Using the data from the spreadsheet of Gilles, the following graph indicates that the payload rises as the OWE increases. There is a sense of diminishing returns as the OWE rises, which reflects the increasing structural weight needed per kg of extra payload as the aircraft gets larger; it might also be related to range.
The black trend line deviates increasingly from the linear as the OWE increases. There is less scatter than I had expected from the discussion above.
There is the obvious exception with the Boeing 747SP.
Then when the aircraft are divided by range, they show the effect of fuel capacity on the OEW/payload relationship.
The aircraft have been divided into short range (<5000 km), long range (5,000 - 10,000 km), and ultra-long range (>10,000km). These are my arbitrary divisions, with divisions approximately at the 737-800 and 767-300ER ranges.
The linear trend lines show the differing OEW/payload relationships. As expected, the longer the range, the less "efficient" is the structural weight needed per kg of extra payload.
