Johannes Dahlke

Asst Professor Digital Innovation @UTwente | CSO @istari_ai | Research Fellow @KOFETH, @ETH_AI_Center | Economics of AI, human x AI, digital transitions.

[Johannes Dahlke @joukoda]

From 🤓🆚🤖 to 🤓🤝🦾:

It is on us to “find out how we can meaningfully enrich #jobs using #AI instead of sacrificing them for short-​term efficiency gains.”

See my full interview on how AI is changing jobs below. Available in 🇩🇪/🇬🇧/🇫🇷/🇮🇹⬇️

[KOF Swiss Economic Institute @KOFETH_en]

Almost exactly a year ago, #artificial #intelligence got a big boost with the release of #ChatGPT. @joukoda categorises the opportunities, challenges and risks of this technology in the latest KOF Bulletin. https://t.co/0E3B7tXF3l

[Johannes Dahlke @joukoda]

How can we separate signal from noise when working with web data to understand socio-economic change?

We discussed all of this at our 3rd annual @istari_ai research workshop with more than 20 of our international researcher partners. 🧙‍♂️🧙

[Johannes Dahlke @joukoda]

Retracing (rewriting?) economic history with the help of machine learning. ⬇️

[Philipp Koch @philippmkoch]

📢New paper in PNAS!

How rich was Vienna at the time of Mozart or Tuscany at the time of Michelangelo? Historical GDPs per capita are scarce, leaving this unanswered.

Here, we provide new estimates using machine learning to augment historical GDPs per capita. /1 🧵

[Johannes Dahlke @joukoda]

@tdienlin https://t.co/mEI6DU5EaY

[Johannes Dahlke @joukoda]

@tdienlin Yes, but the terms imply what we want to evoke from subjects. The term informant implies that the goal is not met by having people participate in an interview, but by having them provide new information. I find this particularly important when teaching students.

[César A. Hidalgo @cesifoti]

Eleven years ago, I quit an abusive academic relationship with my former co-author Ricardo Hausmann, presently a Professor of Practice at the Harvard Kennedy School. This abuse has now returned in a new act of plagiarism. Today, I am pushing back.

A few months ago, I learned about two new working papers that Hausmann and his team posted at the World Intellectual Property Organization (WIPO) [1,2]. The first one claims to introduce a “multidimensional approach” to economic complexity by combining data on trade, patents, and publications, and concludes that these metrics are complementary when explaining economic growth. Yet, this is exactly what my co-authors and I did in a 2023 peer-reviewed journal paper [3] that Hausmann and his co-authors were perfectly aware of (as I document below). Instead of giving proper acknowledgement to our work, however, Hausmann and his team hide it in a footnote on page 18 clouded in a factually incorrect statement. The footnote says that “other studies have developed similar complexity metrics under different names.” That is not true. Our 2023 paper uses the same name as Hausmann et al.’s 2024 “new” working paper, namely, the “Economic Complexity Index (ECI).”

Now, it is reasonable to wonder if this is a misunderstanding or oversight. People sometimes forget to cite others for innocent reasons. But there is clear evidence to the contrary in two other working papers published almost simultaneously with the first one. One of them, is another WIPO working paper by his coauthors but without Hausmann [4]. This paper does properly acknowledge the existence of our work in multidimensional economic complexity in the main text: “Following […] Stojkoski et al. (2023).” So, the team clearly knew about the work. The third working paper [2], also at WIPO, includes Hausmann as a first author making a call for the use of multidimensional economic complexity methods in innovation policy. Here, the only example of multidimensional work cited by Hausmann et al. is his own 2024 working paper [1], this time completely ignoring the 2023 peer-reviewed paper he knew about (“In Hausmann et al. (2024), we measure this for each country in different dimensions (e.g., trade and patents).”). Unsurprisingly, I had made a similar call for multidimensional expansion of the field in a paper that was also published after peer-review in 2023 (Hidalgo CA, Research Policy, 2023 [5]) and that had been available as a pre-print for two years. In that paper, I cite our original multidimensional economic complexity contribution and also two other papers that had introduced multidimensional approaches to relatedness, another key concept in the field, which Hausmann does not cite either.

