Abhishek Nagaraj 🗺️
@abhishekn
Followers
11K
Following
30K
Media
1K
Statuses
7K
Associate Prof @berkeleyhaas, NBER RA. Dad to 2 girls. he/him/his. Interested in using tech, AI and data to drive progress. Using ideas as my maps.
Joined May 2008
Excited to post a new working paper (with @stellayjia) that links LLM performance to the availability of training data via the controversial Books3 dataset. I will do a thread on this paper soon, but check out the link to the full paper if you can't wait!
1
7
36
Quick thread about a mapping trick the @nytimes has adopted in the last few weeks that have dramatically improved their coronavirus maps. The maps are about different topics, but note one other key diff. between maps on left (April 3), middle (wikipedia) & right (NYT today) 1/
20
354
1K
News alert! I will be walking a bit slowly down the hallways in the future! . I heard earlier this semester that I was awarded tenure @UCBerkeley. Thank you to my dear advisors & mentors who set me on this path. Lots more to come. Upward and onward! . (
95
16
654
🚨🚨 Working Paper Alert! 🚨🚨. Excited to share new @nberpubs WP with @MTranchero -- we examine the impact of access to confidential admin data via @uscensusbureau data centers on economics research. A short 🧵 with key findings .
5
116
390
Totally. The chair in the sky all over again.
Completely agree. LLMs being possible should feel like insane magic.
5
47
340
this is a really neat idea that all of us using less granular data to map should adopt. if there were county-level geo-data / shapefiles that make this easy, many would adopt right away and the curse of blotchy maps shading large, sparsely pop counties would go away! /FIN
16
26
303
👋hey - so there's a new associate editor at Management Science excited to start receiving papers in the Entrepreneurship/Innovation desk. see you in Manuscript Central?!. (thank you @tobystuart @ProfAroraAshish #SridharTayur!)
25
6
282
One day I will write something longer about the value of qualitative research methods for quantitative (esp. applied micro) researchers. But 👇 is a short thread. TLDR: Qual methods super useful even if you never present/analyze qual data in published work . #ResearchMethods.
8
67
270
1/ Excited about new paper on Economics of Maps w/ @sstern_mit in the JEP Winter 2020 issue! . Inspired by "the map is not the terrain", we argue that maps matter & examine the economic antecedents & consequences of mapping distortions. tweetstorm below 👇🏾.
The Winter 2020 issue of the Journal of Economic Perspectives (34, 1) is now available online at
8
66
252
In the last 20 years, @BYU has produced more than double the number of undergrads who went on to get a business school PhD as compared to the nearest competitor (@Penn). play with it yourself!: cc: @AOMConnect @AOM_STR
10
44
239
1/ Excited to release a new paper with @ReimersImke looking at the effects of book digitization on the demand for physical copies.
15
91
236
🚨 Working paper Alert! 🚨. New paper (w/ Aruna Ranganathan, @BerkeleyHaas ) on the impact of digitization on gender inequality in the Hindi film industry ("Bollywood"). Our basic story? Digitization *reduced* female share among playback singers. 🧵👇
20
43
219
As the summer break in schools and childcare centers approaches across the country - spare a thought for the toll these breaks takes on working mothers. see new paper by Brendan Price and Melanie Wasserman. (
0
39
209
nice new addition to the "brain drain" literature. @econgaurav (one of my fav economists of my generation!) and @nicomoralesg argue that increased *probability* of immigration leads natives to upskill, even if they dont actually immigrate.
4
43
184
As someone with a very short attention span, in-person seminars forced me to be locked in a room and engage with a paper in detail. I loved it! . Even with the million v.interesting online seminars now, I struggle to stay focused. Do others face this issue? Any suggestions?.
16
4
176
As student in India with no connections, I worked remotely with a grad student at Harvard doing data work for them on the side. Learnt a lot & wrote about experience in my essay. Now, as a faculty member, I love working remotely with promising RAs in a similar fashion.
4
13
183
🚨🚨 New paper alert 🚨🚨. Very excited about new paper with @SayWhatYouFound title:. Improving data access democratizes and diversifies science. forthcoming in @PNASNews . with @esther_shears & @MathijsdeVaan (@BerkeleyHaas) . link: Thread 👇.
