Is internet penetration in constant terms across the whole window? (edited to fix a unit)

Kasper Bakker The restatement in 2008 affected the North region disproportionately, with revisions ranging from 3% to 18%, while the South region saw only minor adjustments. This asymmetry suggests either a localized data-quality issue or a genuine structural change in the North. Should we document this as a known comparability break, or should we attempt to restate the entire time series consistently?