Japanese election forecasting: Classic tests of a hard case

Abstract

We look at the substantive theory behind political forecasting models, which are generally based on theories of electoral behavior. We argue that the theory relied on for developing forecasting models should be as non-controversial as possible; what we call "core theory". We take the lessons of theory and practice in forecasting of western democracies and apply them to the unexplored Japanese case. We discover that, in fact, Japanese elections are quite amenable to forecasting. We observe the continuing cross-democracy generalizability of the core political economy model, a model that has bound together so much election forecasting work. © 2012 .