In my view, these three documents provide clear evidence of an attempt by Hausmann to misappropriate an idea published by a former co-author while knowing of that work. This, violates Harvard’s honor code [6] which requires the “accurate attribution of sources” and considers “plagiarizing or misrepresenting the ideas or language of someone else as one’s own” as a violation of its community standards. In the first working paper [1], he copies the idea while hiding the attribution in a factually incorrect footnote. Simultaneously, he releases a second working paper citing this unpublished work as the only example of prior art [2]. The third paper [4], the one by his co-authors but not him, shows that the team knows about the work and behaves differently when the senior member of the team (Hausmann) is not an author.
My co-authors and I were dismayed by Hausmann’s attempt to misappropriate our ideas. Why bother doing original work and getting it past peer-review (which for both of our papers was a lengthy process), if authors with a platform such as Harvard University take your ideas and pass them as their own? But this is not the first time.
In 2019, Hausmann and his team announced a “first-of-its-kind” and “unique” feature in their trade data visualization website (“Atlas Online”): the “country profiles.” Yet, my group had supported country profiles in the Observatory of Economic Complexity (OEC) since 2012, several months before Hausmann and I parted ways in 2013 (see full story document for details [7]). The reason that the profiles disappeared in Hausmann’s copy of the Observatory in 2013 was that they were deleted by his team, probably because they did not understand that we had conceived these pages for search engine optimization.

In short, the original OEC that I designed with my group and placed in the public domain had country profiles in 2012 and was redesigned in 2015 to make the country and product profiles the main part of the site. Today, https://t.co/2d5FR0ZgGZ [8] serves over 100,000 profiles, a design concept we use prominently when constructing other data distribution platforms (e.g. Data USA, Data Mexico, etc). In fact, in 2016 I published a “how to” guide in Scientific American for this kind of work in 2016 [9]. But when Hausmann and his team at Harvard finally built country profiles in 2019, seven years after we did, they claimed a “revolutionary first-of-its-kind” and “unique” feature, in their lab’s own website [10] and in the Harvard Gazette [11]

So why share this now?

I have three reasons.

First, I broke off my relationship with Hausmann in 2013 after several unpleasant experiences, which included having to fight for the authorship rights of junior authors (which Hausmann wanted to exclude from our joint book: The Atlas of Economic Complexity) and being exploited when we were equal partners in a company that I had registered for us to do joint consulting work (see full story document [7]). From there on, I just wanted to forget about him. The tacit deal was that, if he stopped the abuse, I could put this in the past. But Hausmann has repeatedly broken this tacit deal. His 2024 working paper research agenda seems to involve putting forward, without proper acknowledgement, ideas my team and I published after peer-review in 2023 and made available online in 2022. It has been more than a decade. This has to stop.

The second reason is that because Hausmann has done this before, I am concerned that if I don’t push back, he will do it again. Right now, my research group and I have a great thing going on. Should we worry that in 2026 Hausmann will try to pass off our 2024 and 2025 papers as his own?

The third reason is that this egregious behavior is supported by a power asymmetry that is highly profitable for Hausmann thanks to the Harvard platform. Soon after we split ways in 2013, Hausmann sold this [12] embarrassingly bad economic data visualization website to the Mexican Government as part of a project rumored to be about USD 5,000,000 (paid using an undisclosed and untraceable escrow account (“fideicomiso”)). The online platform was so sloppy that some government officials started to call it “El Mamarracho” (a derogatory term in Spanish for extremely shoddy work). I learned about this in 2019, after a new team entered Mexico’s economic ministry and invited us to build Data Mexico [13], a properly built open data portal integrating dozens of datasets in thousands of profiles for a fraction of the cost of the “Mamarracho.”

Hausmann has engaged in commercial endeavors with many governments (e.g. Albania, Azerbaijan, Colombia, Mexico, etc.). So, in my opinion, the target audience for these misappropriated working papers are not other scholars, but the officials who purchase economic development contracts from him and take the word of the Harvard Professor of Practice at face value.

