Direct economic losses represent a convenient proxy to communicate seismic risk to stakeholders both at the building and at the large-scale level. In the perspective of a performance-based approach, seismic economic losses can be calculated starting from the response of single buildings expressed in terms of engineering demand parameters such as interstory drift ratios and peak floor accelerations. This way the distribution and the extent of earthquake-induced damage can be estimated for various structural and nonstructural components and buildings content. The Stick-IT, which is a MDOF system consisting of a series of masses lumped at the story level connected by nonlinear shear link elements, has been recently proposed and calibrated to predict the response, in terms of engineering demand parameters, for infilled RC building typologies.This paper adopts the Stick-IT model within a probabilistic framework to predict the expected direct losses for a portfolio of buildings damaged and repaired in the aftermath of the L’Aquila 2009 earthquake. The predicted losses are compared with the actual repair costs showing the potentiality of the proposed model when dealing with large scale loss assessment purposes.