Exploring the integration path of ideological and political education and quantitative analysis in financial risk management

The quantitative analysis of financial risk is a fundamental work for the central bank to control financial risk. For the central bank, the quantification of financial risk, to solve two aspects of the problem, one is the degree of financial risk problem, that is, the object of financial supervision - financial enterprises in the end how much financial risk [1-3]. Second, the quantification of financial risks should enable the central bank to compare the risks of different financial institutions, so as to facilitate the implementation of classified management according to the differences in the risk levels of financial enterprises [4-6]. In this regard, the quantification of financial risk is the requirement to make a comprehensive measure assessment of the risk profile of a financial firm [7].

With the development of financial integration, financial risks are becoming more and more complex, especially under the interaction of financial activity innovation and financial regulatory innovation, financial risk events induced by financial participants are becoming more and more frequent, which not only puts forward higher requirements for the professional skills of financial talents, but also puts forward higher requirements for their professional ethics, ideological and political qualities, and social values [8-10]. “Financial Risk Management” is a compulsory course for finance majors, and its training goals are knowledge impartation, ability improvement and value shaping, which is the core moral education carrier in finance professional courses. This requires students not only to understand the process of financial risk management, master the basic methods of identifying and measuring financial risks, and strengthen their awareness of risk management, but more importantly, to guide students to have the professional ethics of “integrity, professionalism, diligence, and compliance” and the sense of social responsibility of “daring to take responsibility and dedication”, so as to enhance students’ sense of identity with the superiority of the socialist system and establish a correct world outlook, outlook on life, and values [11-13]. Therefore, in the teaching of “financial risk management”, it is necessary to dig deep into the ideological and political elements and run the ideological and political work through the whole teaching process. It allows students to improve their professional skills and professional ethics and ideological and political literacy [14-15].

Sun, M et al. built a financial efficiency risk management model based on high-frequency data (including trading volume information) with LSTM model piggybacking as well as the theory of semiparametric mechanism, and its financial risk prediction accuracy is better than the traditional model and the model excluding the trading volume information in the practical comparison [16]. Fouejieu, A et al. proposed a new Keynesian model with endogenous financial bubbles to study the macroeconomic and financial trade-offs with the objective of financial stability and found that a central bank that is inversely related to the market will face a trade-off between inflation/output stabilization and financial stability, and stated that interest rates should serve the macroeconomic objective first and foremost and deal with the build-up of financial risk secondly [17]. Risman, A et al. combined multiple linear regression analysis methods as well as moderated regression analysis models to analyze panel data of 120 financial samples to elucidate that market risk can effectively physically modulate the impact of digital finance on financial stability, and that an increase in systematic risk in the cut can reduce the positive impact of digital finance on financial stability [18]. Baruník, J et al. conceived an analytical model to measure the connectivity between financial variables and introduced a spectral representation framework with variance decomposition as the core logic for the estimation of connectivity across types of financial cycles, finding that the high-frequency connectivity phase is in the phase of calm processing of information in the stock market and shocks to assets are short-term, whereas shocks in the low-frequency connectivity phase are continuous and persistent [19]. Financial risk management has always been the top priority in the financial field, many scholars around the interest rate, digital finance and financial cycle connectivity and other elements of the impact of financial risk elements and the role of the mechanism of the relevant analysis and investigation, but also scholars explore the construction of the financial risk management model and the prediction of financial risk.

Cantoni, D et al. examined the impact of China’s textbook reform on students’ ideological and political awareness, and based on the findings, it was learned that the reform was effective in reshaping the students’ ideological and political awareness, which made the students have a new view of liberal democracy, and they were somewhat skeptical of the proclaimed free economic market, and they could look at liberal democracy with a dialectical perspective [20]. Ao, J et al. empirically explored the integration of Civic and political education in the process of training financial talents, and concluded that the cultivation of financial product design talents needs to be based on international political economy, describing the international political economy teaching model that integrates Civics and finance [21]. Yang, H In order to implement the relevant requirements and policies for the work of Civic and Political Education in national colleges and universities, the path of integrating Civic and Political Education in the classroom of intelligent teaching of internal control and risk management was explored from the dimensions of teaching objectives, contents, methods, and guarantees [22]. Practitioners in the financial industry need to have a good sense of civic and political awareness and correct values, so in the training of financial talents, some scholars have explored the strategy of integrating civic and political education into the financial teaching courses and the role of integrating civic and political education in the training of financial talents through theoretical analysis, literature review, questionnaire survey and empirical analysis.

Taking financial practitioners as the entry point, this paper explores the enhancement of ideological and political education on the legal and ethical awareness and compliance consciousness of financial practitioners, so as to reveal the important role of ideological and political education in financial risk management. The VaR model is used to measure the risk in financial management risk and realize the quantitative analysis of financial risk. Two stock price indices, SSE 180 Index and Shenzhen Chengxin Index, are selected from January 1, 2011 to December 31, 2022, and the yield series are constructed with the results of the yield calculation, and the basic characteristics, smoothness and ARCH effect of the data are examined. Based on the GARCH model, the VaR results of the SSE 180 Index and the Shenzhen Composite Index are calculated, and the accuracy of the model is tested using the failure frequency test. The integration path of ideological and political education and quantitative analysis for financial risk management is proposed from the perspectives of ideological and political education, integration model construction, and empirical research. The financial risk measurement index system is established to quantify financial risk in the application analysis of the path combined with the entropy weight method. Taking two municipal districts and three municipal counties in S city as the study area, according to the actual survey results, we calculate the changes in the comprehensive evaluation results of financial risk before and after the application of the fusion path of this paper for eight key financial management objects, such as banking institutions. The fusion of ideological and political education and quantitative analysis proposed in this paper is shown to have a significant effect on the reduction of financial risk, and the possibility of applying it to financial risk management is explored.

