Language as economic infrastructure

Transaction costs and network effects

Economists came to language comparatively late. One of the first to treat it as an economic object was Jacob Marschak, whose article “Economics of Language” of 1965 is regarded as the starting point of the field.¹ The insight that a shared language saves costs is simple, but its consequences are far-reaching. Every transaction between parties who do not share a language requires translation, interpretation or at least a greater effort of comprehension. Each of these carries costs and creates a risk of misunderstanding that must be priced into the transaction. Ronald Coase’s concept of transaction costs, the costs of searching for partners, negotiating, drafting and enforcing contracts, applies to language with particular force, because language is the medium in which all of these activities take place.²

A second strand of theory concerns network effects. Just as a telephone is worth more when more people own one, a language gains value with every additional speaker. Economists describe goods of this kind as network goods whose value to the individual user increases with the number of other users.³ Language exhibits the property in an extreme form. The value of learning English for a Zurich bank employee depends on how many clients, counterparties, regulators and colleagues also use English. Each additional user raises the value for all others. Network effects explain why a single language tends to become dominant in an international field once a critical mass has been reached, as English did in finance after the Second World War. The same mechanism explains why smaller languages lose ground in international business even when their speakers are wealthy and well educated.

The concept of infrastructure brings the two strands together. Roads, railways and electricity grids are infrastructure because they are shared, because they lower the costs of economic activity for all users and because they require continuous investment. Language has all three properties. Shared by definition, it lowers the costs of every transaction conducted within it and must be maintained through schooling, dictionaries, terminology databases, translation services and the drafting practice of legislators and courts. The difference from physical infrastructure lies in ownership. Nobody owns a language, so that the maintenance of linguistic infrastructure falls partly to the state, partly to professional communities such as the legal profession, and partly to the market, which rewards individuals who invest in the right language skills.

The Swiss research on the value of multilingualism

Switzerland has become one of the main laboratories of the economics of language, largely through the work of François Grin at the University of Geneva. In 2008 Grin, together with François Vaillancourt of the University of Montreal and Claudio Sfreddo, presented the results of a project on foreign languages in professional life conducted within the National Research Programme 56 on linguistic diversity. The team surveyed around 2,500 residents and 205 companies in the French- and German-speaking parts of the country. Its central finding was that the language skills of the Swiss population contributed about nine to ten per cent of gross domestic product, an amount of roughly CHF 46 billion. The study was presented as the first attempt to calculate the linguistic value of a national economy. It also observed that directors and purchasing staff had better language skills than sales personnel and that larger firms had stronger English skills than smaller ones.⁴

The results were developed further in a monograph published in 2010, which proposed a systematic approach to identifying and measuring the relationship between language skills and economic performance and tested the model with data from two countries.⁵ The central argument of this research is that multilingualism is productive in two ways. Individuals earn more when they master languages in demand, which shows up in wage differentials. Firms, in addition, perform better when their workforce as a whole covers the languages of their markets, because they can serve clients, negotiate with suppliers and comply with foreign rules without buying these services from outside. The second effect is the more important for a financial centre, because banks, insurers and law firms sell products whose value lies to a great extent in communication.

Three conclusions for Zurich follow from this research. The first is that English and the national languages are complements in the Swiss labour market. The Zurich private banker who speaks German, English and French is more valuable than one who speaks only English, because each language opens a different client segment. The second conclusion is that linguistic skills are unevenly distributed within firms, so that the linguistic capacity of a financial centre depends on the concentration of multilingual people in the functions that matter, such as client advice, legal and compliance, and senior management. The third conclusion is that the linguistic infrastructure of a country is a public good whose value is created collectively through the school system and through immigration, which is why debates about the order in which foreign languages are taught in Swiss schools have a direct bearing on the financial centre.

The specific demands of a financial centre

Financial services are more language-intensive than most other industries. A manufacturer can export machines whose functioning does not depend on the language of the operating manual. A bank sells promises, contracts and advice, all of which exist only in language. The legal documentation of a syndicated loan, the prospectus for a bond issue, the investment policy of a discretionary mandate and the suitability assessment required by financial services law are texts whose precise wording determines the rights and obligations of the parties. A mistranslation can shift liability, create an unintended guarantee or render a clause unenforceable. The linguistic infrastructure of a financial centre therefore has two levels. The first is the general level of everyday communication, where a good command of a common language suffices. The second is the level of legal and technical precision, where specialised terminology, drafting conventions and knowledge of the legal system behind the words are required.

At the second level, language and law cannot be separated. A legal term names a concept that exists within a particular legal system. Its translation into another language is only exact where the target system has an equivalent concept. The Swiss Verwaltungsrat and the German Aufsichtsrat are both translated into English as “board”, yet they perform different functions within different governance models. The Swiss Betreibung has no exact equivalent in German law, still less in English law. For a financial centre that serves clients from several legal systems, the ability to move between these concepts is part of the infrastructure. It is maintained by lawyers, translators and terminologists whose work rarely attracts public attention but on whose accuracy the reliability of the entire centre depends.

Notes

1. Jacob Marschak, “Economics of Language”, Behavioral Science 10 (1965). On its role as “a reference starting point for this branch of economic science”: New Economic School (Moscow), “How Can Language Policy Affect Economics?”, https://guru.nes.ru/en/kak-yazyikovaya-politika-mozhet-povliyat-na-ekonomiku.html. ↩
2. Ronald H. Coase, “The Nature of the Firm”, Economica 4 (1937), no. 16, 386–405, as summarised in Wikipedia, “The Nature of the Firm”, https://en.wikipedia.org/wiki/The_Nature_of_the_Firm, which lists search and information costs, bargaining costs and policing and enforcement costs as the hidden costs of using the market. ↩
3. Michael L. Katz and Carl Shapiro, “Network Externalities, Competition, and Compatibility”, American Economic Review 75 (1985), no. 3, 424–440, cited in Wikipedia, “Network effect”, https://en.wikipedia.org/wiki/Network_effect, which gives the telephone as a classic example of a network good. ↩
4. SWI swissinfo.ch, Simon Bradley, “Languages generate one tenth of Swiss GDP”, 20 November 2008, https://www.swissinfo.ch/eng/languages-generate-one-tenth-of-swiss-gdp/7050488. The article names the project “Langues étrangères dans l’activité professionnelle” (LEAP) and quotes the finding that people in organisations “easily operate with three, four or five languages”. ↩
5. LINGUIST List 21.2345, announcement of François Grin, Claudio Sfreddo and François Vaillancourt, The Economics of the Multilingual Workplace (Routledge, 2010), https://linguistlist.org/issues/21/2345/. ↩

For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of September 2025.