4
69
167
🚨New paper thread🚨 How to reopen the economy safely & yet not hurt the economy? . Our empirical model leverages cellphone location data to map city-based social networks & finds one size does not fit all! Answer depends on local context. full paper: 👇🏾
2
46
173
🚨New research alert! 🚨. Tired of expensive, time-consuming experiments? 😴 We propose a new way to test and extend social science theories with an LLM-based framework: Generative AI-Based Experimentation (GABE) 🤖 . w/ @MTranchero @_arulm_ & Cecil B.
6
36
179
My article on the effects of copyright on the diffusion of information,now published at @INFORMS Management Science!
5
42
169
Mentioning a non-US country in the title of the paper reduces follow-on interest (experimental evidence).
2
47
168
all three are county level maps of the US -- made with county level data. what makes the last map on the right different? . one key design choice highlighted in the footnote: . "parts of a county with a population density lower than 10 people per square mile are not shaded" 2/
3
13
160
Very excited to team up with Stephen Turban and Megan Gorges (incoming @HarvardHBS PhD students) on a new wiki to provide guidance for those applying to PhD programs in business schools. Check it out! . CC: @MITSloanPhD @phdproject. 1/.
2
44
152
respectfully:
2
18
153
Randomly adding 2 paragraphs and 1 photo to @Wikipedia about Spanish towns increased visits by 9%! @WikiResearch .
0
94
143
super excited to have this come out! (finally!) with @ReimersImke.
Forthcoming in AEJ: Economic Policy: "Digitization and the Market for Physical Works: Evidence from the Google Books Project" by Abhishek Nagaraj and Imke Reimers.
10
12
126
🚨🚨 New Paper Alert! 🚨🚨. 🧵 Excited to share our latest working paper where we build a large new dataset to measure geographic bias in top management research. (with Randol Yao, @MITSloanPhD and ex pre-doc) . Full paper:
4
28
123
I’m all for experiments like the Thiel fellowship but the program is so selective, awardees were likely to do great things regardless. I wish they’d release the names of those who almost made the cut but didn’t so that we have a control group.
Since 2011 Peter Thiel has been giving out $100K to college students to drop out and start a company. Forbes wrote, "Thiel Fellows will create lesser successful companies than degree holders.". 10 yrs later, they've founded some of the most breakthrough companies of the decade:.
5
4
126
Excited about my presentation at the #NBERSI science of science funding session tomorrow -- I will be presenting some new results (w/ @MTranchero) on the impact of Census data at 3:45pm ET/12:45pm PT. prog: stream: see you!
3
14
123
🚨 @nberpubs working paper alert! 🚨. Online reviews 🔼 sales but we show.* almost 20% of places don't even have a listing by 2017.* simply creating a listing increases sales ~5-10%. Getting on the map matters! . w./ @mike_luca @gaurisubramani. Thread👇
4
30
118
So at @AOMConnect 2024, my dear student & co-author @MTranchero won the Newman dissertation award. This is the highest award a dissertation could get in the academy -- so we at @BerkeleyHaas are delighted!. If you are interested in joining in his footsteps apply to our PhD
6
7
113
🚨🚨 Hiring Alert! 🚨🚨 . I'm hiring a new pre-doc to work with me starting next year. Please apply if you're interested in getting experience with applied micro-techniques before grad school! And spread the word! . cc @econ_ra @predoc_org.
5
34
110
I wish referees think about this the next time they ask for every possible interaction effect to get at the "mechanism". better to admit its hard to estimate rather than fish for what interactions are significant" and make up a story around that.
You need 16x the sample size to estimate an interaction rather than main effect. This makes diff-in-diff a little underpowered . Wanted: a high powered causal inference estimator.
0
12
104
RIP #VanHalen. Patent innovation researchers, did you know that he was probably a data point in your dataset? .
4
13
98
🚨🚨🚨.Excited to announce that I'm looking to hire a post-doc for next year as a part of a recent @SloanFoundation grant examining the impact of administrative data on academic science (. Apply now!.
3
29
102
#predoc alert! 🚨🚨. If you are looking for a predoc for next year in topics around labor, development, gender and have strong Stata skills could you please DM your CV? I might have an opportunity for you! . Cc: @econ_ra @predoc_org.
