Memory is fallible. I posted a comment on the Wonkblog yesterday recalling the 1986 Gramm-Rudman-Hollings reductions. Easy to do, since it was agony for us. All fine, except I wrote it was a 5.6 percent reduction. Did a little searching this morning; it was actually 4.3 percent.
So memory is fallible, but I know I walked to and from school uphill both ways.