1 论文标题
Nowcasting China Real GDP

2 作者信息
Domenico Giannone,
Silvia Miranda-Agrippino,
Michele Modugno

3 出处和链接（比如，NBER working paper No.11000）
第九届CIRANO-CIREQ 会议：数据更新（Data Revision）与宏观经济预测和宏观经济政策

4 摘要
In this paper we nowcast mainland China real GDP estimating a Dynamic Factor
Model on a set of mixed frequency indicators to exploit the information content of
more timely data in producing early forecasts of GDP gures. Estimating the model
via quasi maximum likelihood (Doz et al, (2012), Banbura and Modugno (2012)) allows
to eciently handle complications arising from the dissemination pattern and transfor-
mations of Chinese data types; moreover, model-based news can be used to sequentially
update GDP nowcasts allowing for a meaningful interpretation of the revision process.
We evaluate the model running a pseudo real time forecasting exercise over the period
2008 - 2013 and show how the model eciently processes new information increasing
forecasting accuracy as more data releases are taken into account; this ultimately de-
livers a forecast which is comparable to the best judgemental forecasts available within
nancial markets, while signicantly outperforming the standard AR benchmark. We
also evaluate model performance against institutional forecasters in terms of annual
growth rates: we show how model-implied annual rates become extremely accurate
at Q4, beating virtually all the surveyed benchmarks, with performance expected to
further improve towards the end of the year.



这是今天刚结束的会议的论文