Maximilian Göbel (2:00 pm – 3:00 pm)
Forecasting U.S. Recessions. The Yield-Curve – What Else?!
The inversion of the yield-curve has long held up as the single most prominent early-warning indicator of a looming recession in the United States. Yet, recessions are arguably the result of a complex convolution of many economic variables. While such a setting strains the capabilities of orthodox models, machine-learning algorithms are supposed to excel in such environments.
Still, classical probit and penalized regression models turn out to be resilient competitors. Beyond providing a plain point-forecast and quantifying its uncertainty, I address the criticism of ML’s limited interpretability. The results corroborate the standing of the yield-curve as the principal predictor of U.S. recessions, followed by labor-and stock-market indicators. Bundling predictive ability and interpretability within a single package, I propose RecAE, a structural autoencoder-type architecture that leverages the predictive power of the yield curve while conditioning its relation with the probability of an upcoming recession on the state of the economy. The RecAE identifies three consecutive regimes in which recession-probabilities have become increasingly sensitive to a steepening of the yield-curve, hinting at the U.S. economy undergoing several structural changes. The results advocate for carefully designed ML to be a valuable addition to the econometric toolbox.
Clément Gorin (3:00 pm – 4:00 pm)
The emergence, growth and stagnation of cities: France 1760-2020
Historical maps contain rich information about buildings, land-use or transport networks, and provides novel insights to understand the origins of spatial disparities. This paper analyses the evolution of French urban areas from a historical perspective. Specifically, we implement a fully-convolutional neural network to extract pixel-level information from three collections of digitised historical maps covering mainland France in the 18th, 19th and 20th centuries. Using the extracted buildings, we define consistently urban areas and analyse their development trajectories along the urban hierarchy. Our network model performs remarkably well and proves robust to considerable representation heterogeneity both within and across maps collections, as well as severe class imbalance. This approach is efficient, scalable and readily transferable to other historical maps with minimal manual labelling. Our findings highlight increasing urbanisation with fewer and larger urban areas, consistent with agglomerations economies. Disaggregate analysis reveals significant heterogeneity, with the emergence, persistence, disappearance of urban areas.
Join at: http://imt.lu/seminar