**5.1 Cointegration results**

Although other studies do not support the use of PMG cointegration test, this chapter carried the cointegration test using the Kao. The test results suggest the presence of cointegration between economic variables and crime at 1 percent level of significance. Since the cointegration among variables has been established, a long-run and short-run relationship was discussed using the Pooled Mean Group in section 5.2.

### **5.2 PMG results and discussions**

The empirical results show a Hausman test of 0.3182 which accepts the null hypothesis that PMG is the appropriate estimator compared to MG and DFE. Therefore, this chapter used the PMG and compared with other estimators for robustness of the results. The PMG results illustrate that economic growth inversely influence crime activities. A 1 percent increase in economic growth decreases crime

by 0.48 percent. This negative coefficient is line with both the MG and DFE models and this supports the Becker's economic theory that states that economic growth reduces the level of crime activities in an economy. Furthermore, these results are in sync with other studies such as [21] who posit that an increase in economic growth reduces the level of crime in an economy. However, Mulok et al. [22] found interesting results that an increase in goods and services increases the number of crimes committed in an economy. Education is inversely related to crime in the Gauteng municipalities. A 1 percent increase in post-secondary education reduces crime by 0.52 percent. This result was expected and the findings are in line with the study done by Garidzirai and Zhanje [19] and Jonck et al. [20] who concluded that education gives a household exposure that reduces crime rates.

In line with a priori expectations, employment was found to inversely influencing crime in the Gauteng local municipalities. This result confirms the findings of Tang and Lean [14] who concluded that employment keeps individuals busy to the extent that they do not contemplate of crime activities. A 1 percent increase in employment reduces crime activities by 0.93 percent. The PMG results also illustrate that income inequality positively influence crime in Gauteng local municipalities. A 1 percent increase in income inequality increase the crimes by 0.87 percent. This result was expected and confirms the findings by Anser et al. [12] and Kingston [13] who concluded that income inequality increase crime activities in an economy. Poverty was significant and positively influencing crime which suggests that more people living in poverty are likely to commit crime in Gauteng local municipalities. This is shown by a 1 percent increase in poverty which leads to a 0.69 percent in crime activities. Cheteni et al. [7] and Dong et al. [15] also share the same sentiments that poor people have a higher probability of committing crime. This is also in line with the strain theory which stipulates that individuals with the low level of income tend to be frustrated when they are surrounded by those with high level of income [10]. Surprisingly, trade openness was found to be positively influencing crime but statistically insignificant.


*Source: Own compilation.*

*Note: \*, \*\*, \*\*\* represents 10, 5 and 1 percent respectively. Figures in parenthesis are T-statistics. Hausman test p-value 0.3182.*

#### **Table 5.**

*Long-run and short-run results.*

*An Analysis of Economic Determinants and Crime in Selected Gauteng Local Municipalities DOI: http://dx.doi.org/10.5772/intechopen.96339*

**Table 5** also shows an Error Correction Model estimated using the PMG, MG and DFE estimators. Since Hausman test proposed the use of PMG over MG and DFE showing a significant and negative ECT of −0.6184. This means that 61.84 percent of disequilibrium in the Gauteng municipalities is corrected in the upcoming period. Thus, the model moves back to equilibrium after 1 year 6 months (1/0.6184). Banerjee et al., [31] share the notion that the higher ECT the stable the relationship between economic variables and crime in the Gauteng provinces.
