| 2005年中国税收收入预测模型比较 |
| |
| 引用本文： | 尚红云. 2005年中国税收收入预测模型比较[J]. 统计与信息论坛, 2008, 23(3): 27-31 |
| |
| 作者姓名： | 尚红云 |
| |
| 作者单位： | 东北财经大学,统计学院,辽宁,大连,116025 |
| |
| 摘 要： | 为探索一种较为有效的工具来提高税收收入预测精度,利用1985-2004年的样本数据,建立了五个模型来预测中国2005年的税收收入。结果表明：ARMA（1,1）模型中,以GDP为外生变量的自回归模型、以政策因素为虚拟外生变量的自回归模型以及对数线性移动平均模型都是预测税收收入的有效模型,但以GDP为外生变量的自回归模型在预测2005年税收收入时,预测值与实际值的预测偏差仅有1．23％,此模型在预测税收收入时预测精度最高,是预测税收收入的一种较为有效的工具。
|
| 关 键 词： | 中国税收 半对数模型 自回归模型 虚拟变量 |
| 文章编号： | 1007-3116(2008)03-0027-05 |
| 修稿时间： | 2007-10-09 |
The Comparison on Forecast Tax Revenue Model at 2005 |
| |
| SHANG Hong-yun. The Comparison on Forecast Tax Revenue Model at 2005[J]. Statistics & Information Tribune, 2008, 23(3): 27-31 |
| |
| Authors: | SHANG Hong-yun |
| |
| Affiliation: | SHANG Hong-yun (Dongbei University of Finance and Economics, Dalian 116025, China) |
| |
| Abstract: | The paper establishes five models to forecast tax revenue based on the data from the year1985 to 2004. The results demonstrate that： ARMA（1,1） model, autoregressive model with GDP as exogenous vari- able, auto- regressive model with policy as exogenous dummy variable, and logarithm linearity and removal average model are all valid models for forecasting tax revenue. But auto-regressive model with GDP as exogenous variable is the most valid model among the five models and the deviation is only 1.23 %. |
| |
| Keywords: | China＇s tax part- logarithm model auto-regressive model dummy variable |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
