[author]JIANG, SUN

[content]

Regional Disparities in Judicial Justice

—A Blinder-Oaxaca Decomposition of Sentencing Differences

Author Jiang Guohua

Professor, School of Law，Wuhan University

*Author Sun Qiwei

PhD Candidate, School of Law, Wuhan University

Abstract: Regional disparities in judicial decisions pose a risk of undermining judicial justice and have therefore always received significant attention from the Chinese judiciary. Taking criminal justice as an example， the Supreme People’s Court of China has been implementing sentencing standardization reforms since 2008， emphasizing that this is a long-term endeavor. On March 8， 2023， the Supreme People’s Court stated in its work report： “Maintain and promote social fairness； strengthen the unification of adjudication standards， and effectively prevent fairness and justice from being compromised due to regional, urban-rural, and industry differences. ” Despite years of effort and significant progress in sentencing standardization, there remains a risk of sentencing imbalances in China’s judiciary. Existing studies have shown that regional sentencing disparities objectively exist in China. However, due to limitations in research methods, there has been no fact-based response to whether these disparities are within reasonable limits and what factors primarily cause them. This study uses publicly available verdicts of bribery cases nationwide from 2020 as samples and employs the Blinder-Oaxaca decomposition method to quantify the extent of sentencing disparities between regions with different levels of economic development， as well as the contribution of relevant influencing factors to these disparities. Regarding overall differences： the ranking of the severity of bribery offenses from most to least serious is Region 3 > Region 1 > Region 2； the ranking of sentencing outcomes from harshest to lightest is Region 3 > Region 1 > Region 2； the ranking of strictness in judges’ application of the law from strictest to most lenient is Region 3> Region 2 >Region 1. It shows that although judges in different regions do not have uniform standards of discretion， they overall adhere to the principle of proportionality between crime and punishment. In terms of group differences， the ranking of judges’ discretionary power in determining the amount involved in the crime， from strict to lenient， is Region 3 > Region 2 and Region 3 > Region 1，consistent with the levels of regional economic development. This suggests that judges in more economically developed regions assess the social harmfulness of the crime amount as lower. Therefore， the following suggestions are put forward. Firstly， It is necessary to establish a judicial assistance system to provide assistance for judges to reasonably grasp the extent of their discretionary power. Secondly， the amount of crime should be converted based on the level of economic development to reduce the differences in judges’ determination of the amount. Thirdly, It is imperative establish a mathematical model to monitor the extent to which judges exercise their discretion. Fourthly, It is imperative strengthen the writing standards of judicial documents and comprehensively demonstrate the basis for judges’ discretion as much as possible.

1. Presentation of the problem

Regional differences in judicial decisions are a common phenomenon globally, and China is no exception.Article 4 of the Guiding Opinions on Uniform Application of Laws and Enhancing Search for Class Cases (for Trial Implementation) published by the Supreme People's Court in 2020 stipulates that the scope of search for class cases is limited to cases that have entered into force by the Supreme People's Court as well as by court decisions within the administrative division of the province, which means that the Supreme People's Court also recognizes that there is regional disparity in China's judicial decisions. This means that the Supreme People's Court also recognizes the existence of regional differences in judicial decisions in China. Regional differences in judicial decisions carry the risk of eroding justice.2023 The SPC's work report of March 8 emphasized that it “maintains and promotes social fairness. It will strengthen the uniformity of adjudication standards and effectively prevent justice from being compromised due to regional, urban-rural, and industrial differences”.

The CPC Central Committee attaches great importance to judicial justice. General Secretary Xi Jinping pointed out that “impartial justice is the last line of defense in maintaining social justice.” Article 33 of the Chinese Constitution states that “citizens are equal before the law.” “Fairness and equality are all attributable to and dependent on justice.” Accordingly, judicial justice means that the judiciary applies the law equally, i.e., equal treatment for equal situations and different treatment for different situations. Just as there are no two identical leaves in the world, there are no cases in the world that are identical in an absolute sense. Cases have similarities as well as differences. Therefore, judicial fairness requires judges to seek a balance between uniformity and differences when making decisions. The Supreme People's Court has made many efforts to this end, and a series of documents published by the Court have set out what is meant by “seeking a balance between uniformity and differences”. Uniformity” means that judges focus on the uniform application of the law when applying the law; ‘difference’ means that judges focus on the inclusiveness of differences while maintaining the uniform application of the law; and ‘seeking a balance’ means that judges are not to be confused with the uniform application of the same law, but are to seek a balance between the two. “Seeking balance” means that judges should be basically balanced in their decisions on cases with similar circumstances in the same period, in the same region, and on similar cases.