This is what makes misappropriating the multidimensional economic complexity idea important. A strong motivation for any consulting work in this space is the fact that trade-based measures of economic complexity explain future economic growth. This was established in a paper Hausmann and I published together in 2009 [14], based on work we did together while I was a PhD student at the University of Notre Dame. But the 2023 multidimensional economic complexity paper made that idea obsolete. On the one hand, it simultaneously considers multiple expressions of complexity (trade, patents, and research), and on the other hand, it looks at multiple outcomes (growth, inequality, and emissions). Giving advice based on the latest developments in the field requires using a multidimensional framework.

Finally, I want to say that I have nothing against Hausmann’s co-authors, some of whom I know personally and some of whom I don’t. I understand that they are in a situation in which they may depend financially and professionally on Hausmann. His first attempt to spin my publicizing these facts may be to argue that someone of his caliber does not have the time to look at petty things like footnotes and references (pinning the responsibility on them). I don’t think that’s fair.

I also want to say that I do not consider Ricardo’s behavior necessarily representative of Harvard University or of the World Intellectual Property Organization. I have met many people at those organizations and continue to collaborate with some of them. They are outstanding scholars

I have decided to go public with this statement after consulting with several senior colleagues and lawyers specialized in disputes in higher education.

I am absolutely distraught by having to deal with this situation eleven-years after I quit that abusive relationship. But he needs to cut it out.

[Johannes Dahlke @joukoda]

Humbled to share that my dissertation was awarded with the @Suedwestmetall prize in support of junior researchers making societal impact. Thank you. 🙏

This prize means a lot to me because it shows that our research is acknowledged beyond purely academic circles.

[Johannes Dahlke @joukoda]

@cesifoti @LKarbevska 🚀

[Johannes Dahlke @joukoda]

Jetzt schon meine Entdeckung des Jahres: Osnabrück Hbf als Umstiegsbahnhof.

Ansehnlich, übersichtlich, sauber, belästigungsfrei, gut versorgt. Und: Zugänge zu Gleisen auf zwei Ebenen für Umsteiger gut verbunden. Top Bahnhof für Reisende nach NL!

[Johannes Dahlke @joukoda]

Wie viel zusätzlichen Umsatz Fastfood Ketten in und um Bahnhöfen wohl durch Verspätungen der Deutschen Bahn machen? #positiveexternalities

[Johannes Dahlke @joukoda]

Tell me you work in econ without telling me you work in econ.

[Johannes Dahlke @joukoda]

@cesifoti I am so happy to see this. Félicitations!

[Dongmiao Zhang @DongmiaoZhang]

Interesting paper on a novel "AI exposure" that explains the slower roll-out of AI from a cost perspective. One of the takeaways for me is that automation is not necessarily economically attractive - perhaps more efforts could be devoted to use AI to complement workers.

[Johannes Dahlke @joukoda]

But for real: I’m noticing intentional effort to avoid ChatGPT phrasing in my own writing. A first sign of emphasized humanness.

[Johannes Dahlke @joukoda]

ChatGPT ruined the word “intricate” for me.

My students, however, now seem to discover so many intricacies, it’s crazy! 😬🚩

[Johannes Dahlke @joukoda]

Meine Theorie ist ja, dass Weselsky mit der DB gemeinsame Sache macht, damit unter dem Deckmantel des Streiks Weichen unbemerkt repariert werden können.

[Johannes Dahlke @joukoda]

@TBroekel Newspapers may play this role—and/or more dangerously: Influencers as authentically perceived individuals may play it. This would be more in line with current trends of anti-establishment and personality cults.

[Pierre-Alexandre Balland @pa_balland]

Carless cities ==> life

[Lauren Waardenburg @Lauren_wbg]

We are back!! On 16-17 May we will host the second edition of the "AI beyond the hype" workshop. This time at the campus of @essec in Paris! Mark your calendars and submit your best abstracts before Feb. 29. We hope to see you soon in Paris! @cristina_alaimo and @lauren_wbg

[Balázs Lengyel @blengyelb]

The next Geography of Innovation Conference will be in Budapest in 2026. We are thrilled to organize the event with @uni_corvinus and @KRTK_HU. The community of @AnetiLabs and @LearningCCL is looking forward to welcome you in this beautiful city! #geoinno2026