7
29
92
EIGHTY TWO 5 minute videos presenting results from recent field experiments can now be found on youtube via the #AFE2020 conference:. Whatay contribution to the public good! @RDMetcalfe.
3
30
94
Have you been told that the work you are doing is "not really economics"?. well, you are not alone! . Ken Arrow was told the same about his dissertation!!.
7
13
91
Imbens and @Susan_Athey providing a template on what else might we report in analysis beyond treatment effect and SEs at @UCBerkeley @UCBITSS #bitssam2023
0
10
85
🚨🚨 hiring alert! 🚨🚨. I am hiring a pre-doctoral research assistant to work with me next year. Please look at the application below and apply. Please help to spread the word!. cc: @predoc_org @econ_ra .
5
28
85
Excited to be co-organizing the NBER digital Econ and AI spring meetings at Stanford (Feb 29-March 1) with @johnjhorton @avicgoldfarb and @ce_tucker -- . submit here:
I'm co-organizing the NBER Digital Economics & AI spring meeting @ Stanford (1st week of March '24). Submit your best work here --> cc: @ce_tucker @avicgoldfarb @abhishekn.
0
14
90
Its clear that with more granular data (say census block level), you can ignore areas with basically zero human population - but the NYT does not have more granular data & yet they pull off this trick. how? simple: do not "shade" portions of a county below minimum pop density 3/.
3
3
82
Congrats to @MTranchero for winning the @OrganizationSci best dissertation award for 2023! . (Psst: he’s on the job market - .
@rotmanschool @BUsociologydept @MITSloan @EileenSuh7 @BUQuestrom @ChelseaLide @StanfordGSB @jeanjhoh @Columbia_Biz @MiyaDraga Our final presenter @MTranchero @BerkeleyHaas
5
8
80
Excited to announce the lineup for the @BerkeleyHaas Entrepreneurship & Innovation seminar for 2022-23. Stop by if you happen to in/near campus on these dates!. (cc: @TaniaBabina @SharonBelenzon @AdrienMatray @DeepakHegde_EFL @ProfSongMa)
7
7
85
It always amazes me when attending a seminar about a very specialized topic, say the econ. of breakfast cereal, neither the author nor an audience member has spent any serious time talking someone steeped in the industry. "domain knowledge" is critical for good empirical work.
If you want to be a good data scientist, you should spend ~49% of your time developing your statistical intuition (i.e. how to ask good questions of the data), and ~49% of your time on domain knowledge (improving overall understanding of your field). Only ~2% on methods per se.
5
12
81
I’m looking for a part time RA to help with a big data project that involves lot of big data GIS data processing and analysis with @openstreetmap data. If that’s your jam, send me a DM or email (nagaraj@berkeley.edu).
5
54
82
I think I'm going to switch to python + stata for figures.
3
9
80
If your RA did this - you’d let them go on the spot.
Had a sort of a funny experience with OpenAI’s Deep Research tool, which I wanted to share since I think it reveals some of the tool’s strengths and weaknesses.
5
4
86
Just discovered that @Afinetheorem and Heidi Williams have a series of seemingly excellent webcasts/lectures on markets for innovation. Check it out:
0
12
84
Theorist: it depends on market power & shape of demand curve, but I don't know.Experimentalist; A/B tests show prices should be increased, but I don't know why.Structuralist: counterfactual where virus spreads to 10% of pop, prices should increase by 96.3% but no one believes me.
3
9
80
Coolest paper I’ve seen at @strategysci conference- this paper uses changes in library classification to understand how these changes affect serendipity.
4
12
79
im curious to hear from other academics about how you are incorporating LLMs in *writing* academic papers? the regular stuff, like intro, literature reviews, description of results etc. are you using these systems at all? if so, how?. they're clearly useful - but finding it.
14
9
84
wohoo! a tweet of mine inspired a full paper!! need to put that on the CV :) great work @MishaTeplitskiy and team!.
🚨 New study on prestige bias in peer review! 🚨. link: A quick🧵on the findings, and why the study owes its existence to. #ScienceOfScience Twitter!
5
2
77
Economic historians and other digitizers be like 🎉😄🎆.
