The increasing size and diversity of genomewide association studies provide an exciting opportunity to study how the genetics of complex traits vary between diverse populations. Here, we introduce covariateadjusted LD score regression (covLDSC), a method to accurately estimate genetic heritability ( ) and its enrichment in both h g 2 homogenous and admixed populations with summary statistics and insample LD estimates. Insample LD can be estimated from a subset of the GWAS samples, allowing our method to be applied efficiently to very large cohorts. In simulations, we show that unadjusted LDSC underestimates by 10% 60% in admixed populations; in contrast, h g 2 covLDSC is robust to all simulation parameters. We apply covLDSC to genotyping data from approximately 170,000 Latino, 47,000 African American and 135,000 European individuals. We estimate and detect heritability enrichment in three quantitative and h g 2 five dichotomous phenotypes respectively, making this, to our knowledge, the most comprehensive heritabilitybased analysis of admixed individuals. Our results show that most traits have high concordance of and consistent tissuespecific heritability h g 2 enrichment between different ethnic groups. For example, estimates of for BMI are h g 2 0.22 ± 0.01, 0.23 ± 0.03 and 0.22 ± 0.01 in Latino, African American and European populations respectively. However, for age at menarche, we observe populationspecific heritability differences with estimates of . We observe consistent patterns of h g 2 tissue-specific heritability enrichment across populations, for example in the limbic system for BMI, the per-standardized-annotation effect size are 0.16 ± 0.04, 0.28 ± 0.11 and 0.18 ± 0.03 τ in Latino, African American and European populations respectively. Our results demonstrate 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71that our approach is a powerful way to analyze genetic data for complex traits from underrepresented populations.