PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 3, 2026Economics Letters0 citationsOpen Access

Regime-dependent predictive accuracy and structural stability of Eurozone inflation swaps

View Full Paper
RKRafael Kothe

Key Points

  • Results indicate a significant variance in inflation swap accuracy based on maturity and volatility regime.
  • 12-month inflation swaps are unbiased but lose forecast precision significantly during high-volatility states.
  • Longer maturities ranging from 24–36 months show notable positive biases and calibration losses in both volatility periods.
  • Structural failures are identified in longer tenors using moving block bootstrap analysis across regimes.

Abstract

This paper investigates the predictive accuracy of Eurozone inflation-linked swaps (ILS) across volatility regimes using a Markov-switching framework and regime-specific Mincer–Zarnowitz regressions. Results show a sharp divergence by maturity. While 12-month ILS remain approximately unbiased, their forecast precision (RMSE) deteriorates sharply in high-volatility states. In contrast, longer maturities (24–36 months) develop statistically significant, large positive biases (up to 378 basis points) and calibration losses evident across both volatility periods — a finding masked by standard asymptotic inference. These findings highlight the structural, persistent nature of the mispricing at medium horizons and the risk of policy misinterpretation. • Euro area inflation swap accuracy is highly dependent on maturity and volatility. • Short-term 12-month swaps remain unbiased but lose precision during market stress. • Longer maturities (24–36 months) exhibit large, statistically significant biases. • Moving block bootstrap reveals structural failure in longer tenors across regimes. • Findings suggest policy adjustment is needed when using market-based indicators.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Rafael Kothe (2026) studied this question.

synapsesocial.com/papers/69a7674fbadf0bb9e87e06cfhttps://doi.org/10.1016/j.econlet.2026.112826
Ask AI
Helpful
Bookmark
Share
View Full Paper