Summary:
As GDP can comprehensively reflect the economic condition of a country or a region, GDP predictions are carefully scrutinized by many institutions. However, because GDP is usually only calculated on a quarterly frequency and released after a delay of 3 weeks, classical forecasting models cannot provide accurate and timely GDP predictions. However, some macroeconomic variables that are highly correlated with GDP, such as industrial added value, import and export volumes, and the total retail sales of consumer goods, are released monthly with a much smaller delay. The incorporation of this monthly information into GDP prediction could therefore improve the timeliness of GDP forecasting, enable the correct evaluation of economic conditions, and facilitate the formulation of appropriate macroeconomic regulations. However, the incorporation of these monthly indicators into economic forecasting models will require the solution of key problems associated with these indicators' underlying data, namely its mixture of data frequencies, the ragged-edge behavior of real-time data, the presence of data revision, and the periodic absence of data. To deal with the problems of absent data and ragged-edge data, we nowcast China's GDP based on Zheng and Wang's (2012, 2013) mixed-frequency dynamic factor model for year-on-year growth rates. Compared with mixed data sampling (MIDAS) and mixed-frequency vector autoregression (MFVAR) models, the mixed-frequency dynamic factor model accounts for missing data in addition to dealing with ragged-edge data, and thus makes full, accurate, and timely use of the data. In addition, a year-on-year growth rate model is more useful in China, as the National Bureau of Statistics announces only year-on-year growth rates for most macroeconomic indicators, and policymakers focus on year-on-year GDP growth rates. Moreover, as year-on-year growth rates are based on data for the same month or quarter each year, they can mitigate effects due to seasonality, which is not generally accounted for in official year-on-year economic growth data released in China (and is another reason why a year-on-year growth rate model is appropriate for China). Thus, once new data are released, we can immediately update our nowcasting result. This means that we can nowcast the quarterly GDP growth rate in real time using the most up-to-date data and thus provide a reliable and timely economic prediction for decision-makers. Our results also show that the mixed-frequency dynamic factor model provides more accurate predictions than the MIDAS and the MFVAR models. In addition to developing GDP nowcasting, we derive some other important results. First, in contrast to GDP data, which are announced quarterly, in April, July, October, and January, with an approximately 3-week delay, we can estimate the quarterly GDP growth rate on a monthly basis. For example, we estimate the year-on-year growth rate of GDP from February to April, which is crucial information for decision-makers and for economic modeling. Second, in addition to determining the smoothed estimator of a common factor, we also obtain the smoothed estimator of the idiosyncratic factor. The sum of these factors is then used to derive a smoothed estimate of the monthly year-on-year growth rate of real GDP. Third, we circumvent the missing data problem, which is due to the restriction of statistical rules, the effect of the Spring Festival, and other factors, by estimating the missing data from the observations of other indicators, which affords a complete time series of data. In summary, our GDP nowcasting method enables daily (rather than quarterly) forecasting of China's quarterly GDP growth rate, which means that we can incorporate the latest information about economic conditions into our forecasts. Thus, we can provide more timely and reliable economic predictions to policymakers. From the perspective of macroeconomic regulation, these economic predictions may allow policymakers to generate real-time updates of its judgments on current economic conditions and thereby formulate more timely and suitable macroeconomic policies. From the perspective of microeconomic decision-making, these economic predictions may enable enterprise managers to understand the current economic situation more clearly and in a timely fashion, and thereby to efficiently adjust investment plans and development strategies. Thus, we believe that our nowcasting-based economic predictions will be invaluable for developing more effective national-level macroeconomic control and enabling better market-level microeconomic decisions.
王霞, 司诺, 宋涛. 中国季度GDP的即时预测与混频分析[J]. 金融研究, 2021, 494(8): 22-41.
WANG Xia, SI Nuo, SONG Tao. Nowcasting China's Quarterly GDP Using Mixed-Frequency Data. Journal of Financial Research, 2021, 494(8): 22-41.
