Jason Abaluck
@Jabaluck
Followers
12,126
Following
384
Media
196
Statuses
10,400
Professor of Economics at Yale SOM
New Haven, CT
Joined April 2009
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
UCLA
• 352966 Tweets
#1MAYIS
• 170238 Tweets
Happy New Month
• 107690 Tweets
İşçi
• 98793 Tweets
かしわ記念
• 62530 Tweets
Newey
• 42389 Tweets
Happy Workers
• 41755 Tweets
Paul Auster
• 40099 Tweets
Red Bull
• 35101 Tweets
Polise
• 33864 Tweets
#ErdenTimursuzOlmaz
• 18355 Tweets
シャマル
• 17134 Tweets
PABLO MOONLIGHT PHOTO
• 15660 Tweets
#seibulions
• 14152 Tweets
#県民スター栄誉賞
• 11337 Tweets
Victoria Prego
• 10136 Tweets
Last Seen Profiles
Pinned Tweet
I thought I'd try to reconstruct Bryan's discussion with a 13-year-old about the minimum wage based on the below tweet:
(
@bryan_caplan
let me know if anything is inaccurate)
Took me under 5 minutes to turn a normal, smart 13-year-old against the minimum wage.
Contrary to almost everyone, the textbook argument IS intuitive.
It's just emotionally unappealing.
638
50
1K
48
536
3K
With dozens of researchers at Yale, Stanford, Berkeley and IPA and several other organizations, we ran a cluster randomized trial involving almost 350,000 people and 600 villages in Bangladesh to assess the impact of community masking on COVID.
309
3K
8K
My preferred policy solution to gun violence is to hold gun purchasers liable if the gun they purchase is ever used in a crime, then to require gun owners who can't otherwise demonstrate a capacity to pay to purchase insurance.
399
193
2K
The reduction was larger in villages where we (randomly) used surgical masks than those where we used cloth masks; in surgical mask villages, we saw a 12% reduction in COVID overall and a 35% reduction among those aged 60+.
69
373
2K
With this 29 percentage point increase in mask-wearing, we saw a 9% drop in serologically confirmed COVID.
13
280
1K
Since severe morbidity and mortality are concentrated among the elderly, this suggests that community-wide masking can be an extremely effective tool to combat COVID.
28
189
1K
If going from 13/100 to 42/100 people wearing masks leads to reductions of the magnitudes above, near universal mask-wearing (as is possible with enforced mandates in some areas) might lead to substantially larger reductions.
22
207
1K
As noted, we find especially convincing evidence that surgical masks are effective. Cloth masks reduce COVID symptoms, but the effect we find on symptomatic infections (confirmed via blood tests) is driven by surgical masks.
24
202
1K
Cloth masks are likely better than nothing, but surgical masks or masks with higher filtration efficiency should be preferred to cloth masks where available.
63
191
1K
Hi
@benshapiro
, our study DOES suggest benefits of mask-wearing for people under the age of 50:
Lot of focus on this study in Bangladesh on masking. Essentially, it shows that mandatory masking -- particularly with a surgical mask -- can decrease transmission among those who are older. But the same study shows no statistical significance for those below age 50.
129
318
2K
27
167
1K
"The Case for Universal Cloth Mask Adoption & Policies to Increase the Supply of Medical Masks for Health Workers" by me
@judy_chevalier
@NAChristakis
@thehowie
E Kaplan, A Ko (Chair of Epidemiology
@yale
) S Vermund (Dean of School of Public Health
@yale
)
40
625
989
There is near unanimity among economists I know that a war-level investment is needed on policies designed to directly counteract the spread of the virus. Any investment that will marginally reduce R_0 may have a return in the hundreds of billions or trillions of dollars.
27
327
877
We conducted an intervention that increased mask-wearing by 29 percentage points using the techniques described here:
We ran a massive randomized controlled trial investigating many different strategies to get communities of people to wear masks -- we're now scaling up the strategies that worked in many regions throughout South Asia.
7
102
281
12
114
798
13-YEAR OLD: I'm familiar with the first-welfare theorem you dolt, do I look 12 to you? Even if there were efficiency consequences (and we will return to this), the minimum wage redistributes money from firms to existing workers.