From the viewpoint of judicial practice, judging whether the regional differences in judicial decisions are in line with the principle of balance is of great significance to the in-depth promotion of judicial justice. The law of regional differences is the law reflected in the regional differences of judicial decisions, grasp this law can provide a basis for the quantitative assessment of judicial justice, but also for the promotion of judicial reform and judicial policy intervention to provide decision-making reference. Tracing back the existing literature, some scholars have confirmed the existence of regional differences in judicial decisions in China. However, is the existence of such differences premised on the adherence to the uniform application of the law? Do judges achieve a balance between uniformity and variability in the adjudication process? Currently, no research has been conducted to answer these questions on the basis of facts.

In order to explore whether this discrepancy is reasonable, taking criminal justice as an example, the depth of previous studies is further expanded by analyzing all publicly available criminal judgments on bribery in 31 provincial administrative regions of China in 2020. First, with reference to the conclusions of existing studies, the 31 provincial administrative regions were divided into three regions based on their rankings in terms of the level of comprehensive economic development, and the correlation between the statutory factors (amount and circumstances) and the sentencing outcomes in each of the three regions was quantified using the OLS method, and possible endogeneity problems were tested using the instrumental variables method; then, based on the analytical results of the regression model, the The Blinder-Oaxaca decomposition bisects differences in offense severity between districts and differences in judges' offense determinations (judge discretion) to support how and to what extent judge discretion affects sentencing district differences. Finally, the law of district differences in judicial justice is summarized, and corresponding recommendations are made for the further maintenance of judicial justice.

2. Literature Review and Research Hypotheses

Judicial justice is a concept with rich connotations. Before quantitative analysis, firstly, a brief review of the meaning and categorization of judicial justice based on existing research is conducted to clarify the context of judicial justice to be explored according to the mainstream viewpoints in the academic community. Second, the assessment schemes of existing studies on the quantification of judicial justice are sorted out, and then appropriate assessment schemes are proposed. Finally, a brief review of relevant research findings on the factors affecting regional disparities in sentencing is conducted, and research hypotheses are proposed.

2.1 Meaning and categorization of judicial impartiality

The meaning of “judicial justice” can be categorized into the following doctrines: application of the law, dispute resolution, combination, the results of individual cases, and the operation of judicial power. From different research perspectives, “judicial justice” mainly includes the following classifications: substantive justice and procedural justice, overall justice and individual justice, social justice and legal justice. Drawing on the results of previous research, judicial justice can be regarded as the fairness and justice that the judiciary follows and embodies when applying the law, and this fairness and justice refers to substantive justice, overall justice and legal justice.

2.2 Quantitative Assessment of Judicial Justice

In the past, the quantitative assessment of judicial justice program to “the same case, the same sentence” as the standard, test whether the results of judicial decisions in line with the “uniformity” principle. There are certain limitations to this evaluation standard: first, the inference of this evaluation standard is not rigorous enough. From the perspective of adjudication results, although the adjudication results of cases with similar facts in different regions have certain regional differences, we cannot conclude that justice is not fair on this basis. There may not be only one unique outcome for any case, and under the premise of adjudication in accordance with the law, the same judgment in the same case is only a possible result of the application of the law. Therefore, from the “same case, same judgment”, can not be reversed to deduce the conclusion that the judge unreasonable use of discretion. Secondly, this evaluation standard cannot directly touch on judicial justice. The same judgment reflects the public's expectation of judicial justice based on the simple concept of justice, which belongs to social justice. The same case with the same sentence as the standard for the trial differences in the verification, direct proof of the problem of judicial authority, can only indirectly verify the judicial justice.

In this regard, judges need to seek a balance between uniformity and differences in order to achieve judicial justice. “Uniformity” is manifested in the fact that the more serious the crime, the heavier the outcome of the decision, which should meet the requirements of legality, i.e., ‘the punishment fits the crime’; ‘disparity’ is manifested in the fact that the judge's discretionary power over Difference” manifests that the judge's discretionary impact on sentencing regional differences, although subjective determination, but not arbitrary, arbitrary, should meet the requirements of reasonableness, i.e., with an objective basis; ‘balanced’ manifests that the sentencing regional differences are mainly affected by the statutory factors, non-statutory factors on the sentencing regional differences in the impact of limited.

2.3 Factors affecting regional disparities in sentencing

Previous studies have concluded that both statutory factors (factors affecting the outcome of adjudication clearly stipulated in the law) and non-statutory factors (factors affecting the outcome of adjudication not clearly stipulated in the law, such as region, gender, age, etc.) can cause regional disparities in sentencing outcomes. Among them, the statutory factors cause regional differences in sentencing results because the severity of crimes varies from region to region; the non-statutory factors cause regional differences in sentencing results because judges in different regions determine the degree of harm of the same crime differently, i.e., there are differences in the discretion of judges in different regions. Research hypothesis 1 is thus obtained:

Research hypothesis 1: The sentencing result of the passive bribery offense is influenced by statutory and non-statutory factors.

