This work studies three research objectives covering the following aspects:
- Impact of pending sanction for prosecutions on the independent professional work of filing charge sheets by Indian ACAs by applying simple linear regression technique.
- Impact of attitudinal and organizational factors on different categories of supervisors in punishing their agents on corruption charges by applying Chi-square test for independence
- Differences in appreciating the question of prima-facie evidence by different categories of supervisors in allowing the ACAs for investigations by employing two factors ANOVA test without replication.
6.1. Prior sanctions for prosecutions and performance of anti corruption agencies
Null hypothesis H01: Filing charge sheets by anti-corruption agencies is not impacted by pending sanction for prosecutions.
Alternative hypothesis: The null hypothesis if false.
In order to estimate the impact of the identified predictor on the dependent variable, year-wise percentage of cases pending for sanction for prosecutions, treated as predictor variable were regressed with the number of cases charge sheeted by the anti-corruption agencies (dependent variable Y). Correlation coefficient of -77208 shows high negative correlation between the variables.
As per regression results, multiple correlation coefficient of the model is 0.5961 indicating that about 60 % variability in the dependent variable is explained by the predictor. The remaining 40% of the variability could be due to various other factors. For example, in addition to pending sanction for prosecutions, filing charge sheets can be delayed by factors like, late receipt of experts’ reports, orders by the higher judiciary etc. This study used analysis of variance (ANOVA) technique for testing the significance of the linear regression model. As per results of analysis, the value of F statistics is 19.1868 and its associated P value is 0.000744. For the purpose of this analysis, level of significance was kept at 0.05. Since P value of the model is less than the alpha value of 0.05, it can be concluded that the relationship that exists between the predictor and outcome variables at 5% significance level is not by chance and has statistical reasons. In other words, the inference that the issue of pending sanctions for prosecutions impacts the functioning of anti-corruption agencies in India is not random. As per results of analysis, t-statistic for the intercept and slope are 7.778375 and -4.38027 respectively. In both cases, the corresponding P-values are less than the alpha value of 0.05. Thus, it is concluded that 1. Values of intercept and slope are not zero and 2. The regression line does not through the origin. Data analysis further shows that 95% confidence interval for the predictor is (-100.301, -34.0422). This indicates that rate of change in the dependent variable lies in the interval (100.301, -34.0422) 95% of times when there is a unit change in the predictor. Further, data analysis in this work has helped to evolve a statistical model for quantifying the impact of predictor on the dependent variable. As per regression results, a = 6268.424 and b = - 67.1714. Thus, the regression model derived from data analysis is: Y = - 67.1714 X + 6268.424.
Thus, from data analysis it can be inferred that delay caused in granting sanctions for prosecutions in corruption cases negatively impacts the professional work of anti-corruption agencies in India. In other words, merely by investigating a corruption case, the effectiveness of anti-corruption agencies cannot be improved because, filing charge sheets by the ACAs in India is the function of permissions granted by administrative authorities and the decision to decide prior permissions is external to the anti-corruption enforcement.
6.2 Attitudinal and Organizational factors as determinants of corruption control
Null hypothesis H02: Distribution of responses of supervisors in letting off their juniors in their corruption related misconduct is independent of comparison groups. In other words, there is no difference in the distribution of responses to the outcome across comparison groups.
Alternative hypothesis: The null hypothesis is false.
In order to test the null hypothesis, Chi-square tests for independence were done for the responses collected for nine identified factors that possibly influence the decisions of supervisors. The identified factors were independently studied and Chi-square values calculated. Decisions were taken on the basis of critical region approach by comparing the calculated Chi-square values with critical values at 5% significance level. For studying the null hypothesis, following test statistic was found appropriate as the sample size was large and the expected frequencies were above five for each of the response categories in each group.
![](https://myfiles.space/user_files/58653_1b1c6aeb34a62c68/58653_custom_files/img1624376048.jpg)
Decision rule: Decision rule in this case was framed on the basis of level of significance and degrees of freedom (df). At 5% level of significance, the critical chi square value for df=3, is: 7.815. If Chi-square values are less than the critical value, the null hypothesis will be true. Thus, the decision rule framed is as follows: Reject H02 if the calculated χ2> 7.815.
Results of data analysis can be tabulated as below: (see Figure 7)
Since, the calculated Chi- square values for the factors under study are greater than the critical value of 7.815, the null hypothesis is rejected at 5% significance level. Thus, it is claimed that the distribution of responses are not independent of comparison groups. In other words, there exists statistically significant relationship between the identified factors impacting the decision of supervisors in letting off their juniors of their corruption related misconduct and comparison groups. From data analysis, it can be inferred that factors prompting the decision of supervisors to let off their juniors of their corruption related misconduct need not be based on merits of the case. Such decisions can be subjective and driven by attitudinal and organizational factors. From data analysis, it can further be inferred that, in Indian bureaucratic organizations, getting the instances of corruption related allegations investigated is not the natural choice for supervisors and such decisions are the outcome of a variety of considerations that need not be in consonance with the requirements of rules.
6.3: Exercising options differently in allowing the anti corruption agencies to investigate instances of corruption
Null hypothesis: H03 - A: Average effect of responses of various categories of public servants in deciding a prima facie corruption case for criminal investigation is the same.
Null hypothesis: H04 - B: Average effect of different options available for public authorities in deciding a prima facie corruption case for criminal investigation is the same.
Alternative hypotheses: Null hypotheses are false.
Results of analysis of variance are as follow: (see Figure 8)
As per results, F value is larger than the corresponding F critical value with respect to the categories of public servants. Similarly, in the case of different options exercised, F value is larger than the corresponding F critical value. In both the cases, P value is less than 0.05. Thus, the null hypotheses, H03- A & H03- B that the average effects of responses of 1. Categories of public servants and 2. The options exercised by them in deciding a prima facie corruption case for criminal investigation are the same are rejected at the significance level of 0.05. Since, both the null hypotheses are rejected, pair wise comparison was done to study the statistically significant difference between various pairs of factors, if any. Pair wise comparison results below using randomized block design technique show significant difference between various options weighed by different categories of public authorities in allowing the ACAs for conducting investigations into suspected instances of corruption. (see Figure 9)
From results, it can be inferred that issue of allowing the ACAs to conduct investigations into corruption related allegations is subject to interpretations and appreciation of facts differently by different categories of public authorities. Institutional arrangements that enable different public authorities to interpret the same facts differently and to apply their discretionary powers to weigh convenient options can result in unregulated rent allocations and consequent rent seeking. Such arrangements can distort the performance of anti-corruption enforcement by not allowing them to conduct investigations into the instances of corruption.