This article contains an analysis of the impact channels through which credit rating agencies may affect systemic risk, while raising the research question of whether reputational risk is a sufficient mechanism for disciplining credit rating agencies and motivating them to care for the high quality of ratings. Research into reputational risk for rating agencies has led to the conclusion that credit rating agencies manage reputational risk. During a boom period, they perceive this risk as low and hence they tend to offer overestimated ratings. It is difficult then to show errors in rating decisions, and both issuers and investors benefit from high ratings. It is only during a bearish period that assessments become more diverse and better reflect the actual credit risk, sometimes overestimating it. During such a period, the likelihood of insolvency and therefore the need for a gradual reduction in ratings are higher. This proves the existence of reputation cycles and puts into question the claim that agency ratings are through-the-cycle ratings. Due to the possibility of an independent credit risk assessment of single-name securities, ratings are not usually overestimated for these securities, while the reputation of rating agencies built on the single-name instruments market is transferred to the structured instruments market where risk is underestimated. In order to reduce systemic risk, in both the United States and the European Union, decisions have been made to subject rating agencies to civil liability. There was also a need to define a regulatory framework for rating agencies and to intensify supervision.