Page 12 - 陳慧光

表 2: PCE vs. PPI 及其細項之交叉驗證表

組別 小組 變數 A 之主項 / 子項 變數 A 變數 B 之主項 / 子項 變數 B
13-1 PCE PPI
13
13-2 PPI PCE
14-1 Core PCE Core PPI
14
14-2 Core PPI Core PCE
15-1 Core PCE PPI
15
15-2 PPI Core PCE
16-1 Core PPI PCE
16
16-2 PCE Core PPI

17-1 PCE Goods PPI Final demand goods
17
17-2 PPI Final demand goods PCE Goods
18-1 PCE / Goods Durable goods PPI Final demand goods
18
18-2 PPI Final demand goods PCE / Goods Durable goods
19-1 PCE / Goods Nondurable goods PPI Final demand goods
19
19-2 PPI Final demand goods PCE / Goods Nondurable goods
20-1 PCE Services PPI Final demand services
20
20-2 PPI Final demand services PCE Services
Final demand services less
21-1 Core PCE PPI / Final demand services trade, transportation, &
warehousing
21
Final demand services less
21-2 PPI / Final demand services trade, transportation, & Core PCE
warehousing

3.2 研究模型

3.2.1 單根檢定模型

統計學上常見透過最小平方法 (ordinary least squares, OLS) 找出

將殘差平方和降到最低的斜率項與截距項係數，進而產生擬合數據的

迴歸線。由於許多金融和總經數據存在著時間趨勢特徵或均值的非定

態現象，例如: 隨機漫步 (random walks) 或存在漂浮項 (drift element)

10