Frontiers in Business, Economics and Management ISSN: 2766-824X | Vol. 10, No. 3, 2023 122 The Impact of China's Urban Per Capita Disposable Income on Education, Culture and Entertainment Services Lingyi Zhao*, Binbin Xue, Xunzhang Tong, Xinxuan Dong School of Economics and Management, Communication University of China, Beijing 100024, China *Corresponding author: zhaoly_01@163.com Abstract: This study is from the perspective of China's cultural consumption, to study the impact of China's urban per capita disposable income on education, culture and entertainment service expenditure, with local financial budget revenue, consumer price index, regional GDP as control variables, so as to more accurate China's urban per capita disposable income on education, culture and entertainment service expenditure, Then explore how to expand China's cultural consumption in order to enhance the country's soft power. In this paper, the multiple regression model of economic variables and expenditure on education, culture and entertainment services is established, and the data of each year are found to form time series data, called panel data, and the necessary tests are carried out to obtain a more suitable and reliable model and the final regression results, which can provide direction for the government's economic policy. Keywords: Disposable income per capita, Expenditure on education, Culture and recreation services, Time series. 1. Introduction Since the reform and opening up, with the rapid growth of China's economy, the living standards of urban and rural residents have been continuously improved. According to the report of the National Bureau of Statistics, the living standards of urban and rural residents in China have been from basically eliminating poverty, to solving the problem of food and clothing, to realizing the overall well-off society, and are moving forward to the goal of building a well-off society in an all-round way. The income of Chinese urban and rural residents has increased significantly. The per capita net income of rural residents increased from 133.6 yuan in 1978 to 9892 yuan in 2014, with an average annual growth rate of 7. 1%. The per capita disposable income of urban residents increased from 343.4 yuan to 28,844 yuan, with an average annual growth rate of 7.2 percent in real terms. The consumption level of urban and rural residents has increased significantly, and the consumption structure of urban and rural residents has improved significantly. The Engel coefficient (the proportion of food expenditure in total consumption) has decreased significantly, and the proportion of development and enjoyment consumption has been increasing. The Engel's coefficient for rural households dropped from 67.7% in 1978 to 43.3% in 2014, and for urban households from 57.5% to 38.6%. People's living standards have improved significantly. As the problem of food and clothing has been solved, the proportion of development and enjoyment consumption in urban and rural residents' living consumption has increased. The per capita expenditure on cultural, educational and recreational goods and services of rural households rose from 5.1% in 1980 to 9.5% in 2014. The per capita expenditure on educational, cultural and recreational services of urban households also rose from 8.4% to 13.3% in consumption. The consumption market of education, culture and entertainment of urban residents has shown a trend of rapid development, which has become a new bright spot in stimulating household consumption and expanding domestic demand. The per capita disposable income of urban residents refers to the part of income that reflects the total cash income of a household that can be used to arrange the daily life of the household. Expenditure on education, culture and entertainment services consists of three major parts: cultural and entertainment articles, cultural and entertainment services and education expenditure. Expanding cultural consumption is an important measure to expand domestic demand, raise people's consumption level, promote the development of the cultural industry and enhance the country's soft power. Among them, increasing the expenditure on cultural, educational and entertainment services is of great significance in expanding cultural consumption. 