Below is a study from 2013 that has cfd and wind tunnel results that shed further light on the aerodynamic interaction between drafting cyclists. Eyeballing figure 6b indicates that at a separation distance of ~15cm (I'd consider this typical) the lead cyclist can see around a 1-2% reduction in CxA. A few things here when placing these CFD and wind tunnel result into context:
1) a simplified inviscid analysis yields similar results
2) using 1:10 scale model race car data yields similar results
3) table 5 is something to consider when observing follow car separation distance during TT's:
4) if you are reporting 4x-8x+ what is suggested by these multiple independent lines of inquiry, check your experimental setup, and test again, and again, and again...
5) if you are using a powermeter to quantify this leading cyclist drafting effect, you've got your work cut out for you
Here's the link to the study: