Variable as confounding if it influences other factors in opposite directions?

I examine the relationship between population density (PD) and the insurance density (ID) taking into account different market exploitations (ME) of an insurance company in municipalities. The correlation matrix (Pearson) shows the following relationships (all high signifikant): PD vs. ID (-) (according to the theory a positive relationship is expacted) ME vs. ID (+) ME vs. PD (-) Is the ME variable a confounding factor if it effects PD and ID in different directions (positive and negative)? Of course it is a controll variable in my regression models but I´m not sure if I can name it as confounding cause normally it should influence both variables in the same direction?