This is one of the engineering questions I would always have nightmares about. Remember that the impendance to airflow from your 120mm is not just in the 120mm-to-80mm funnel/fandapter constriction, but also at the end where the heatsink is, ie. the heatsink must be included for calculations/measurements. You can bet that if your HS has a high fin density, then the flowrate of air will seriously drop (esp. for axial fans) - 90cfm to just 10cfm, say. Calculations are difficult for this but if you can get hold of a anemometer (device that detects air speed), try and get an average speed of the air coming out of your heatsink and multiply that with the calculated opening area available on the HS for the airflow exit. If the fins are distributed evenly (ie. equally spaced, same heights), then the opening area is half of the area of that side no matter what fin density. Remember that the cfm of fans are rated at free-flow (near zero static pressures). Once obstructions come up, air flows slower and the 'accumulated' air ups the pressure -which the fan has the spend power to overcome.
To recap I think what your trying to say is, even if the funnel causes no loss, different fans even though they may claim similar specs 50mm x 50mm x 38mm and 50cfm (hypothetical fans) they may perform quite differently in application.
I should have realized this, in retrospect I don't know what I was thinking when I posted this question. The variables in this problem are mind boggling. I think I'll go back to my nuclear power plant it's less confusing.