Metrology professional thesis

econometrics

Final experimental report

Experimental name: Analysis of per capita consumption expenditure of urban residents in large and medium-sized cities and its influencing factors.

Last name:

Student number:

Class level:

Instructor:

Time:

Per capita consumption expenditure of urban residents in 23 cities

Analysis of its influencing factors

I. Background of economic theory

In recent years, China's economy has maintained a rapid development momentum, and investment, export and consumption have formed a "troika" to promote economic development, which has been recognized by all walks of life. By establishing an econometric model and using econometric analysis methods, this paper analyzes the factors affecting the per capita consumption expenditure of urban residents, finds out the key influencing factors, and provides some references for policy makers, so as to finally promote the "carriage" of consumer demand to become the cornerstone of China's healthy, rapid and sustainable economic development.

Second, the per capita consumption expenditure theory and its influencing factors

We mainly analyze the influencing factors of China residents' consumption expenditure from the following aspects:

(1) The expected increase of residents' future expenditure affects the growth of residents' spot consumption.

Residents' passive savings directly lead to a huge diversion of purchasing power, thus weakening the immediate demand for consumer goods, seriously affecting the growth of residents' immediate consumption, leading to insufficient effective demand, and ultimately leading to weak economic growth. Since the late 1990s, a series of reform measures such as medical care, pension, unemployment insurance and education have been introduced one after another, and the original system has been broken, but the new system has not yet been established and improved. Therefore, the current medical care, pension, unemployment insurance and education systems have great pressure on residents' personal expenditures, which are basically rigid expenditures, and the uncertainty of expenditures is also great, which leads to the increase of residents' expectations for future expenditures.

② The structural contradiction between supply and demand of commodities is still outstanding.

From the perspective of consumption structure, China's consumer goods market has undergone new fundamental changes: residents' low-level consumption is almost saturated, while higher-level consumption has not yet reached. After more than 20 years of reform and opening up, urban and rural residents have gone through the popularization stage of mid-range durable consumer goods. At present, the income consumption of ordinary people is not enough to form a new leading consumption hotspot with high-end products as the content, such as cars and houses, which is far from being included in the mainstream consumption of most people, and the existing purchasing power of residents cannot form the driving force to promote the upgrading of leading consumer goods.

(3), the overall price level continues to run at a low level, and the pressure of deflation is greater, which is not conducive to the growth of consumption.

After China's entry into WTO, with the reduction of tariffs and the expansion of import scale, the impact of foreign products on China's market will further increase, and international price tightening will have a negative impact on domestic price changes. The continuous decline in prices is not conducive to the consumption growth of residents. Because from the consumer psychology of residents, it is the habitual psychology of residents to buy up and not buy down. Because residents expect prices to fall further, they often delay consumption, which is not conducive to the growth of residents' consumption. In addition, statistically speaking, due to falling prices, nominal consumption growth is often lower than actual consumption growth, which is not conducive to the improvement of consumption growth to some extent.

At present, there is no big consumption hotspot in China, which makes it difficult to drive the rapid growth of consumption.

After cultivation and development in recent years, China has formed some consumption highlights such as housing consumption, residents' automobile consumption, communication and electronic products consumption, holiday consumption and tourism consumption, which can promote the steady growth of consumption, but it has never formed a big consumption hotspot, so it cannot drive the rapid growth of consumption.

Three. Relevant data collection

Relevant data are all from China Statistical Yearbook in 2006:

Basic situation of urban households in 23 large and medium-sized cities

Regional average employed population per household (person) Average burden of employed population per household (person) Average actual income per person per month (yuan) Per capita disposable income (yuan) Per capita consumption expenditure (yuan)

Beijing1.61.81865.11633.21187.9

Tianjin1.42.02010.61889.8 939.8

Shijiazhuang1.42.01061.310/0.0722.9

Taiyuan1.32.21256.91159.9 789.5

Hohhot1.51.91354.21279.8 772.7

Shenyang1.32.1148.51048.7812.5438+0

Dalian1.61.81269.81133.1.946.5

Changchun1.81.7111016.1690.2

Harbin 1.4 2.0 992.8 942.5 727.4

Shanghai1.61.91884.01686.11505.3

Nanjing1.42.01536.41394.0 920.6

Hangzhou1.51.91695.01464.91264.2.

Ningbo1.51.81759.41543.21271.4.

Hefei1.61.81042.5950.1686.9

Fuzhou1.71.91172.51059.4 942.8

Xiamen1.51.91631.71394.3 998.7.

Nanchang1.41.81405.01321.665.4

Jinan1.71.71491.31356.81071.4.

Qingdao1.61.81495.61378.51020.7.

