Chapter 6.5 · The lost years

Violence and the life-expectancy gap

A rough accounting, not an age-specific decomposition, attributes 4.8% of the gap of 8.3 years in life expectancy between the most and the least deprived tenths of tracts to homicide, and 18.5% of the difference in the clusters with the highest homicide rates.

Author: Vishal Singh
Affiliation: NYU Stern
Published: 9 October 2026

Homicide accounts for 4.8% (95% CI 4.1% to 5.4%) of the 8.3 (95% CI 7.9 to 8.7) years by which life expectancy at birth, for deaths in 2010 to 2015, is lower in the most deprived tenth of tracts than in the least deprived. The figure comes from a rough accounting that takes each tenth's homicide rate for 2022 to 2024 and assumes that each death takes 41.5 years of life. The remaining 7.9 (95% CI 7.5 to 8.3) years, 95.2% (95% CI 94.6% to 95.9%) of the gap, are left to every other cause of death. CDC publishes one homicide rate for each cluster of about five neighboring tracts. In the clusters with the highest rates, which hold 1.3% of residents, homicide accounts for 18.5% (95% CI 16.5% to 20.3%) of a difference of 8.1 (95% CI 7.7 to 8.4) years from the clusters with the lowest rates. That group spans only 54 to 67 counties.

Neither source gives death rates by age for a tract, so this is a rough accounting and not an age-specific decomposition of the kind Arriaga described.[1] It multiplies a crude homicide rate by life expectancy and by an assumed number of years of life per homicide. The two sources also cover different periods, life expectancy for deaths in 2010 to 2015 and homicide rates for 2022 to 2024, and come from different models. The share is proportional to the assumed years, reaches 8.5% (95% CI 7.3% to 9.7%) at the largest value that one setting shared by every tenth could take, and ranges from 3.9% to 6.2% across the other changes tested.

The concentration of gun homicide in a few clusters, and the mix of causes of early death, belong to Chapter 6.2, Gun homicide and the causes of early death. The gradient of life expectancy across deprivation tenths belongs to Chapter 6.1, Place and the length of a life, whose tract sample and gap this chapter reuses. Overdose, alcohol and suicide are in Chapter 6.3, Overdose, alcohol and suicide, and their change since 2018 is in Chapter 6.4, Mortality trends, 2018 to 2024. The unit here is the tract, with each tract carrying the rate of its cluster, and nothing below describes an individual.

Homicide rates rise across deprivation tenths, with the steepest step at the top

This chapter counts a violent death as a homicide by any means, as the CDC injury and violence dashboard defines it, and says "homicide" for what it counts. The National Violent Death Reporting System, the source of the years lost per death used below, counts suicide among violent deaths; this chapter does not. Suicides of every means are not counted. The dashboard's data dictionary names deaths in legal intervention only inside its all-firearm measure, which this chapter does not use. One check below adds firearm suicide. Firearms account for 75.1% of the homicide deaths that the rates imply.

The dashboard's tract file gives one rate for each cluster of about five neighboring tracts and repeats it on every tract of the cluster; the rate is a modeled estimate for clusters with a few deaths, zero for none and a direct rate for many, and Chapter 6.2, Gun homicide and the causes of early death describes the file. The rates are crude rates per resident. For the clusters where CDC publishes an exact count of deaths for 2022, the rate times the cluster's residents divided by the count has a median of 0.99, as in the check in that chapter. A tract here carries the mean of its cluster's annual rates for 2022, 2023 and 2024, and the 2024 values are provisional.

The tenths are those of Chapter 6.1, Place and the length of a life: 77,936 tracts ranked on the Area Deprivation Index and cut into ten groups with equal numbers of tracts, not of residents. The least deprived tenth holds 32.2 million residents and the most deprived 23.0 million. The mean rate rises from 1.8 (95% CI 1.3 to 2.2) homicide deaths per hundred thousand residents a year in the least deprived tenth to 14.8 (95% CI 12.8 to 16.9) in the most deprived, a ratio of 8.4 (95% CI 6.5 to 11.3) (Figure 1). The rate is higher in each tenth than in the one before, and the smallest step among the first nine is 0.4 (95% CI 0.1 to 0.7). The step from the ninth tenth to the most deprived, 6.4 (95% CI 4.8 to 8.0), is larger than the next-largest step by 5.3 (95% CI 3.6 to 7.0).

Figure 1

Homicide deaths per hundred thousand residents a year rise from 1.8 in the least deprived tenth of tracts to 14.8 in the most deprived

Mean annual homicide deaths per 100,000 residents, 2022 to 2024, by tenth of tracts on the Area Deprivation Index, tracts outside New York City with a published rate

Homicide deaths per 100,000 residents a year

Show data

Data for Homicide deaths per hundred thousand residents a year rise from 1.8 in the least deprived tenth of tracts to 14.8 in the most deprived
conf_high	conf_low	estimate	n_units	tenth
2.16	1.35	1.76	6,479	1
3.18	2.07	2.62	6,827	2
3.48	2.63	3.03	7,072	3
4.18	3.29	3.70	7,220	4
4.69	3.97	4.33	7,178	5
5.63	4.68	5.16	7,602	6
7.09	5.55	6.27	7,317	7
8.32	6.48	7.34	7,495	8
9.59	7.41	8.43	7,425	9
16.85	12.76	14.81	7,610	10

