Some of Donald Trump’s policy proposals could have profound macroeconomic implications, but there is large uncertainty around the (net) economic effects a second Trump term would have. For instance, would the US dollar appreciate due to new tariffs, or would it fall in the face of Trump’s repeated vocal opposition to a strong dollar? And what would a Trump victory mean for US growth or the disinflationary process underway in both the US and the euro area? As the election draws closer, this uncertainty has already led to heightened volatility in financial markets, as documented by Albori and coauthors in a recent VoxEU contribution (Albori et al. 2024). Extending their analysis, in this column we use betting market data in a VAR model to assess the economic implications of a second Trump term from the perspective of financial market participants.

Measuring Trump’s victory odds using betting data

We directly measure the market’s evolving assessment of Trump’s victory prospects using data from prediction markets. These markets allow participants to bet money on certain events, including election outcomes. As with other financial market prices, betting quotes then contain all sorts of information that might affect the outcome of the bet, and have been used by Moramarco and coauthors to quantify political risk in a Vox contribution (Moramarco et al. 2020). For our analysis, we use implied probabilities of a Trump victory in the upcoming US presidential election from PredictIt and PolyMarket, as averaged and provided by Bloomberg.

Relative to election polling data – which were used by Albori et al. (2024) – betting odds come with several advantages. First, they account for the particularities of the US electoral system such as the Electoral College. ¹ An improvement, say, in polling numbers does not necessarily translate into better chances of actually winning the election. ² Second, polling data are gathered over several days and published with a lag.  ³ In contrast, betting odds respond to election-relevant news almost immediately in an information efficient way.

However, the odds of a Trump election win, and by implication betting quotes, will generally respond to all sorts of news and economic developments, giving rise to an identification problem. For instance, the publication of surprisingly high US inflation readings might lower the odds of a Democratic win because they could signal continued price pressures that weigh on the current administration’s perceived economic performance. To the extent that an inflation surprise also signals more restrictive US monetary policy, asset prices might fall. Therefore, an observed co-movement of betting odds and asset prices is not necessarily informative about what we are ultimately interested in, namely, the causal effect of changes in Trump’s likelihood to win the election, as interpreted by financial markets.

We overcome this identification problem by exploiting the real-time nature of betting quotes. Specifically, we measure the high-frequency movements of Trump betting odds around key election-related events (see Table 1). ⁴ These events clearly affected the markets’ assessment of the likelihood of a Trump victory, but were independent of other factors such as the state of the economy. This allows us to use these high-frequency movements as an instrumental variable in a financial market VAR, which we describe below. ⁵...

more at CEPR