Is this a common (ecological) fallacy?

You have data on two levels (individuals and countries) for an outcome variable (e.g. ‘trust’) and a predictor (e.g. ‘wealth’). Supppose that the pooled and within-country individual-level correlations between the two variables are strongly positive but the between-country (country-level) correlation is zero. You build a regression with individual-level ‘trust’ as the dependent variable and individual-level wealth AND the country-level average of individual welath as predictors. Is there a well-known mechanism that induces a significantly negative coefficient for the country averages?