2. Theoretical Analysis The explained variable is the per capita consumption expenditure of culture, education and entertainment services, which can directly reflect the changes in the industry of education, culture and entertainment services. The explanatory variable is selected as China's urban per capita disposable income, which is increasing year by year. One of the control variables is the local financial budget income. The higher the local financial budget income is, the higher the government's investment in cultural, educational and entertainment service consumer goods will be. The more and more perfect the types of consumer goods are, the higher the per capita expenditure on cultural, educational and entertainment service consumption will be. The second control variable is the consumer price index, which reflects the change in the price level of consumer goods. The higher the price of consumer goods of cultural, educational and entertainment services is, the lower the per capita consumption of cultural, educational and entertainment services will be. The third control variable is the regional GDP, which can directly reflect the local economic development. The higher the regional GDP is, the more perfect the regional consumption level of cultural, educational and entertainment 123 services will be. Where variable subscripts i and t represent the data of the ith individual in the TTH year respectively, is the intercept term to be estimated, is the marginal value, is the coefficient corresponding to the explanatory variable to be estimated, and is the random error term. Where, ln represents the data after logarithmic transformation. In order to make the linear trend of the data more significant and reduce the heteroscedasticity and collinearity of the model, the natural logarithm of the data is processed, where expand represents the per capita consumption expenditure of cultural, educational and entertainment services, income represents the per capita disposable income of urban China, fiscal represents the local fiscal budget revenue, and cpi stands for consumer price index, gdp stands for gross regional product, as shown in Table 1. Table 1. Variable selection and expected symbols Properties of indicators Variable symbol Variable name Expected symbol Explained variable lnexpand Log of per capita consumption expenditure on cultural, educational and entertainment services Explanatory variables lnincome Logarithm of per capita disposable income in urban China + Control variables lnfiscal Log of local fiscal budget revenue + lncpi Log of the consumer Price index - lngdp Log of gross regional product + 3. Data Description (1) Data collection and sorting Cross-section data intelligently analyzes the changes of a certain year, while time series data is easy to cause collinearity of variables and cannot accurately analyze the characteristics of data. Therefore, Therefore, this group downloaded the average consumption expenditure of cultural, educational and entertainment services, China's urban per capita disposable income, local fiscal budget income and consumer price index of 10 cities including Beijing, Tianjin and Guangdong in four years from 2009 to 2012 in 16 quarters from Zhonghong Statistical database of China Macro Economic Network. gdp represents the time series data of the five indicators of regional gross domestic product (GDP), which is integrated into panel data. (2) Descriptive statistics of the data Scatter plot was presented to preliminarily understand the relationship between explanatory variables and explained variables Figure 1: Figure 1. Scatter plot It can be seen that all explanatory variables and control variables have a positive linear trend with the explained variables, except that the scatter plots of lncpi and lnexpand are scattered, and the linear trend is not obvious. After the original data were log-transformed, descriptive statistics were performed on the original data and the log- transformed data, including the number of observations, mean, standard deviation, minimum value, maximum value, skewness and kurtosis, as shown in Table 2. It can be seen that due to the large size of the original sample data, the standard deviation of the data is reduced after the logarithmic processing of the original data, which reduces the heteroscedasticity generated by the model. However, considering that the data used in this group is panel data, the time and individual differences of panel data are large, so the preliminary judgment of this group needs to use the standard deviation of the robust form of heteroscedasticity for estimation. 