协整检验及误差修正模型的建立.doc

《协整检验及误差修正模型的建立.doc》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、协整检验及误差修正模型的建立　1978年-2003年度北京农村居民人均纯收入与人均消费支出的统计表单位：（元/人）年份197819791980198119821983198419851986人均纯收入R193.25191.01191.47221.34241.64238.22229.29239.05253.45人均消费支出P184.89175.97155.07180.99197.75197.81186.63187.26210.6年份198719881989199019911992199319941995人均纯收入R285.77343.54386

2、.02399.17408.09406.86486.01584.39661.74人均消费支出P243.52290.85315.9356.92336.45329.9403.66504.15552.83年份19961997199819992000200120022003人均纯收入R743.29811.19848.58904.13925.54964.711015.921046.33人均消费支出P594.86593.65620.57626.11617.22633.6671.1692.73一．为了方便计算，首先将数据取对数化得到新的统计表：obsLNPL

3、NR19785.2197610536422615.26398468746753219795.170313525963675.25232578269704419805.0438766278760585.25473113838511419815.1984417811207815.39969998083663919825.2870036066477645.4874490152705419835.2873069740279845.47319461645275519845.2291280471950385.43498757792290219855.23

4、24980255107435.47667273507211219865.3499605996998755.53516656491835919875.4951990746733215.6551872914767419885.6728076703567715.8393035532088819895.755425707808045.95588918159406119905.8775116670319515.98938739131278219915.8184495496343266.01148773831344919925.7987895782345

5、366.00846914592380619936.0005729394719736.18622919982049519946.2228738428392246.3705685681627219956.315050540156056.4948727295461119966.3883260837419796.61108627812265419976.3862899201118016.69850230534476319986.4306384105300266.74356436426046419996.439526074513196.80697315

6、536291920006.4252255243709946.83037735096074520016.4514178405002166.87182753803964820026.5089181571641946.92354988488077720036.5406403128824456.95304408244083二．检验时间序列的平稳性从图中发现，LNP和LNR有大致相同的增长和变化趋势，说明二者可能存在协整关系。此时间序列存在随时间增长的趋势，因此，模型应存在时间项。三．对LNP进行ADF检验NullHypothesis:LNPhasau

7、nitrootExogenous:Constant,LinearTrendLagLength:3(Fixed)t-Statistic  Prob.*AugmentedDickey-Fullerteststatistic-2.226120 0.4533Testcriticalvalues:1%level-4.4407395%level-3.63289610%level-3.254671从中我们看到t统计量是-2.2261,大于显著性水平为1%(5%,10%)的临界值,表明序列LnP是不平稳的。四．对LNR进行ADF检验NullHypothesi

8、s:LNRhasaunitrootExogenous:Constant,LinearTrendLagLength:3(Fixed)t-Statistic  Prob