According to article 386 of the Criminal Code, statutory factors include the amount and circumstances. Non-statutory factors are factors other than statutory factors, which are ultimately reflected in the exercise of the judge's discretion. It is worth noting that the magnitude of the judge's discretion is the proportion of the variance caused by the non-statutory factors to the total variance. The discrepancy caused by statutory factors is the explainable part, also known as the discrepancy caused by the endowment factor, which refers to the discrepancy in sentencing results caused by the difference in the degree of malignancy of the crime, i.e., the discrepancy in sentencing results caused by the difference in the size of the amount and the severity of the circumstances. Non-statutory factors are the unexplainable component, also known as variance due to circumstantial factors, which refers to variance due to factors other than amount and circumstances.

It has also been shown that the main reason for differences in judges' discretion between regions is the different levels of economic development. Different levels of economic development between regions lead to different judges in different regions to determine the degree of social harm in cases of the same amount of money differently.

This leads to research hypothesis 2: Research hypothesis 2: The determination of the degree of harm by judges in cases of the same amount of money is mainly affected by the level of regional economy.

Dividing regions based on the level of economic development can create economic differences in the external environment between geographic regions, i.e., it is possible to control the level of economic development in the environmental factors in the model, and therefore the differences caused by the environmental factors can be approximated as being mainly caused by the level of economic development. Among other factors, there are other factors in the environmental factors that affect the discretion of the judges, and their effects cannot be separated from the effects of the environmental factors due to the limitations of the data and research methods.

3. Research design

3.1 Sample selection and data sources

Based on the publicly available passive bribery judgments in 2020, the analysis of the passive bribery offense was conducted to explore the reasons for regional differences in sentencing. The reason is that it is easier to measure and analyze the statutory factors affecting criminal sentences than civil and administrative sentences. The search was conducted on March 1, 2021. According to the provisions of the time limit for the public disclosure of legal documents of the people's courts, the closed cases should be disclosed on the China Judgment and Decision Network within 15 days, theoretically, the search results on March 1, 2021 should have covered all the bribery cases closed in 2020 across the country. The specific steps of sample selection are shown as follows: first, 2256 judgment documents were retrieved by setting advanced search conditions, and all the judgment documents were downloaded and then uniformly numbered; second, the researchers determined the information needed for the empirical analysis through discussion and transformed it into the corresponding variables and options; third, the manual reading of the judgment documents was used to record the relevant information in the form of options into the Stata 16.0 software, and in order to guarantee the accuracy of the data, all data entry was completed and checked by two researchers individually. After deleting 189 judgments with blank or missing information and obtaining 2,067 judgments, the judgment information of 2,201 defendants was finally extracted (some of the documents contained information of multiple defendants).

In order to obtain different economic sub-districts, it is necessary to categorize the 31 provincial administrative regions based on the level of economic development. Drawing on the results of Wang Bingshen's study, they are categorized into five categories based on the comprehensive level of China's economic development. Due to space constraints and to simplify the analysis process, the five types of regions are consolidated into three categories: developed and more developed regions (Region I), better developed regions (Region II), and generally developed and more backward regions (Region III).

3.2 Research methodology and model construction

First of all, the ordinary least square method (ordinary least square method) is applied to regression analysis of the relevant factors affecting sentencing by region, and the following model is constructed with the first and second regions as an example:

(1)

(2)

The contribution (rate) of each influencing factor to regional differences in sentencing was measured on the basis of the regression analysis using the Blinder-Oaxaca decomposition. Combined with the results of the regression analysis, the formula for the Blinder-Oaxaca decomposition is shown below:

(3)

3.3 Descriptive Statistics

As the main sentence sentencing results include both years and months, for ease of operation, the use of rounding all converted to years, such as: sentenced to six months of imprisonment assigned to 1, sentenced to one year and nine months of imprisonment assigned to 2, and the rest of the set accordingly. The final sample size totaled 2,123 persons, with 533, 381 and 1,209 persons in districts I, II and III, respectively. As shown in Table 1, the amount and circumstances, the two statutory factors affecting sentencing for the offense of passive bribery, are the independent variables. In order to make the data satisfy normal distribution and facilitate statistical analysis, the bribe amount takes the natural logarithmic value. In terms of bribe amount, the average bribe amount of different economic regions in China in 2020 was ranked from highest to lowest: Zone 3 > Zone 1 > Zone 2. The selected circumstances covered all statutory sentencing circumstances, discretionary circumstances that appeared with high frequency in the judgment, and the nine other aggravating circumstances set out in paragraphs 2 and 3 of article 1 of the Interpretation on Several Issues Concerning the Application of Law in Handling Criminal Cases of Embezzlement and Bribery.