The rate is higher in each tenth than in the one before, and the step from the ninth to the most deprived tenth is larger than any other. The line plots the population-weighted mean of tract values, each carrying the rate of its cluster, with a 95% interval from a bootstrap over counties.Note: The homicide rate is CDC's modeled estimate for a cluster of about five neighboring tracts, repeated on each tract; a tract's rate is the mean of its cluster's annual rates for 2022 to 2024. 2024 is provisional.Note: Homicide is deaths from homicide by any means, as CDC's dashboard defines it. Suicides of every means and deaths in legal intervention are not counted; the panel publishes legal intervention only inside all firearm deaths.Note: New York City is left out of every rate: its tracts hold 2.6% of residents in the gradient and carry rates near zero in CDC's tract file, a mean of 0.17 per hundred thousand against 4.9 in the county file, with 75.0% of its tracts at exactly zero. A further 4.7% of residents live in clusters where CDC withholds the rate in at least one year, and 92.8% are in tracts with a rate.Note: Intervals are 95% intervals from a bootstrap that draws whole counties with replacement (replicates and seed are in the methods note).Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024; modeled estimates for tract clusters; population-weighted

Rates are available for 92.8% of the residents in the gradient. New York City holds 2.6% of them and is left out, because its tract rates are near zero: the mean is 0.17 per hundred thousand in the tract file against 4.9 in the county file for the same boroughs, and 75.0% of its tracts sit at exactly zero. The city holds 12.1% of the least deprived tenth, where the index, which understates deprivation in high-cost cities, ranks many of its tracts. Giving its tracts the county file's rate for their borough raises the least deprived tenth's rate from 1.8 to 2.1 (95% CI 1.5 to 2.6) and moves the share of the gap from 4.8% to 4.6% (95% CI 4.0% to 5.3%). A further 4.7% of residents live in clusters where CDC withholds the rate in at least one of 2022, 2023 and 2024. The gap uses all of the gradient's tracts, the rates cover 72,225 of them, and each tenth's rate is applied to the whole tenth. Measuring the gap over only the tracts with a rate gives 4.7% (95% CI 4.0% to 5.3%).

A homicide rate becomes years of life through two assumptions

For a group of tracts $g$ , the years of life expectancy lost to homicide are approximated by

\text{years lost}_g \;=\; r_g \times e_{0,g} \times L,

where $r_g$ is the group's homicide rate per resident per year, $e_{0,g}$ is its life expectancy at birth and $L$ is the years of life a homicide takes. In a population with constant births and constant age-specific death rates, the crude death rate times life expectancy at birth equals the lifetime probability of dying of that cause. Under this approximation a newborn's lifetime risk of dying of homicide is 0.14% (95% CI 0.11% to 0.18%) in the least deprived tenth and 1.10% (95% CI 0.95% to 1.24%) in the most deprived. Multiplying the risk by the years a death takes gives, to a first approximation, the life expectancy that removing the cause would add. The approximation ignores competing risks, the chance that a person dies of another cause first, and the age mix of each tract. Residents of working age are 61.4% (95% CI 60.1% to 62.3%) of the least deprived tenth and 58.5% (95% CI 58.2% to 58.7%) of the most deprived, and scaling the rates to a common working-age share moves the answer to 5.0% (95% CI 4.3% to 5.7%).

The value of $L$ comes from outside both files. Zimmerman, Fridel and Trovato subtract the age at death from a life expectancy specific to the victim's race, ethnicity and sex for homicide victims in the National Violent Death Reporting System between 2003 and 2019, and find a mean of 41.5 years, floored at zero for a victim who died older than the life expectancy.[2] Most of the records come from 17 states. The accounting applies that national mean to every tenth. In that study Hispanic, Asian and Black victims lose more years on average than white victims, so the years lost would be above the mean where most victims come from those groups and below it elsewhere. Scaling the mean by each tenth's life expectancy over the national one, which lowers it for the most deprived tenth, gives 4.4% (95% CI 3.8% to 5.0%). The sweep below varies the setting.

Most of the gap belongs to causes of death other than homicide

Across all tracts with a rate, the accounting puts the average loss to homicide at 0.18 (95% CI 0.16 to 0.20) years of life expectancy. Tenth by tenth, the loss is 0.06 (95% CI 0.05 to 0.07) years in the least deprived tenth and 0.45 (95% CI 0.39 to 0.52) in the most deprived. The difference of 0.39 (95% CI 0.33 to 0.46) years is 4.8% (95% CI 4.1% to 5.4%) of the gap of 8.3 (95% CI 7.9 to 8.7) years between life expectancies of 82.3 (95% CI 82.0 to 82.6) and 74.0 (95% CI 73.7 to 74.3) years that Chapter 6.1, Place and the length of a life reports. The other 95.2% (95% CI 94.6% to 95.9%) is the remainder, which includes every other cause of death, among them the overdose and suicide deaths in Chapter 6.3, Overdose, alcohol and suicide (Figure 2). The two lines rest on different tracts: the observed line uses every tract of each tenth, and the homicide line uses those with a rate.