2

Exploration of the relationship between financial risk management and ideological and political education

2.1

Financial risk management

Financial risk management effectively identifies, evaluates, monitors and controls the various types of risks faced by financial institutions in the course of their operations through a series of scientific and systematic methods to minimize the adverse impact of risks on the institutions’ operations and the financial market [23]. This management process encompasses not only traditional risk types such as market risk, credit risk, and liquidity risk, but also emerging risk areas like operational risk and legal risk. The objective of financial risk management is to make sure that financial institutions can operate smoothly, safeguard the safety of customer assets, and maintain the stability and healthy development of the financial market. At the same time, financial risk management is also a dynamic management process that requires financial institutions to continuously adjust and optimize their risk management strategies according to changes in the market environment.

Financial risk management theory is composed of four core links: risk identification, risk assessment, risk control and risk response, the specific structure is shown in Figure 1. Each link is interconnected, interacts with each other, influences each other, and constrains each other, and together they provide support for the risk management system of financial institutions. Taking risk identification as the starting point, potential risk points are discovered in a timely manner through in-depth analysis of the internal and external environment of the enterprise. Risk assessment involves quantitatively analyzing the various risks identified and evaluating the probability of their occurrence and the degree of impact. Risk control refers to the development and implementation of specific management measures to reduce risks. Risk response is the timely implementation of remedial measures to reduce losses and restore normal operations when risks occur. Providing scientific methods and tools for financial institutions to apply in risk management is of great practical significance to improve risk management and safeguard operations.

2.2

The role of ideological and political education in financial risk management

Ideological and political education has an important guiding role in financial risk management [24]. It guides financial practitioners to comply with laws, regulations, and ethical norms, and enhance their sense of social responsibility by establishing correct values and professional ethics. Through the ideological and political education and case warning education for financial industry practitioners, it enables them to always remain vigilant and firm in their thoughts, effectively build a firm ideological line of defense, and ensure that the risks are controllable, so as to escort the financial risk management. Regularized, long-term ideological and political education for financial industry practitioners, so that financial practitioners can always hold the fear of the law and ideological vigilance, so as to avoid financial crimes, leading financial practitioners to balance the relationship between material life and spiritual life, and constantly enriching their own spiritual and cultural life, to promote the free and comprehensive development of individuals. The specific performance is as follows: First, ideological and political education helps to cultivate the professional quality of financial practitioners, so that they consciously adhere to professional ethics in their work. Secondly, through ideological and political education, it can enhance the risk prevention consciousness of financial practitioners and avoid financial crises caused by moral risks. Thirdly, ideological and political education can improve the ideological awareness and political quality of financial employees and enhance their understanding and implementation of State policies.

2.3

Interaction mechanism between financial risk management and ideological and political education

There is a close interaction between financial risk management and ideological and political education. On the one hand, ideological and political education provides valuable guidance for managing financial risks and ensures that financial activities are carried out within the framework of legal compliance. On the other hand, the practical achievements of financial risk management enrich ideological and political education, enriching the content and methods of ideological and political education through actual cases. This interactive mechanism requires financial managers to pay attention not only to economic benefits, but also to social benefits, to realize a win-win situation for both economic and social benefits.

3

Application of quantitative analysis in financial risk management

The foundation and core of risk management is the measurement of risk. As the size, complexity and dynamics of financial markets and financial transactions have increased, the techniques for measuring risk in financial markets have become more varied and complex.In the late 1970s and 1980s, numerous large financial institutions began to use internal model measurement systems. The most well-known of these systems is the risk measurement system developed by JPMorgan Chase, and they use the VaR method as their measurement.

3.1

Overview of the VaR model

VaR refers to the maximum possible loss of a financial asset over a specific period of time in the future at a certain level of probability (confidence level) [25]. It is expressed in a mathematical formula as: (1) $\Pr o b (Δ P > V a R) = 1 - c$ $$\Pr ob(\Delta P > VaR) = 1 - c$$

where ΔP is the loss of the portfolio over the holding period Δt and the VaR value is the value at risk at the confidence level c.

3.1.1

Calculation of VaR with general distribution

Assuming that the initial value of the portfolio position is P₀ and R is the rate of return over the holding period, the value of the portfolio position at the end of the period is P = P₀(1 + R). Assuming that the expected value of the rate of return R is μ and the volatility is σ, and that the minimum value of the portfolio at a given confidence level d is P^* = P₀(1 + R^*), the relative VaR value of the portfolio over the future holding period L at that confidence level c can be defined, by definition, as: (2) $V a R = E (P) - P^{*} = - P_{0} (R^{*} - μ)$ $$VaR = E(P) - {P^*} = - {P_0}({R^*} - \mu )$$

The absolute VaR value can be defined as: (3) $V a R = P_{0} - P^{*} = - P_{0} R^{*}$ $$VaR = {P_0} - {P^*} = - {P_0}{R^*}$$

It follows that calculating VaR is equivalent to calculating the minimum value of the portfolio P^* or the minimum return R^*. Assuming that the probability density function of the distribution of the portfolio’s future returns is f(p), the minimum value of the portfolio P^* for confidence level c has: (4) $1 - c = \int_{- \infty}^{p} f (p) d p$ $$1 - c = \int_{ - \infty }^p f (p)dp$$

This expression is suitable for any form of distribution.