124 Table 2. Descriptive statistics stats N mean sd min max skewness kurtosis expand 160 469.632 221.078 203.590 1191.880 1.159 4.074 income 160 6053.841 1819.134 3379.410 12151.700 0.829 3.086 fiscal 160 529.126 306.113 157.500 1748.390 1.404 5.070 cpi 160 102.716 2.474 96.570 106.670 0.59 2.408 gdp 160 5129.955 3042.257 1309.210 16382.200 1.625 5.632 lnexpand 160 6.052 0.441 5.320 7.080 0.268 2.261 lnincome 160 8.666 0.287 8.130 9.410 0.312 2.243 lnfiscal 160 6.125 0.538 5.060 7.470 0.198 2.424 lncpi 160 4.632 0.024 4.570 4.670 0.607 2.450 lngdp 160 8.397 0.531 7.180 9.700 0.261 2.796 4. Model Setting and Empirical Analysis (1) Selection of functional forms In the above selection of data forms, this paper chooses to process the natural logarithm of the data, and then verify the interaction terms and polynomials of the data to see whether there are interaction terms and polynomials in the data, as shown in Table 3: Table 3. Various forms of regression analysis log square interact VARIABLES lnexpand lnexpand lnexpand lnincome 1.382 * * * (0.092) 0.126 (4.552) 1.204 (1.332) lnincome2 0.087 (0.261) lnfiscal 0.312 * * * (0.062) 0.315 * * * (0.063) 0.332 * * * (0.065) lncpi 1.328 (1.041) 1.282 (1.070) 0.978 (1.105) lngdp 0.243 * * * (0.056) 0.246 * * * (0.058) 2.468 * (1.401) c.lnincome#c.lngdp 0.312 * (0.162) Constant 0.100 (4.747) 6.435 (19.328) 21.066 * (11.190) Observations 160 160 160 R-squared 0.598 0.598 0.607 r2_a 0.588 0.585 0.594 F 69.63 54.75 54.47 Note: *** p<0.01, ** p<0.05, * p<0.1 Since the quadratic term is prone to collinearity, the coefficient of the quadratic term is completely different from that of the linear term, and the interaction term is not the research purpose of this paper. Therefore, the immediate interaction term is significant, so this paper does not include the interaction term analysis, and this paper still chooses the logarithmic model. (2) The mixed cross section analysis method is used to analyze the data Next, this paper uses the method of mixed OLS estimation to estimate the data, which is not used as panel data, but as mixed data. The standard deviation in the form of heteroscedasticity file is also used to estimate the mixed OLS, as shown in Table 4: Table 4. Mixed cross section regression Robust lnexpand Coef. Sts.Err t P>|t| [95% Conf. Interval lnincome 1.3833 0.0919 15.05 0.000 1.2017 1.5649 lnfiscal 0.3132 0.0622 5.04 0.000 0.4360 0.1904 lncpi 1.3231 1.0352 1.28 0.203 3.3681 0.7219 lngdp 0.2436 0.0568 4.29 0.000 0.1314 0.3557 _cons 0.0650 4.7017 0.01 0.989 9.2226 9.3527 R-squared 0.5980 F (4155). 69.17 Prob>F 0.0000 Note: *** p<0.01, ** p<0.05, * p<0.1 125 Through the estimation of regression results, the R-squared is 0.5980, the goodness of fit of the model is 59.80%, the F value is 69.17, and the corresponding p value is 0.0000, which is less than 0.01. Therefore, the effects of explanatory variables and control variables on the explained variables are jointly significant. The influence coefficient of lnincome is 1.3833, which has a significant positive effect. Every 1% increase in income will cause the explained variable per capita consumption expenditure of cultural, educational and entertainment services to increase by 1.3833% on average. 