Real-time GDP forecasting for Japan: A dynamic factor model approach

• We develop a single-index dynamic factor model using mixed-frequency data.
• We obtain early estimates of Japan’s quarterly GDP growth in real time.
• The model performs well in terms of its accuracy to predict actual GDP estimates.
• The model performs well in terms of its ability to signal turning points in GDP.
• The model is able to assess ongoing economic activities in Japan early on.

Accurate and timely information on GDP is important to gauge the overall state of the economy and is thus essential for economic policymaking. A single-index dynamic factor model is estimated using mixed-frequency data on GDP, industrial production, employment, private consumption and exports to obtain early estimates of Japan’s quarterly GDP growth in real time. The results of a real-time forecasting exercise suggest the model performs well in comparison to the consensus forecasts, in terms both of its accuracy, measured by the size of forecast errors, to predict actual GDP estimates, and of its ability to signal turning points in GDP, showing the advantage of using the model for early assessment of ongoing economic activities in Japan. An equally important goal of this study is to share with other forecasters the results of ongoing real-time GDP forecasts for Japan, aiming at increasing knowledge regarding Japan’s GDP forecasting.