3.1.2

Calculation of VaR under normal distribution

If the return distribution of the portfolio is normal and if the density function of the standard normal distribution is ϕ(ξ), then there is: (5) $1 - c = \int_{\infty}^{β} f (p) d p = \int_{\infty}^{| R |} f (r) d r = \int_{\infty}^{α} ϕ (ξ) d ξ$ $$1 - c = \int_\infty ^\beta f (p)dp = \int_\infty ^{|R|} f (r)dr = \int_\infty ^\alpha \phi (\xi )d\xi$$

where α(α > 0) is the quantile of the standard normal distribution at a specified confidence level of c. The formula is expressed as: (6) $- α = \frac{R^{*} - μ}{σ}$ $$ - \alpha = \frac{{{R^*} - \mu }}{\sigma }$$

This yields a minimum return R^* of: (7) $R^{*} = μ - α σ$ $${R^*} = \mu - \alpha \sigma$$

Under the standard normal distribution, the confidence level corresponds to the quantile. When a confidence level such as 95% is given, it corresponds to quantile α = 1.65, and the corresponding R^* and VaR values can then be calculated.

The above calculations are based on a one-day time interval basis. If the VaR value is to be calculated for a holding period length of Δt, and assuming that the returns in consecutive time intervals are uncorrelated, the relative VaR value for a time interval of Δt is: (8) $V a R = - P_{0} (R^{*} - μ) = P_{0} α σ \sqrt{Δ t}$ $$VaR = - {P_0}({R^*} - \mu ) = {P_0}\alpha \sigma \sqrt {\Delta t}$$

Under normal market conditions, for a given confidence level of c, the corresponding critical value is the statistically maximum possible loss for that financial market or portfolio. While it is possible for actual losses to exceed the VaR, according to sampling distribution theory, the probability of losses exceeding the VaR does not exceed l − c, with l − c representing the probability of the worst-case scenario occurring.

The assessment period for VaR is usually 1 day, and a confidence level of 95% indicates that there is a 95% probability that the average maximum amount of loss of the asset on a future day will be less than that VaR. The length of the assessment period is closely related to the magnitude of the VaR, usually the longer the assessment period, the greater the VaR.

VaR is a prediction of the possible loss of an asset over a given time interval and can therefore be calculated using the predicted distribution of future returns on a financial position. For example, applying daily returns R_t, the VaR value of an asset over a 1-day holding period can be calculated by R_t+1 using the predictive distribution given the information known at a given t moment in time. Thus, the study of VaR centers on the prediction of the distribution of future returns of an asset, and different assumptions about the characteristics of the return distribution will lead to different VaR risk management models.

3.2

Common Calculation Methods for VaR Models

At present, most of the VaR calculations are centered around the distributional characteristics of asset returns, and the commonly used methods are historical simulation method, Monte Carlo simulation method, variance one covariance method, and GARCH model method. This paper is mainly based on the GARCH modeling method to develop the VaR calculation of financial risk, so the GARCH modeling method will be introduced in this section.

Aiming at the problems of spiky thick-tailedness, volatility agglomeration and conditional heteroskedasticity that often occur in financial return series, researchers have proposed conditional heteroskedasticity models. The first one proposed was the ARCH model. Subsequently, some scholars generalized the ARCH model, i.e., the GARCH model.

The standard GARCH (p, q) model is of the form: (9) ${\begin{matrix} r_{t} = μ_{t} + ε_{t} \\ ε_{t} = σ_{t} Z_{t}, Z_{t} ~ N (0, 1) \\ σ_{t}^{2} = ω + \sum_{i = 1}^{p} β_{i} σ_{t - i}^{2} + \sum_{j = 1}^{q} α_{j} ε_{t - j}^{2} \end{matrix}$ $$\left\{ {\begin{array}{*{20}{c}} {{r_t} = {\mu _t} + {\varepsilon _t}} \\ {{\varepsilon _t} = {\sigma _t}{Z_t},\:{Z_t}\sim N(0,1)} \\ {\sigma _t^2 = \omega + \sum\limits_{i = 1}^p {{\beta _i}} \sigma _{t - i}^2 + \sum\limits_{j = 1}^q {{\alpha _j}} \varepsilon _{t - j}^2} \end{array}} \right.$$

Where Z_t is an independent and identically distributed random variable, usually assumed to follow a standard normal distribution, μ_i and $σ_{t}^{2}$ $$\sigma _t^2$$ are the conditional expectation and conditional variance (also called volatility) of the returns under historical information, respectively, and the third equation describes the relationship between the conditional variance and its lag term. In the model, ϖ, β_i, α_i are all parameters greater than 0 to ensure that the conditional variance is always positive, $\sum_{i = 1}^{p} β_{i} + \sum_{j = 1}^{q} α_{j} < 1$ $$\sum\limits_{i = 1}^p {{\beta _i}} + \sum\limits_{j = 1}^q {{\alpha _j}} < 1$$ is to make the volatility series have smoothness. the lag order of the GARCH model p is q do not need to be taken too large, it can be able to better describe the characteristics of the volatility of the yield data clustering, usually take 1 or 2.

The conditional mean μ_i of the yields in the model can be simply assumed to be constant. In order to better filter the autocorrelation of the yield series, time series models can also be used to characterize the correlation between the conditional mean and the historical information. Commonly used models are the autoregressive model (AR), the sliding average model (MA), and the autoregressive sliding average model (ARMA) which is a combination of the two.The general form of the ARMA(P, O) model is: (10) $r_{t} = ϕ_{0} + \sum_{i = 1}^{P} ϕ_{i} r_{t - i} + \sum_{j = 1}^{Q} θ_{j} ε_{t - j} + ε_{t}$ $${r_t} = {\phi _0} + \sum\limits_{i = 1}^P {{\phi _i}} {r_{t - i}} + \sum\limits_{j = 1}^Q {{\theta _j}} {\varepsilon _{t - j}} + {\varepsilon _t}$$

In the ARMA model, the error term ε_t is the white noise series, ϕ_t, θ_t is the coefficient of the corresponding variable, and P and Q are the lag orders. ARMA(0, Q) = MA(Q) when the order is P = 0 and ARMA(Q, 0) = AR(Q) when the order is Q = 0.

