國際能源組織 Berkeley lab 有研發 PAMS model 以協助開發中國家預估其自身國家的家電產品飽和度，並輔助政府制定相關的節能政策。鑒於台灣近幾年夏季耗電量持續攀升，從 2015 年的 695 億度、2016 年的 709 億度、一路增至 2017年的 724 億度，為結合 PAMS model 使政府將來可以做節能政策的預期評估，故本研究運用 2001 年至 2018 年主計處家庭收支調查資料，以建立冷暖氣機的擁有率和擁有數量之預估模型。其中，使用 probit model 來建立冷暖氣機擁有率的預估模型，使用 ordered probit model 來建立冷暖氣機擁有數量之預估模型。 依 probit 模型實證結果顯示，戶長所得、住宅面積和住宅所屬等家庭特性為冷暖氣機擁有率預估模型之重要變數；依 ordered probit 模型實證結果顯示，家戶人口數、戶長所得、住宅面積和住宅所屬等家庭特性為冷暖氣機擁有數量預估模型之重要變數。接續，利用 Chow test 找出上述兩個模型中之重要變數發生結構性轉變的情況，並利用迴歸模型預估出重要參數未來的變化。 最終，將本研究有考量參數隨時間變化之預估結果，與 PAMS model 採用固定參數之預估方式做比較，分析此兩種預估擁有率和擁有數量上的差異。;The International Energy Organization Berkeley Lab has developed a PAMS model to assist developing countries in estimating the saturation of home appliances in their own countries and assist the government in formulating relevant energy-saving policies. In view of the fact that Taiwan′s power consumption in summer has continued to rise in recent years, from 69.5 billion degrees in 2015 to 72.4 billion degrees in 2017. I used the survey data of income and expenditure of the household from 2001 to 2018 on my study to establish a model for estimating the ownership and number of air-conditioners in order to make the government expect energy-saving policies in the future easily. According to the empirical results of the probit model, household head’s income,residential area, and tenure of dwelling are important variables for the forecasting model while in ordered probit model, number of household persons, household head’s income, residential area, and tenure of dwelling are important variables. Then, find out structural changes of important variables through Zou test and predict the future parameters by multiple regression model. Finally, estimated results of this study using the parameters changing with time would be compared to the results of using fixed parameter like PAMS model and analyze the difference between the results of these two methods.
