using BREl to nowcast the Belgian business cycle : the role of survey data
评估比利时国家银行调查数据对预测GDP等季度宏观指标的效用，基于BREL即时预测平台和弹性网络回归，发现调查数据特别是细分指标能显著提升早期预测精度。
The article assesses the usefulness of indicators taken from surveys carried out by the National Bank of Belgium for predicting Belgian GDP and other important quarterly macroeconomic aggregates. To this end, the authors use the recently created BREL now-casting platform that consists of targeted bridge models for different data availability scenarios. BREL is based upon an elastic-net regression approach that takes into account the ragged-edge nature of the data set. The results of their empirical analysis suggest that survey data clearly help to predict Belgian (but also European) macroeconomic developments, in particular for earlier estimates, when the relevant hard data, notably firms’ turnover and industrial production, are not yet available. They also show that forecast accuracy is higher when using disaggregated survey results, rather than just the headline consumer confidence and business sentiment indicators. In this connection, demand expectations in the manufacturing industry and the unemployment expectations in the consumer survey consistently feature among the best predictors for real GDP growth.