Since the ARMA model is based on the variance is a constant established, however, most of the financial yield series have the “variance of time-varying” phenomenon exists, that is, the variance changes over time, in the time series plot shows that the larger fluctuations and smaller fluctuations will be clustered in a certain period of time phenomenon, so confined to the ARMA model is no longer able to Solve the problem. It is usually a combination of an ARMA model to describe the autocorrelation of expected returns and a GARCH model to describe the heteroskedasticity of conditional volatility. Fitting the return data by ARMA-GARCH model can lead to higher accuracy of the fitted model, and investors can more accurately grasp and quantify the risk.

In the GARCH model, since the conditional variance $σ_{t}^{2}$ $$\sigma _t^2$$ of the error ε_i varies over time, the calculation of VaR based on the GARCH model first separates σ_t and analyzes the remaining residuals Z_t. The distribution assumptions for the residuals Z_t can be described by the t distribution, generalized error distribution (GED), mixed normal distribution, etc., in addition to the commonly used normal distribution. The VaR value of the residuals Z_t is first calculated based on the distributional characteristics of Z_t and then converted to the VaR value of the returns based on the assumptions of the GARCH model, which can be expressed by the formula: (11) $V a R (r_{t}) = μ_{t} + σ_{t} V a R (Z_{t})$ $$VaR({r_t}) = {\mu _t} + {\sigma _t}VaR({Z_t})$$

The GARCH model method is more advanced in calculating VaR than the variance-covariance method, because the GARCH model can filter the “heteroskedasticity” and “fluctuation agglomeration” of the yield series, and the ARMA model can also filter the “autocorrelation” of the yield, so that the calculation results are more accurate.

3.3

Tests of the VaR model

After calculating the VaR value of the rate of return, it is necessary to perform statistical hypothesis testing on the results obtained to verify the validity of the model. The method used in this paper is the failure frequency test, also known as the likelihood ratio test. Specifically, assuming a confidence level of 1 − α, the probability that the actual loss of an asset at a certain unit of time in the future exceeds the value of VaR should be equal to α. In practice, the value of VaR will be affected by a variety of factors, such as randomness, which will make the probability that the actual loss at a certain unit of time is greater than the VaR have a small difference from α, but it should be fluctuating in the vicinity of α within a certain range. If it is far away from α, the assumptions of the model are not reasonable.

Let the number of days of actual observation be T and the number of days when the actual loss is greater than the VaR be N, then the frequency of failure is $\hat{P} = N / T$ $$\hat P = N/T$$. Consider hypothesis testing: (12) $H_{0} : P = α H_{1} : P \neq α$ $${H_0}:\:P = \alpha \quad {H_1}:\:P \ne \alpha$$

If the failure rate $\hat{P}$ $$\hat P$$ is close to α, then support also the original hypothesis is valid. If $\hat{P}$ $$\hat P$$ is far from α, the original hypothesis is rejected and the opposing hypothesis is supported. The opposing hypothesis is divided into two scenarios, if P > α, it indicates that the modeling underestimates the actual risk, and if P < α, it overestimates the risk for investors or financial institutions. The likelihood ratio test statistic (LR) constructed for this test problem is as follows: (13) $L R = - 2 \ln [{(1 - α)}^{T - N} α^{N}] + 2 \ln ⌊ {(1 - \hat{P})}^{T - N} {\hat{P}}^{N} ⌋$ $$LR = - 2\ln \left[ {{{(1 - \alpha )}^{T - N}}{\alpha ^N}} \right] + 2\ln \left\lfloor {{{(1 - \hat P)}^{T - N}}{{\hat P}^N}} \right\rfloor$$

Under the original hypothesis, the statistic obeys a χ² distribution with 1 degree of freedom. If the test level is 0.05, the rejection domain of the test is ${L R > χ_{0.05}^{2} (1) = 3.841}$ $$\left\{ {LR > \chi _{0.05}^2(1) = 3.841} \right\}$$. If LR > 3.841, it means that the original hypothesis of the model is rejected, i.e., the model is poorly fitted.

3.4

Empirical analysis

3.4.1

Basic analysis of financial market data

1)

Selection of data

The stock price index is an indicator that describes the change of the total price level of the stock market, which is obtained by selecting a representative group of stocks, weighting their prices on average and through certain calculations. In this paper, the selection of the SSE index (also known as SSE Component Index), Shenzhen Component Index, due to the SSE index is the Shanghai Stock Exchange of the original SSE index was adjusted and renamed into the sample stocks in all shares of the stock to extract the most representative of the market in the kind of sample stocks, since July 1, 2002 since the official release. So the selected time period for the January 1, 2011 to December 31, 2022 of the SSE index, January 1, 2011 to December 31, 2022 of the Shenzhen Constituent Index, a total of 3,584 trading days of data for each index.

2)

Calculation of Returns (1)

Simple yield calculation formula

Let p_t be the price of the financial asset at moment t and p_t−1 be the price at moment t − 1, then the single period return on the financial asset can be defined as:

(14)

R_{t} = \frac{p_{t} - p_{t - 1}}{p_{t - 1}}

$${R_t} = \frac{{{p_t} - {p_{t - 1}}}}{{{p_{t - 1}}}}$$

In fact, returns are relative price changes, and returns as defined here are expressed in percentage terms. (2)

Logarithmic return formula

Given a nominal rate of return of R_n, a one-year effective rate of return of R_e, and m as the number of times a profit is earned in a year, then: (15) $R_{e} = {(1 + \frac{R_{n}}{m})}^{m} - 1$ $${R_e} = {(1 + \frac{{{R_n}}}{m})^m} - 1$$

When m → ∞, continuous compounding is obtained: (16) $R_{e} = \lim_{m \to \infty} {(1 + \frac{R_{n}}{m})}^{m} - 1 = e^{R_{n}} - 1$ $${R_e} = \mathop {\lim }\limits_{m \to \infty } {(1 + \frac{{{R_n}}}{m})^m} - 1 = {e^{{R_n}}} - 1$$

Denote by R_c the rate of return calculated using successive compounding, then: (17) $e^{R_{c}} = (1 + \frac{R_{n}}{m^{m}})$ $${e^{{R_{\text{c}}}}} = (1 + \frac{{{R_{\text{n}}}}}{{{m^m}}})$$

Combining equations (15) and (17) gives: (18) $R_{c} = \ln (R_{c})$ $${R_c} = \ln ({R_c})$$

Define equation (18) as the logarithmic rate of return.

