Epidemiological studies have revealed that schizophrenia and autoimmune diseases co-occur in the general population at higher than expected rates. Here, we evaluated whether the epidemiologic correlation between immune diseases and schizophrenia might be explained by shared genetic risk factors. We first evaluated the association of 581 variants previously reported to be associated with 19 different immune diseases at genome-wide significance in a recent genome-wide association study (GWAS) of schizophrenia (N=35,476 cases and 46,839 controls). We identified nine variants with pleiotropic effects, associated with both schizophrenia and autoimmunity. Five of these pleiotropic variants were located outside of the HLA region, and mapped to genes with known roles in calcium signaling. We then evaluated whether polygenic risk scores for immune diseases, which take into account the collective effects of all SNPs (p<1 in association with the immune disease of interest), predicted schizophrenia. Among 14 immune diseases with available genome-wide summary statistics, higher polygenic risk scores for narcolepsy (liability-scale R2=0.02%, p=4.1x10-4), primary biliary cirrhosis (R2=0.04%, p=1.4x10-8), psoriasis (R2=0.02%, p=3.6x10-5), systemic lupus erythematosus (R2=0.04%, p=2.2x10-8), type 1 diabetes (R2=0.03%, p=2.0x10-6), and ulcerative colitis (R2=0.02%, p=4.3x10-4) were significantly associated with schizophrenia. We also observed suggestive evidence of sex-dependent pleiotropy between schizophrenia and multiple sclerosis (interaction p=0.02), with genetic risk scores for multiple sclerosis associated with greater risk of schizophrenia among males but not females. Our findings reveal the presence of modest pleiotropy between schizophrenia and several autoimmune diseases, which in some cases may be sex-dependent.