What is the best way to forecast the outcome of the Scottish referendum on independence? This is a vitally important question for individuals, organisations and businesses trying to plan for the post-referendum period. This paper introduces an alternative way of forecasting the referendum outcome, which under certain assumptions will produce a forecast with good statistical properties. It is based on the use of what are now commonly referred to as “prediction markets”. In turn, the superiority of the use of prediction markets in forecasting over other types of forecast is based on a somewhat abstruse theorem in financial economics that is known as the “efficient markets hypothesis” (EMH).

This paper begins by describing prediction markets and how they relate to the EMH. It then uses these arguments to develop a current forecast of the likely outcome of the referendum before going on to discuss how these forecasts relate to opinion polls which measure voting intentions in the referendum.