ABSTRACT This paper attempts to explore dynamicity and nonlinearity in the association between financial development and macroeconomic volatility, considering different aspects of financial development. The data to this end comprises a panel of 98 countries from 1991 to 2016. This paper uses an objective trichotomous development taxonomy to classify the countries according to their overall financial development and income. Dynamic panel regressions ensure that the role of overall financial development in smoothing macroeconomic volatility dwindles as the level of overall financial development increases. Overall financial development begins to amplify volatility after a certain point. Other aspects of financial development also affect volatility in a similar fashion. Nonlinear estimation results show that most of the measures of financial development are associated with macroeconomic volatility in a U-shaped manner. On the contrary, we find an inverted U-shaped association between financial institution efficiency and macroeconomic volatility. The findings are found to be robust to an alternative country classification system based on income and an alternative measure of volatility. To minimise macroeconomic volatility, policymakers should try to increase financial institution efficiency beyond its threshold point and keep all other aspects of financial development at their optimal level.