Do housing price cycles reflect herd behaviour in developing and emerging economies?

Abstract

This study aims to examine whether housing price cycles in 16 developing and emerging economies reflect herd behaviour. This study used annual panel data from 2005 to 2023 sourced from the Bank for International Settlements and the World Bank’s World Development Indicators. The main model uses real housing price growth (inflation-adjusted) as the dependent variable and nominal housing price growth as a robustness check. Grounded in herd behaviour theory and information cascade theory, the study uses two-way fixed-effects panel regression, an instrumental-variables fixed-effects (FE-IV) estimator related to System GMM, and a mean group estimator, complemented by the Chang et al. (2000) cross-sectional absolute deviation (CSAD) herding test. Macroeconomic fundamentals were found to be the dominant drivers of housing price growth, but the role of inflation shrinks sharply. In the preferred two-way fixed-effects specification, gross domestic product growth (β = 1.223, p < 0.01) is a robust predictor, while the coefficient of inflation falls from β = 1.816 (p < 0.01) under nominal growth to β = 0.388 (p = 0.015) under real growth, consistent with much of the previously observed inflation effect being mechanical rather than behavioural (R2 = 0.154 for real growth versus R2 = 0.444 for nominal growth in the same specification). The herding proxy (lagged HP growth) is not statistically significant in the preferred two-way fixed effects or FE-IV specifications under either deflator, and the CSAD dispersion test is not significant and changes sign depending on the deflator used, together pointing to limited and episodic, rather than broad or structural, evidence of herd behaviour. These findings contribute to the limited cross-country behavioural literature on emerging market housing, demonstrating that herd behaviour is episodic rather than structural, and most pronounced when global financial conditions generate the public signals necessary for cascade formation. This study contributes to the behavioural real estate finance literature by providing one of the first cross-country study adopting panel data to understand herd behaviour in housing markets across developing and emerging economies.