Thank you for the reply.
The flow was set using the frame motion, and the force outputs are in the correct directions. Below are additional outputs including velocity contour. The pathlines show less interaction than expected, and forces are too low.
The second approach was to use a stationary frame and rotate the flow around the domain. This resulted in better forces, although the flow was only ideal around the car. The pathlines and contours sohw turbulence not seen with the frame motion.
Based on the results and the fact that only the fluid zone was changed from moving to stationary, I think the problem stems from the enclosure + moving frame pairing. The car seems to be present when flow first makes contact in moving frame, but then pathlines look like they stop, and force values are far too low.