Ludwig-Maximilians-Universität München’s Post

The 17th CFE-CMStatistics 2023 joint conference will take place in December in Berlin. It will take place in a hybrid format, live and online. You are invited to submit abstracts in either the statistical (CMStatistics) or econometric (CFE) part of the conference. The deadline for submission of contributed abstracts is set for September 8. As a member of the scientific committee of the econometric part of the conference, I will organize invited sessions on the topics of Econometrics of textual data, Novel paths in difference-in-differences approaches and Household finance using SHARE data (joint with Dr. Giacomo Pasini). The CFE-CMStatistics conferences are usually very well attended and a regular place where mathematically oriented researchers from the fields of statistics and econometrics meet. #econometrics #statistics #computation #financial #methodology #methodologies #causalinference #causality #bayesian #machinelearning #datascience #nonparametric #functionaldata #objectbased #symbolicdata #extremevalue #highdimensional #textanalysis #textanalytics #robuststatistics #latentvariable #mixturemodelling #networkdata #spatialdata #graphicalmodels #timeseries #paneldata #modelselection #regression #probability #stochasticprocesses #mathematics #cfecmstatistics2023

Interested in macroeconometrics and in the estimation of dynamic causal effects through external instruments in #macroeconomics? Take a look 👇 at our paper, joint with the amazing Giuseppe Cavaliere and Giovanni Angelini, Alma Mater Studiorum - Università di Bologna, freshly accepted by @jeconometrics! We show that, if your instruments for the target shocks are "weak", you can still recover the dynamic causal effects using "strong" proxies for the non-target shocks. To ensure validity of this procedure, we also develop a novel bootstrap test for relevance. This is (i) computationally simple (you need to bootstrap the proxy-SVAR and subsequently apply simple normality tests under specified conditions); (ii) it controls size; and (iii) it is not subject to the usual pre-testing issues. Takeaway: there are situations in which it is possible to bypass the need for weak-instrument robust methods when estimating the dynamic causal effects of interest. This can be achieved by using the (pre-tested) strong instruments for the non-target shocks, employing standard asymptotic methods. The advantges are evident in our empirical illustrations 👇 Hope you will like it - comments welcome! ⭐

Join us on October 17 at noon for a talk by Nobel Laureate Lars Peter Hansen. Professor Hansen's talk is titled "Risk, Ambiguity, and Misspecification." The Weidenbaum Center on the Economy, Government, and Public Policy is cosponsoring this event with the Department of Statistics and Data Science, the Center for Dynamic Economics, and the Transdisciplinary Institute in Applied Data Sciences. Learn more at https://lnkd.in/gPhZxuU5

I wrote up a first part explainer to a paper by Tymon Słoczyński (published last year in Review of Economics and Statistics) that I think is maybe my favorite paper of the last year. It's a decomposition of OLS estimated coefficients on a treatment variable when treatment effects are heterogenous with respect to covariates and treatment itself satisfies unconfoundedness. This is only part 1 and it just focuses on the algebraic break down of the theorem Tymon found which is that OLS estimates a number which is a weighted average of two other numbers. But the weights are not intuitive, and they cause bias (though today I hold off on discussing bias). Given the ubiquity of estimating causal effects with OLS and "controlling for X", the marginal benefit to marginal cost ratio of this paper is, in my opinion, extremely high. Come check it out! https://lnkd.in/gfsNsxEP

Thanks to a Robert Summers Economic Measurement Fellowship, I could visit the Conference on Research on Income and Wealth (CRIW) meeting at the National Bureau of Economic Research Summer Institute last month. Robert Summers was a pioneer in the study of international price and output comparisons. With his colleagues, he developed the Penn World Table. This dataset provides, among others, information about input, output, productivity and income by country. Data in different countries is measured in the same way to make them comparable. The database grew from 10 to 183 countries and spans the period 1950-2019. Many of us use data, fewer compile it themselves, even fewer have the vision to make the data a public good and even fewer manage to attract the necessary funding and accomplish it. Robert Summers did it all and now the Penn World Table is freely accessible. The most recent version can be found at https://lnkd.in/eGb-D9JE This data provides many kinds of research opportunities; I used it myself last month. The goal of the fellowships was to promote research on economic measurement and to bring the academic world and compilers of economic data together. The fellowships were awarded to five economic statisticians working in government statistical agencies and international organisations. For me, this goal was definitely reached. Having the opportunity to chat with researchers one previously knew from the literature, seeing the latest developments in research and data (e.g., disaggregating economic flows, https://lnkd.in/euuvdKKX, and patents, https://lnkd.in/eafmkKgN), speaking to colleagues from different statistical agencies and bringing them into touch with their counterparts in the Netherlands. Very inspiring and very beneficial for me and Centraal Bureau voor de Statistiek! Therefore, I recommend all economic statisticians to apply next year; rumour is there will be fellowships again. See https://lnkd.in/eD28wZhK for the very diverse programme of this year, links to papers and attendants. #economicmeasurement #academics #economicstatistics  

📊 Interactive dashboards are a great way to help non-academic audiences understand research. Our Centre for Economic Performance's Real Estate Economics and Finance (REEF) index is a prime example of interactive research in action, showing property price trends across England and Wales from 2010 to 2020: https://lnkd.in/eA7NeTcA Read about how dashboards are making research accessible to the public and policymakers on LSE Impact Blog ⬇️

Every exceptional scientist questions himself about the fundamentals of his/her science, thus positing epistemological questions, thus methodological questions. Keynes was such a scientist. But not only John Maynard, John Neville (his father) as well. This is an extraordinary extract taken from "The Scope and Method of Political Economy", which can be found for free also here: https://lnkd.in/dSkum6eH "It is true that it is one thing to establish the right method for building up a science, and quite another thing to succeed in building it up. It is also true, as the Austrian economist Menger has remarked, that sciences have been created and revolutionized by those who have not stopped to analyse their own method of enquiry. Still their success must be at- tributed to their having employed the right method, even if they have employed it unconsciously or without going out of their way to characterize it. Their method must, moreover, be subjected to careful analysis before the value of their contributions to the science can be properly estimated. Economics is not in any way peculiar in requiring that its method should be discussed. The logic of other sciences is, however, for the most part sufficiently dealt with in general works on logic or methodology. There are special reasons, partly to be found in the nature of the subject itself, and partly due to extrinsic causes, why the logic of political economy needs a more detailed consideration. In the first place, economic science deals with phenomena that are more complex and less uniform than those with which the natural sciences are concerned; and its conclusions, except in their most abstract form, lack both the certainty and the universality that pertain to physical laws. There is a corresponding difficulty in regard to the proper method of economic study; and the problem of defining the conditions and limits of the validity of economic reasonings becomes one of exceptional complexity. It is, moreover, impossible to establish the right of any one method to hold the field to the exclusion of others. Different methods are appropriate, according to the materials available, the stage of investigation reached, and the object in view; and hence arises the special task of assigning to each its legitimate place and relative importance."