3
10
75
For the final case discussion of my #entstrat class at.@HarvardHBS we invited the enormously talented.@BabaBrinkman for a "rap up" - a fantastic live zoom performance that summarizes the key ideas of the course and the case discussion. Check it out!!.
9
8
78
Jane Wu's job market paper is just terrific. ( . (a) the introduction of crash test dummies reduced car fatalities but (b) fatalities were particularly lower among people who looked like the dummy!. Policy-relevant empirics and great theoretical idea!
2
12
78
I’m going to try to tweet from the inaugural management/analytics/data (mad) conference today and tomorrow at @Columbia_Biz. (. First up @k_mcelheran on AI adoption in America with @erikbryn and others
3
10
79
🚀 Just 1 week until the @nberpubs Digital Economics and AI Meeting that I'm helping to organize along with @johnjhorton and @avicgoldfarb and @ce_tucker at .Stanford GSB. quick 🧵👇 on 8 great papers that are coming up.
1
18
77
Last week, my fantastic @BerkeleyHaas student and collaborator @MTranchero presented a preview of his job-market paper for the first time at @1994ccc last week #CCC2023 . a quick 🧵 about why he/his work is so amazing and why you should all be making a bee-line to hire him!.
The death of theory thanks to #BigData ? Not so fast! Matteo Tranchero @MTranchero documents rampant false positives and explores which firms can avoid investing in false positives!! #ccc2023
3
12
76
But at least now, only 20% of this county is shaded, which does not make you think immediately that "things are terrible in this big part of California" . That read of the data would be wrong, but that is precisely what a county-level map like this one on the right would tell 5/
3
4
68
What are the best new working papers in the area of entrepreneurship and innovation in 2021? Reply below! Do no hesitate to suggest your own work!.
6
14
72
Didn't get into your favorite PhD program? don't despair! Consider applying for this predoc to work with me and try again next year? I promise it will be a fun & rewarding experience. We are doing another review of apps this week so please get in touch soon! (email is ok!).
🚨🚨 Hiring Alert! 🚨🚨 . I'm hiring a new pre-doc to work with me starting next year. Please apply if you're interested in getting experience with applied micro-techniques before grad school! And spread the word! . cc @econ_ra @predoc_org.
6
18
74
Excited to finally start on the faculty at @BerkeleyHaas with @guoxu_econ & others. Dean @richlyons welcoming us in our brand new building!
7
3
72
Excited to hear that despite the challenging times, @BerkeleyHaas has a new open position for senior/junior scholars working in the area of entrepreneurship. If you or your friends/students work in this area ask them to apply!.
1
23
70
Amazing paper forthcoming in @QJEharvard reports RCT findings from China where business owners were randomly assigned to be part of networking groups for a year. Treatment increased revenues by ~8% and was persistent! increase in innovation is one channel
5
32
69
I do this often! I cant remember who I copied this idea from, but I find it to be a very effective way to end talks. See example.
Try writing the *paper in a slide.* ie, your research question, identification, results, takeaway, etc. all fit on one slide before results. You don't need to use this slide in the end. But may help you keep only what's essential for people to understand your key contribution.
1
2
70
So many amazing research questions here for those at the intersection of AI / Innovation / Strategy research via @a16z.
1
22
66
Very exciting lineup of the remote work conference by @I_Am_NickBloom, @prithwic @emma_k_h @Jose_MariaRD etc. Emerging research areas:.1- measurement & selection.2-spatial & urban impacts.3-impact on workers & firms.4-impacts for innovation. Cool stuff! .
0
13
67
Graduate school in economics/management really needs to teach (quant empirical) students a lot of this material as a rite of passage.
I'm teaching a "data science for economists" course this semester. If you're interested in learning more about #rstats, Git(Hub), programming, databases, cloud computation, ML, etc., I'll be making all of my course material publicly available here:
2
9
66
Hello I'm a newly-minted innovation economist. You might know me from my ongoing productions "copyright does really stop the flow of digital info.", "good maps are the key to new discoveries" & "data are not unbiased representations of ground truth, but are socially constructed".
Hello I’m a historian of economics. You might know me from my Greatest Hits including "no we"re not extinct (yet)", "economic facts & models are constructed", "no, it doesn't mean your discipline is no science, only that your data, theories & tools are historically situated".