The single-period logarithmic rate of return on a financial asset, r_t, is defined as the single-period simple rate of return, R_t, plus one taken as the natural logarithm, i.e: (19) $r_{t} = \ln (1 + R_{t}) = \ln (\frac{P_{t}}{P_{t - 1}})$ $${r_t} = \ln (1 + {R_t}) = \ln (\frac{{{P_t}}}{{{P_{t - 1}}}})$$

(3)

Properties of logarithmic rates of return

Since the range of values of the logarithmic rate of return extends to the entire domain of real numbers, it does not violate the principle of finite liabilities. Since it is easier to derive the properties of the sum of a time series than to derive the properties of the product of a time series, the definition of the logarithmic rate of return makes the statistical modeling of the rate of return much simpler.

The yield formula used in this paper is $r_{t} = 100 \times \ln (\frac{p_{t}}{p_{t - 1}})$ $${r_t} = 100 \times \ln (\frac{{{p_t}}}{{{p_{t - 1}}}})$$, and the resulting logarithmic yield series is denoted as {r_t}.

3)

Basic Characterization

All analyses in this section are done using R language statistical software programming. The results of the daily returns of SSE and SZCI are shown in Figure 2, SH180 denotes the SSE 180 index and Shenzhen denotes the SZCI, and the two stock price indices used below are the same. (a) and (b) denote the daily returns of the SSE 180 Index and the daily returns of the Shenzhen Composite Index, respectively. As can be seen from the figure, the daily return of the Shenzhen Composite Index is less volatile relative to the SSE 180 Index. There is volatility aggregation in both returns.

Table 1 shows the basic statistics of the daily returns of the two indices, and P-value is the corresponding J-B statistic P-value, which is calculated by R software and is not a specific value. From the table, it can be seen that the excess kurtosis of SSE 180 index is 0.9852, and the excess kurtosis of SZCI index is 6.3847. The excess kurtosis of both stock price indices is greater than 0, which indicates that the distribution of their series of returns are both more concentrated than the normal distribution, and that the distribution of returns in the Shenzhen stock market is more concentrated than that in the Shanghai stock market.

Table 1.

Basic statistics of daily returns for two indices

	SH180	Shenzhen
Mean	0.2846	0.0348
Std	1.8359	1.3574
Min	-5.8956	-4.7852
Max	5.9483	4.9543
Skewness	-0.3649	0.0586
Excess Kurtosis	0.9852	6.3847
Jarque Bera	1598.4873	8129.9035
P-value	<2.3e-18	<2.3e-18

The skewness of the return series of SSE 180 is -0.3649 and the skewness of the return series of SZCI is 0.0583, then the return series of these two markets do not obey the normal distribution, and the skewness of the return series of SSE 180 is less than 0, which indicates that this return series has a long left trailing tail. The skewness of the return series of the Shenzhen Composite Index is greater than 0, indicating that this return series has a long right trailing tail.

If the distribution is normal, then its skewness should be 0 and kurtosis equal to 3. The Jarque Bera test measures the degree to which the skewness and kurtosis deviate from 0 and 3. The JB test statistic is: (20) $J B = \frac{T}{6} [{\hat{S}}^{2} + \frac{1}{4} {(\hat{K} - 3)}^{3}]$ $$JB = \frac{T}{6}[{\hat S^2} + \frac{1}{4}{(\hat K - 3)^3}]$$

where T denotes the number of observations, and $\hat{S}$ $$\hat S$$ and $\hat{K}$ $$\hat K$$ are the skewness and kurtosis of the sample data estimates, respectively. Under the assumption that the observations are independent and follow a normal distribution, the JB statistic follows a χ²(2) distribution.

The JB statistic p values of the return series of SSE 180 and SZCI are <2.3e-18, indicating that the return distributions of the two markets are significantly different from the normal distribution.

The QQ plots of the two indices are shown in Figure 3. If the two indices have logarithmic returns obeying a normal distribution, their QQ plots should be basically consistent with the theoretical QQ plots (i.e., the straight lines in the figure), so the returns of the two markets are non-normal.

In summary, it can be obtained that the distribution of the return series of SSE 180 index and SZSE index are non-normal. 4)

Smoothness Test

Generally, before conducting time series analysis, the smoothness of the time series used should be tested, otherwise the meaning of the analysis will be lost. The standard method to test the smoothness of the time series is the unit root test. In this section, the lagged 20-order ADF test will be used to carry out the unit root test on the stock index return series, and the results are shown in Table 2. From the table, we can get that the Dickey-Fuller value of SSE 180 is between [-9.478,-9.007], and the Dickey-Fuller value of SZCI is between [-9.927,-9.511], and the corresponding P-value of all lagged orders is less than 0.01, which rejects the original hypothesis of the existence of a unit root, and indicates that the return series of this market are smooth.

