The alternate universe I live in is where codes regarding duct designs exist, are generally enforced, inspectors keep track, and contractors understand the requirements. I get that this not universal across the US, but it's mostly true in MA & NH. In CA under Title 24 you BET they install jump ducts or other return paths from doored-off rooms- it's cheaper to do it right ahead of time than to rectify it after the inspectors call you on it. (They seal and test duct leakage there too.) In states that have adopted IRC2009 or IRC2012 (or comparable) standards they pressure-test the house as part of the commissioning too, but I s'pose
that isn't happening in AZ yet. But even where local codes may ignore the issue doesn't mean it isn't worth addressing.
In the alternate universe I live in, air that's blown out of the house from one location creates a low pressure drawing air in elsewhere, and this is
YOUR universe too! By pressurizing one room to above atmospheric, it depressurizes another below atmospheric. All exfiltration volume has an equal and opposite infiltration volume. Just because the undefined place where the air is infiltrating isn't the room with the inadequate return doesn't mean that increasing the impedance of return path isn't responsible for creating a larger pressure difference to drive infiltration- it absolutely DOES!
Of course it's the pressure difference between the room with the supply duct(s) an the room(s) with the returns that causes the air motion, but by keeping the pressure requirements low within the pressure boundary of the house it keeps the exfiltration/infiltration path from becoming a large fraction of the return. The impedance of a jump duct or 2" door cut is still substantially higher than a path completely within a room with a dedicated return or an open door, but still sufficiently low to keep the exfiltration/infiltration paths from moving a lot of air in and out of the building.
The leakier the building envelope, the larger the fraction of air volume taking paths outside the thermal & pressure envelope of the house. IRC 2009 calls out 7 air changes per hour @ 50 pascals (7ACH/50), which isn't super-tight, but tighter than you'd be in places where air-sealing is given little or no attention during construction. Almost all housing built in New England after 1980 would pass, or could pass with a very modest effort at air sealing. IRC 2012 calls out 3ACH/50 which still isn't super-tight, but more difficult to retrofit in houses that were built with little regard to air-tightness. But installing jump-ducts is still cost effective just on utility savings in a 3ACH/50 house. Canadian R2000 standard housing calls out 1.5ACH/50 max, which may or may not be leaky enough to warrant guaranteeing very low impedance return paths from a total heating cost point of view, but comfort might suffer in rooms that don't get the designed amount of flow.
The leakage paths will vary- windows are the dead-obvious but not always the largest exfiltration/infiltration paths. Electrical penetrations in walls and ceilings (particularly recessed/pot lighting fixures) can often be bigger paths than the windows & doors. From the whole house perspective the foundation sills & band joists can be HUGE air leaks that become part of the infiltration/exfiltration path.
In chefwong's D.C. location anything that reduces infiltration has a comfort benefit, both winter and summer, given the relatively high summertime dew points and not-exactly-balmy overnight low temps in winter. It's much less of a comfort (or energy use) issue in say, San Diego CA.