Table 1 Distribution of the amount and circumstances of passive bribery offences nationwide in 2020

4. Empirical analysis

4.1 Analysis of the main factors influencing sentencing outcomes in each region

In order to verify the robustness of the estimation results, influencing factors were included in stages to test the correlation between statutory factors and sentencing outcomes. Model 1 included only the amount of bribe for single-factor analysis; Model 2 included the amount of bribe, surrender, merit, complicity, confession or truthful statement, guilty plea, guilty repentance, return of stolen goods, solicitation of bribe and other aggravating circumstances in the regression model for multifactor analysis.

As shown in Table 2, within the same region, the estimation results of Models 1 and 2 are relatively robust; within different regions, the comparative analysis of the estimation results of Models 1 and 2 reveals that there are regional differences in the main penalty sentencing for the offence of passive bribery in China in 2020, which are reflected in the obvious differences in the regression coefficients between the amount and the main penalty sentencing results.

Table 2 OLS analysis of the main penalty sentencing results by region

Note: Due to space constraints, decomposition results for other variables are not shown; *, **, and *** indicate significant at the 10%, 5%, and 1% levels, respectively; robust standard errors are in parentheses.

The results of the OLS analyses in zones one, two, and three all show significant associations between bribe amount and the main sentencing outcome. Model I, with the inclusion of only the amount of the bribe, shows that for every 1% increase in the amount of the bribe, the main sentencing outcome increases by 0.01448, 0.01567 and 0.01638 years in Zones I, II and III, respectively. Model 2, after incorporating the bribe amount and the circumstance factors, was found to have regression coefficients that were both significant at the 1% level, with each 1% increase in the bribe amount increasing the main sentencing outcome by 0.01487, 0.01505 and 0.01641 years in Regions I, II and III, respectively, which made the model estimation results more robust. By way of comparison, it was found that judges' sentencing outcomes for the amount of the unit's bribe were ranked in descending order: zone 1 < zone 2 < zone 3, indicating that judges in more economically developed regions were less sensitive and more tolerant of the amount of the bribe.

4.2 Possible endogeneity problems

Endogeneity problems may arise from bias due to omitted variables. Although the model has incorporated, to the extent possible, the statutory influences on sentencing for passive bribery, there are still unobservable variables, such as judicial disclosure, which may have an impact on both the “outcome of sentencing for the main offense of passive bribery” and the “confessions of suspects for the offense of passive bribery”. “On the one hand, judicial openness has the potential to prevent judicial corruption. On the one hand, judicial openness has the effect of preventing judicial corruption. Judicial transparency is a manifestation of judicial openness. Based on the sociological theory of punishment, Wu Yuhao suggests that under China's existing trial management model, judicial transparency may affect the outcome of sentencing by judges, and judicial transparency means that sentencing decisions are more likely to be noticed, which is conducive to prompting judges to make prudent decisions. On the other hand, judicial openness has the function of publicizing and educating the rule of law. Practical circles believe that job crime suspects have special characteristics, have a high level of cultural knowledge, personal qualities and abilities are strong, the way to look at the problem is more rational, with the basis of education and penitentiary. Therefore, judicial openness is conducive to the education and sensitization of job crime suspects, thus successfully realizing the purpose of truthful confessions.

