<p>Uncovered interest parity (UIP) is routinely rejected at short horizons, motivating departures toward heterogeneous-expectations frameworks. We estimate a behavioral UIP (BUIP) law of motion for monthly exchange-rate returns, where expectations form as a profitability-weighted average of a chartist (trend-following) rule and a fundamentalist rule anchored to absolute purchasing power parity (PPP). Building on Proaño (<CitationRef CitationID="CR68">2011</CitationRef>, <CitationRef CitationID="CR69">2013</CitationRef>), rule shares evolve through a discrete-choice mechanism. We compare the structural BUIP against a reduced-form logistic smooth-transition regression across fourteen bilateral USD exchange rates (seven advanced, seven emerging) from the early 1980s to 2025. We reject linearity for 13 of 14 currencies, with nonlinear models reducing in-sample fit. However, the Akaike information criterion frequently prefers the linear benchmark in advanced economies, suggesting nonlinearity is most rewarded in emerging markets. The structural BUIP identifies stabilizing PPP-consistent mean reversion in seven countries. While embedding heterogeneous beliefs into UIP partly reconciles the parity condition with nonlinear dynamics, the mixed evidence on PPP correction highlights the limitations of a “one-model-fits-all” approach. Ultimately, the choice of fundamental anchor appears to be highly market- and time-dependent.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Nonlinear exchange rate dynamics and behavioral expectations: New evidence from selected emerging and advanced economies

  • Lorenzo Biondi,
  • Christian R. Proaño

摘要

Uncovered interest parity (UIP) is routinely rejected at short horizons, motivating departures toward heterogeneous-expectations frameworks. We estimate a behavioral UIP (BUIP) law of motion for monthly exchange-rate returns, where expectations form as a profitability-weighted average of a chartist (trend-following) rule and a fundamentalist rule anchored to absolute purchasing power parity (PPP). Building on Proaño (2011, 2013), rule shares evolve through a discrete-choice mechanism. We compare the structural BUIP against a reduced-form logistic smooth-transition regression across fourteen bilateral USD exchange rates (seven advanced, seven emerging) from the early 1980s to 2025. We reject linearity for 13 of 14 currencies, with nonlinear models reducing in-sample fit. However, the Akaike information criterion frequently prefers the linear benchmark in advanced economies, suggesting nonlinearity is most rewarded in emerging markets. The structural BUIP identifies stabilizing PPP-consistent mean reversion in seven countries. While embedding heterogeneous beliefs into UIP partly reconciles the parity condition with nonlinear dynamics, the mixed evidence on PPP correction highlights the limitations of a “one-model-fits-all” approach. Ultimately, the choice of fundamental anchor appears to be highly market- and time-dependent.