Determinants of the Urban Investment Bonds in China

Abstract

This chapter investigates the determinants of the issuance of urban investment bonds (UIBs) across 31 provinces in China during the period of 2005–2013 using a spatial autoregressive model. We find that the provincial governments tend to compete with or imitate their neighboring provinces in bond issuance. The neighborhood can be defined in terms of either geographical proximity or economic performance. We also find that the fiscal gap, governments’ investments in housing and public welfare, and bond level in the previous year have positive association with the issuance of UIBs in China.

Appendix

1.1 Weight Matrix Using GDP Ranking

To get the weighting matrix based on GDP ranking of a certain year, we first sort the 31 provinces by their GDP scales. Then we assign the two provinces, which are immediately above a certain province in GDP ranking with value 1, otherwise 0.⁶ Specifically, each one of the four municipalities (Beijing, Tianjin, Shanghai, and Chongqing) is compared with their three counterparts separately, and the same principle applies to other 27 provincial administrative units. In this way, we get the first matrix. Matrix 1 is an example, which shows the initial GDP ranking matrix in 2004. For example, in 2004, the two provinces that were immediately beyond Tianjin in terms of GDP were Beijing and Shanghai. Therefore, the first element of the second row and the ninth element in the same row are assigned 1, with others being 0.

Matrix 1

Rank matrix in 2004

After obtaining the initial matrixes, we can standardize them to weight matrixes whose sum of row elements is equal to 1. To do this, we first calculate the sum of elements in each row, and then divide each element with the corresponding row sum.

1.2 Weight Matrix Using Trade Flow

The Wy_ trade is the weighted average of UIB amount based on the annual inter-provincial trade flows in every year. As data on trade flows is not available from official records and the only data on the goods transportation through waterway, roadway, and railway is available, we calculate the trade flows through the following steps. We firstly define that:

Trade_flowijt = The trade flow between Province i and Province j in year t

= Goods transportation through railway, waterway and road between Province i and Province j in year t;

Trans_railit = The goods transportation through railway of province i in year t;

Trans_waterit = The goods transportation through waterway of province i in year t;

Trans_roadit = The goods transportation through road of province i in year t.

With assumption as:

$$ \frac{Trade\_ rai{l}_{ijt}}{Trans\_ rai{l}_{it}}=\frac{Trade\_ flo{w}_{ijt}}{Trans\_ rai{l}_{it}+ Trans\_ wate{r}_{it}+ Trans\_ roa{d}_{it.}} $$

Therefore, we get:

$$ Trade\_ flo{w}_{ijt}=\frac{Trade\_ rai{l}_{ijt}\times \left(\begin{array}{l} Trans\_ rai{l}_{it}+ Trans\_ wate{r}_{it}\\ {}\kern4em + Trans\_ roa{d}_{it}\end{array}\right)}{Trans\_ rai{l}_{it}} $$

Matrix 2

Weight matrix based on GDP rank in 2004 with row standardization

Therefore, we can get the trade flow matrix of every year. Then we standardize the trade flow based weight matrix of every year. By multiplying the weight matrix with the corresponding UIB amount in that year, we can finally get the weighted average of UIB amount.

1.3 Weight Matrix Using Railway Mileage

The calculation of weighted average of UIB amount is basically the same as the above two except a significant difference is that the weight matrix is static. We adopt the railway distance between capital cities in province i and province j to construct the weight.