt is ‘die Schuld’. It’s both the financial debt and the fault. There’s a kind of guilt in debt.” The Economist (2015) similarly observed that “this mental link helps explain Germany’s peculiar concern with borrowing money”.
Philosophers and linguists have long argued that language does more than describe reality – it shapes how individuals perceive and interpret it, including in economic choices (Sapir 1929, Whorf 1956). Yet, debt aversion is also shaped by many other forces, including economic conditions, institutional structures, and cultural norms (Almenberg et al. 2021, Martínez-Marquina and Shi 2024). Whether language itself influences how people think about borrowing or merely reflects deeper cultural values has remained an open question.
In a new paper (Aksoy et al. 2025), we provide the first causal evidence that language affects economic attitudes and behaviour toward debt. Drawing on large-scale cross-country surveys, incentivised experiments, and text analysis, we show that the moral connotations embedded in debt-related language are powerful. They reduce individuals’ willingness to borrow, lower approval for government debt, influence firms’ stated financing plans and advertising strategies, and shape how politicians frame fiscal policy.
The experiments
To identify the causal effect of language on attitudes and behaviour, individuals must be randomly exposed to different words, something not feasible with observational data. We therefore conducted four large-scale survey experiments – three in household surveys and one in a firm survey – in seven countries. In four of them (Germany, the Netherlands, Sweden, and the German-speaking part of Switzerland), the term for debt inherently carries a sense of guilt; for example, the German word Schulden shares its root with Schuld (“guilt”). In these countries, respondents were randomly assigned to one of two otherwise identical questions about debt: one using the standard, morally loaded term and the other a neutral synonym. This design isolates the causal effect of moral connotations within a shared cultural and institutional context. The remaining three English-speaking countries (Australia, the UK, and the US) serve as a placebo group, where wording should have no effect in the absence of such moral associations.
We also tested whether linguistic framing influences actual financial behaviour rather than attitudes alone. In a borrowing-to-invest experiment in Switzerland, participants could borrow up to CHF100 at a 5% interest rate and invest it in an opportunity yielding a safe return of 10%. Borrowing the full available amount was therefore the profit-maximising choice. We randomised the description of the borrowing opportunity, again varying only the terminology for debt. This way, differences in behaviour between the treatment and control groups capture the causal impact of words alone.
Can a single word affect how people think about debt?
Our results show that using a guilt-laden term for ‘debt’ significantly reduces willingness to incur debt among households and business managers, reduces support for public borrowing, and lowers actual borrowing in incentivised settings. Figure 1 summarises the main findings from our survey experiments.
Figure 1 Effects of guilt-connoted wording on the willingness to incur household debt, public debt attitudes, and firms’ debt management strategies
Notes: The figure shows ordinary least squares (OLS) estimates from survey experiments regressing different outcome variables on a treatment indicator equal to one for guilt-connoted wording. Top-left: Surveys in Germany (DE), the Netherlands (NL), and Sweden (SE) on willingness to borrow for consumption, and in Germany on general willingness to borrow, for consumption, and for real estate. ‘Borrowed to Invest’: whether debt was incurred in an incentivised borrowing-to-invest decision in Switzerland (CH). Top-right: Survey on German firms’ plans for debt financing and the perceived importance of quick repayment. Bottom: Support for government borrowing to finance general spending, social transfers, or infrastructure, and voting intentions in hypothetical Swiss referenda on joint EU debt and the Fiscal Budget Act.
Household debt
We find that simply using a guilt-laden term for debt reduced the willingness to borrow for consumption (e.g. a mobile phone or a car) by about seven percentage points (an 18% decline relative to the control group that received the neutral wording). By contrast, we find no effect in English-speaking countries, where the word ‘debt’ carries no moral connotation.
In a separate German survey, we expanded the scope of questions to include willingness to borrow in general, for consumption, and for real estate. Guilt-framing reduced overall willingness to borrow by 12 percentage points (20% of the control mean). The effect was again significant for consumption borrowing, but not for mortgages. This suggests that moral framing is particularly powerful in domains perceived as financially imprudent, such as consumer debt, while mortgages are viewed as a socially acceptable form of debt.
Similar patterns emerged in the incentivised borrowing-to-invest experiment in Switzerland. In this experiment, participants made real financial decisions. Each participant could take on debt to invest, with a guaranteed, risk-free return that exceeded the cost of borrowing. Even though borrowing was financially optimal, participants confronted with guilt-laden language were nine percentage points less likely to borrow at all and borrowed 23% less on average. In other words, language connoting guilt discouraged people from taking an action that would have earned them more money. Words alone made people leave money on the table.
Public debt
Turning to public debt, we find that support for government borrowing is approximately five percentage points (11% in relative terms) lower when the survey question refers to public debt using a guilt-connoted term. Notably, this negative framing effect disappears when the purpose of the debt is explicitly mentioned (for example, when respondents are informed that the debt will finance a specific investment, such as social transfers or infrastructure).
This pattern suggests that while moral language can sway general attitudes against public debt, a concrete context can mitigate the semantic bias. In Switzerland, where citizens regularly vote on fiscal referenda, we observed a similar pattern: support for reforms that increased public debt without specifying its purpose was significantly lower when described using a guilt-connoted term.
Business debt
Finally, we extended the survey design to firms. In the ifo Business Survey of German managers, exposure to guilt-connoted terms reduced the likelihood that managers planned to expand debt financing and increased the likelihood that they saw rapid repayment as essential.
Strategic use of language
We also examined whether politicians deliberately employ morally charged language when debating public debt in the German Parliament. To this end, we assembled a dataset covering more than 30 years of Bundestag speeches. We identified speeches referring to public debt and determined the speaker’s stance toward borrowing by combining manual annotations with automated text classification. We then constructed a measure of guilt slant, defined as the frequency of guilt-connoted debt terms per 10,000 words.
Figure 2 plots the average intensity of guilt-related language in speeches against and in favour of public debt (lines), alongside the share of speeches in favour of public debt (bars). Politicians arguing against public borrowing – advocating fiscal consolidation or stricter budget rules – use guilt-laden terms, such as Schulden, significantly more often than those promoting expansionary fiscal policy.
Figure 2 Evolution of guilt slant in parliamentary speeches concerning debt
Notes: The green line (left y-axis) shows the intensity of using guilt-connoted terms in speeches supporting public debt, while the red line (left y-axis) shows the same measure for speeches opposing public debt. Intensity measures the frequency of guilt-connoted words in a speech, scaled per 10,000 words. The grey bars (right y-axis) represent the percentage of speeches in favour of public debt in each electoral term.
This divergence becomes particularly pronounced in the second half of the sample period, when speeches defending higher levels of borrowing become more frequent. This shift occurred against the backdrop of historically low interest rates and Germany’s declining debt-to-GDP ratio until 2019, following the introduction of the constitutional debt brake in 2016. Even after controlling for time-invariant characteristics of individual politicians, such as ideology, changes in stance are systematically associated with changes in guilt slant.
Implications
Our findings have several practical implications. First, they suggest that cross-country differences in debt levels may, in part, be rooted in linguistic framing, adding a new dimension to the determinants of fiscal and financial behaviour. Second, the findings highlight the potential for carefully chosen language to influence borrowing decisions in contexts where debt aversion may be either beneficial, by reducing over-indebtedness, or detrimental, by constraining productive investment. Understanding and potentially harnessing these linguistic effects could inform communication strategies for policymakers, regulators, and financial institutions seeking to guide borrowing toward socially optimal levels.
Source : VOXeu