In order to solve the endogeneity problem, the number of times the 31 provincial discipline inspection commissions and supervisory committees carry out the “four forms” of supervision and discipline in 2020 is proposed to be used as an instrumental variable. First, there is a strong correlation between the number of times the discipline inspection commissions and supervisory committees carry out the “four forms” and the number of job crime suspects' confessions or truthful confessions. The “four forms” of supervising and enforcing discipline is a collection of theoretical and practical innovations and experiences in the fight against corruption since the 18th National Congress of the CPC, and it is the “basic anti-corruption policy” of the CPC and the state. General Secretary Xi Jinping emphasized that “the precise use of the ‘four forms’ allows Party members and cadres to feel the Party's strict control and generous love.” Practical circles believe that job crime suspects have long been educated and inculcated by the Party, and it is easier for them to identify with the interrogators' education and sensitization, so as to awaken the correct values of the suspects, and successfully achieve the purpose of the suspects' truthful and voluntary confessions. The results of the weak instrumental variables test also show that the number of “four forms” does have a significant correlation with whether the suspect chooses to confess or truthfully confess. Second, the “four forms” of supervision and discipline enforcement by discipline inspection commissions and supervisory committees are exogenous. General Secretary Xi Jinping pointed out, “We should effectively prevent corruption from growing, move the anti-corruption line of defense forward, strengthen day-to-day management and supervision, accurately apply the ‘four forms’, catch the smallest at an early stage, prevent the smallest from the smallest, and set up defenses at all levels.” In terms of corruption governance, the party's disciplinary work is different from the judiciary's trial work, so there are reasons to believe that supervision and discipline “four forms” is born out of the main criminal sentencing results of bribery. The analytical results of the over-identification test also do not reject the original hypothesis that “all instrumental variables are exogenous”, so there are reasons to believe that the selected instrumental variables satisfy the prerequisite of exogeneity. Third, the prerequisite for regression analysis using the instrumental variable method is that the confessions or truthful confessions of job crime suspects are endogenously selected. The number of times the provincial discipline inspection commissions and supervisory committees carry out the “four forms” of supervising and enforcing discipline as an instrumental variable is included in the corresponding regression model by economic region, and the P-values of the Durbin-Wu-Hausman (DWH) test are all significantly greater than 0.1, so the original hypothesis that the confessions of job crime suspects are exogenous is not rejected. hypothesis (see Table 3 for details). It is reasonable to believe that it is unlikely that the results of the regression analysis shown in Table 3 are affected by endogeneity issues leading to distorted results.

Table 3 Test results for instrumental variables

4.3 Blinder-Oaxaca method to decompose the influencing factors of regional differences in sentencing for the offense of passive bribery in China

As shown in Table 4, District 1 has 0.164 units more than District 2 overall, indicating that District 1 has a heavier sentencing outcome than District 2; the contribution of the interpretable component is 0.030, indicating that the offense is more serious in District 1 than in District 2; and the contribution of the non-interpretable component is 0.134, indicating that the judges in District 1 are more stringent in sentencing than those in District 2. Zone II has 0.423 units less overall than Zone III, indicating that the sentencing outcome in Zone II is less severe than in Zone III; the contribution of the explainable component is -0.462, indicating that the offense of passive bribery is less severe in Zone II than in Zone III; and the contribution of the unexplainable component is 0.039, indicating that the judges in Zone II are stricter in their sentencing than in Zone III. The overall reduction of 0.259 units in District I compared to District III indicates a lighter sentencing outcome in District I than in District III; the contribution of the explainable component of -0.411 indicates a lesser offense in District I than in District III; and the contribution of the unexplainable component of 0.152 indicates that the judges in District I were stricter in their sentencing than in District III.

Table 4 Comparison of the overall differences in the main criminal sentencing between zones

As shown in Table 5, between Districts I and II, the contribution of the bribe amount in the explainable and unexplainable parts was 0.087 and -0.072, respectively, neither of which was statistically significant under the 5% test level; between Districts II and III, the contribution of the bribe amount in the unexplainable part was -0.534, which was statistically significant under the 10% test level, suggesting that there were differences between the two There was a difference between the judges' determination of the amount of the passive bribe, resulting in a 1369.23 per cent expansion of the sentencing result between the two divisions; between Divisions I and III, the contribution of the amount of the passive bribe to the unexplained component was -0.617, which is statistically significant under the 5 per cent test level, indicating that there was a difference in the determination of the amount of the passive bribe by the judges in Divisions I and III, resulting in a 405.92 per cent expansion of the sentencing result between the two divisions.

Table 5 Contribution of the amount of the bribe to the difference in sentencing for the main penalty between divisions

Note: Due to space constraints, the decomposition results of other variables are not shown; *, **, and *** indicate significant at the 10%, 5%, and 1% levels, respectively.

5. Refinement of the Law of District Difference

By comparing and analyzing the information related to the 2020 national public verdicts on the crime of passive bribery, the Blinder-Oaxaca decomposition method was used to quantify the contribution of statutory factors (amount and circumstances) and non-statutory factors (judge's discretion) to the sentencing district differences on the basis of the regression analysis, so as to verify whether the sentencing district differences are in line with the requirements of judicial justice. Comparing the decomposition results of the three groups of regions, the law of regional differences in sentencing for passive bribery in China can be summarized:

Conclusion 1: The main sentencing results in each region correspond to the severity of their offenses. Comparison of the findings in Table 4 reveals that: the severity of the offense of passive bribery is ranked from heaviest to lightest: three regions > one region > two regions; the main penalty sentencing results for passive bribery are ranked from heaviest to lightest: three regions > one region > two regions. The ranking of the degree of severity of the main penalty sentencing results is the same as the ranking of the severity of the crime, indicating that the more serious the crime, the heavier the main penalty sentencing results in the region.