5)

ARCH effect test

The Lagrange multiplier test, or ARCH-LM test, is used to test for the presence of the ARCH effect in the residual series.The original hypothesis of the ARCH-LM test is that there is no ARCH effect in the residual series up to the P order. In this section, the first 20 orders are selected to calculate the F-statistic values of the return series of the SSE 180 Index and the Shenzhen Composite Index to test whether there is an ARCH effect in the return series, and the results are shown in Table 3. From the table, it can be seen that when the 20th order is reached, the F-value of the return series of SSE 180 index and SZCI index is 197.843 and 678.493, respectively, and their corresponding concomitant probabilities are all <2.3e-18, so the original hypothesis that the residual series do not have ARCH effect is rejected, i.e., the return series of the two markets have a significant heteroskedasticity, which can be modeled by a GARCH-type model for data fitting.

Table 2.

ADF test of yield sequence

Lags	SH180	Shenzhen
1	-9.411(<0.001)	-9.529(<0.001)
2	-9.008(<0.001)	-9.655(<0.001)
3	-9.211(<0.001)	-9.942(<0.001)
4	-9.305(<0.001)	-9.789(<0.001)
5	-9.007(<0.001)	-9.889(<0.001)
6	-9.141(<0.001)	-9.840(<0.001)
7	-9.126(<0.001)	-9.763(<0.001)
8	-9.365(<0.001)	-9.781(<0.001)
9	-9.325(<0.001)	-9.588(<0.001)
10	-9.369(<0.001)	-9.739(<0.001)
11	-9.401(<0.001)	-9.920(<0.001)
12	-9.045(<0.001)	-9.877(<0.001)
13	-9.337(<0.001)	-9.685(<0.001)
14	-9.292(<0.001)	-9.553(<0.001)
15	-9.240(<0.001)	-9.541(<0.001)
16	-9.478(<0.001)	-9.927(<0.001)
17	-9.26(<0.001)	-9.567(<0.001)
18	-9.079(<0.001)	-9.529(<0.001)
19	-9.017(<0.001)	-9.622(<0.001)
20	-9.358(<0.001)	-9.511(<0.001)

Table 3.

ARCH-LM test of yield series

ARCH-LM	SH180	Shenzhen
1	198.631(<2.3e-18)	665.099(<2.3e-18)
2	198.981(<2.3e-18)	664.345(<2.3e-18)
3	189.063(<2.3e-18)	680.327(<2.3e-18)
4	197.476(<2.3e-18)	666.001(<2.3e-18)
5	186.311(<2.3e-18)	665.704(<2.3e-18)
6	181.904(<2.3e-18)	676.155(<2.3e-18)
7	197.224(<2.3e-18)	668.832(<2.3e-18)
8	190.624(<2.3e-18)	683.513(<2.3e-18)
9	178.073(<2.3e-18)	675.633(<2.3e-18)
10	195.854(<2.3e-18)	683.333(<2.3e-18)
11	187.703(<2.3e-18)	681.439(<2.3e-18)
12	188.342(<2.3e-18)	678.003(<2.3e-18)
13	186.013(<2.3e-18)	674.255(<2.3e-18)
14	198.126(<2.3e-18)	675.393(<2.3e-18)
15	191.694(<2.3e-18)	679.701(<2.3e-18)
16	184.751(<2.3e-18)	678.973(<2.3e-18)
17	177.697(<2.3e-18)	682.884(<2.3e-18)
18	193.612(<2.3e-18)	665.907(<2.3e-18)
19	186.970(<2.3e-18)	682.650(<2.3e-18)
20	197.843(<2.3e-18)	678.493(<2.3e-18)

3.4.2

VaR calculation results and analysis

In this paper, the closing prices of each of the 3,584 trading days of the SSE 180 Index and the Shenzhen Component Index between January 1, 2011 and December 31, 2022 are selected to calculate the VaR values. In order to simplify the calculation, this paper assumes that the returns of the SSE 180 Index and the Shenzhen Component Index are normally distributed, so the formula for calculating VaR by applying the GARCH family model is: (21) $V a R_{t} = N_{a}^{- 1} σ_{t} \sqrt{Δ T}$ $$Va{R_t} = N_a^{ - 1}{\sigma _t}\sqrt {\Delta T}$$

Where VaR_t denotes the insured value on a particular trading day, $N_{a}^{- 1}$ $$N_a^{ - 1}$$ denotes the insured value on a day at a confidence level of (1 − α) normal distribution quantile, with a 95% confidence interval selected in this paper, then $N_{a}^{- 1} = 1.644854$ $$N_a^{ - 1} = 1.644854$$, ΔT is the square root of the holding period, then ΔT = 1.36 of the daily insured value.

The descriptive statistics of VaR calculation results for each stock index are shown in Table 4. Analyzing the data in the table, it can be seen that in terms of the SSE 180 index, the daily VaR value calculated by the GARCH model is higher than the other two models, and its standard deviation is also the largest. In terms of SZSE index, the ARMA-GARCH model calculates higher VaR values than the other two models, and its standard deviation is also the largest. And from the comparison of the two indexes, SSE 180 and SZSEI, it can be seen that no matter which calculation method, the VaR value of SSE 180 is higher than that of SZSEI, which shows that the volatility of SSE market is stronger than that of SZSE market, and the corresponding risk is also higher.

Table 4.

Descriptive statistics of VaR for various stock index

Index and model	Mean	Max	Min	Std
SH180GARCH-VaR	0.287	0.046	0.016	0.008
SH180MA-GARCH-VaR	0.253	0.042	0.014	0.006
SH180ARMA-GARCH-VaR	0.231	0.044	0.011	0.006
ShenzhenGARCH-VaR	0.015	0.042	0.010	0.009
ShenzhenMA-GARCH-VaR	0.017	0.041	0.010	0.011
ShenzhenARMA-GARCH-VaR	0.020	0.043	0.011	0.012

3.4.3

Accuracy tests of the model

Failure frequency test was applied to test the accuracy of the models, and the results are shown in Table 5, where E, T, Er, and Tr are the expected number of days of failure, the actual number of days of failure, the expected failure rate, and the actual failure rate, respectively. From the table, it can be seen that the corresponding LR values of VaR values calculated by all indices using GARCH-type models are less than the critical values, indicating that these models have passed the model accuracy test. In the SSE 180 index, the actual failure rate of VaR calculated by the GARCH and ARMA-GARCH models is higher than the expected failure rate, thus underestimating the risk.The actual failure rate of VaR calculated by the MA-GARCH model is lower than the expected failure rate, thus overestimating the risk. In the SZSE, all models calculate VaR lower than the theoretical expected failure rate, indicating that all models underestimate risk. Comparison of the LR values of each model reveals that for the SSE 180 index, the models with the highest accuracy are the GARCH and ARMA-GARCH models, while for the SZSE index, the model with the highest accuracy is the ARMA-GARCH model.

