Alberto Simpser has a theory about extreme vote manipulation:
Why do politicians manipulate elections excessively? The conventional wisdom associates electoral manipulation with close elections and small margins of victory. In fact, however, many manipulated elections are won by overwhelming margins of victory, and some elections are manipulated even though the result is scarcely in doubt. I present a theory about the incentives that shape electoral manipulation under conditions that often characterize developing countries. The central idea is that in such settings, electoral manipulation, in addition to directly affecting vote totals, can influence expectations and consequently impact patterns of political participation. This simple idea goes a long way toward explaining observed patterns: When large-scale manipulation can help to deter opponents in the future, politicians may purposefully use it beyond the point necessary for victory. Evidence from a variety of regions and time periods suggests that large-scale manipulation and overwhelming margins of victory have often had such an effect.