Zhengzhou1.42.1.1012.2 954.2 750.3

Wuhan1.5 2.01052.5 972.2 853.438+0

Changsha1.42.11256.91148.9 986.8

Guangzhou1.71.81898.61591.1215.1

Fourthly, the establishment of the model

According to the data, we establish the general model of multivariate linear regression equation as follows:

These include:

-Per capita consumption expenditure

-Constant term

-Parameters of regression equation

-Average number of employed persons per household

-Average number of persons employed.

-Real monthly income per capita

-Per capita disposable income

-Random error term

Experimental process of verb (abbreviation of verb)

(A) Regression model parameter estimation

Establishing a multivariate linear regression equation according to the data;

Firstly, the model is estimated by OLS using Eviews software, and the sample regression equation is obtained.

Using Eviews, the output results are as follows:

Dependent variable: y

Methods: Least square method.

Date: 12/ 1 17 Time: 16:08

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c- 1682. 180 13 1 1.506- 1.282633 0.2 159

x 1 564.3490 395.2332 1.427889 0. 1704

X2 569. 1209 379.7866 1.498528 0. 15 13

X3 1.5525 10 0.62937 1 2.466766 0.0239

x4- 1. 180652 0.742 107- 1.590947 0. 1290

R squared 0.72 1234 mean correlation var 945.29 13.

The adjusted R-squared value is 0.659286 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 130.8502 information standard of red pool 12.77564

The sum of residual squares is 308 19 1.9 Schwartz criterion 13.02249.

Logarithmic likelihood-141.9199f-statistics 1 1.64259

Durbin-Watson statistics 2.047936 proband (F statistics) 0.000076

According to the output results of Eviews multiple linear regression, the estimated values of parameters can be obtained as follows:,,,

Therefore, the regression equation obtained preliminarily is:

se =( 13 1 1.506)(395.2332)(379.7866)(0.62937 1)(0.742 107)

t =(- 1.282633)( 1.427889)( 1.498528)(2.466766)(- 1.590947)

f = 1 1.64259 df = 18

Model test: because the test P values of explanatory variables,, and are all greater than 0.05, the variables are not significant, which indicates that there may be multiple * * * linearities in the model, and then the model is modified.

(B) dealing with multiple * * * linearity

We use the stepwise regression method to test and deal with the multilinear of the model:

X 1:

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:28

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 153.8238 5 18.6688 0.296574 0.7697

x 1 523.0964 34 1.4840 1.53 1833 0. 1405

R squared 0. 100508 mean correlation var 945.29 13.

The adjusted R-squared value is 0.057675 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 2 17.6 105 information standard of red pool 13.68623

Residual sum of squares 99444 1.2 Schwartz criterion 13.78497

Logarithmic likelihood-155.39 17 F- statistic 2.3465 1 1

Durbin-Watson statistics 1.770750 probability (f statistics) 0. 14049 1

X2:

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:29

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 1756.64 1 667.2658 2.63296 0.0 156

X2-424. 1 146 347.9597- 1.2 1886 1 0.2364

R squared 0.066070 average dependent variable 945.29 13

Adjusted R squared 0.02 1597 standard deviation dependent risk value 224. 17 1 1

The standard deviation of regression is 22 1.737 1 akachi information standard 13.72380.

Residual sum of squares 10325 15. Schwartz standard 13.82254

Logarithmic likelihood-155.8237 F- statistics 1.485623

Durbin-Watson statistics 1.887292 probability (f statistics) 0.2364 12

X3:

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:29

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 182.8827 137.8342 1.32683 1. 1988

X3 0.540400 0.095343 5.667960 0.0000

R squared 0.6047 12 mean correlation var 945.29 13.

The adjusted R-squared value is 0.585888 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 144.2575 information standard of red pool 12.86402

Residual sum of squares 4370 14.5 Schwartz criterion 12.96276

Logarithmic likelihood-145.9362 F- statistic 32. 12577

Durbin-Watson statistics 2.064743 proband (F statistics) 0.0000 13.

X4:

Dependent variable: y

Methods: Least square method.

Date: 12/ 1 17 Time: 16:30.

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 184.7094 16 1.8 178 1. 14 1465 0.2665

x4 0.596476 0. 12423 1 4.80 1338 0.000 1

R squared 0.523300 average dependent variable 945.29 13

The adjusted R-squared 0.500600 standard deviation depends on the risk value of 224. 17 1 1

Standard deviation of regression 158.4 178 red pool information standard 13.05 129.

The sum of residual squares is 527020. 1 Schwartz criterion 13. 15003.