Conclusion 2: The level of economic development mainly affects the judges' determination of the degree of harm of the amount of the passive bribery offense. Based on the level of economic development, the 31 provincial-level administrative regions were divided into three regions, in descending order: Region 1 > Region 2 > Region 3. An analysis of the contribution of judges' discretion (the contribution of the unexplainable part) reveals that the factor with a significant impact on regional differences and with the highest overall contribution is the amount of passive bribery, indicating that differences in judges' determination of the degree of harm caused by the amount are the main reason for the differences in discretion between regions. The contribution of the unexplained component due to the amount of the bribe was -0.072 between Districts I and II, indicating that judges in District I were more lenient than judges in District II in their determination of the degree of harm of the same amount. The difference in the judge's determination of the level of harm of the amount could not significantly affect the difference in the outcome of the main sentence between District I and District II, probably due to the small difference in economic development between District I and District II. The contribution of the unexplained component due to the amount of the bribe was -0.534 between District II and District III, indicating that judges in District II were more lenient than judges in District III in determining the degree of harm for the same amount. The contribution of the unexplained part due to the amount of the bribe was -0.617 between District I and District III, indicating that the judges in District I were more lenient than the judges in District III in determining the degree of harm for the same amount. Accordingly, it can be seen that the order of judges' determination of the degree of harm of the amount of the crime from leniency to strictness is as follows: District 1 > District 2 > District 3, and compared with the order of the level of economic development of the three regions, it is not difficult to find out that the more economically developed the region, the lower the degree of harm of the same amount of the crime is determined by the judges.

Conclusion 3: The impact of judges' discretion on regional differences in the sentencing of the main penalty for the offense of passive bribery is limited. The overall difference between the main penalty sentencing in Region 1 and Region 2 was 0.164, of which 18.29% was caused by the amount and circumstances, and 81.71% by the judge's discretion, indicating that the regional differences in the main penalty sentencing between Region 1 and Region 2 were mainly caused by the judge's discretion. Despite the fact that non-statutory factors contributed 81.71% of the variance in principal sentencing between the two districts, the overall variance in principal sentencing between Districts I and II was relatively small, and it can be assumed that the variance due to judge discretion is still limited. The overall difference in principal sentencing between Regions 2 and 3 is -0.423, of which 109.22% of the difference is due to amount and circumstance, and -9.22% of the difference is due to judge's discretion, indicating that the regional difference in principal sentencing between Regions 2 and 3 is mainly due to amount and circumstance.The overall difference in the main penalty sentencing between Zone 1 and Zone 3 was -0.259, of which 158.69% was caused by the amount and circumstances, and -58.69% by the judge's discretion, indicating that the regional difference in the main penalty sentencing between Zone 1 and Zone 3 was mainly caused by the amount and circumstances.

By comparing the differences between judges' discretion and sentencing results for the offense of passive bribery in the same period and in different regions, it can be found that, on the whole, Chinese judicial decisions are fair. Conclusion 1 confirms that the more serious the offense, the heavier the sentencing result in the region, indicating that the results are in line with the principle of “the punishment fits the crime”. This is consistent with the purpose of “uniformity”, and the judicial decisions have been made in such a way that the same situation is treated equally, i.e., the same crime is punished equally. Conclusion 2 confirms that, in terms of the relationship between the non-statutory factor of region and the outcome of the decision, the judge's decision corresponds to the objective reality of the level of economic development of the region. This is consistent with the purpose of “difference”, and the regional differences in sentencing resulting from the judge's discretion are consistent with the level of regional economic development and are reasonable. The regional disparity in sentencing resulting from the judge's discretion, although subjective, is based on objective grounds. Conclusion 3 confirms that regional differences in sentencing due to the severity of the crime are higher than regional differences in sentencing due to judges' discretion, with the exception of the set of comparisons between Regions I and II. This is consistent with the purpose of “balance”, i.e., regional differences in sentencing are mainly caused by regional differences in the severity of the crime, and the impact of the judge's discretion on regional differences in sentencing is limited, i.e., the judicial decisions have taken into account the differences in accordance with the premise of uniformity.