1
20
635
A longer discussion of our intervention is available here, along with the underlying working paper:
6
139
633
An intuitive way to grasp the effectiveness of masks: extrapolating from our results, every 600 people who wear masks for a year in public areas prevents 1 person from dying of COVID given status quo death rates in the US.
125
156
599
It is sad.
@DrJBhattarcharya
is the worst example I have personally seen of someone who was previously a scholar but who now engages in repeated misrepresentation of scientific results to serve a partisan agenda.
48
171
557
Does health insurance save lives? While there are dozens of studies with a wide range of estimates, the best studies consistently suggest that the answer is yes and the magnitude is considerable.
14
117
549
It is good for the world that virtually no one can earn a living wage as a telegraph operator, and it would have been bad if we had prevented people from using telephones in order to protect telegraph workers.
114
60
528
People puzzled by Trump's near-win forget how hard it is to form beliefs about the world. Even for professional policy analysts, your views about nearly all topics are heavily shaped by deferring to those around you. If you had the wrong epistemic community, you'd be sunk too!
8
94
516
In the next few weeks, we'll post a public GitHub package with all of our data and analysis (with identifiers removed).
32
26
505
How large are the benefits? Even if masks reduce transmission probabilities by only 10% (and as you'll see, that is likely very conservative), the value of *each cloth mask* is between $3,000 and $6,000. Our best estimate is that their protective value is closer to 40-50%.
10
226
487
Our 350,000 person, 600 village RCT evaluating the impact of masking on COVID builds on a large, but not entirely unambiguous literature studying the impact of masks on respiratory diseases, including COVID.
4
102
479
A great moment in Econometric history: Jim Heckman develops the Heckman selection correction and offers to share the code to compute correct asymptotic standard errors. For the next two years after publication. For a small fee.
14
53
458
Bad news everyone! We've been way off about this whole science thing
43
13
446
For those keeping track at home, this is definitely not what a confidence interval is. A 95% CI is a function of the data such that, given the data generating process with an unknown true parameter, the CI constructed in this way will contain the true parameter 95% of the time.
@Akustronique
2/ What most people dont understand is that any value within the interval x and y is fair game with uniform distribution. There is NO guarantee that the prior distribution (the thing you are measuring) would follow a gaussian distribution or any other.
8
1
10
7
26
438
What do all successful people have in common? They don't condition on the dependent variable when modeling success.
6
84
427
There is a selection bias generated from the fact that researchers choose to study a topic that causes us to overstate the strength of the evidence in nearly all historical studies and many others.
11
63
429
"People respond to incentives" is a 2nd order description of human behavior (important in many cases).
To 1st order, people always do the same stuff. Behavior is hard to change.
The most powerful interventions change contexts so that the same behavior leads to better outcomes.
6
74
406
Know your elementary models of competition:
1) Bertrand (e.g. fast food) -- compete on prices
2) Cournot (e.g. airline capacity) -- compete on quantities
3) Hotelling (e.g. vendor location) -- compete on horizontal attributes
4) Hoteling (only hotels) -- compete on # of pillows
1
43
402
BRYAN: Oh shit, I just wrote a WHOLE BOOK entitled, "Don't be a Feminist." Boy am I going to have egg on my face for this one.
11
6
370
A critical error that I see many grad students make: they try to estimate Frankenstein's model. Rather than viewing a model as answering a research question, they view a model as an arbitrary hodgepodge of models they learned about in their classes.
5
64
367
This is not a
@mattyglesias
point. This is a Pigou point. But it's astonishing that this point has penetrated the public discussion so little that
@NateSilver538
feels it should be attributed to
@mattyglesias
. What are millions of econ students learning if not this?
@MattZeitlin
This is piggybacking off a
@mattyglesias
point but I think it shows the problem is that a lot of COVID behavior which may be rational from an individual's standpoint can nonetheless cause negative externalities for society.
19
14
423
36
15
354
In subsequent posts, which I'll link here when available, I'll say more about how our study fits into the existing literature, as well as caveats and policy implications.
