ABSTRACT
Referendums on European integration and EU/EEC membership have received a fair bit of coverage (Mendez, Mendez and Triga, 2014; Hobolt, 2009; Atikcan, 2015; Wellings and Vines 2015; Hainsworth, 2006; Criddle, 1993). While political scientists have addressed the “puzzle that governments …voluntarily introduce another veto player in the decision-making process” (Hobolt, 2006: 157; see also: Morel, 2007), less work has been devoted to understanding why the voters have rejected or endorsed further integration and EU/EEC membership. Although there has been comparative writings as to what determines the outcome of these referendums (Franklin, van der Eijk and Marsh, 1995; Qvortrup, 2016), psephologists have not yet devised a predictive model for forecasting EU referendums of the type developed by academics studying parliamentary and presidential elections (Lewis-Beck, 1990, 2014). In this regard, the study of referendums lags behind that of elections. In 1991, Saunders correctly predicted the outcome of the 1992 British Parliamentary election – which many pollsters got wrong (Saunders, 1991). Similar results are missing in the realm of referendum studies. To be sure, in an aside one researcher predicted the outcome of the 2016 Brexit referendum in January 2016, when he observed, “based solely on this statistical analysis, we would expect the current government to lose the referendum by 4 per cent” (Qvortrup, 2016: 65). However, this article did not spell out the methodology, nor did it rely on rigorous testing of hypotheses. Further, this article was not tested against the outcomes of previous referendums. The present article expands the argument of the said model, but develops a general model for predicting EU referendums and tests it on all the 46 previous referendums held on EU membership or treaty changes 1972–2016.
