We report the prevalence of subjective cognitive impairment (SCI) and cognitive impairment no dementia (CIND), their socio-demographic profile, and the contribution of genetic background and shared familial environment to SCI and CIND. Subjects were 11,926 dementia-free twin individuals aged ≥ 65 from the Swedish Twin Registry. SCI was defined as subjective complaint of cognitive change without objective cognitive impairment and CIND was defined according to current criteria. Overall prevalence of SCI and CIND was 39% (95% CI 38-39) and 25% (95% CI 24-25). Among those with CIND, 57% had subjective cognitive complaints; 43% did not. In multivariate GEE models, both SCI and CIND were older compared with people without any cognitive impairment. CIND were also less educated, more likely to be unmarried and to have lower socioeconomic status (SES). SCI individuals differed from persons with CIND as they were older, more educated, more likely to be married, and to have higher SES. Co-twin control analysis, which corrects for common genetic and shared environmental background, confirmed that low education was still associated with CIND. Probandwise concordance for SCI and CIND was 63% and 52% in monozygotic twins, 63% and 50% in dizygotic same-sex twins, and 42% and 29% in dizygotic unlike-sex twins. Tetrachoric correlations showed no significant differences between monozygotic and dizygotic same-sex twins. We conclude that subjective and objective cognitive impairment are both highly prevalent among nondemented elderly yet have distinct sociodemographic profiles. Shared environmental influences rather than genetic background play a role in the occurrence of SCI and CIND.