3) Panel regression analysis In this paper, the individual fixed effect model, time fixed effect model and individual time point fixed effect model are analyzed respectively, and the following results are obtained, as shown in Table 5: Table 5. Fixed effects model Individual fixed effects Time fixed effects Two-way fixed effects VARIABLES getiguding lnexpand lnexpand lnincome 0.906 * * * (0.135) 1.146 * * * (0.058) 0.507 * * * (0.147) lnfiscal 0.561 * * * (0.063) 0.178 * * * (0.052) 0.023 (0.070) lncpi 0.855 (1.033) 1.247 (1.572) 0.260 (1.331) lngdp 0.392 * * * (0.074) 0.085 * (0.044) 0.043 (0.104) 20092.season 0.003 (0.060) 0.099 * (0.053) 20093.season 0.599 * * * (0.071) 0.506 * * * (0068). 20094.season 0.088 (0.061) 0.092 (0.090) 20101.season 0.169 * * (0.073) 0.028 (0.067) 20102.season 0.112 (0.108) 0.022 (0.102) 20103.season 0.515 * * * (0.091) 0.581 * * * (0.095) 20104.season 0.175 * (0.105) 0.029 (0.116) 20111.season 0.240 * * (0.116) 0.054 (0.105) 20112.season 0.177 (0.112) 0.045 (0.128) 20113.season 0.395 * * * (0.127) 0.620 * * * (0.138) 20114.season 0.199 * (0.106) 0.044 (0.130) 20121.season 0.291 * * * (0.094) 0.090 (0.101) 20122.season 0.198 * * (0.082) 0.074 (0.117) 20123.season 0.473 * * * (0.083) 0.635 * * * (0.103) 20124.season 0.256 * * * (0.085) 0.086 (0.127) 2.code 0.524 * * * (0.088) 0.353 * * * (0.069) 3.code 0.925 * * * (0.117) 0.703 * * * (0.100) 4.code 0.595 * * * (0.109) 0.571 * * * (0.096) 5.code 0.454 * * * (0.098) 0.313 * * * (0.073) 6.code 0.471 * * * (0.100) 0.427 * * * (0.093) 7.code 0.333 * * * (0.098) 0.125 (0.077) 8.code 0.776 * * * (0.122) 0.509 * * * (0.089) 9.code 0.173 (0.114) 0.075 (0.144) 10.code 0.633 * * * (0.112) 0.527 * * * (0.093) Constant 5.125 (4.509) 10.036 (7.310) 3.281 (6.266) Observations 160 160 160 R-squared 0.733 0.914 0.959 r2_a 0.710 0.902 0.950 F 46.66 102.3 114.3 Number of sode Note: *** p<0.01, ** p<0.05, * p<0.1 126 It can be seen that most of the dummy variables of the individual fixed effect are significant, and the dummy variable of the time fixed effect is also significant, but it is not so significant compared with the individual fixed effect. In order to test whether there is individual fixed effect, the individual fixed effect model is tested by F test: F test that all u_i=0: F(13, 146) = 47.18Prob > F = 0.0000 Through the self-provided F test, the statistic of F test is 47.18, and the corresponding p value is 0.0000, indicating that the individual fixed effect exists. In order to test the existence of time fixed effect, wald test is used to test the time fixed effect model: (1) 20092.season = 0; (2) 20093.season = 0; (3) 20094.season = 0; (4) 20101.season = 0; (5) 20102.season = 0; (6) 20103.season = 0; (7) 20104.season = 0; (8) 20111.season = 0; (9) 20112.season = 0; (10) 20113.season = 0; (11) 20114.season = 0; (12) 20121.season = 0; (13) 20122.season = 0; (14) 20123.season = 0; (15) 20124.season = 0 F(15, 140) = 32.72 Prob > F = 0.0000 It can be seen that the F-value is 32.72 and the corresponding p-value is 0.0000, indicating that time fixed effects exist. The fixed effect model is directly regression through the panel data estimation command in stata, as shown in Table 6: Table 6. Fixed effects model VARIABLES Fixed effects model lnexpand lnincome 0.906 * * * (0.084) lnfiscal 0.561 * * * (0.080) lncpi 0.855 * (0.382) lngdp 0.392 * * * (0.093) Cnstant 5.163 * * * (1.468) Obeservation 160 R-aquared 0.358 r2_a 0.314 F 135.3 Number of code 10 Note: *** p<0.01, ** p<0.05, * p<0.1 5. Model Selection and Evaluation (1) Summary of research results, as shown in Table 7 (2) Model selection and research results After the previous test, both the individual fixed effect and the time fixed effect exist, and according to the table data, the significance of the dummy variable under the individual fixed effect model is higher. This group also noticed that the dummy variable in the two-way fixed effect model was not significant, which was probably because time diluted the significance of the individual. Therefore, the individual fixed effects model is the most suitable for this study. The goodness of fit of the model was 71%, indicating a high goodness of fit. Among them, in addition to the negative change of local budget revenue and per capita consumption expenditure of cultural, educational and entertainment services, the per capita disposable income, consumer price index and GDP of China's urban areas all show positive change with per capita consumption expenditure of cultural, educational and entertainment services. For every 1 unit increase in urban per capita disposable income, the per capita consumption expenditure of cultural, educational and entertainment services in China will increase by 0.906 units. The influence coefficients of the control variables of local budget revenue, consumer price index and GDP are -0.561, 0.855 and 0.392 respectively. 