The reason behind the law of regional differences in judicial justice lies in the differential order of judicial justice. Differential order of justice refers to the imbalance in the process of realizing justice, which is embodied in the links, levels, and parts of the priority in time and space. It requires that, in terms of space, the non-homogeneity and imbalance of the socio-economic characteristics of different regions should be taken into account from a practical point of view, and that, in the current situation where the scales of adjudication in different regions cannot be completely equalized, the spatial order of cases in different regions should be maintained in such a way that they are moderately different from each other in terms of the requirements and levels of justice achieved. Judicial decisions, as the result of judicial operation, are both the product of the judge's application of the law and the result of the interaction between justice and society. Therefore, the judicial decision is both absolute and objective and “timely”, relative or subjective. Differential justice is not contradictory to the principles of justice and equality; it is not a strategic choice to abandon the principles of justice and equality, but rather a strategic choice for socialist society to realize justice and social equality. China's vast territory, unbalanced economic and social development between regions, cases in different regions have both similarities and differences. The complexity of the object of adjudication determines that universal justice and differential justice in judicial adjudication should not be given preference. Therefore, the need for judges to maintain the uniformity of the ruling scale at the same time tolerate the differences between the cases, and seek a different order of justice.

6.Relevant Suggestions

In order to prevent the erosion of judicial justice due to regional differences, based on the above law, a few suggestions are made.

First, establish an adjudication support system to provide assistance to judges in rationally grasping the range of discretionary use. After analyzing the results of sentencing for the offence of passive bribery in different economic regions of China, it is found that the application of the law by judges basically conforms to the principle of “uniformity”. In order to consolidate the existing results and enhance the uniformity of adjudication scales between regions, it is necessary to further accelerate the in-depth integration of artificial intelligence technology and judicial work.

After more than a decade of exploration, China's judicial organs at the local level have accumulated some experience in the construction of adjudication support systems: in 2006, the Shandong court system developed a “computer-assisted sentencing system for 100 commonly used criminal law offenses”; in 2011, the Fujian Procuratorate system developed the “Sentencing Recommendation Management System”; and in 2016, the Guizhou Procuratorate system developed the “Sentencing Recommendation Management System”. In 2011, the Fujian Procuratorate system developed the “Sentencing Recommendation Management System”; in 2016, the Guizhou court system developed the “Big Data Sentencing System”; and in 2017, the Shanghai court developed the “Big Data Trial Support System”. In the Opinions on Standardizing and Strengthening the Judicial Application of Artificial Intelligence published by the Supreme People's Court in 2022, it was pointed out that “by 2025, a relatively complete judicial artificial intelligence technology application system will be basically completed, providing all-round intelligent support for the administration of justice for the people and the fair administration of justice”. Unlike the adjudication assistance system established at the local level, if a national adjudication assistance system is to be built, the differences in adjudication scales between regions need to be considered, so the difficulty of building the system is predictable. When utilizing artificial intelligence for sentencing estimation in the adjudication assistance system, not only does it need to take into account the collective experience of the judges, but it also needs to take into full consideration the regional idiosyncrasies.

Secondly, the amount of the crime is discounted according to the level of economic development, which reduces the discrepancy in the judge's determination of the amount. Discretion is a process of the judge's mind, and the process of discretion is not intuitive. The adjudication results of cases with similar merits often have differences, so their adjudication results are easily questioned.In 2012, the Supreme People's Court put forward in Article 3 of the Guiding Opinions on Practically Regulating the Exercise of Discretionary Powers in Trial Execution to Guarantee the Uniform Application of Laws, “efforts are being made to enhance the certainty and predictability of the exercise of discretionary powers. “. The greater the certainty and predictability of discretion, the less unreasonable suspicion judges will incur. Conclusion II has confirmed that judges are influenced by the level of economic development in determining the degree of harm in cases of the same amount of money, which can directly lead to regional differences in sentencing. Considering that the relevant indicators of the level of economic development are known and quantifiable, it is possible to convert the amount of the crime into a reduction.

From the point of view of the specific method of abatement, previous studies have given two ways: Wang Jianbo believes that the provincial people's courts and provincial people's procuratorates should, according to the state of economic development of the region and the state of local crime, within the range of the amount stipulated in the judicial interpretation, determine the specific amount of the standard for the implementation of the region, and report it to the Supreme People's Court and the Supreme People's Procuratorate for approval. Chen Lei believes that the legislative approach of the State Compensation Law can be borrowed to take the provincial administrative division as a unit and adopt the approach of multiplying the “same economic indicator (per capita disposable income of urban residents or average annual salary of employees) of each province” by a multiple, and initiate a legislative assessment every five years to determine whether and how to adjust on the basis of the research. Considering from the perspective of convenience, clarity and operability, Chen Lei's conversion standard has possibilities. In terms of the operational process, there are two options: one is to regularly publicize the commutation standard in the same way as the State Compensation Law, where judges do not use monetary units but time units when measuring the degree of harm of the amount of the crime; the other is to enter the sentencing information of the verdicts that have already come into effect into the adjudication assistance system, where the system will do the commutation, and provide a reference for the commuted sentencing range. Regardless of which approach is taken, converting the amount of the crime according to the level of economic development can reduce the differences in the determination of the degree of harm of the amount of the crime by judges in different regions, thus enhancing the certainty and predictability of the judge's discretion.