2
4
66
About 36% of top 1000 ranks in JEE migrate abroad - number rises to 62% for top 100. Cc: @prithwic @inaganguli @patrick_gaule .
4
13
64
I just spent 3 hours assembling this IKEA duktig kitchen for my toddler’s birthday. AMA.
6
0
68
Achievement unlocked - got a picture with Dick Nelson (will go right next to my picture with Arrow) — and @sstern_mit he’s very excited about the economics of maps! 😃
0
1
62
I think I found the killer app for ChatGPT from our local parenting forum. "We use it to create bedtime stories for the kids. Example: tell me a 3 minute story about a pink unicorn running on the beach.".
9
3
60
Our digitization webpage ( has reading lists from previous workshops. Check it out:
PhD Student Workshop in the Economics of Digitization. March 21-23 at Stanford. Application deadline Nov. 16. Speakers: Matthew Gentzkow, Dina Mayzlin, Shane Greenstein, Joel Waldfogel. @dmayzlin @JWaldfogel.
3
13
62
Almost unfair that you can pack so many exciting speakers into one lineup. Awesome work @digitaleconorg . With all talks recorded no excuse to not "attend". CC: @JorgeGuzmanCBS @johnjhorton @robseamans @Prof_Yildirim @CageJulia @DianeCoyle1859
1
18
59
🚨🚨new paper alert!🚨🚨 . Using national brand policies to instrument for local chains store closings we show that chain closures cause nearby independents to close as well. link: w/ @MathijsdeVaan #sameersrivastava @saqibmtz & @SafeGraph data. thread👇
2
9
59
Hard to say something new about the oft remarked fact that map projections warp our view of how the world looks/works. this article by @tomaspueyo does it tho:. my fav bit was this analogy:
1
16
57
Interesting set of slides on how strategy is taught from Juan Alcacer based on a review of syllabi of core strategy classes in 20 top business schools. Here are the most common topics. full slides: . at @strategysci conference:.
1
18
57
My dept. at @BerkeleyHaas is hiring. You are probably applying if you are on the strategy/organization-theory job market. You should!!. I'm here to encourage a few other fields to consider applying to us if you aren't thinking of doing so already . .
1
23
57
Spent a ton of time on two cool papers this spring. Was waiting for 2-3 months for reports. Both got rejected at good outlets within an hour of each other last month. 🤷♂️ reports were more positive than usual but still reject. Revising now to resubmit elsewhere!.
Academics, let's talk more about rejections! We see a lot of announcements of acceptances and published papers but very few rejections. Even if (at least for me) rejection is more common than getting papers accepted. Let me start in the reply. #LifeInAcademia #AcademicTwitter.
7
0
57
check it out! excited to have this out this week.
Exploring whether to use past data or start with a blank slate finds that under some conditions data can lead agents to look under the lamppost, constraining search and harming payoffs, from Hoelzemann, Manso, @abhishekn, and @Matteo Tranchero
1
6
56
One day, I will write a paper with an introduction one tenth as good as this one, from Stigler's classic 1961 piece . Also, while the economics of information is no longer in a slum in the town of economics, still feel like it occupies only a small one-bed apartment .
1
16
54
Uber, Lyft drivers manipulate fares at Reagan National causing artificial price surges
3
21
52
an email audit shows that when it comes to random pitches over emails, investors are *more* likely to respond to women and minority (asian) entrepreneurs. why, then, do we see huge biases against female entrepreneurs in practice?. likely because in practice, only 10% of deals
1
7
54
Congrats to @nberpubs on doing a great job with a zoomified #NBERSI2020. But, can we have the recorded sessions posted on Youtube permanently? pretty please?. The thought of all that amazing content rotting away on a disk somewhere makes me want to cry!.
1
5
56
Started #entstrategy class today with live stock @GME ticker -- perfect ice breaker and a good way to show that I'm finally making good use of OBS! (thanks @lukestein!). cc: @sstern_mit @erin_l_scott
2
3
57
Congratulations to @gaurisubramani who is graduating from @BerkeleyHaas with a PhD this year and is illuminating new drivers and solutions for the gender gap in innovation. Read all about it!.
2
8
54