6. Conclusion In this paper, the time series data are integrated into panel data, descriptive statistics are carried out after the logarithmic processing of the data, the data are analyzed by the mixed cross section analysis method, and three models of individual fixed effect, time fixed effect and two-way fixed effect model are estimated. Various tests are carried out to select the appropriate model, and the final results are relatively reliable. It has certain academic value, and more comprehensive research can be carried out in the future, such as adding control variables, selecting more provinces and time points, etc. 127 Table 7. Comprehensive model selection Mixed Cross section OLS Fixed effects model Individual fixed effects model Time fixed effects model Two-way fixed effects model VARIABLES lnexpand lnexpand lnexpand lnexpand lnexpand lnincome 1.382 * * * (0.092) 0.906 * * * (0.084) 0.906 * * * (0.135) 1.146 * * * (0.058) 0.507 * * * (0.147) lnfiscal 0.312 (0.062) 0.561 * * * (0.080) 0.561 * * * (0.063) 0.178 * * * (0.052) 0.023 (0.070) lncpi 1.328 (1.041) 0.855 * (0.382) 0.855 (1.033) 1.247 (1.572) 0.260 (1.331) lngdp 0.243 * * * (0.056) 0.392 * * * (0.093) 0.392 * * * (0.074) 0.085 * (0.044) 0.043 (0.104) 20092.season 0.003 (0.060) 0.099 * (0.053) 20093.season 0.599 * * * (0.071) 0.506 * * * (0.068) 20094.season 0.088 (0.061) 0.092 (0.090) 20101.season 0.160 * * * (0.073) 0.028 (0.067) 20102.season 0.112 (0.108) 0.022 (0.102) 20103.season 0.515 * * * (0.091) 0.581 * * * (0.095) 20104.season 0.175 * (0.105) 0.029 (0.116) 20111.season 0.240 * * (0.116) 0.054 (0.105) 20112.season 0.177 (0.112) 0.045 (0.128) 20113.season 0.395 * * * (0.127) 0.620 * * * (0.138) 20114.season 0.199 * (0.106) 0.044 (0.130) 20121.season 0.291 * * * (0.094) 0.090 (0.101) 20122.season 0.198 * * * (0.082) 0.074 (0.117) 20123.season 0.473 * * * (0.083) 0.635 * * * (0.103) 20124.season 0.256 * * * (0.085) 0.086 (0.127) 2.code 0.524 * * * (0.088) 0.353 * * * (0.069) 3.code 0.925 * * * (0.117) 0.703 * * * (0.100) 4.code 0.595 * * * (0.109) 0.571 * * * (0.096) 5.code 0.454 * * * (0.098) 0.313 * * * (0.073) 6.code 0.471 * * * (0.100) 0.427 * * * (0.093) 7.code 0.333 * * * (0.098) 0.125 (0.077) 8.code 0.776 * * * (0.122) 0.509 * * * (0.089) 9.code 0.173 (0.114) 0.075 (0.144) 10.code 0.633 * * * (0.112) 0.527 * * * (0.093) Constant 0.100 (4.747) 5.613 * * * (1.468) 5.125 (4.509) 10.036 (7.310) 3.281 (6.266) Observations 160 160 160 160 160 R-squared 0.598 0.358 0.733 0.914 0.959 r2_a 0.588 0.341 0.710 0.909 0.950 F 69.63 135.3 46.66 102.3 114.3 Note: *** p<0.01, ** p<0.05, * p<0.1 128 References [1] Shao Minzhi; An empirical study on the relationship between disposable income and consumption expenditure of urban residents in Shanghai Based on grey system Modeling [J]. Operations Research and Management,2015 (3). [2] Cheng Yan: ARIMAX model and its application: Disposable income and consumption expenditure of urban residents in Chongqing [J]. Journal of Chongqing Technology and Business University (Natural Science Edition), 2015 (11). [3] Wu X Q, Si F N, Tian P: An empirical analysis of regional differences in the relationship between consumption and income of urban residents in China based on non-parametric panel data model [J]. Journal of Mathematical Statistics and Management,2016 (4). [4] Han X & Chen B B; Analysis on the relationship between consumption expenditure and disposable income of urban residents in Jilin Province [J]. Economic Forum, 2016 (6). [5] Liang W X. Empirical analysis of urban residents' income and consumption expenditure in Gansu Province [J]. Cooperative Economics and Technology,2018 (7). [6] Liu Zhexi & Chen Yanbin: The mystery of weak consumption and the policy of expanding consumption [J]. Research on Financial and Economic Issues,2018(11).