Third, a mathematical model was developed to monitor the extent of judges' discretion. The Blinder-Oaxaca decomposition method was used to quantify the contribution of statutory and non-statutory factors to regional differences in sentencing for passive bribery, confirming that China's sentencing for passive bribery takes into account “differences” while adhering to the premise of “uniformity”. However, in terms of methodological application, there are other environmental factors beyond the level of socio-economic development that affect judges' discretion, and the existing model fails to separate their effects from those of the overall environmental factors. In order to facilitate monitoring, it is necessary for future research to further explore the above issues and gradually optimize the monitoring model.

From the results of the study, the construction of a mathematical model to monitor judicial justice can be specifically divided into four steps: (1) the monitor classifies all cases in the country for the whole year according to the crime at the end of the year and analyzes each type of case one by one; (2) taking the crime of passive bribery as an example, taking a certain provincial administrative region as one whole and all the remaining provincial administrative regions as the other, the mathematical model is utilized to arrive at the total sentencing results differences, statutory factor contributions and non-statutory factor contributions, and repeat the same operation for each provincial administrative region and each type of case; (3) conduct research on provincial administrative regions with large total differences in sentencing outcomes and high non-statutory factor contributions, and specifically analyze whether there is an objective basis for the regional differences in sentencing resulting from judge discretion; (4) if there is not a sufficient basis and the contributions from non-statutory factors are (4) If there is no sufficient basis and the contribution of non-statutory factors is too high, it is necessary to control the extent of the use of judges' discretion accordingly.

The Blinder-Oaxaca decomposition method is used to construct a monitoring model to understand whether the existence of discrepancies is in line with the requirements of justice and whether the trend of discrepancies is reasonable. If the Blinder-Oaxaca decomposition result shows that the contribution of the explainable part is 100%, it means that the judge's discretion has been completely excluded, and “the same case and the same judgment” has been realized in the inter-district adjudication results. However, as Han Shiyan and Tian Hongjie said, “It is hard to imagine that the justice requirement of the law can be realized by completely eliminating discretion.” The use of mathematical modeling to monitor the results of sentencing is aimed at controlling the judge's discretion within a reasonable range, rather than pursuing a high degree of uniformity in the results of adjudication, which is an important guideline to be followed to promote judicial justice.

Fourthly, the standardization of the writing of decision documents should be strengthened, so that the basis of the judge's discretion can be shown as comprehensively as possible. Whether it is the use of adjudication support system to carry out sentencing reference, or to build mathematical model to monitor the judge sentencing results, all need to judge the information in the document from the text form into digital form, passive bribery judgment writing there are some irregularities lead to the judges to decide the basis of the extraction of the more difficult to monitor the use of judges' discretionary power to bring about obstacles, in particular: (1) part of the instrument of the bribe (1) Some of the instruments are inconsistent in their expression of the circumstances of bribery. For example, the term “first-time offender”, some instruments are expressed as “no prior criminal record” or “have not intentionally committed a crime”; the term “guilty plea and repentance”, some instruments are expressed as “no prior criminal record” or “have not intentionally committed a crime”; the term “guilty plea and repentance”, some instruments are expressed as “guilty plea and repentance”. For the term “guilty plea and repentance”, some instruments refer to “a sincere attitude in confessing guilt and a genuine intention to repent”; for the term “return of stolen goods”, some instruments refer to “return of illegal proceeds”; and for the term “first-time offender”, some instruments refer to “no previous criminal record” or “not having committed a crime intentionally”. The term “return of stolen goods” is expressed in some instruments as “return of illegal proceeds”. (2) The legal provisions cited in the “Court's opinion” section of some of the instruments did not correspond exactly to the circumstances mentioned in the “description of the case” section. Through interviews with the first-line judges learned that the judgment, “the court that” part of the law quoted may include both the defendant's claim, may also include the prosecution's claim of the answer; judgment quoted in the law may include both the judge that a certain plot is established on the basis of the judge may include the basis of the establishment of the case, may also include the basis of the case is not established. (3) Insufficient reasoning in the instrument. Sentencing reasoning is not simply repeat the facts and law, but has a strong theoretical and logical, to maintain the balance and consistency of sentencing, sentencing justice has important practical significance. Strengthening the reasoning of the instrument can prevent the alienation of discretionary power into judicial arbitrariness and improve the credibility and authority of judicial decisions.