Scholarlink

<p>This study re-examines the Feldstein–Horioka puzzle using Adaptive Elastic Net regression with ten-year rolling-window estimation, applied to a panel of over 100 countries observed between 1960 and 2022. The Feldstein–Horioka puzzle has proved remarkably durable, yet uniform specifications may conflate structural capital immobility with transitory policy effects—particularly in emerging markets characterised by frequent regime changes. The global savings-retention coefficient falls substantially below the original Feldstein–Horioka estimate, but this aggregate attenuation conceals a sharp dichotomy. OECD economies exhibit a gradual secular decline consistent with progressive financial integration. Emerging markets, in contrast, display pronounced within-country volatility driven by discrete policy regime changes and crisis episodes rather than smooth liberalisation. Türkiye exemplifies this regime dependence, with its coefficient moving sharply across phases of pre-liberalisation, post-liberalisation instability, crisis, and post-crisis restructuring; India, Thailand, Brazil, and Argentina display comparable policy-driven trajectories. Two-way fixed-effects estimation, PCA-augmented specifications that partial out common macroeconomic shocks, lead-savings placebo tests, within-country permutation exercises, and sensitivity analyses across alternative regularisation configurations confirm the robustness of these dynamics. The findings indicate that the emerging-market ‘puzzle’ reflects transitory policy configurations rather than genuine capital immobility, and that time-varying estimation is a methodological necessity—not an optional refinement—for capturing capital mobility in economies characterised by structural discontinuity.</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

Regime-dependent savings retention across emerging markets rolling-window adaptive elastic net evidence from a cross-country panel

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Insu Choi

摘要

This study re-examines the Feldstein–Horioka puzzle using Adaptive Elastic Net regression with ten-year rolling-window estimation, applied to a panel of over 100 countries observed between 1960 and 2022. The Feldstein–Horioka puzzle has proved remarkably durable, yet uniform specifications may conflate structural capital immobility with transitory policy effects—particularly in emerging markets characterised by frequent regime changes. The global savings-retention coefficient falls substantially below the original Feldstein–Horioka estimate, but this aggregate attenuation conceals a sharp dichotomy. OECD economies exhibit a gradual secular decline consistent with progressive financial integration. Emerging markets, in contrast, display pronounced within-country volatility driven by discrete policy regime changes and crisis episodes rather than smooth liberalisation. Türkiye exemplifies this regime dependence, with its coefficient moving sharply across phases of pre-liberalisation, post-liberalisation instability, crisis, and post-crisis restructuring; India, Thailand, Brazil, and Argentina display comparable policy-driven trajectories. Two-way fixed-effects estimation, PCA-augmented specifications that partial out common macroeconomic shocks, lead-savings placebo tests, within-country permutation exercises, and sensitivity analyses across alternative regularisation configurations confirm the robustness of these dynamics. The findings indicate that the emerging-market ‘puzzle’ reflects transitory policy configurations rather than genuine capital immobility, and that time-varying estimation is a methodological necessity—not an optional refinement—for capturing capital mobility in economies characterised by structural discontinuity.