頁籤選單縮合
題 名 | 臺灣地區綜合所得稅收入預測方法之研究=Studies on Methods of Individual Income Tax Revenues Prediction in Taiwan |
---|---|
作 者 | 古坤榮; 林全; 鄭光甫; | 書刊名 | 國家科學委員會研究彙刊. 人文及社會科學 |
卷 期 | 8:2 1998.04[民87.04] |
頁 次 | 頁276-287 |
分類號 | 567.218 |
關鍵詞 | 納稅單位; 綜合所得稅; Filing cases; Individual tax; |
語 文 | 中文(Chinese) |
中文摘要 | 過去國內、外預測稅收入的一般研究文獻,通常包括兩大類型。其一是建立經 濟結構模型,利用傳統計量方法進行估測,此種估計往往受限於模型的簡化性。因此,不易 對個別稅收水準加以細分,或由於計量方法的限制,估計的信賴難以令人滿意。其次一種分 析方法,是採取時間序列分析方法,此種方式的缺點是缺乏對經濟或環境因素的考慮,因此 對長期間的估計水準,較難掌握。上述兩種估計方法也同時面臨一個共同缺陷,即難以預測 稅制變動對稅收水準的影響。 本文根據鄭光甫、林全( 1992 )過去的研究基礎,結合統計分析的有關技巧,應用 在綜合所得稅的稅收估計模型建立工作上。其中主要特色包括:( 1 )建立綜合所得稅收 入估計模型,並以估計不同年度綜合所得稅收入的成長率(Δ TR △△)為對象;( 2 ) 利用最適化模型及計量方法,先建立「次模型」,從而建立一更具實用性及精確性的估計模 型。 |
英文摘要 | In the past, the tax models were usually developed by employing economic structural models or univariate time series models. However, in general, on hand one finds that the economic models are not refined enough to have accurate prediction. On the other hand, the latter approaches involve no exogenous variables and hence can not provide adequate economic explanation. Further, these methods have the disadvantage that the effects of the changes in tax laws can not be properly evaluated. In the prese t project, we follow the previous results by Cheng and Lin (1993) to establish better models for predicting individual income tax revenues. The basic approach is to first develop models for individual income tax and the estimate the growth rates of these taxes revenues. Regarding the individual income tax model, we consider the estimation of the conditional distribution of the filing cases with respect to the income brackets given exogenous variables. Then apply usual regression analysis methods to estimate the total tax. |
本系統中英文摘要資訊取自各篇刊載內容。