TL;DR
A Swiss National Bank working paper by Severin Joho and Enzo Rossi examines 1,063 speeches by senior SNB officials from 1928 through 2024. It reports that speeches became more frequent, shorter and more accessible, while inflation gained prominence in policy discussion during the flexible-exchange-rate era.
Severin Joho and Enzo Rossi have published a Swiss National Bank working paper analyzing 1,063 speeches by high-ranking SNB officials from 1928 to 2024. The study finds that annual speech frequency rose over time, speeches grew shorter and language became more accessible, while inflation took a more prominent place in policy discussion in the era of flexible exchange rates.
The paper, titled “The evolution of frequency, language complexity and policy narrative: benchmarking Python-based NLP and ChatGPT on a century of SNB rhetoric,” draws on a dataset the authors describe as novel and comprehensive. It examines how often officials spoke, the length and readability of their speeches, and changes in their thematic content.
According to the report, annual speech frequency has followed an upward trajectory since the mid-1950s, with a more marked acceleration after Switzerland moved to flexible exchange rates. The authors also identify a shift toward shorter speeches and increasingly accessible language beginning in the 1960s.
The thematic analysis finds that inflation moved to the forefront of SNB rhetoric in the flexible-exchange-rate era. By comparison, cyclical economic development and unemployment received less emphasis. The authors also compare traditional, Python-based text-mining methods with ChatGPT, discussing how the two approaches can complement each other and where language-model analysis presents risks.
How SNB Communication Changed
The findings offer a long-run view of how a central bank communicates as its policy environment changes. More frequent public speeches can give audiences additional material to assess, while shifts in length and readability affect how accessible that material may be to the public and specialists. The study reports these patterns across nearly a century; it does not establish that any one change caused another.
The shift in topics also matters for interpreting institutional rhetoric. The authors’ account links inflation’s greater prominence to the flexible-exchange-rate era, while noting reduced emphasis on cyclical developments and unemployment. Tracking such changes can help researchers study which economic concerns feature in public communication, although the paper’s reported patterns do not by themselves show how audiences understood or responded to the speeches.
The methodological comparison has relevance beyond this dataset. Researchers increasingly use generative AI to process large collections of text. The paper argues that language models and conventional text-analysis frameworks have complementary roles, while warning that algorithmic pitfalls and open research questions remain.
Python natural language processing books for text analysis
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A Century of Central Bank Speeches
The study covers speeches delivered from 1928 through 2024 by high-ranking officials of the Swiss National Bank. Its timeline includes the move to flexible exchange rates, which the authors identify as a notable point in the development of both speech frequency and policy themes.
Rather than focusing only on what officials said, the researchers also examine features such as speech length and readability. They benchmark Python-based natural-language-processing methods against ChatGPT, a comparison intended to test how computational approaches can be used to analyze historical institutional communication. The source summary does not specify the individual speeches, the detailed scoring methods, or the exact size of changes in each measure.
„The frequency of annual speeches has followed an upward trajectory since the mid-1950s, accelerating notably after the transition to flexible exchange rates.“
— Severin Joho and Enzo Rossi, authors of the SNB working paper
text mining and computational linguistics research books
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits of the Reported Findings
The source summary does not provide numerical estimates for how much speech frequency, length or readability changed, nor does it detail the measures used to assess accessibility. It also does not identify specific speeches or show whether the findings hold across different categories of speakers or audiences.
The paper presents broad associations across time, including a change after the move to flexible exchange rates, but the summary does not establish that the exchange-rate regime caused the change in speaking patterns or subject matter. The authors also flag algorithmic pitfalls and open questions in applying large language models, but the available description does not enumerate those issues or provide a full account of ChatGPT’s performance against the Python methods.
central banking and monetary policy books Switzerland
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Further Testing of Text Analysis
The working paper outlines methodological lessons and identifies open questions for future research into the use of large language models for central-bank communication. Further work could test the approaches on other institutions or speech collections and examine how results change with different models and analytical choices. The supplied report does not announce a specific follow-up study or schedule.
ChatGPT generative AI for academic research guide
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What does the new SNB paper examine?
It analyzes the frequency, length, readability and themes of 1,063 speeches by high-ranking Swiss National Bank officials delivered from 1928 to 2024.
What changes did the authors report?
The authors report that speeches became more frequent, particularly after the mid-1950s and the move to flexible exchange rates; they also became shorter and more accessible over time.
How did the themes of SNB speeches shift?
According to the paper, inflation gained prominence during the flexible-exchange-rate era, while cyclical economic development and unemployment received less emphasis.
Why did the researchers compare Python methods with ChatGPT?
The comparison was intended to assess approaches for analyzing a large historical text collection. The authors describe traditional Python-based methods and generative AI as complementary, while noting algorithmic pitfalls and unresolved questions.
Does the study prove that exchange-rate changes caused the shifts?
No causal conclusion is stated in the supplied summary. It reports changes in speech patterns and themes around the flexible-exchange-rate era but does not establish that the policy regime caused them.
Source: primary
