TL;DR
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A Swiss National Bank working paper by Severin Joho and Enzo Rossi examines 1,063 speeches by senior SNB officials from 1928 to 2024. It reports more frequent speeches, shorter and more accessible language, and a stronger focus on inflation in the flexible exchange-rate era; the authors also compare conventional Python text analysis with ChatGPT.
Severin Joho and Enzo Rossi have published a Swiss National Bank working paper analyzing 1,063 speeches by senior SNB officials from 1928 through 2024. The study traces changes in how often officials spoke, how long and accessible their language became, and which policy subjects received attention, while comparing Python-based text analysis with ChatGPT.
The paper reports that the annual number of speeches has risen since the mid-1950s, with the increase accelerating after Switzerland moved to flexible exchange rates. It also finds a long-term shift toward shorter speeches and language that has become more accessible since the 1960s. The source summary does not provide the size of these changes or a year-by-year count, so the findings are qualitative rather than a basis for calculating a specific rate of increase.
The researchers also identify a change in the topics emphasized. In the flexible exchange-rate era, inflation moved to the forefront of SNB speech themes, while cyclical economic development and unemployment received less emphasis. These are findings about the content of speeches, not direct measurements of policy outcomes or proof that the bank’s priorities changed in the same way in every period.
Alongside the historical analysis, Joho and Rossi compare traditional, deterministic Python methods with ChatGPT, a generative AI system, for analyzing central bank communication. The authors describe the approaches as complementary and discuss algorithmic pitfalls and open research questions. The provided summary does not specify which model version was tested or give detailed performance scores.
How SNB Communication Changed
The speech record offers a way to examine how a central bank communicates as its economic and monetary setting changes. The reported increase in speech frequency, movement toward brevity and accessibility, and changing attention to inflation provide historical evidence about the public language of the SNB across nearly a century. For readers, the subject matters because central bank speeches are a channel through which officials explain economic conditions and policy concerns.
The methodological comparison also has relevance beyond this particular archive. Researchers increasingly use language models to process large collections of documents, but the paper cautions that generative AI should not simply replace established methods. Its emphasis on combining LLM analysis with Python frameworks points to a need to check model outputs and account for their limitations, especially when studying long-term changes in language.
Python natural language processing books text analysis
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A Century of SNB Speeches
The dataset spans 1928 to 2024 and contains speeches by high-ranking officials of the Swiss National Bank. The authors say they assessed the material’s quantity, length, readability and thematic content. The study’s long time span allows the researchers to compare communication across different monetary arrangements, including the period after the transition to flexible exchange rates.
The paper is a working paper published through the SNB’s research publications. Its stated focus is both substantive and methodological: it reports patterns in the bank’s rhetoric and examines how conventional text-mining tools and a generative model can be used to study that rhetoric. The available source summary gives the headline findings but not the full statistical detail or examples from individual speeches.
“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, in the working paper’s summary
natural language processing with Python NLTK book
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Summary Leaves Open
The source summary does not provide statistical estimates, detailed charts or uncertainty ranges for the reported trends. It also does not explain how the researchers defined readability, brevity or thematic prominence, or provide a detailed account of how speeches were selected and processed. Those details would be needed to assess the methods and the strength of individual comparisons.
The summary identifies algorithmic pitfalls and open questions but does not list them individually. It gives no model version, prompt design, validation results or direct comparison scores for ChatGPT and the Python methods. It is therefore not possible from the supplied information to determine how closely the two approaches agreed, or which tasks each performed best.
The findings describe patterns in speeches and their themes. The summary does not establish why those patterns occurred or show that language changes caused shifts in policy. It also does not say whether the researchers’ dataset or analysis will be updated after 2024.
large language models for text analysis research book
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Questions for Further NLP Research
The paper points to open questions about using large language models to analyze central bank communication. Further work could clarify how model choices and prompts affect results, how AI-generated classifications can be checked against reproducible methods, and where conventional text processing is more reliable. The source does not announce a specific follow-up study or release schedule.
For readers seeking a fuller assessment of the findings, the next step is the complete working paper, which may provide the methods, examples and quantitative results absent from the short summary. Until those details are examined, the reported trends should be understood as the authors’ conclusions from their analysis of the historical speech collection.
computational text analysis software Python
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did the SNB working paper examine?
It analyzed 1,063 speeches delivered by high-ranking Swiss National Bank officials between 1928 and 2024, looking at frequency, length, readability and themes.
What changes did the researchers report?
The authors report that speeches became more frequent from the mid-1950s, with faster growth after the move to flexible exchange rates. They also found a shift toward shorter, more accessible language.
How did the themes of speeches change?
The study says inflation received greater prominence in the flexible exchange-rate era, while cyclical economic development and unemployment received less emphasis.
Why did the researchers compare Python methods with ChatGPT?
They compared conventional Python-based text mining with a generative AI approach to study central bank language. The authors describe the methods as complementary and discuss limitations and open questions in applying language models.
Does the summary show how accurate ChatGPT was?
No. The supplied summary does not include model versions, validation results or performance scores, so it does not establish how accurately ChatGPT analyzed the speeches.
Source: primary
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
