Forecasting the production side of GDP

Abstract

We evaluate the forecasting performance of time series models for the production side of gross domestic product (GDP)-that is, for the sectoral real value-added series summing up to aggregate output. We focus on two strategies to model a large number of interdependent time series simultaneously: a Bayesian vector autoregressive model (BVAR) and a factor model structure; and compare them to simple aggregate and disaggregate benchmarks. We evaluate point and density forecasts for aggregate GDP and the cross-sectional distribution of sectoral real value-added growth in the euro area and Switzerland.We find that the factor model structure outperforms the benchmarks in most tests, and in many cases also the BVAR. An analysis of the covariance matrix of the sectoral forecast errors suggests that the superiority can be traced back to the ability to capture sectoral comovement more accurately.

Maldives Macroeconomic Forecasting:

“…17. Following the specification of Bäurle et al (2018), we denote quarter-on-quarter growth of a single sector or component at time by and the stacked vector of in all sectors by . The vector of macroeconomic variables is denoted by .…”

“…In this environment, an econometric approach developed specifically to address the "curse of dimensionality" may be highly relevant. In particular, the class of Bayesian VAR models that has been recently developed (De Mol et al [2008], Banbura et al [2010], Giannone et al [2012a], Banbura et al [2014], and Bäurle et al [2018]) is known to produce adequate results even when a relatively large number of variables, compared with the sample time series, are included in the model simultaneously.…”

“…6. This study substantially benefited from the literature survey, methodology, and findings of a similar study conducted by Bäurle et al (2018). Section II presents relevant literature on component-driven Bayesian VAR; section III sets up the model; section IV describes the dataset and corresponding patterns; section V reports the forecast results; and section VI concludes.…”

“…However, Drechsel and Scheufele showed that in certain cases, the production-side approach produces statistically significant smaller forecast errors than the direct GDP forecasts. More recently, Martinsen et al (2014) and Bäurle et al (2018) found that disaggregate survey data at regional and sectoral levels improve the performance of factor models in forecasting overall output growth.…”

Unemployment and Gross Domestic Product: Evidence from Papua New Guinea

“…Liu et al (2020) find that the effect of GDP competition on firm investments is significantly positive and also find that GDP competition destroys investment efficiency significantly, especially by increasing overinvestment. Bäurle et al (2020) found that the factor model structure outperforms the benchmarks in most tests, and in many cases also the Bayesian vector autoregressive model (BVAR), and the analysis of the covariance matrix of the sectoral forecast errors suggests that the superiority can be traced back to the ability to capture sectoral co-movement more accurately. Caporale & Gil-Alana (2021) shows that the behaviour of US GDP can be captured accurately by a model incorporating both stochastic trends and stochastic cycles that allows for some degree of persistence in the data, both appear to be mean reverting, although the stochastic trend is nonstationary, while the cyclical component is stationary, with cycles repeating themselves every 6-10 years.…”