4
22
359
There are two facts which make correlations controlling for confounds uninformative about many (but not all) causal effects: 1) the R^2 of the mechanisms we understand is low, 2) our uncertainty about not well-understood mechanisms should be high. (1 / about 13-15)
7
98
352
Except the snarky economists are completely right. It's not a good paper. It's almost certainly false that going to the opera once a year reduces mortality by 14%. There are dozens of possible confounds not controlled for.
I'm seeing a lot of snarky economist tweeting about this paper. If you read the paper itself it's very well written, doesn't claim causality; however, the authors have collected an exceptional amount of data with rich controls. 1/N
4
37
159
6
32
337
@federicolois
The best analogy I can give is that your use of statistical language is like a person who says -- "I speak Spanish fluently -- mi favorito food is hamberguesas because is the best of foods". If you think you speak Spanish, it shows that you don't know what Spanish is.
3
11
332
We have two principal recommendations: 1) everyone should immediately begin wearing cloth masks in public, 2) The govt. should immediately use all available means to increase the supply of medical masks, especially by heavily rewarding producers.
8
162
305
I hope this is the last time I have to remind the people of the internet: please stop making extremely confident claims about technical subjects where you would likely fail introductory level coursework.
4
22
297
Except:
1) Academic economists were unanimous about the need to contract the economy to fight the virus:
2) You said this in March, 2020 about masks: , arguing against the academic economists calling for universal mask adoption
Watching econ twitter lose their gd minds over being called “supportive of the current capitalist system” as an epidemiologist who has spent the past 18 months being called a literal baby killer for advocating face masks & lockdowns to control a pandemic is a mood.
24
92
1K
7
26
299
Academics need to engage with people spreading dangerous misinformation, especially under the guise of technical jargon.
If you do not respond, no one will and the world will be less informed.
The kind of engagement
@jhausdorfer
wants is *not possible* on the needed scale.
I'm sad and annoyed that my community is humiliating and ridiculing someone who has a deficient understanding of econometrics and was not modest about it. This is not the
#EconTwitter
I know and love.
A rope: 1/n
17
48
483
7
25
280
Note that this is not very different from what we already do in the US with car insurance. This does not mean it is politically realistic in the US, but political feasibility is harder to forecast than pundits realize.
40
5
281
Some people need reminding: when you die, God doesn't care how many Twitter followers you have! He cares about the total number of likes on your tweets weighted by their source. This is a *much* better measure of whether the value of your tweets outweighs other bad stuff you did.
6
22
279
We ran a massive randomized controlled trial investigating many different strategies to get communities of people to wear masks -- we're now scaling up the strategies that worked in many regions throughout South Asia.
A randomized controlled trial (N=341,830) conducted by IPA,
@mushfiq_econ
,
@Jabaluck
, doctors and mask engineers from
@StanfordMed
with Bangladesh health researchers and officials increased mask wearing by a lot, and the behavior maintained after the intervention ended
3
50
128
7
102
281
While it would certainly be nice to have more RCTs, this article misses the key point
@LizHighleyman
: there are powered studies -- like our study in Bangladesh and many quasi-experimental studies -- which find effects. There are underpowered studies that don't find effects.
Here's my latest for
@Slate
on the
@CochraneLibrary
review of masks for
#COVID
prevention. Thanks
@JenniferNuzzo
,
@Jabaluck
&
@VPrasadMDMPH
for comments (& apologies to those who ended up on the cutting room floor due to space limitations). 1/
253
137
684
6
65
275
No, it's not that. We're worried that if the economics PhD students experience a taste of luxury, they'll lose the will to spend 18 hrs a day in an infested, underground cement bunker producing the lagrangians that fund our decadent lifestyles.
Most incredible part of columbia business school is that you don’t have access to it if you’re an economics student in the school of arts and sciences. Are they afraid we’re gonna come in and start graffitiing lagrangians or something?
23
14
410
3
8
273
13-YEAR OLD: Saying that men are more likely to die violent deaths doesn't engage *at all* with whether it would be normatively desirable to have more equal representation of women in positions of power and authority, or whether earnings gaps reflect undesirable norms.