Table 5.

Kupiec test results for various stock index

Index and model	E	T	Er	Tr	LR	χ25%(1)
SH180GARCH-VaR	180	192	5%	5.36%	0.017	3.732
SH180MA-GARCH-VaR	180	137	5%	3.82%	0.284	3.732
SH180ARMA-GARCH-VaR	180	194	5%	5.41%	0.018	3.732
ShenzhenGARCH-VaR	180	173	5%	4.83%	0.298	3.732
ShenzhenMA-GARCH-VaR	180	162	5%	4.52%	0.537	3.732
ShenzhenARMA-GARCH-VaR	180	171	5%	4.77%	0.089	3.732

4

Exploring the integration path of ideological and political education and quantitative analysis

4.1

Integration pathways

1)

Strengthen ideological and political education and improve professional ethics

In financial risk management, the moral standards and professional conduct of employees are key factors. Through regular ideological and political education and training, the legal awareness and compliance consciousness of practitioners can be enhanced, so that they can make choices in favor of the overall interests of the society in the face of conflicts of interest. In addition, a long-term mechanism has been established to incorporate ideological and political education into the daily management and assessment system to ensure its continuity and effectiveness. This not only helps to reduce moral risks arising from the pursuit of personal interests, but also enhances the stability and credibility of the entire financial industry.

2)

Construction of integration model and data collection

On the basis of the traditional quantitative analysis model, variables related to ideological and political education, such as moral standard and compliance awareness, are introduced. These variables are quantified by means of questionnaires and expert assessments, etc., and used as model input parameters to comprehensively assess financial risks. Meanwhile, a sound data collection and processing mechanism is established to ensure the reliability and validity of historical transaction data, financial statements, risk event records, and data related to ideological and political education of practitioners. Data processing includes cleaning, outlier processing, and standardization to ensure the accuracy of the model. In this way, financial risks can be assessed and managed comprehensively, and the accuracy and applicability of predictions can be improved.

3)

Continuous optimization and empirical research

Future research and practice should further optimize the design of the fusion model to improve its practicality and operability. Explore more variables that reflect still behavioral characteristics and psychological states, combining qualitative analysis and expert experience to cope with the changing market environment. Meanwhile, the validity of the model is verified through empirical studies and case studies. The study shows that the quantitative model integrating ideological and political education performs more prominently in high-risk situations and can significantly improve the accuracy of risk assessment. In addition, supervision and guidance of financial institutions’ risk management practices should be strengthened, and financial institutions should be encouraged and supported to integrate elements of ideological and political education in risk management, so as to enhance the comprehensive effectiveness of financial risk management.

4.2

Application effects

4.2.1

Quantitative measurement of financial risk

In order to analyze and quantify the regional financial risk in a more in-depth way, so as to calculate the change of financial risk before and after the application of the integration path of ideological and political education and quantitative analysis in this paper. In this study, the entropy weight method is finally used as the core method for quantifying regional financial risk, taking into account factors such as mathematical logic, degree of complexity, and assessment effect. Entropy weight method is one of the common methods for comprehensive analysis of multiple indicators [26], which assigns the indicators to the initial data after measuring the implied and associated information between the indicators, in order to reduce the subjective bias in the process of assigning, and to obtain a comprehensive evaluation index to reflect the comprehensive information.

Assuming that there is M year or project, N evaluation indicators, forming the original data matrix X = (x_ij)_{m × n}(i = 1, 2 ⋯, m, j = 1, 2, ⋯n), for a certain evaluation indicator x_j, the greater the difference in the value x_ij of the indicator, the greater the role of the indicator in the comprehensive evaluation. The specific steps for comprehensive evaluation using the entropy weight method are as follows:

Step 1: Normalize the raw data to eliminate the influence of different quantitative outlines: (22) $x_{i j} \frac{x_{i j} - \min (x_{i j})}{\max (x_{i j}) - \min (x_{i j})}$ $${x_{ij}}\frac{{{x_{ij}} - \min ({x_{ij}})}}{{\max ({x_{ij}}) - \min ({x_{ij}})}}$$

Step 2: Calculate the weight of the jnd indicator value for the ist year P_ij: (23) $P_{i j} = \frac{x_{i j}}{\sum_{i = 1}^{m} x_{i j}}$ $${P_{ij}} = \frac{{{x_{ij}}}}{{\sum\limits_{i = 1}^m {{x_{ij}}} }}$$

Step 3: Calculate the entropy value e_j and the coefficient of variation g_j for the indicators of item j: (24) $e_{j} = - k \sum_{i = 1}^{m} P_{i j} \ln P_{i j}$ $${e_j} = - k\sum\limits_{i = 1}^m {{P_{ij}}\ln {P_{ij}}}$$

where $k = \frac{1}{\ln m}, 0 \leq e_{j} \leq 1$ $$k = \frac{1}{{\ln m}},0 \leq {e_j} \leq 1$$. In information theory, the entropy value e_j indicates the degree of orderliness of the system, and the larger its value, the higher the orderliness of the system. Define the coefficient of variation g_j − 1 − e_j, the larger its value indicates the higher the disorder of the system, the greater the weight of the corresponding indicator in the composite index.

