Premium
UK regional nowcasting using a mixed frequency vector auto‐regressive model with entropic tilting
Author(s) -
Koop Gary,
McIntyre Stuart,
Mitchell James
Publication year - 2020
Publication title -
journal of the royal statistical society: series a (statistics in society)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.103
H-Index - 84
eISSN - 1467-985X
pISSN - 0964-1998
DOI - 10.1111/rssa.12491
Subject(s) - nowcasting , autoregressive model , aggregate (composite) , econometrics , computer science , exploit , vector autoregression , economics , geography , meteorology , composite material , computer security , materials science
Summary Output growth data for the UK regions are available at only annual frequency and are released with significant delay. Regional policy makers would benefit from more frequent and timely data. We develop a stacked, mixed frequency vector auto‐regression to provide, each quarter, nowcasts of annual output growth for the UK regions. The information that we use to update our regional nowcasts includes output growth data for the UK as a whole, as these aggregate data are released in a more timely and frequent (quarterly) fashion than the regional disaggregates which they comprise. We show how entropic tilting methods can be adapted to exploit the restriction that UK output growth is a weighted average of regional growth. In our realtime nowcasting application we find that the stacked mixed frequency vector‐autoregressive model, with entropic tilting, provides an effective means of nowcasting the regional disaggregates exploiting known information on the aggregate.