BetData's historical political betting odds have been used as a data source in a number of academic papers. Below is a list of published research that cites data from betdata.io.
Journal of Financial and Quantitative Analysis, 53(6), 2663–2683
This paper develops a method for forecasting exchange-rate movements around major political events by combining two sources of information: implied probabilities from betting markets and risk-neutral densities from currency option prices. Applied to the 2016 Brexit referendum and the 2016 US presidential election, the model produced accurate conditional exchange-rate forecasts for each possible outcome. A key finding is that markets were able to separately assess both the likelihood of a political event and its expected impact on exchange rates.
Sveriges Riksbank Working Paper Series, No. 364
This paper introduces a probabilistic voting model in which macroeconomic forecasters act as political agents with economic stakes and potential influence over voters. The theory predicts that such forecasters will publish strategically biased forecasts before a referendum. Testing this with high-frequency forecaster-level data around the 2016 Brexit referendum, the authors find that forecasters with financial stakes released significantly more pessimistic GDP growth forecasts than other forecasters, and were also more often incorrect. The estimated propaganda bias explains up to 50% of these forecasters' forecast error. The paper uses daily average odds recorded by BetData as a measure of the betting market's implied probability of Brexit.
CESifo Working Paper, No. 11221
A revised and expanded version of Cipullo & Reslow (2019), this working paper further develops the theoretical model of strategic forecast bias in referendum settings. The authors develop a model that yields three testable predictions: forecasters are more likely to bias forecasts for policies that are less likely to win or carry greater economic uncertainty; forecast bias increases with the forecaster's influence on voter choice; and biased forecasters maintain their distorted predictions even after the referendum. Empirical evidence from the Brexit referendum, using BetData's UK EU Referendum 2016 historical odds, supports all three predictions.
Centre for Macroeconomics Discussion Paper, No. 2019-21
This paper examines why the UK economy did not experience the widely predicted sharp recession after the Brexit referendum, but instead exhibited a slow, persistent decline in production. Using a heterogeneous firms model with capital adjustment costs, the authors model the referendum as an uncertain news shock about future fundamentals. BetData's implied probability series for Brexit outcomes is used as part of the paper's illustration of Brexit policy uncertainty. The estimated long-run effects of Brexit imply GDP losses of 4.8% under a Soft Brexit scenario and 7.7% under a Hard Brexit scenario.