由於國內外經濟活動日益頻繁,且資訊之取得與流通愈發迅速與便利,使得整體經濟體系的變化更加快速且複雜。因此,高頻率資訊的蒐集與分析,變成預測研究中非常重要的一個課題,而探討用高頻率資料來改進低頻率預測的研究也日益增加。 本研究以Klein教授建立的即期季模型(Current Quarterly Model, CQM)爲藍本,建立以台灣資料爲依據的「總體經濟即期季模型」,藉以改進現有的總體經濟季預測。本研究分別從GDP支出面,以及Principal Component統計分析,估測名目與實質GDP,研析判斷兩者之估測值,而得到模型之研究結果。 CQM模型的預測過程,主要在於得到初步預測值,以爲其它模型爰用,如作爲傳統總體計量模型預測修正調整方向的參考,或做爲其基準解(benchmark),並以此結果設定常數項調整值(add factor)之參考。亦即利用CQM模型之資料導向(data driving)特點,將其視爲傳統總體計量模型之輔助工具。另外,本模型且可納入每月最新資訊,重新估測經濟成長率,充分利用當前資訊快速流通之優點,提供國內最新景氣預測資訊。
The increased interaction between international and domestic economic activities along with the ease of access to information have made changes in the economy more rapid and complicated, whilst also making economic forecasting more difficult than ever. The only real way to deal with the problem is to grasp and analyze current changes promptly, revising forecasts as necessary. Therefore, the collection and utilization of high frequency data to improve low frequency forecasts has become an important issue in recent forecasting literature. Our study is based on this trend and follows Nobel laureate, Lawrence Klein's approach to the construction of a Current Quarterly Model (CQM) for Taiwan. The CQM model predicts quarterly GDP growth from the GDP expenditure side and from the approach of principal components. On the expenditure side, we build Bridge equations for NIPA (which includes private consumption, private investment, government sector and exports imports) and the high frequency leading indicators. In the principal components approach, we abstract some leading indicators to predict the nominal GDP, real GDP and GDP deflator. We then average out the predicted values to obtain the final forecasting results. The purposes of the CQM are to obtain the initial predicted values as a benchmark, and to add factors for a traditional macroeconometric model in order to improve the forecast results, which should then enable us to release the latest GDP growth forecast and utilize recent developments in information networks within the international institutes.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。