This paper studies whether institutions matter for economic performance. For this aim, we first construct a simple framework illustrating how to examine the interaction between institutions and economic performance from a different point of view. Then, using this framework, we introduce an innovative estimation approach including cutting-edge econometric techniques so-called Johansen et al. (2000) co-integration methodology with structural breaks to empirically investigate the interaction between institutions and economic performance in Turkey. Co-integration analysis finds a long-run relationship between institutions and economic performance in the presence of structural breaks. Also, the estimate of structural breaks reveals the effect of noteworthy changes in institutional structure on investments and economic growth. The findings confirm that institutions matter for economic performance in Turkey, even though the institutional quality of the country is not satisfactory. Lastly, the results suggest that approach employed in this paper is useful and convenient to empirically investigate whether institutions matter for economic performance in the study of a country-level time-series data.