[1] 陈磊、隋占林和咸金坤,2016,《我国经济增长速度的实时预测》,《吉林大学社会科学学报》第9期,第43~51页。 [2] 耿鹏和齐红倩,2012,《中国季度GDP实时数据预测与评价》,《统计研究》第1期,第8~14页。 [3] 刘汉和刘金全,2011,《中国宏观经济总量的实时预报与短期预测—基于混频数据预测模型的实证研究》,《经济研究》第3期,第4~17页。 [4] 王维国和于杨,2016,《基于混频回归类模型对中国季度GDP的预报方法研究》,《数量经济技术经济研究》第4期,第108~125页。 [5] 张劲帆、刚健华、钱宗鑫和张龄琰,2018,《基于混频向量自回归模型的宏观经济预测》,《金融研究》第7期,第34~48页。 [6] 郑挺国和尚玉皇,2013,《基于金融指标对中国GDP的混频预测分析》,《金融研究》第9期,第16~29页。 [7] 郑挺国和王霞,2012,《一种基于混频数据的宏观经济景气一致指数》,《21世纪数量经济学》第12卷。 [8] 郑挺国和王霞,2013,《中国经济周期的混频数据测度及实时分析》,《经济研究》第9期,第58~70页。 [9] Bai J., 2003. “Inferential Theory for Factor Models for Large Dimensions”, Econometrica, 71:135~172. [10] Banbura M., D. Giannone, and L. Reichlin, 2011. “Nowcasting”, in Clements M. P. and Hendry (eds),The Oxford Handbook of Economic Forecasting, New York: Oxford University Press. [11] Camacho M., and G. Perez-Quiros, 2010. “Introducing the EURO-STING: Short-term Indicator of Euro Area Growth”, Journal of Applied Econometrics, 25:663~694. [12] Clements M.P. and A. Galvão, 2008. “Macroeconomic Forecasting with Mixed-frequency Data”, Journal of Business and Economic Statistics, 26: 546~554. [13] Clements M.P. and A. Galvão, 2009. “Forecasting US Output Growth Using Leading Indicators: an Appraisal Using MIDAS Models”, Journal of Applied Econometrics, 24: 1187~1206. [14] Croushore D. and T. Stark, 2001. “A Real-time Data Set for Macroeconomists”, Journal of econometrics, 105(1): 111~130. [15] Croushore D. and T. Stark, 2003. “A Real-time Data Set for Macroeconomists: Does the Data Vintage Matter?”, Review of Economics and Statistics, 85(3): 605~617. [16] Diebold F. X. and R. S. Mariano, 1995. “Comparing predictive accuracy”, Journal of Business and Economic Statistics, 13: 253~265. [17] Giannone D., L. Reichlin and D. Small, 2008. “Nowcasting: The Real-time Informational Content of Macroeconomic Data”, Journal of Monetary Economics, 55(4): 665~676. [18] Kim C. and C. Nelson, 1999. “State-Space Models with Regime Switching”, London: the MIT press. [19] Kuzin V., M. Marcellino and C. Schumacher, 2011. “MIDAS vs. Mixed-frequency VAR: Nowcasting GDP in the Euro Area”, International Journal of Forecasting, 27(2): 529~542. [20] Marcellino M., M. Porqueddu and F. Venditti, 2016. “Short-term GDP Forecasting with a Mixed-frequency Dynamic Factor Model with Stochastic Volatility”, Journal of Business & Economic Statistics, 34(1): 118~127. [21] Mariano B. S. and Y. Murasawa, 2003. “A New Coincident Index of Business Cycles Based on Monthly and Quarterly Series”, Journal of Applied Econometrics, 18: 427~443. [22] Newey W. K. and K. D. West, 1987. “A Simple, Positive Semi-definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix”, Econometrica, 55: 703~708. [23] Nunes L.C., 2005. “Nowcasting Quarterly GDP Growth in a Monthly Coincident Indicator Model”,Journal of Forecasting, 24: 575~592. [24] Schorfheide F. and D. Song, 2015. “Real-time Forecasting with a Mixed-frequency VAR”, Journal of Business & Economic Statistics, 33(3): 366~380. [25] Stock J. H. and M. W. Watson, 1989. “New Indexes of Coincident and Leading Economic Indicators”, NBER macroeconomics Annual: 351~394. [26] Stock J. H., and M. W.Watson, 1991. “A Probability Model of the Coincident Economic Indicators”, In Lahiri, K. and Moore, G.H. (eds), Leading Economic Indicators: New Approaches and Forecasting Records, (Cambridge University Press, Cambridge): 63-89. [27] West K. D., 1996. “Asymptotic Inference about Predictive Ability”, Econometrica, 64: 1067~1084.