台灣總體計量月模型之人口結構與股票價格探討

Abstract

整體社會中人口年齡結構的分布，是影響總體經濟的重要議題。不同的年齡層有不同的消費需求，年齡結構的改變也會影響民眾消費、投資與儲蓄的意願，進而影響經濟體系中的物價、薪資水準與匯率。隨著醫療水準的進步，過去幾十年間台灣人口的平均壽命有顯著的延長現象，而生育率的下降也可能在未來會影響經濟的永續發展。此外，隨著民眾對理財觀念的重視，股價指數的變動可以評估人們財富水準的變化。 本研究的目的即冀望應用統計回歸、時間序列的方法來建構總體計量模型，並計算當扶幼比、扶老比、台灣集中市場股價指數大幅變動衝擊下，對總體經濟造成的影響。 總體經濟模型之研究，主要有從需求面與供給面分析兩種，考量到本模型的衝擊類型多為需求面的影響，因此與大多數國內文獻相同，從需求面開始分析。由於部分資料的起始時間較晚，為了增加樣本數，因此嘗試建立總體計量月模型來對經濟進行評估。並將季資料與年資料，針對不同資料的特性，利用不同的方法轉換成月資料。此外，在模型中對消費的分類有別於大部分總體計量模型，而主要分成食物衣服、能源交通、教育醫療三大類型，以期能捕捉人口結構改變對消費的影響。 模型中共有87 條方程式，其中有55 條結構方程式與32 條定義式，87 個內生變數與23 個外生變數。模型靜態測驗的期間為1994 年4 月到2009 年12 月，並預測2010 年1 月到2012 年12 月的國內經濟走勢。樣本外的預測與情境分析的結果大多符合經濟直覺，文末則點出未來維護及修正模型的方向。

The distribution of population age structure is an important issue affecting the overall economy. Since different age groups have different consumption demands , age structure change will affect the public consumption, investment and savings at first , then affect the economy and thus the price, wage level and the exchange rate. With advances in medical standards, average life expectancy of Taiwan's population over past few decades have significant extension, while decline in fertility may also affect economy’s sustainable development in the future. In addition to , as public spend more time managing their wealth now, stock index is a useful instrument to evaluate changes in people's wealth level. The purpose of this study is to apply statistical regression, time series methods to construct monthly macroeconomic model, and simultaneous impact of large changes from young dependency ratios, elderly dependency ratios, and stock price index , to the overall economy. While most of macroeconomic data are built on season or year , consider different datas’ characteristic ,I translate them into month and year data by different methods. In addition, in order to capture the demographic changes on consumption. I divided consumption into three main types , which are food and clothing consumption , energy and transportation expenditure, education and health care expenditure . There are 87 model equations in this article , including 55 structural equation and 32 definition equations , 87 endogenous variables and 23 exogenous variables. The model’s period is from April 1994 to December 2009, and forecast in January 2010 to December 2012 for domestic economic trends. Sample forecasts and scenario analysis results are in keeping with economic intuition .