Economic Time Series Predictions and the Illusion of Support Recovery

28 Pages Posted: 30 Jan 2022 Last revised: 13 Apr 2024

See all articles by Philipp Adämmer

Philipp Adämmer

University of Greifswald

Rainer Alexander Schüssler

University of Rostock - Department of Economics; University of Konstanz

Date Written: April 12, 2024

Abstract

We investigate to which extent popular forecasting methods are capable of recovering the true number and identities of relevant predictors for macroeconomic and financial time series data. The key feature of our simulation-based approach is that we infer realistic signal-to-noise ratios based on the degree of (out-of-sample) predictive accuracy from real-data forecasts. Our main finding is that popular forecasting methods cannot reliably detect the true number and identities of the relevant predictors when using realistic signal-to-noise ratios.

Keywords: Signal-to-noise ratio, Variable selection, Shrinkage, High-dimensional data

JEL Classification: C53, C55

Suggested Citation

Adämmer, Philipp and Schüssler, Rainer Alexander, Economic Time Series Predictions and the Illusion of Support Recovery (April 12, 2024). Available at SSRN: https://ssrn.com/abstract=4019646 or http://dx.doi.org/10.2139/ssrn.4019646

Philipp Adämmer

University of Greifswald ( email )

Friedrich-Loeffler-Strasse 70
D-17487 Greifswald, 17489
Germany

Rainer Alexander Schüssler (Contact Author)

University of Rostock - Department of Economics ( email )

Ulmenstr. 69
Rostock, 18057
Germany

University of Konstanz ( email )

Fach D-144
Universitätsstraße 10
Konstanz, D-78457
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
177
Abstract Views
708
Rank
313,835
PlumX Metrics