Ever wanted to measure anger in political speeches? Wondered how political reform can affect anger? Do MPs use anger strategically?
My working paper (undergoing revisions, comments welcome) with Kuniaki Nemoto https://t.co/LVAOnx2k4n tries to answer these questions for NZ.
In short, when it comes to trends in countries, the two obvious approaches, econometrics and poli sci, suggest that any permanent change will take a long time to be ascertained. Optimistically, at least a few years! More likely, only after 2028. Be wary of claims before that.
I've just made a new post on my new-ish Substack:
https://t.co/3q0IeqwxGT
It's about the challenges of how soon we know when trends permanently change. E.g., the US.
(Note that I mostly keep away from social media nowadays, I'm only posting here in case anyone is interested.)
@Noahpinion !
You're literally a couple of minutes away from where I am, interesting coincidence! Wouldn't have expected many Econ PhDs in Harajuku.
Not sure if you care, but there's some really nice coffee around, I recommend Lattest or Be a Good Neighbour for very nice espresso.
Anyway, the point is this is a good way to illustrate some of the challenges of establishing causality in social sciences. I definitely aim to show parts of this video to my students in the future, so they can (hopefully!) understand these points. 8/8
#EconTwitter
An interesting example of how evidence for causality can be found without the use of regressions:
https://t.co/JdPQVVdT2B
It explores how the name Tiffany has fared over time by CPG Grey. At the timestamp, he video discusses the explosion of the name in the 80s. 1/n
(One note: I will say that the evidence for the effect of the VHS is the weak part in the video. I'm a bit skeptical that VCR penetration was sufficiently high to have the effects the video posits. But this just highlights my points above of how evidence is never conclusive.) 7/n
I am very happy that a letter of mine is in this week's edition of the Economist:
https://t.co/TT4Xj9bf0L
And despite my critiques, I should be clear that I recommend @sarahmaslin 's report.
@JonSteinsson Very interesting! I definitely experienced a lot of the changes you write about. I might have a comment or two, how would you prefer to receive them, via email or here?
@MichaelFranca Sobre os pesos, tenho uma ou outra ideia, mas nada muito especial. Todo modo, se eu tiver tempo, te mando um email sobre isso!
Obrigado novamente!
@MichaelFranca Obrigado! Acho vou precisar ler o Firpo et al e Ransom e Sutch para entender melhor algumas coisas, e.g. peso igual para cada dos 3 componentes do indice (que me parece ser a escolha obvia/correta, mas nao é "neutra").
Todo modo, muito revelador ter um indice desses, parabens!
@MichaelFranca Oi Michael, muito interessante e relevante o indice, confesso que fiquei surpreso com alguns dos valores!
Uma pergunta, voces tem/terao um paper discutindo a metogologia? Estou curioso de saber um pouco mais sobre sua construcao.
(Corrigindo de um tweet apagado com autocorrect)
@michaelxpettis@WSJ Yes, I agree, I was playing a bit of devil's advocate.
I am curious to what extent it's each of these factors (and others!). I think the risk aversion factor is smallish overall, but might explain a significant, but not sure how much, part of the overall jump in the last year.