mikli made some interesting analysis of top 100 averages progression in different events a few years back:viewtopic.php?f=1&t=31194
It seems that long jump is one of the events where the improvement in yearly averages stopped in the late 80s or early 90s. The averages did not get better any more, but neither did they get worse. Instead, the apparent slump was caused by the fact that the likes of Powell, Pedroso and especially Lewis were rare individuals who were still nearly as much ahead of their time as Bob Beamon's 890 was in 1968, and after they retired the very top level marks fell.
It would be great if mikli found the time to update his calculations as years go by - perhaps that could give some idea if the general level of the event starts improving again after staying flat for over 20 years.