4
5
264
Suppose you purchase a gun and it is later used to kill 10 people (by you or anyone else). You can be sued for $100 million (VSL x 10). If an insurance company thinks there is a 1 in 10,000 chance of this happening, they'll charge you $10,000 in premiums when you buy that weapon.
43
3
254
Economists who are minimizing the danger of AI risk:
1) Although there is uncertainty, most AI researchers expect human-level AI (AGI) within the next 30-40 years (or much sooner)
2) Superhuman AI is dangerous for the same reason aliens visiting earth would be dangerous
55
42
264
This solution achieves the first-best in the sense that insurers would then be incentivized to figure out the size of the externality imposed by any given gun purchase and charge the would-be purchaser a premium equal to that externality.
6
7
253
One response I haven't seen yet: your existential crisis is probably correct, your research is probably wrong or ill-motivated, and you should fix this by asking better questions, getting better data, or finding a better identification strategy (probably all three).
how do people deal with emotional ups + downs of research. like one minute everything is awesome and exciting and you love your job, and the next minute your variation is useless and you're having an existential crisis about your agenda. How do you people live like this!?
83
59
972
8
27
248
What's your degree in?
Me: Economics. Are you familiar with natural experiments?
"No, b..."
Me: Okay, suppose a college admits all applicants with an SAT score above...
[20 minutes later]
"Sir, there are other people waiting to get through customs, you have to leave."
What’s your degree in?
Me: Economics
Ooh, what stock should I buy! What’s bitcoin? Don’t you hate fiat money?
Me: fakes a heart attack
18
71
473
3
13
242
Since 2010, there were 350 articles from AER, QJE, ReStud, JPE or ECTA which estimated discrete choice models. 90% assumed consumers are fully informed. 126/350 do welfare analysis. 86.5% assumed consumers are fully informed.
5
52
240
In philosophy, there is a resource called the Stanford Encyclopedia of Philosophy which is: a) written by leading scholars, b) up to date, c) free, d) comprehensive. The closest things I know of in econ are the JEP and JEL, but these are lacking on b)-d). (ht:
@ben_golub
)
9
38
250
If a rich person spends $200 million on a painting, someone else gains ~$200 million. If a rich person spends $200 million to build a yacht, many resources are exhausted on yacht building rather than being allocated to something that helps people. This distinction matters.
31
36
241
One of the best feelings in the world is when you wake up, from a good night's sleep, look out the window, the sun is shining and you think, "What a beautiful day for some Twitter diatribes."
8
5
237
Terrible news: some economists developed a way to signal that all authors deserve equal credit for a paper without listing the names alphabetically.
Love this solution to ordering the names of authors in economics, where alphabetical is the norm. Special symbol indicating certified randomization. Now endorsed by
@AEAjournals
13
73
304
11
21
237
Finally, papers that "don't claim causality" but are only of any interest if the correlation they report is causal don't get a free pass. Either the result is causal or the paper is completely useless and misleading.
2
18
231
@federicolois
Again, I want to assure readers who may not know better that the above is literally gibberish -- it is technical language used without regard for the underlying meaning.
2
1
234
This paper reminds me of the scene from Chernobyl.
"Our first-stage F-stat is 15, we have no weak instruments problem."
"The threshold for valid inference isn't 10. It's 104.7."
"God forgive us."
For all people doing IV in their work, this brand new paper seems *very* relevant!
One important takeaway msg is: Rule of 1st stage F stats > 10 can be very misleading.
Authors also give nice alternative inference tools that can “rescue” us all!
22
280
1K
3
28
219
13-YEAR OLD: So we should start our discussion by understanding the *incidence* of the minimum wage, then we can return to the efficiency consequences.
1
1
208
If you are a farmer in Wyoming and your rifle is exceedingly unlikely to be used in a crime, insurers will charge you very little to insure your firearm. This policy would make no difference to the price you pay to own a gun.
5
7
207
What regulatory options make sense to reduce risk from AI? My tentatively preferred option is to allocate at least $100 billion a year or more for rewards and grants for AI safety innovations, assessed by a board of relevant CEOs and AI researchers (i.e. people with inside info).
