iiinova — where your digital self keeps evolving

Reposted by Jonas Frey

Randi Hjalmarsson @profhjalmarsson.bsky.social · 02/10/2026

The Economics department at the University of Gothenburg is hiring two tenure track Assistant Professors - one in Applied Micro (especially labor, health, education, inequality, and crime) one in Environmental Economics. Spread the word! econjobmarket.org/positions/12747

econjobmarket.org

EJM - Econ Job Market

102

Jonas Frey @jfrey.bsky.social · 01/10/2026

What else would you expect from a College that is not even a hundred years old. Give them a couple of hundred years more.

000

Jonas Frey @jfrey.bsky.social · 28/09/2026

In "What Analysts Miss: Predicting Consensus Forecast Errors from Earnings Calls" we fine-tune the model to predict analyst earnings forecast errors using earnings calls and show that these predicted errors forecast returns at the next earnings announcement. papers.ssrn.com/sol3/papers.... (3/3)

papers.ssrn.com

What Analysts Miss: Predicting Consensus Forecast Errors from Earnings Calls

We fine-tune an open language model on two decades of earnings-call transcripts to predict the consensus analyst forecast error, evaluating it on 2024 and 2025

000

Jonas Frey @jfrey.bsky.social · 28/09/2026

In "The Return Impact of Form 10-K Releases: Evidence from a Large Language Model" we fine-tune the model to predict returns after the release of Form 10-K filings and find evidence of a delayed reaction to the filings' information. papers.ssrn.com/sol3/papers.... (2/3)

papers.ssrn.com

The Return Impact of Form 10-K Releases: Evidence from a Large Language Model

Using the text of Form 10-K filings, we fine-tune a large language model to explain contemporaneous announcement-window returns and to predict post-filing retur

100

Jonas Frey @jfrey.bsky.social · 28/09/2026

I have two new working papers on how market participants react to text-based information. Both fine-tune Llama 3.1 8B (an open LLM trained on data up to 2023) to predict financial variables, and we evaluate the model on data from 2024–2025 to ensure that the analysis is out-of-sample. (1/3)

100

Reposted by Jonas Frey

Department of Economics, University of Gothenburg @gothenburg-econ.bsky.social · 02/09/2026

📢 We are hiring an Assistant Professor in Economics! The Department of Economics at the University of Gothenburg is recruiting an Assistant Professor as part of a major international initiative. 📅 Application deadline: 30 October 2026 👉 Read more and apply: web103.reachmee.com/ext/I005/103...

004

Jonas Frey @jfrey.bsky.social · 26/01/2026

I had this issue before and if I recall correctly, asking Claude to calculate the relevant coordinates using python before making the chart fixed it

010

Jonas Frey @jfrey.bsky.social · 31/10/2025

The solution to this is using open-source/weights models, which will remain available indefinitely. I feel like there should be a norm that researchers need to use open models at least as a robustness check, but I am wondering what others think about that. (2/2) #EconSky

010

Jonas Frey @jfrey.bsky.social · 31/10/2025

I recently realised that using proprietary LLMs like OpenAIs GPT models to analyse text data is a disaster for replication. OpenAI discontinues its models after 2-3 years, after which replication will be impossible. This is less than what it takes to publish a paper. (1/2) #EconSky

110

Jonas Frey @jfrey.bsky.social · 06/10/2025

If they still manage to be admitted despite optimizing less, they were stronger candidates to begin with. The way to get a perfect score on the GRE verbal is to memorize around 1000 obscure English words. There are no additional skills required, and I doubt this correlates with writing ability.

000

Jonas Frey @jfrey.bsky.social · 06/10/2025

My intuition would be that the only reason GRE verbal predicts anything is that everyone who knows how the admission process is supposed to work ignores verbal. Therefore, the candidates with the high verbal scores are those who optimize less. (I/II)

100

Jonas Frey @jfrey.bsky.social · 19/08/2025

Great to hear that you had a good time. I went to the middle school equivalent of a "Schhülerakademie" (Juniorakademie) more than 15 years ago, and it was a life-changing experience. I have not found the time to teach at a Schülerakademie yet, but it is on my bucket list.

110

Reposted by Jonas Frey

Department of Economics, University of Gothenburg @gothenburg-econ.bsky.social · 04/06/2025

Lasse Pedersen (Copenhagen Business School) gave a keynote lecture at the CFF conference on Asset Pricing and Machine Learning. The conference is organized by @gothenburg-econ.bsky.social and funded by Andra AP-fonden and Handelsbanken.

043

Jonas Frey @jfrey.bsky.social · 24/03/2025

I will run a workshop on how to code experiments with oTree on April 29th & May 6th, 2025. The suggested workshop fee is 200€, but you can pay less or attend for free. Signup: forms.gle/6oe5qXmHTK1M... More information: jonasfrey.net/workshops #EconSky

forms.gle

oTree Workshop 29 April 2025 & 6 May 2025

Prerequisites: none Fee: €200, fee reduction / waiver possible if needed Time: 09:00 to 18:00 CET (UTC+1) including breaks The time from 09:00 to 11:00 on the 29 th will be used for a Python / HTML...

021

Reposted by Jonas Frey

Department of Economics, University of Gothenburg @gothenburg-econ.bsky.social · 20/11/2024

We're very excited to introduce three candidates on the #econjobmarket this year! Meet Vorada Limjaroenrat, Jimmy Karlsson, and Jens Ewald, and read their job market papers here: gu.se/en/school-bu... #econsky #econjmp #econ #jmp @jensewald.bsky.social @jimmykarlsson.bsky.social

Three job market candidates at the Department of Economics at the University of Gothenburg in Sweden.

0208

Jonas Frey @jfrey.bsky.social · 05/10/2023

Do you know if candididates with a background in finance would be considered for these positions?

100