FRED-SD: A real-time database for state-level data with forecasting applications

Abstract

We construct a real-time dataset (FRED-SD) with vintage data for the U.S. states that can be used to forecast both state-level and national-level variables. Our dataset includes approximately 28 variables per state, including labor market, production, and housing variables. We conduct two sets of real-time forecasting exercises. The …rst forecasts state-level labor-market variables using …ve di¤erent models and di¤erent levels of industrially-disaggregated data. The second forecasts a national-level variable exploiting the cross-section of state data. The state-forecasting experiments suggest that large models with industrially-disaggregated data tend to have higher predictive ability for industrially-diversi…ed states. For national-level data, we …nd that forecasting and aggregating state-level data can outperform a random walk but not an autoregression. We compare these real-time data experiments with forecasting experiments using …nal-vintage data and …nd very di¤erent results. Because these …nal-vintage results are obtained with revised data that would not have been available at the time the forecasts would have been made, we conclude that the use of real-time data is essential for drawing proper conclusions about state-level forecasting models.

Forecasting regional industrial production with novel high‐frequency electricity consumption data

“…Our study also relates to the literature on regional economic forecasting, which has developed rapidly in recent years; Lehmann and Wohlrabe (2014) provide an overview of studies conducted up to the mid‐2010s that are based on standard time series models. In an international context, recent articles apply more sophisticated econometric techniques such as factor models (Chernis et al, 2020; Gil et al, 2019) or some form of vector autoregressions (Bokun et al, 2023; Koop, McIntyre, & Mitchell, 2020; Koop, McIntyre, Mitchell, & Poon, 2020). We contribute to this fast evolving international literature in general and to the literature on Germany in particular.…”

EasyData: A Monthly Dataset for Macroeconomic Research on Pakistan

“…Fourth, our paper aligns with the current trend in the forecasting literature, where researchers assemble a macroeconomic research database. For example, see Bokun, et al (2023). Finally, we add to the growing literature on putting econometric models against the latest machine-learning techniques for macroeconomic forecasting, using data from different countries of the world.…”

Quarterly GDP Estimates for the German States: New Data for Business Cycle Analyses and Long-Run Dynamics

“…Their chosen methodology is a dynamic factor model with mixed-frequencies to bring together weekly, monthly, and quarterly observations. Bokun et al (2023) instead compiled a data set with real-time indicators for the U.S. states and use this information for regional and national forecasting experiments. As the data situation at the German state level is definitely expandable, we provide one piece of that puzzle-namely quarterly real GDP growth-to stimulate further research activities with a particular macroeconomic focus.…”