Why Economists Get Predictions Wrong (and Why They Don't)
GDP forecasts announcements receive ample media attention as they provide future expectations about the economy as a whole. Economic agents like managers and capital providers use this information in their investment decisions. However, GDP forecasts are often wildly different from actual GDP data, casting doubts on whether managers and investors should use them when considering their investment decisions. Although forecasting macroeconomic trends is not an easy task, there is concern that economists strategically reduce the accuracy of GDP forecasts. For instance, anecdotal evidence suggests that US economists release more optimistic forecasts when Presidential Elections are approaching and they are making donations to either candidate. Then, if the funded candidate loses the election, the same forecast turns out to be pessimistic. In "Economists' Political Donations and GDP Forecast Accuracy", recently published in the European Accounting Review, Andrea Bafundi (University of Padua) and I tackle this issue analyzing whether economists that donate to politicians release GDP forecasts that are less (or more) accurate.
On the one hand, economists making political donations could make less accurate forecasts as political donations reflect personal orientation and activism. As a result, when individuals make political donations, they may be biased in interpreting information according to their partisan beliefs, and the same is true for economists' political donations. On the other hand, political donations may foster interaction with politicians, facilitating access to political information. Political information is highly relevant for economists as it shapes economic policy (e.g., fiscal reforms and budget cuts) and thus affects macroeconomic trends they attempt to predict.
Analyzing a sample of U.S. GDP growth forecasts included the Wall Street Journal Economic Forecasting Survey (WSJ) from 2003 to 2020, we find that that economists with political donations are more accurate than their peers. Thus, although economists' political donations reflect their ideological orientation and can lead to a partisan bias, they can provide an information advantage in predicting GDP. We also document that the informational benefits of donating are stronger when the party financed by economists controls both the Senate and House of Representatives and when their employer also finances the same political party. Instead, the information advantage fades in periods of uncertainty and when economists are exposed to heterogeneous and more impartial political information. For instance, we document that economists making political donations issue less accurate GDP forecasts when their employers finance candidates from both political parties. In this case, economists receive more inputs but these are more likely to be discordant and difficult to process.
Thus, our evidence challenges anecdotal evidence suggesting that political donations can be a driver of forecasting error in estimating future GDP. At the same time, we delve into the potential circumstances under which donations can have downside effects. In this way, we have managed to deepen our understanding on why GDP forecasts are often inaccurate.