21
33
214
13-YEAR OLD: I want more redistribution and you want less. In our 5 minute discussion, we can't resolve this normative question. However, we can understand whether a world is possible that is superior either way.
1
2
205
Alternatively, if you purchase an automatic weapon, insurers will have a strong incentive to make damn sure you don't plan to use it to kill a bunch of people, otherwise they will charge you astronomical premiums.
32
8
198
We summarize the what worked with the acronym NORM, because this reminds implementers to do four things despite us being bad at acronyms:
N: No-cost mask distribution
O: Offering information
R: Reinforcement, public and in-person
M: Modeling, by religious and political leaders
2
61
205
2) The benefits we document for age > 50 may come from mask-wearing from age < 50. Masks protect the elderly partly by preventing young, infected people from transmitting the virus. We estimate total village-level impact, not the protective impact for specific individuals.
5
21
199
Worst date EVER
Her: Libras have more professional success but they are also conceited.
Me: Put aside the theoretical implausibility, what empirical test would convince you that planets *don't* impact personality?
Her: Planets? I just think quarter of birth is exogenous.
10
7
204
Some follow-ups: I should note that the PIs on this project were myself and
@mushfiq_econ
(economists at Yale),
@Kwong_Laura
, Steve Luby and Ashley Styczynski, epidemiologists and environmental scientists at Stanford (and in Laura's case, now Berkeley!).
2
8
196
I recently testified pro bono in a trial about masking in schools (in my view, a complex question) for the sole purpose of explaining that the court should not trust
@DrJBhattarcharya
because he is deliberately misleading people about our study and others.
4
27
195
The basis for our recommendation is simple: anything that combats the spread of the virus is absurdly valuable due to the resulting reduction in mortality risk (not to mention accelerating resumption of normal economic activity).
3
46
187
BRYAN: I'm glad to hear you learned something. I'm off to convince a 9-year old that "feminism" means "Women have harder lives than men" and that this is wrong.
3
3
196
The promised policy / cost-benefit analysis thread is here:
In summary, the average value of masks as a policy to counter COVID remains high in the US, especially in places where deaths and severe morbidity remain high, where vaccination rates are low, or where caseloads are growing.
11
33
82
24
27
191
The promised thread on how this fits into the existing literature is here:
The bottom-line is that, as of March 2020, we had RCTs showing effects in hospitals for medical masks, lab evidence suggesting masks were effective as source control for related viruses, and suggestive cross-country evidence for COVID
3
14
61
2
26
190
This paper from
@instrumenthull
,
@aislinnbohren
and
@alexoimas
(BHI) very helpfully clarifies how to define discrimination, which is essential to interpret “taste” vs. “statistical” discrimination and other currently popular distinctions in econ.
Very excited about this new working paper with
@aislinnbohren
&
@alexoimas
, "Systemic Discrimination: Theory and Measurement"
We develop theoretical & empirical tools to model & measure the systemic drivers of discrimination in many settings
Summary 🧵:
15
261
807
3
48
193
13-YEAR OLD: But as a starting point, we should note that redistribution *always* has a shadow cost, so we want to compare the minimum wage to other policies like a negative income tax that redistribute at some efficiency cost.
1
1
184
@federicolois
This is an RCT, there is no omitted variable bias. Multicollinearity is so far from being relevant here that it is totally obvious to anyone with an undergraduate stats education that you have no idea what you are talking about.
3
3
181
The totality of the evidence on masking -- from laboratory studies (and modeling building on those studies), quasi-experimental studies, and powered RCTs -- is now extremely compelling and suggests that masks can be an essential complement to vaccines in containing COVID.
5
70
180
There is a very bad idea popular in the humanities, among historians, and in popular culture that trying to massively extend human lifespans is immoral hubris, & that more generally, society must be short-signed due to the urgency of our current problems.
9
24
182
To frame this as "mask advocates" vs. "vaccine advocates" is to thoroughly miss the point.
Everyone on earth with access to vaccines should get vaccinated.
