Two prominent emotions in borderline personality disorder (BPD) are shame and anger. Rumination has been demonstrated to occur in response to shame and to escalate anger, and rumination, particularly anger rumination, has been shown to predict BPD symptoms. The present study tested a structural equation model in which shame leads to the features of BPD via increased anger and anger rumination. A sample of 823 undergraduates completed self-report measures of shame, trait-level anger, anger rumination, and BPD features. The hypothesized model of shame to anger and anger rumination to BPD features was largely supported. Bootstrapping was used to establish significant indirect effects from both situational and global forms of shame via anger rumination to BPD features, and from global shame via anger to most BPD features. The alternative hypothesis that anger and anger rumination contribute to BPD features via increased shame was also examined, with no significant indirect effects found. Recognizing this function of anger and anger rumination may be important in understanding the relationship between shame-proneness and BPD features and may have implications for treatment. Further research into determining other ways individuals maladaptively respond to shame, and understanding the functions of anger and anger rumination, is recommended.