Nowcasting GDP growth rate: a potential substitute for the current flash estimate
Currently Statistics Netherlands provides a first estimate of the GDP growth rate 45 days after the end of the reference quarter. This paper investigates whether nowcasting could provide a good substitute for the current estimation method. We found that the best performing nowcasting model has a margin of uncertainty very close to the current ’flash’ estimates. Importantly, it is demonstrated to suffer from a significantly smaller bias than the current method does.
Currently Statistics Netherlands provides a first estimate of the GDP growth rate 45 days after the end of the reference quarter. Since a lot of information on both the consumption and production side is still missing by that time, this first estimate often deviates substantially from the final growth rate. Therefore, this paper investigates whether nowcasting could provide a good substitute for the current estimation method. We found that the best performing nowcasting model has a margin of uncertainty very close to the current ’flash’ estimates. Importantly, it is demonstrated to suffer from a significantly smaller bias than the current method does.
