From what I've heard, airlines typically use 55-65% reliability in their block times. This means that of 100 flights, 55-65 of them arrive at or under the published block time. Northwest I heard went to 50% block times, where half of all flights are automatically late. This was to save pilot costs.
It's a toss-up though, since lowering the block time makes for less crew costs (most crews are paid either the scheduled hours or flown hours, whichever is greater) while hurting on-time reliability and likely missed connections. Raising the block time makes for better on-time at the expense of crew costs and connections.
Don't forget connections in this equation. Say you are flying LAX
-Europe, and your connection is planned right at the minimum connection time, say 45 minutes. If your airline uses 50% reliability, then you have a 50% chance of being late for your European departure at JFK
, even if you leave LAX
on time. Raising that to 65% means raising the block time, but now your connection time is only 38 minutes (example), so your routing is no longer legal. Retiming the LAXJFK flight might work, but now you may miss the inbound traffic from HNL
. Cutting the minimum connection time is also something you don't want to do (especially to International flights) since that will cause in increase in DB
's and general bad feelings by the passengers that missed it.
This is the stuff that airline managers have to figure out regularly.
Measure to the millimeter, mark with a crayon, cut with an axe.