My copy of Hucho shows very clearly what you are talking about, but they use a standard sedan, not a "wagon" type vehicle. What we are shooting for is CdA reduction, not just Area reduction. If the re-shaping reduces the Cd enough, the CdA will reduce enough to offset or overcome the increased Area. In Hucho's book, it shows how increasing the height of a standard sedan's roof impacts the overall drag force (CdA) encountered. It's Section 4.4.4 Roof, and Figure 4.41. In his example, l_r is used for the length of the roof, and a_r is used for the increased height of the roof. He calls a_r/l_r the "roof camber". He finds the lowest Cd at roof camber = 0.065 (ish), but the lowest CdA is 0.0175 (ish) or less. Remember, this was for a sedan, not a wagon, so YMMV.
Do you have attached flow on the roof of your vehicle now? If so, it is less likely that messing with the roof will help, and more likely that adding some boat tailing onto the rear hatch will help. If your vehicle has poor attachment/no attachment on the roof, you must correct that before a boat tail will do anything positive.
If you try "raising the roof", document what you do, and what the results are. I would start out doing a tuft test on the roof of the car. Video tape it with your phone and see what's happening.
Good attachment?
If your roof is 9' long (108"), I would try a 2" rise, peaking at the center. This will give your tail end a bit of a downward angle that should help your boat tail be a bit more effective.
Poor attachment?
Keep adding height and rounding the FRONT edge of the car over until you get good attached flow. Following a gentle arc, take the rounding you started at the front all the way to the back of the car. The Ideal Template would be a good place to start
.