Whether There Is a Competition between the Interprovincial Governments on Fiscal Expenditure —From the Detection of Spatial Correlation

The fiscal decentralization gives the local government economic participation status. Is there a spatial spillover effect in the process of local government financial expenditure? This paper introduces the spatial econometric model, and applies the 2005-2016 data to test the spatial correlation of the whole country by calculating the Moran’s I index and the Geary index C. At the same time, we use the local Moran’s index to detect the degree of agglomeration between regions. The results show that there is a significant spatial spillover effect between regions. There are policy imitation behaviors in the region and adjacent areas, and there is a high value and high value agglomeration among regions, as well as a trend of low value and low value agglomeration.


Introduction
Whether there is a competition between Chinese governments on fiscal expenditure? Tax competition has always been the main means of fiscal competition, but with the change of society and the development of time, the preferential tax policy has shifted from region to industry, and the means of local fiscal competition have gradually shifted to the field of expenditure competition, which provides public goods and services [1]. Different scholars have chosen different methods to verify the existence of financial expenditure competition and test the effects of financial expenditure competition. The research on spatial interaction ists "horizontal strategic interaction" among local government and "vertical common reflect" among central and local governments in the field of fiscal expenditure [3]. Fiscal expenditure competition can cause factor flow. However the viewpoints of the effect of fiscal expenditure competition among scholars are not quite the same. Some believe that the basic economic construction expenditure can have suppressive effect on economic growth, and have positive effect on science and education [4]. Fiscal expenditure competition distorts the structure of fiscal expenditure, leading to the low level of public service expenditure [5].
Therefore scholars make further research on the effect of fiscal expenditure competition on provincial capital flows and consider that in the field of social and livelihood, fiscal expenditure competition can have an effect on capital inflows [6]. Some scholars further detect the invalidity of fiscal expenditure competition that there is U linear relationship between fiscal expenditure competition and economic growth and excessive competition is not favorable for economic growth.

General Introduction of Fiscal Expenditure Structure
The financial expenditure is divided into four categories, public management expenditure, science and education expenditure, social security expenditure and economic construction expenditure. In case of the annual population differences and in order to reflect the reality of the situation, per capital level to reflect fiscal expenditure situation is applied.
From Figure 1 above, there are four shades from deep to shallow, representing public management expenditure, economic construction expenditure, social security expenditure and the science and education expenditure respectively. The entire circle represents the whole local fiscal expenditure and each part represents the proportion of fiscal expenditure. The 12 figures represent the structure order of fiscal expenditure from 2005 to 2016 respectively; dynamic changes can be seen while observing 12 figures continuously.
As shown in Figure 2, as time goes on, the proportion of per capita public management expenditure is gradually reduced, from 36% in 2005 to 14% in 2016 m while the proportion of per capita economic construction expenditure shows a gradually rise, from 31% in 2005 to 41% in 2016. Science and education expenditure contribute the largest proportion of 42% in 2006, while the rest years are maintained at around 27%. The social security expenditure proportion has been maintained at 15%. Considering this, we can see that with the development of society, the fiscal expenditure structure with the largest proportion of economic construction, public management expenditure and the lowest social security expenditure has been gradually formed. The structure of fiscal expenditure is biased towards economic construction expenditure.

Spatial Correlation Standard
There are a lot of spatial correlation standard, Moran index method, Gil coeffi- The n is the total number of regions in the study; W ij is the spatial weight of the area I and the area J is the adjacent area, which is divided into geographical proximity and economic neighbor. So W ij is divided into geographical weight and economic weight. X i and X J are the observation variables of region I and region J respectively, which are the financial expenditure items. X is the average attribute observation variable and S 2 is the variance of observed variables.
The range of Moran's I index is generally between −1 to 1; more than 0 says positive correlation; value close to 1 represents that similar values bond together and less than 0 indicates negative correlation values, close to −1 indicates different attributes bond together. If close to 0, then there is no spatial autocorrelation, no spatial dependence, and no expenditure competition.
A Geary index C is also an index of global clustering tests. A median deviation If I is greater than 0, it represents a high value-high value, low value-low value cluster; and if the I is less than 0, the low value-high value, high value-low value cluster is expressed.  Table 1 below is the result of spatial correlation test of national fiscal expenditure in 2016. According to the results, in general, from 2005 to 2016, fiscal expenditure spatial correlation index has significant correlation, indicating the existence of spatial dependence of fiscal expenditure in adjacent areas with competition. Specifically, the paper selects two matrixes of geographical weight and economic weight to calculate the spatial dependence of geographically spaced adjacent areas and similar economic regions. Comparing Moran index,  more than 0 and close to 1, P values are less than 0.01 which strongly rejects the null hypothesis, therefore, among the total fiscal expenditure competition, economic space correlation strategy behavior contributes a larger impact on the region. The region selects strategy imitation. Considering Geary s C index, from 2005 to 2016, there is significance, but economic spatial similar regions is more significant than the geographically adjacent areas, indicating that in the total financial expenditure competition, the area is more affected by economic spatial area. The index is less than 1, indicating that there is a significant positive correlation in similar regions. The strategic mimic behavior is selected in the area whose result is consistent with the Moran index.

Conclusion
The calculation of Moran index, Gil index and local Moran index can draw the conclusion that the total fiscal expenditure of 31 provinces in China has strong spatial dependence, which is not only affected by the geographically adjacent area, but also affected by the economically adjacent area, and the fiscal expenditure behavior is competitive.