Log-likelihood-148.0898 F- statistic 23000.000000000005

Durbin-Watson statistics 2.037087 proband (F statistics) 0.000096

As can be seen from the obtained data, the determination coefficient of adjustment is the largest, so the adjustment equation is introduced first, and then the variables,, and are introduced respectively, and OLS obtains:

X 1、X3

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:32.

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c-222.899 1 345.908 1-0.644388 0.5266

x 1 289.8 10 1 227.2070 1.275533 0.2 167

X3 0.5 172 13 0.095693 5.404899 0.0000

R squared 0.634449 average dependent variable 945.29 13

The adjusted R-squared value is 0.597894 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 142. 15 10 red pool information standard 12.87276

Sum of squares of residuals 404 138.2 Schwartz criterion 13.02087

Logarithmic likelihood-145.0368 F- statistics 17.35596

Durbin-Watson statistics 2.032 1 10 probability (f statistics) 0.000043.

X2、X3

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:33.

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 239.5536 53 1. 1435 0.45 10 15 0.6568

X2-27.0098 1 244.0392-0. 1 10678 0.9 130

X3 0.536856 0. 102783 5.22322 1 0.0000

R squared 0.604954 mean dependent variable 945.29 13

The adjusted R-squared value is 0.565449 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 147.7747 information standard of red pool 12.95036

Residual sum of squares 436747.0 Schwartz standard 13.09847

Logarithmic likelihood-145.9292 F- statistics 15.3 1348

Durbin-Watson statistics 2.063247 proband (F statistics) 0.000093

X3、X4

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:34.

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 33 1.70 15 142.5882 2.326290 0.0306

X3 1.766892 0.553402 3. 192782 0.0046

x4- 1.47372 1 0.656624-2.244390 0.0363

R squared 0.684240 mean dependent variable 945.29 13

Adjusted R-squared 0.652664 standard deviation dependent risk value 224. 17 1 1

Standard deviation of regression 132. 1 157 red pool information standard 12.72634

Sum of squares of residuals 34909 1.0 Schwartz criterion 12.87445

Logarithmic likelihood-143.3529 F- statistic 2 1.66965

Durbin-Watson statistics 2.11635 probability (f statistics) 0.0000 10.

From the data results, it can be seen that the adjustment coefficient of the equation is the largest when X4 is introduced, and all explanatory variables pass the significance test, and then X 1 and X2 are introduced respectively for analysis.

X 1、X3、X4

Dependent variable: y

Methods: Least square method.

Date: 12/ 1 17 Time: 16:37

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 193.6693 403.8464 0.479562 0.6370

x 1 89.29944 243.65 12 0.366505 0.7 180

X3 1.652622 0.646003 2.558228 0.0 192

x4- 1.34500 1.757634- 1.775265 0.09 19

R squared 0.686457 mean dependent variable 945.29 13.

The adjusted R-squared value is 0.636950 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 135.07 12 red pool information standard 12.80625

Residual sum of squares 346640.3 Schwartz standard 13.00373

Logarithmic likelihood-143.27 19 F- statistics 13.669 1

Durbin-Watson statistics 2.082 104 probability (f statistics) 0.000050.

X2、X3、X4

Dependent variable: y

Methods: Least square method.

Date:12/1/07 Time: 16:38.

Sample: 1 23

Comments included: 23.

Variable coefficient standard. Error t- statistics problem.

c 62.60939 489.2088 0. 12798 1 0.8995

X2 134. 1557 232.9303

X3 1.886588 0.600027 3. 144 175 0.0053

x4- 1.596394 0.70 10 18-2.27725 1 0.0345

R squared 0.689658 mean dependent variable 945.29 13

The adjusted R-squared value is 0.640657 standard deviation and the dependent variable is 224. 17 1 1

Standard deviation of regression 134.3798 information standard of red pool 12.79599

Sum of squares of residuals 343 100.8 Schwartz criterion 12.99347

Logarithmic likelihood-143. 1539 F- statistics 14.07429

Durbin-Watson statistics 2. 143 1 10 probability (f statistics) 0.000046

From the output results, it can be seen that at the level of, the test of P value sum of explanatory variables is greater than 0.05, and explanatory variables cannot pass the significance test, so it can be concluded that only two variables, X3 and X4, can be introduced into the model. The adjusted multiple linear regression equation is:

se =( 142.5882)(0.553402)(0.656624)

t =(2.326290)(3. 192782)(-2.244390)

F=2 1.66965 df=20

(3). Heteroscedasticity test

White test on the model:

White heteroscedasticity test:

F statistics 1.07 1659 probability 0.399378

Obs*R squared 4.423847 Probability 0.35 1673

Test equation:

Dependent variable: RESID^2

Methods: Least square method.

Date:12/1/07 Time: 16:53.