國內生產毛值 (Gross Domestic Product, GDP) 是建構景氣基準循環時一個最主要的參考依據。 由於 GDP 數列本身為季資料, 因此進行總體模型分析時必需要將季資料轉換成月資料以符合計算上的需求。 本研究主要目的是依據狀態空間(state-space)計量模型,嘗試重新建構適當的GDP月資料指標以供政府經濟決策之參考。 從計量經濟的角度來看, 此一模型的優點在於其能包含文獻上常用的其它模型, 並且放寬即有共整合的假設, 以建構出適當的GDP月資料指標。 若從實證結果來看, 本研究所估計出的 GDP 月指標與 GDP 季資料有許多相似的特徵。故我們相信, 利用此模型所建構出的月指標確實可以提供政府短期經濟決策時之參考。

In this paper, we describe a framework that nests a wide range of
interpolation/distribution setups but relaxes the co-integration con-
dition of temporal disaggregation that has been reported in the lit-
erature. Our goal is to evaluate alternative interpolation/distribution
models and then generate the monthly deseasonalized real gross do-
mestic product of Taiwan. Our empirical results show that themonthly
estimates, incorporated in the information obtained from the indus-
trial production index, are highly consistent with quarterly figures.
These estimates should be invaluable to researchers and practitioners
for short-run policy analysis in that they signal any emerging eco-
nomic problems.