This analysis uses IHME county-level estimates of deaths from interpersonal violence (all homicide mechanisms) for 24 north-central Texas counties, 2000\u20132019. The merged file holds 60,480 modeled rate estimates (both sexes combined) disaggregated by county, six ethnicity groups (Total, White, Black, Latino, Asian, AIAN), and 21 age bands plus All Ages and Age-standardized aggregates, each with 95% uncertainty bounds. Roughly a third of records are suppressed for small populations \u2014 Black estimates are available in only about eight counties with sufficient population (all 20 years each), and AIAN\/Asian coverage is sparser still. All rates below are age-standardized deaths per 100,000 unless an age band is named.<\/p>\n\n\n\n
The region traced a two-phase arc: a fifteen-year decline from 4.54 per 100,000 (2000) to a low of 3.79 (2014), then a sharp gun-driven reversal to 4.46 by 2016, settling at 4.22 in 2019. Only 9 of 24 counties ended the period above their 2000 level, meaning the long-run decline held in most places despite the recent surge.<\/p>\n\n\n\n
The headline disparity \u2014 Black residents at 10.76 per 100,000 on average versus 3.55 for White residents \u2014 conceals a more troubling dynamic when tracked over time. The three largest groups followed fundamentally different twenty-year paths:<\/p>\n\n\n\n
Black residents<\/strong> experienced a genuine improvement through the 2000s, with rates falling from 12.89 in 2000 to 8.98 by 2010 \u2014 a 30% decline. That progress then fully reversed: by 2019 the rate had climbed back to 12.43, essentially erasing two decades of gains. The post-2014 surge was disproportionately a Black community crisis.<\/p>\n\n\n\n Latino residents<\/strong> show the opposite story \u2014 the largest sustained improvement of any group. Rates fell from 6.35 in 2000 to 3.85 in 2019, a 39% decline that held through the post-2014 surge with only a modest uptick. Latino rates converged toward the White rate over the period.<\/p>\n\n\n\n White residents<\/strong> were essentially flat for twenty years (3.46 \u2192 3.35), with a small bump in 2016.<\/p>\n\n\n\n The consequence: the Black\u2013White disparity ratio is now the widest it has been in the entire series.<\/strong> The ratio narrowed from 3.7\u00d7 in 2000 to 2.6\u00d7 in 2010, then widened to 4.1\u00d7 by 2019. The region’s overall progress and regression in homicide is, to a first approximation, the story of what happened in its Black communities.<\/p>\n\n\n\n AIAN rates (4.6\u20135.0) drifted slightly upward and Asian rates declined (3.5 \u2192 2.3), though thin coverage limits confidence in both.<\/p>\n\n\n\n Where Black populations are large enough to estimate, their rates are elevated everywhere \u2014 not just in Dallas. Dallas County’s Black rate averages 19.72 per 100,000, but Hunt (16.76) and Kaufman (15.08) \u2014 smaller counties east of the metroplex \u2014 are nearly as high, and even suburban Rockwall’s Black rate (9.43) is nearly triple the regional White average. The disparity is regional and structural, not an artifact of urban geography.<\/p>\n\n\n\n The interaction of ethnicity and age is where the burden concentrates most extremely. Among ages 20\u201324, Black residents face a rate of 30.3 per 100,000<\/strong> \u2014 over five times the White young-adult rate (5.81) and more than triple the Latino rate (9.65). A young Black adult in this region faces a homicide rate roughly seven times the all-ages regional average.<\/p>\n\n\n\n Risk follows a distinctive double peak. The single highest age-specific rate belongs to infants under one year (9.23 per 100,000)<\/strong> \u2014 higher even than ages 20\u201324 (8.66) \u2014 reflecting fatal child abuse and neglect rather than community violence. Risk collapses through childhood (0.6\u20130.7 at ages 5\u201314), surges in young adulthood, then declines steadily with age (5.14 at 45\u201349, 2.43 at 65\u201369, 1.52 at 85+). The infant peak also carries a racial gradient: Black infants (14.14) face rates about 1.5\u00d7 White infants (9.63). This bimodal pattern implies two nearly separate public health problems \u2014 child maltreatment and young-adult violence \u2014 requiring different intervention systems.<\/p>\n\n\n\n Long-run improvement was led by the counties that started worst: Dallas fell from 10.49 in 2000 to 8.08 in 2019 (\u22122.41, though up from its 7.23 floor in 2014), with Kaufman (\u22121.54), Ellis (\u22120.86), and Rockwall (\u22120.82) also improving substantially. The counties that worsened<\/em> over twenty years are small rural ones \u2014 Stephens (+0.48), Eastland (+0.41), and Jack (+0.32) \u2014 the same western rural belt that carries excess maternal mortality and weak health system resources in your companion datasets. Eastland County now appears as a deteriorating outlier across all three health domains.<\/p>\n\n\n\n Four conclusions emerge. First, the post-2014 homicide surge was not demographically neutral \u2014 it was concentrated among Black residents, reversing a decade of progress and pushing the racial disparity to its twenty-year maximum. Second, Latino convergence toward White rates demonstrates that large, sustained declines are achievable within this same region and period, sharpening the question of why Black communities did not share in it. Third, the extreme concentration among Black young adults (30+ per 100,000 at ages 20\u201324) and the separate infant maltreatment peak define two distinct intervention targets \u2014 community violence prevention and child welfare surveillance. Fourth, the slow deterioration in the rural western counties, invisible in regional averages, aligns with the broader pattern of rural disadvantage across the maternal mortality and health system performance data.<\/p>\n\n\n\n These are modeled small-area estimates with widening uncertainty for small counties and sparsely covered groups; suppressed values (about a third of records, concentrated in Black, Asian, and AIAN estimates outside the urban core) mean racial comparisons rest primarily on the eight largest counties. Rates are per total population, unadjusted for socioeconomic covariates, poverty, or policing differences, so contrasts are descriptive, not causal. The series ends in 2019, before the 2020\u20132021 national homicide surge, so current disparities are likely wider than shown here.<\/p>\n\n\n\nEthnicity \u00d7 Geography<\/h3>\n\n\n\n
Ethnicity \u00d7 Age: The Core of the Crisis<\/h3>\n\n\n\n
Age Structure: A Bimodal Curve<\/h3>\n\n\n\n
County Dynamics<\/h3>\n\n\n\n
Interpretation<\/h3>\n\n\n\n
Limitations<\/h3>\n\n\n\n