The purpose of this study is two-fold. In the first part, we examine how to estimate and forecast Gyeonggi-do's real Gross Regional Domestic Product(GRDP) based on a regression analysis. In addition to a simple linear regression with a time trend, and...
The purpose of this study is two-fold. In the first part, we examine how to estimate and forecast Gyeonggi-do's real Gross Regional Domestic Product(GRDP) based on a regression analysis. In addition to a simple linear regression with a time trend, and an autoregressive error term, this study also considers Time Varying Coefficient-Error Correction Model(TVC-ECM).
In the second part, we develop a short-run forecasting model for the Gyeonggi-do regional economy based on Factor Augmented Vector AutoRegressive model (FAVAR) to measure the impact on the Gyeonggi-do economy of an external shock such as a change in domestic monetary policy or foreign exchange rates.
Considering the untimely manner in which the GRDP data is released, we propose that a simple regression analysis with Korea's real Gross Domestic Product as an explanatory variable and an autoregressive error term is helpful in estimating and forecasting the Gyeonggi-do's real GRDP.
In addition, this study finds that the explanatory power of TVC-ECM utilizing the data of the Gyeonggi-do's regional electricity demand is relatively high as compared with other specifications and hence, can be a good candidate for forecasting purpose.
From the FAVAR exercise analysing the effect of the monetary policy shock on the Gyeonggi-do economy, we find that a contractionary monetary policy shock has a bigger negative impact on consumer price index, employment measures, sales of large-scale retailers, etc in the national level than the Gyeonggi-do regional level. On the other hand, the same shock has a g reater negati ve effect on construction starts, and unemployment rate in the Gyeonggi-do region than nationwide.
This study shows that the FAVAR approach can be useful in investigating the short-run effect of an external shock on a regional economy in sense that it not only exploits relatively limited regional macroeconomy data but also makes an analysis richer since the model can usually include more various economic variables than the conventional VAR analysis.