Archive for econometrics

efficient measures?

Posted in Books, Statistics, University life with tags , , , , , , , , , on July 24, 2022 by xi'an

When checking the infographics of the week highlighted by Nature, I came across this comparison of France and Germany for the impact of their respective vaccination mandates on health and economics. And then realised this was from a preprint from a Paris Dauphine colleague, Miquel Oliu-Barton (and co-authors). The above graphs compare the impact of governmental measures towards vaccination, short of compulsory vaccination (unfortunately).  Between Germany and France, it appears as if the measures were more effective in the latter. Which may be interpreted as either a consequence of the measures being more coercive in [unruly] France or an illustration of the higher discipline of the German society [despite the government contemplating compulsory vaccination for a while]. As an aside, I am very surprised at the higher death rate in Germany but, beside a larger percentage of people over 65 there and a lower life expectancy, the French curve is interrupted in December 2021. Looking at 2022, the peak was reached at 3.3 cases per day per million people.

Concerning the red counterfactual curves, I did not find much explanation in the preprint, apart from

“Our results are supported by the well-established econometric method of synthetic control.³⁰ We construct counterfactuals for each treated country based on a weighted average of countries that did not implement the COVID certificate and find consistent trajectories for the time period where this method is feasible, i.e., until the end of September 2021.”


“constructing counterfactuals ( i.e., by modelling vaccine uptake without this intervention), using innovation diffusion theory.⁶Innovation diffusion theory was introduced to model how new ideas and technologies spread”

which is not particularly helpful without further reading.

Nature on U.S. abortion laws

Posted in Books, Kids with tags , , , , , , , , on November 4, 2021 by xi'an

The 26 October issue of Nature has a news article on the involvement of US scientists and scientific organisations in fact-checking the dubious arguments made by anti-abortion supporters, incl. several US States. None of them (arguments) are convincing or objective, but providing data and statistical models to counter them is welcome, especially in a scientific journal like Nature.

“…an initiative to compare women who had abortions with those who wanted them, but were turned away from clinics for various reasons, including state restrictions or a lack of doctor availability. Called the Turnaway Study, the effort followed about 1,000 women in the United States for five years after they sought abortions. The women were similar in terms of physical, mental and economic well-being initially, but diverged over time (…) on average, receiving an abortion didn’t harm women’s mental or physical health, but being denied an abortion resulted in some negative financial and health outcomes.”

“Allowing states to ban abortion might even increase maternal and infant mortality rates (…) Unwanted pregnancies are associated with worse health outcomes for several reasons, including that people who plan their pregnancies tend to change their behaviour — drinking less alcohol, for example — and receive prenatal medical care long before those who are surprised by their pregnancy and don’t want it.”

“statistical methods developed over the past 30 years allow researchers to isolate and measure the effects of abortion policies (…) Abortion legalization in the 1970s helped to increase women’s educational attainment, participation in the labour force and earnings — especially for single Black women.”

“The United States is alone among wealthy nations in not mandating paid maternity leave (…) a single parent earning the minimum wage would need to spend more than two-thirds of their income on childcare, with care for the average infant costing about US$10,400 per year (…) two main reasons that women give for seeking abortions are concerns about money and caring for existing children.”

causal inference makes it to Stockholm

Posted in Statistics with tags , , , , , , , on October 12, 2021 by xi'an

Yesterday, Joshua Angrist and Guido Imbens, whose most cited paper is this JASA 1996 article with Don Rubin, were awarded the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel for 2021. It is one of these not-so-rare instances when econometricians get this prize, with causality the motive for their award. I presume this will not see the number of Biometrika submissions involving causal inference go down! (Imbens wrote a book on causal inference with Don Rubin, and is currently editor of Econometrica. And Angrist wrote Mostly Harmless Econometrics, with J.S. Pischke, which I have not read.)

ERC descriptors

Posted in Statistics, Travel, University life with tags , , , , , , , , , , , , on November 9, 2020 by xi'an

Here are the descriptors (or keywords) validated by the (European Research Council) ERC for submitting grant proposal. The recent addition of PE1_15 in the Mathematics panel should help when submitting more methodological projects:

PE1_14 Mathematical statistics
PE1_15 Generic statistical methodology and modelling
PE1_19 Scientific computing and data processing

even though other panels could prove equally suited for some, as in Computer Science and Informatics,

PE6_7 Artificial intelligence, intelligent systems, natural language processing
PE6_10 Web and information systems, data management systems, information retrieval and digital libraries, data fusion
PE6_11 Machine learning, statistical data processing and applications using signal processing (e.g. speech, image, video)
PE6_12 Scientific computing, simulation and modelling tools
PE6_13 Bioinformatics, bio-inspired computing, and natural computing

in Systems and Communication Engineering,

PE7_7 Signal processing

in Integrative Biology,

LS2_11 Bioinformatics and computational biology
LS2_12 Biostatistics

in Prevention,Diagnosis and Treatment of Human Diseases,

LS7_1 Medical imaging for prevention, diagnosis and monitoring of diseases
LS7_2 Medical technologies and tools (including genetic tools and biomarkers) for prevention, diagnosis, monitoring and treatment of diseases

and in Social Sciences and Humanities,

SH1_6 Econometrics; operations research
SH4_9 Theoretical linguistics; computational linguistics

focused Bayesian prediction

Posted in Books, pictures, Statistics, Travel, University life with tags , , , , , , , , , , , , , on June 3, 2020 by xi'an

In this fourth session of our One World ABC Seminar, my friend and coauthor Gael Martin, gave an after-dinner talk on focused Bayesian prediction, more in the spirit of Bissiri et al. than following a traditional ABC approach.  because along with Ruben Loaiza-Maya and [my friend and coauthor] David Frazier, they consider the possibility of a (mild?) misspecification of the model. Using thus scoring rules à la Gneiting and Raftery. Gael had in fact presented an earlier version at our workshop in Oaxaca, in November 2018. As in other solutions of that kind, difficulty in weighting the score into a distribution. Although asymptotic irrelevance, direct impact on the current predictions, at least for the early dates in the time series… Further calibration of the set of interest A. Or the focus of the prediction. As a side note the talk perfectly fits the One World likelihood-free seminar as it does not use the likelihood function!

“The very premise of this paper is that, in reality, any choice of predictive class is such that the truth is not contained therein, at which point there is no reason to presume that the expectation of any particular scoring rule will be maximized at the truth or, indeed, maximized by the same predictive distribution that maximizes a different (expected) score.”

This approach requires the proxy class to be close enough to the true data generating model. Or in the word of the authors to be plausible predictive models. And to produce the true distribution via the score as it is proper. Or the closest to the true model in the misspecified family. I thus wonder at a possible extension with a non-parametric version, the prior being thus on functionals rather than parameters, if I understand properly the meaning of Π(Pθ). (Could the score function be misspecified itself?!) Since the score is replaced with its empirical version, the implementation is  resorting to off-the-shelf MCMC. (I wonder for a few seconds if the approach could be seen as a pseudo-marginal MCMC but the estimation is always based on the same observed sample, hence does not directly fit the pseudo-marginal MCMC framework.)

[Notice: Next talk in the series is tomorrow, 11:30am GMT+1.]

%d bloggers like this: