FX forecasters show anti-herding, not consensus convergence

An ECB working paper finds mixed evidence of herding in foreign exchange rate forecasts, with results often pointing to anti-herding. The study suggests forecaster differences reflect heterogeneous views rather than systematic convergence.

Unpacking FX forecaster behavior

An ECB working paper explores herding behavior in foreign exchange (FX) rate forecasts, where individuals align with consensus.

The study uses a comprehensive dataset of nine major currencies from 1995 to 2024, featuring 40-50 forecasters per currency—a broader scope than typical literature.

Results show mixed evidence, largely favoring anti-herding, where forecasters move away from the consensus.

While some revision-based tests suggest herding with current consensus forecasts, this weakens significantly with lagged information.

Conversely, forecast-error based tests and over-reaction regressions more often indicate anti-herding, particularly at longer horizons.

The findings suggest forecaster differences stem from heterogeneous views, noise, or idiosyncratic error, not systematic convergence.

This interpretation holds across various uncertainty measures and FX predictors.

Beyond consensus: Testing for divergence

The paper's empirical framework analyzes individual-level forecasts against the consensus, employing regression-based tests and Bernhardt et al. statistics.

It uses both individual forecaster regressions and panel mean group estimation, accounting for cross-sectional dependence.

The extensive dataset, from Eikon Refinitiv, covers nominal exchange rate forecasts for 1, 3, 6, and 12-month horizons across nine major G10 currencies relative to the US dollar.

This data spans March 1995 to December 2024, with 40-50 forecasters per currency.

The study also investigates alternative explanations for expectation formation, including various measures of uncertainty, volatility, and FX predictor variables.

This approach extends existing literature by applying these methods to a large panel of individual exchange rate forecasts.

Challenging the herd mentality

This study provides a robust, data-rich analysis that challenges conventional assumptions about forecaster convergence in FX markets.

While some tests initially suggest herding, the overall balance of evidence strongly favors anti-herding, highlighting the role of diverse individual perspectives.

For policymakers, this implies that market expectations are less prone to groupthink, potentially leading to more resilient and less predictable market dynamics.