Abstract:
We perform a quantitative synthesis of 807 estimates of the effect of macroeconomic news announcements on exchange rates, as reported in 25 studies. Estimates are tested for publication selection using visual examination of funnel plots, linear asymmetry tests, and recent non-linear testing techniques. Our analysis reveals that after the inclusion of the moderator variables, publication bias has a moderate but statistically significant effect on the reported estimates, although it does not strongly influence their magnitude. The primary sources of heterogeneity are driven by economic conditions, particularly interest rate differentials and inflation regimes. Surprisingly, the type of macroeconomic announcement does not systematically affect the variation in estimated effects, indicating that differences in the nature of the announcements have little impact on the reported results.