The impact of China’s Internet Finance on the banking systemic risk – an empirical study based on the SCCA model and stepwise regression

ABSTRACT

The disorderly development of Internet Finance impacts both the banking industry and the macroeconomy and opens a risk contagion channel, which easily generates the banking systemic risk. Based on current banking systemic risk measured by Systemic Contingent Claims Analysis (SCCA) model and a stepwise regression, we verify the impacts of China’s Internet Finance on the banking industry and predict that the risk will rise in the future.

KEYWORDS:

• Internet Finance
• the banking systemic risk
• SCCA model
• stepwise regression method
• risk prediction

JEL CLASSIFICATION:

• G21
• G24
• G28

Disclosure statement

No potential conflict of interest was reported by the authors.

1. Weight determination

The whole index system is divided into four levels. Before compiling the index, we need to determine the weight of indicators at all levels. We adopt the method of combining subjective qualitative method and objective quantitative method to determine the weight.

Weight setting of level 4 indicators (breadth index and depth index): transaction penetration rate is 50%, transaction amount per capita is 25% and transaction number per capita is 25%.

Weight setting of level 3 indicators (Alibaba and other institutions): we use a quantitative method to determine the weight. At the same time, in order to reduce the impact of monthly weight changes on index fluctuations, we calculate the respective weights according to the proportion of the three-month moving average of the transaction amount actually obtained by Alibaba and other institutions.

Weight setting of level 2 indicators: the weight among major businesses is mainly determined subjectively based on the development maturity of each business. Maturity is a measure that takes into account the duration and stability of the business. Therefore, the business weight of the six sections is set as follows: 30% Internet payment, 25% Internet money fund, 15% Internet credit and loan, 15% Internet insurance, 10% Internet investment and finance and 5% Internet credit.

Due to the Internet gene, the development and change of each business segment are fast. Therefore, we plan to re-examine the weight of each business segment in the first month of each year according to the business maturity and other standards. In order to avoid index jumping caused by weight change (nonbusiness self-development) when adjusting the weight of each business section, we will refer to the divisor adjustment method commonly used in index compilation literature for correction.

In order to ensure the expansion and representativeness of the index, if there are new business segments in the process of index operation, we will start to include them in the index in the second month after the new business is generated (month-on-month data can be calculated in the second month). As the business segment changes from $\mathrm{N}$ to $\mathrm{N}+1$ at this time, the weight between each business should be re-evaluated according to the expert method after the new business is included.

The new business that just produces is in a fluctuant bigger level. In order to avoid the fluctuation of new business and the large impact of each business weight on the total index, we stipulate that the new business weight should not exceed 10% (set ${\mathrm{W}}_{\mathrm{N}+1}$). The relative weight between the original businesses remains unchanged; that is, the original $\mathrm{i}$th business weight is updated to ${\mathrm{W}}_{\mathrm{i}}\times \left(1-{\mathrm{W}}_{\mathrm{N}+1}\right)$.

2. Index calculation

After determining the weights of all levels, the calculation process of the index is to sum up the weighted average of each level from the bottom up. We first calculate the month-on-month index of all levels and then obtain the level-fixed-base index by chain multiplication based on the level-on-month index. The specific calculation process is as follows:

2.1 Sequential index calculation at all levels

The first-order sequential index is the weighted average of the month-on-month index at all levels from bottom to top. The specific formula is as follows:

$i_t=\frac{I_t}{I_{t-1}}=\sum _{i=1}^6{W}_i\frac{L_{i,t}}{L_{i,t-1}}=\sum _{i=1}^6{W}_i\left(\sum _{j=1}^2{P}_{i,j,t}\frac{K_{i,j,t}}{K_{i,j,t-1}}\right)=\sum _{i=1}^6{W}_i\sum _{j=1}^2{P}_{i,j,t}\sum _{k=1}^3{m}_k\frac{X_{i,j,k,t}}{X_{i,j,k,t-1}}$

where $\frac{I_t}{I_{t-1}}$ represents the month-on-month index of Internet Finance development (level 1 month-on-month index).$\frac{L_{i,t}}{L_{i,t-1}}$ represents the month-on-month index (level 2 month-on-month index) of the $\mathrm{i}$th business in the period $\mathrm{t}$. $\frac{K_{i,j,t}}{K_{i,j,t-1}}$ represents the month-on-month index of part $\mathrm{j}$ of the $\mathrm{i}$th business (Alibaba and other institutions) in the period t (level 3 month-on-month index).$\frac{X_{i,j,k,t}}{X_{i,j,k,t-1}}$ represents the sequential relative number of the $k$th level 4 indicator of the part j of the $\mathrm{i}$th business (Alibaba and other institutions) in the period t (level 4 month-on-month index).$W_i\left(i=1,2,3,4,5,6\right)$ is the weight of the $\mathrm{i}$th business. $P_{i,j,t}$ represents the weight of part j of the $\mathrm{i}$th business in the period of t. $m_j\left(j=1,2,3\right)$ represents the weight of the $j$th level 4 index.

2.2 Fixed-base index calculation at all levels

The fixed-base index at all levels in the period $\mathrm{t}$ is the product of the month-on-month index at all levels in the period $\mathrm{t}$ and the fixed-base index at the corresponding level in the period $\mathrm{t}-1$.

$I_t=\sum _{i=1}^6\left(W_i\frac{L_{i,t}}{L_{i,t-1}}\right)I_{t-1}=\left[\sum _{i=1}^6{W}_i\sum _{j=1}^2{P}_{i,j,t}\sum _{k=1}^3{m}_k\frac{X_{i,j,k,t}}{X_{i,j,k,t-1}}\right]I_{t-1}$

Among them, the fixed-base index in the period $\mathrm{t}-1$ is obtained by successive multiplication of the month-on-month indexes of each period. We set the benchmark value for each business and composite index base period (January 2014) at 100. The calculation formula is as follows:

$I_{t-1}=100\times i_1\times i_2\times \cdots \times i_{t-1}=100\times \frac{I_1}{I_0}\times \frac{I_2}{I_1}\times \frac{I_3}{I_2}\times \cdots \times \frac{I_{t-2}}{I_{t-3}}\times \frac{I_{t-1}}{I_{t-2}}$

where $i_1,i_2,\cdots i_t$ respectively, represent the month-on-month index of each period.

Notes

¹ Transaction penetration reflects the breadth of business coverage. Specifically, the overall penetration rate of each business in China is obtained by dividing the total number of people who have purchase records in the last month of this business by the total number of people in the country in the current period.

² The transaction amount per capita reflects the depth and quality of each business development. It is obtained by dividing the actual total transaction amount of a certain business in the last month by the actual number of transactions.

³ The transaction number per person reflects the depth and quality of a business development. It is obtained by dividing the actual total number of transactions of a business in the last 1 month by the actual number of transactions.

Additional information

Funding

This work was supported by the China’s National Philosophy and Social Science Fund under Grant 17ZDA037, Jiangsu’s Philosophy and Social Science Fund under Grant 18EYD005, Jiangsu Provincial Graduate Training Innovation Project under Grant KYZZ16_0448.