Grüß Gott, Welcome, Benvenuti, Bem-vindo!


Hi, I'm Max, a 6th-year PhD Student in Economics.

Currently visiting UQAM - Université du Québec à Montréal



My research interests are:

climate econometrics, machine learning, networks in macroeconomics, macroeconomics and many more!





Related Topics

machine learning, neural networks, recession, forecasting, deep learning

Forecasting U.S. Recessions

The Yield-Curve – What Else?!

The inversion of the yield-curve, i.e. the spread between the yield on long-term U.S. Government bonds and short-term Treasuries, has long held up as the single most prominent early-warning indicator of a looming recession in the United States. Even though the sole reliance on the power of the yield-curve comes with several caveats, it formed an integral part of orthodox forecasting models. These models quickly run into trouble when faced with a large set of predictors.


Equipped with multiple macroeconomic indicators, I make use of modern Machine Learning (ML) tools, such as boosted trees and neural networks, which only begin to live up to their full potential in high-dimensional and highly nonlinear environments. The combination of ML and a plethora of information about economic conditions provides enhanced predictive ability relative to classical models. Defying the criticism of limited interpretability of ML models, the results corroborate the standing of the family of yield-curves as the leading indicators of U.S. recessions.


Lastly, I propose YSNet, a neural network architecture that leverages the predictive power of the yield-curve while conditioning its time-varying relation with the probability of a recession on the state of the economy. At the one-year ahead forecasting horizon, it outperforms all other competitors

in terms of the AUC score.