This paper investigates recent Italian regional election (2018)(2019)(2020) by assessing if and to what extent regional elections present nationalized or localized features. The extreme volatility of national and regional elections posits problems in explaining changes in multilevel contemporary party systems. Nevertheless, it also allows us to reconsider classical works on nationalization and territorialization of the vote and update some of their traditional assumptions and/or empirical expectations. We argue that Italian regional elections never perfectly stuck to the Second Order Elections' (SOE) expectations and often produced unclear or mixed results in this respect. Nonetheless, regional elections are not completely "localized," especially in Ordinary Status Regions (OSRs). On the contrary, they mirror the national climate in many aspects that are still captured by classical Reif and Schmitt's empirical expectations on statewide parties. Our contribution deploys in two parts. First, we descriptively measure and discuss to what extent high volatility rates are due to local (de-nationalization and/or localization) or national (nationalization and second-order effects) factors. We do so by separating and separately inspecting electoral change produced by different electoral supply across the national and the regional arenas or by the variations in vote shares received by parties in the two different arenas. Secondly, the latter component of volatility as a dependent variable is explored in order to adequately test Reif and Schmitt's classic expectations. In short, we show that recent Italian regional elections can still be considered as second-order elections, as they feature lower turnout, and drops for large and governing parties compared to legislative, first-order elections. However, on top of the impact of national politics, regional and local peculiarities are clearly visible; and they (mostly) follow classic historical trends and features.