Trade in English only? The case of Swedish in Finland

Technical Summary: "Trade in English only? The case of Swedish in Finland"

Problem Statement
The paper investigates whether a common language facilitates international trade even between two highly developed, culturally similar nations (Finland and Sweden) where English proficiency is exceptionally high. While English serves as a lingua franca in both countries, Finland is officially bilingual, with approximately 5.5% of the population speaking Swedish as a first language. The authors question whether trade between these nations is conducted frictionlessly in English or if the presence of Swedish-speaking personnel within Finnish firms provides a distinct advantage in exporting to Sweden. This inquiry is particularly relevant given the shift in Finnish export markets away from Russia toward Nordic neighbors following geopolitical changes.

Methodology
The study utilizes a comprehensive firm-level panel dataset covering the entire universe of Finnish firms from 2010 to 2019. The data is constructed by linking administrative registers from Statistics Finland (providing employer-employee information, including mother tongue) with firm-level goods export data from Finnish Customs.

• Variables: The dependent variable is the log of annual export volume to specific destinations (Sweden and China). The key independent variable is the share of Swedish-speaking personnel within a firm.
• Control Variables: The model controls for firm location (latitude and longitude of the municipality center) to account for geographical distance and proximity to the Swedish border, as well as total firm exports and year dummies.
• Econometric Approach: The authors employ a gravity-style model.
  1. Pooled OLS: Initial regressions are run without firm fixed effects.
  2. Fixed Effects (FE): To address unobserved, time-invariant firm heterogeneity and move closer to causal interpretation, the authors estimate models with firm fixed effects.
  3. Subgroup Analysis: Regressions are stratified by firm size (Micro: 5–9 employees; Small: 10–49; Medium: 50–249; Large: 50+).
  4. Robustness: The authors attempted instrumental variable (IV) regressions using municipal demographic shares as instruments but noted limitations due to low within-municipality variation in demographics.

Key Results

1. Positive Correlation with Sweden: There is a statistically significant positive relationship between the share of Swedish-speaking employees and export volumes to Sweden. In the fixed-effects models, a higher share of Swedish-speaking personnel correlates with increased exports to Sweden, even after controlling for firm-specific time-invariant characteristics.
2. Firm Size Heterogeneity: The positive effect of the common language is most pronounced for smaller firms. In the fixed-effects specification, the relationship is statistically significant primarily for "Micro" firms (5–9 employees) and the aggregate "All" category, but not consistently for medium or large firms.
3. No Effect on China: When the same analysis is applied to exports to China, the share of Swedish-speaking personnel shows no positive relationship; in some OLS specifications, the coefficient is negative and significant. This suggests the observed effect is specific to the linguistic and cultural link between Finland and Sweden, rather than a general firm characteristic.
4. Geographical Factors: In pooled OLS models, firms located further west and south (closer to Sweden) export more to Sweden. However, in fixed-effects models, geographical coordinates lose significance, likely due to the lack of within-firm variation in location over time.

Key Contributions

• Micro-level Evidence: Unlike previous gravity model studies that operate at the country level, this paper provides firm-level evidence on the role of language in trade.
• The "English-Only" Counter-Argument: The study challenges the assumption that high English proficiency renders minority languages irrelevant for trade. It demonstrates that even in a context where English is widely spoken, a shared native language (Swedish) still confers a trade advantage.
• Small Firm Dynamics: The paper highlights that language barriers may be more binding for small and medium-sized enterprises (SMEs) than for large corporations, possibly due to the reliance on personal relationships and direct communication in smaller firms.
• Regional Trade Implications: The findings suggest that minority language speakers within a country (Swedish speakers in Finland) may possess a comparative advantage in trading with neighboring countries sharing that language, supporting arguments for the economic value of minority language rights.

Significance and Limitations
The authors conclude that while they cannot fully rule out cultural proximity as a confounding factor, the specific effect on exports to Sweden (and not China) strongly points to language as a distinct driver of trade volume. The paper argues that for small firms, the common language facilitates trade more effectively than English, even between highly English-proficient nations.

The authors modestly note several limitations:

• Causality: While firm fixed effects control for time-invariant heterogeneity, the precise causal mechanism remains partially uncovered.
• Technological Context: The data predates the widespread integration of Artificial Intelligence (AI) in business, which the authors suggest could potentially diminish the need for local languages in future trade.
• Policy Implications: The results raise questions for language training policies in small countries, suggesting that investing in languages other than English (e.g., French, German) may still hold economic value for international trade, particularly for smaller enterprises.