Step 4: Calculate the weights, i.e. the weights determined by the entropy weighting method w_j: (25) $w j = \frac{g_{j}}{\sum_{j = 1}^{n} g_{j}}$ $$wj = \frac{{{g_j}}}{{\sum\limits_{j = 1}^n {{g_j}} }}$$

Step 5: Obtain the comprehensive evaluation index FI_i of the entropy weighted hair (indicating the comprehensive evaluation index of the ind year): (26) $F I_{i} = (\sum_{j = 1}^{n} w_{j} P_{i j}) \times 10000$ $$F{I_i} = (\sum\limits_{j = 1}^n {{w_j}{P_{ij}}} ) \times 10000$$

4.2.2

Description of indicators and data

Two municipal districts, A and B, and three municipal counties, C, D, and E, of S city were selected. Due to the complexity of financial risk factors and different quantification standards, based on the distribution of financial risks in the jurisdiction understood and mastered in practice, this study selected eight key objects containing banking institutions and other key objects, as well as 14 financial risk indicators such as the bank non-performing loan ratio, to quantify the financial risks, and the results are shown in Table 6. According to the principle of entropy weight method, M in the sample is 5 and N is 14.

Table 6.

Index statistical analysis (Trillion %)

Selected object	Reference index	Mean	Max	Min
Banking organization	Non-performing loan balance(X1)	3.68	8.29	1.82
Banking organization	Non-performing loan ratio(X2)	1.36	2.78	0.43
Small loan company	Non-performing loan balance(X3)	1.89	2.46	1.09
Small loan company	Non-performing loan ratio(X4)	58.03	76.93	50.87
Guarantee company	Compensatory quantity(X5)	363	584	186
Guarantee company	Compensation amount(X6)	0.68	1.03	0.29
Farmers’ mutual aid society	Overdue amount(X7)	0.07	0.15	0.03
Enterprise	Risk quantity(X8)	11	19	5
Enterprise	Amount involved(X9)	4.12	6.86	1.92
Private financing	Number of cases(X10)	16	38	3
Private financing	Case amount(X11)	5.29	12.83	1.67
Real estate	Deposit ratio(X12)	16.83	22.54	14.91
Government finance	Notch rate(X13)	41.22	53.49	30.05
Government finance	Debt ratio(X14)	59.63	86.78	38.41

4.2.3

Analysis of results

Figure 4 shows the changes of financial risk in S city before and after the application of the fusion path of this paper, and O1-O8 denote the 8 research areas such as banking institutions, respectively. According to the financial risk measurement index constructed in this paper, combined with the comprehensive evaluation value calculated by the entropy weight method, it can be obtained that before the application of the fusion path of ideological and political education and quantitative analysis, the comprehensive evaluation value of the financial risk of the eight research objects is 10.40, 17.72, 11.91, 17.26, 10.71, 14.38, 16.14, 17.42, respectively, and the microfinance company (O2) has the financial risk is the highest, followed by government finance (O8), and banking institutions (O1) have relatively low financial risk. And after adopting the fusion path of this paper, the financial risk of eight research objects such as banking institutions in S city is reduced by 26.54%, 69.36%, 39.46%, 58.34%, 30.25%, 58.21%, 68.09%, and 67.91% compared with the pre-application, respectively, and the financial risk of microfinance companies has the highest rate of reduction, and the financial risk of real estate (O7) has the second highest rate of reduction. It is fully verified that the integration path of ideological and political education with quantitative analysis proposed in this paper can effectively reduce financial risks and improve the quality and efficiency of financial risk management.

5

Conclusion

This paper explores the integration of ideological and political education with quantitative analysis through the application of ideological and political education methods in financial risk management. The proposed integration path in this paper is applied to eight key research objects in financial management, including banking institutions in S city, to test its effectiveness. For the eight financial management key research objects, before the application of this paper’s integration path, the financial risk of small loan companies is the highest comprehensive evaluation value of 17.72, followed by the government finance of 17.42, and the financial risk of banking institutions is the lowest among the eight objects of 10.40. The main reason for this phenomenon is that the small loan companies are generally pursuing the supremacy of interests, and even for the sake of self-interests at the expense of the overall interests of the community, the ideological and political education for the employees of the company and the quantitative analysis. There is a significant lack of ideological and political education for company employees. In contrast, banks operate in a more formal way and actively undertake their own social responsibility, and most of them are able to provide ideological and political education for their staff’s professional conduct and legal and moral cultivation. After the application of this paper’s path, the financial risk assessment value of microfinance companies and government finance is 5.43 and 5.59, respectively, compared with a decrease of 69.36% and 67.91%, respectively. The integration path of this paper adds the element of ideological and political education in the quantitative analysis, so that each financial risk management object accepts the ideological and political education subconsciously in the quantitative analysis of its own financial risk. The synergistic effect of quantitative analysis and ideological and political education enables efficient and high-quality management of financial risks.

Language:: English

Publication timeframe:: 1 times per year
Journal Subjects:: Life Sciences, Life Sciences, other, Mathematics, Applied Mathematics, General Mathematics, Physics, Physics, other

Journal RSS Feed

Exploring the integration path of ideological and political education and quantitative analysis in financial risk management

Xiaolong Li

Ying Deng

Published Online: Mar 21, 2025

Received: Oct 16, 2024

Accepted: Feb 12, 2025

DOI: https://doi.org/10.2478/amns-2025-0673

KeywordsVaR model, Entropy weight method, Quantitative analysis, Ideological and political education, Financial risk management

© 2025 Xiaolong Li et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Keywords
VaR model, Entropy weight method, Quantitative analysis, Ideological and political education, Financial risk management