Masks are also a powerful tool against COVID, and you've misunderstood what our study says about them.
@MartinKuldorff
Odd that mask advocates are excited by this study. As a vaccine advocate, I would be horrified if a vaccine trial showed 11% efficacy. Based on the 95% confidence intervals, we do not even know if surgical mask efficacy is more than 0%.
163
1K
4K
7
24
176
We have very good evidence that universal adoption of cloth masks will combat the spread of the virus. Specifically, we know that 1) asymptomatic people spread the virus, 2) mask wearing by infected people prevents them from transmitting the virus (the report provides citations).
5
64
169
1) We only studied adults, but for both surgical and cloth masks, we find impacts on COVID symptoms at all ages tested
4
18
168
13-YEAR OLD: What on earth does that claim have to do with whether it's good for women to have voting rights, the right to enter contracts without male approval or access to education and labor markets?
7
2
175
A striking fact about people with high expected lifetime earnings: very few with low early-career wages take out large loans for consumption purposes as opposed to investment. Mortgages are the one big exception (although maybe because housing can be viewed as an investment).
20
18
171
There are a lot of settings where a Netflix-style recommendation engine seems like it might improve welfare -- e.g. which college classes to take, job search, what city you might like living in, what recreational activity to engage in.
9
12
172
NEW PAPER: Some insurance plans increase beneficiary mortality, but consumers (mostly) can’t tell which ones. Getting rid of the worst 5% of Medicare Advantage plans could save 10,000 lives/year (w/
@autoregress
,
@amandastarc1
, and
@mmcaceresb
7
64
170
This is a totally misleading assessment that is mostly just repeating the same exact errors they made in 2020 (the vast majority of the studies reviewed are influenza studies from before that period).
Update on the 2020 Cochrane review on the effect of physical interventions of influenza / SARS‐CoV‐2 transmission. The results do not suggest medical/surgical masks reduce viral spread. N95/P2 respirators might have some marginal impact.
1/
137
725
2K
33
49
152
13-YEAR OLD: Those groups will have higher income-today than non-wage workers. Tax data could help us understand just how redistributive the min wage is in a lifetime PDV sense, accounting for the fact that min wage workers are younger and might have higher wages in the future.
1
0
162
The judge disallowed him as an expert witness because of his repeated misrepresentations:
11
28
164
This is a common and fundamental problem in medical journals: researchers pretend that all they care about is an association, but their entire research question is premised on their result being causal.
A downside of social science in medical journals is that authors are forced to dance around their real goal — causal inference — thereby obfuscating what the estimand is, why particular analyses were done, etc.
eg
3
5
29
2
35
166
13-YEAR OLD: And ultimately, this understanding will shape the strength of our political advocacy: we may support a 2nd-best solution as better than nothing, but we should push especially hard for policies that achieve political goals while minimizing harm to others.
1
4
160
@DrJBhattarcharya
's behavior is especially disgraceful because he has the credentials and appearance of a person doing this type of science, when in practice, he is engaged in sophistry to justify his predetermined conclusions.
8
19
158
The idea that publication decisions should be based only on design and not on findings is not sensible. Most things in the world don't work. Given this prior, papers that demonstrate effectiveness are all else equal much more informative.
I'm here to report that in the year of our lord 2023 papers are still being rejected based not on their designs, but on their findings.
19
45
502
40
13
160
Ignorant comments like
@kareem_carr
's degrade the public trust in experts and equate "academic economist" with "WSJ editorial writer". This reduces the influence of actual experts and is massively counterproductive science communication.
2
9
158
13-YEAR OLD: when conducting a meta-analysis, we need to be more discerning about the quality of the underlying data and the research design. Many of the best identified recent studies use more compehensive data and highlight the inadequacies of earlier studies.
1
5
152
When you think, "What could go wrong if I assume my demand function is invertible?", it's time to reread Berry, Gandhi and Haile. Worst-case scenario: your mixed logit model summons Beelzebub.
7
17
153
13-YEAR OLD: The purpose of a Kaldor-Hicks type analysis is not to forecast what will happen, but to help understand the space of possible political compromises given our differing normative commitments.
1
0
146