Figure 2

Homicide accounts for 0.39 of the 8.3 years by which life expectancy is lower in the most deprived tenth of tracts than in the least deprived

Years of life expectancy at birth below the least deprived tenth, by tenth of tracts on the Area Deprivation Index: observed for deaths in 2010 to 2015, and the part attributed to homicide by a rough accounting with homicide rates for 2022 to 2024

Years of life expectancy below the least deprived tenth

Show data

Data for Homicide accounts for 0.39 of the 8.3 years by which life expectancy is lower in the most deprived tenth of tracts than in the least deprived
conf_high	conf_low	decile	estimate	n_units	series
n/a	n/a	1	0	7,820	All causes (observed)
n/a	n/a	1	0	6,479	Homicide (rough accounting)
1.55	1.03	2	1.33	7,860	All causes (observed)
0.040	0.015	2	0.028	6,827	Homicide (rough accounting)
2.25	1.65	3	1.98	7,737	All causes (observed)
0.052	0.029	3	0.041	7,072	Homicide (rough accounting)
2.82	2.23	4	2.54	7,757	All causes (observed)
0.077	0.048	4	0.062	7,220	Homicide (rough accounting)
3.44	2.85	5	3.17	7,613	All causes (observed)
0.097	0.066	5	0.082	7,178	Homicide (rough accounting)
4.16	3.59	6	3.89	8,031	All causes (observed)
0.128	0.088	6	0.108	7,602	Homicide (rough accounting)
4.87	4.29	7	4.60	7,722	All causes (observed)
0.172	0.115	7	0.142	7,317	Homicide (rough accounting)
5.69	5.09	8	5.41	7,844	All causes (observed)
0.209	0.143	8	0.174	7,495	Homicide (rough accounting)
6.70	6.09	9	6.42	7,708	All causes (observed)
0.245	0.170	9	0.205	7,425	Homicide (rough accounting)
8.66	7.87	10	8.28	7,844	All causes (observed)
0.459	0.331	10	0.395	7,610	Homicide (rough accounting)

Both lines rise with deprivation, and the homicide line stays far below the observed line, with its largest step at the most deprived tenth. The observed line subtracts each tenth's life expectancy from that of the least deprived tenth. The homicide line multiplies each tenth's homicide rate, its life expectancy and an assumed 41.5 years of life lost per homicide, and subtracts the least deprived tenth's product. Intervals are 95% intervals from a bootstrap over counties.Note: Life expectancy is the USALEEP estimate for deaths in 2010 to 2015, area-weighted from 2010 onto 2020 tracts; tracts where group quarters hold a large share of residents are left out. Tenths hold equal numbers of tracts and means are weighted by residents.Note: The accounting is a rough one, not an age-specific decomposition: it multiplies a crude homicide rate by life expectancy to get the lifetime risk of dying of homicide, and by the years of life a death takes, assumed to be 41.5 for every tenth. The rates are for 2022 to 2024 and life expectancy is for deaths in 2010 to 2015.Note: The two lines rest on different tracts. The observed line uses every tract of each tenth, and the homicide line uses the tracts with a published rate, applying each tenth's rate to the whole tenth.Note: New York City is left out of every rate: its tracts hold 2.6% of residents in the gradient and carry rates near zero in CDC's tract file, a mean of 0.17 per hundred thousand against 4.9 in the county file, with 75.0% of its tracts at exactly zero. A further 4.7% of residents live in clusters where CDC withholds the rate in at least one year, and 92.8% are in tracts with a rate.Note: The homicide rate is CDC's modeled estimate for a cluster of about five neighboring tracts, repeated on each tract; a tract's rate is the mean of its cluster's annual rates for 2022 to 2024. 2024 is provisional.Note: Intervals are 95% intervals from a bootstrap that draws whole counties with replacement (replicates and seed are in the methods note).Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024 (modeled estimates for tract clusters); USALEEP life expectancy, 2010-2015; Area Deprivation Index 2023; years of life lost per death assumed from published estimates; population-weighted

The homicide part is concentrated at the top of the ranking. The homicide rate in the most deprived tenth is 6.4 (95% CI 4.8 to 8.0) higher than in the ninth, so the step in life expectancy between them contains 0.19 (95% CI 0.14 to 0.24) years of homicide, 10.2% (95% CI 8.0% to 12.3%) of that step's 1.9 (95% CI 1.6 to 2.1) years. That is a larger share than across the gradient as a whole. The accounting says how many years of life homicide removes, and says nothing about which of the other causes of death account for the rest.

Because the accounting multiplies by $L$ , the share rises in a straight line with that setting (Figure 3). At 30 years per death homicide accounts for 3.4% (95% CI 2.9% to 3.9%) of the gap, and at 55 years for 6.3% (95% CI 5.4% to 7.2%). A death cannot take more years than a person has. The sweep applies one value to every tenth, so the largest value it can take is the life expectancy of the most deprived tenth, 74.0 years, which would mean a death at birth in that tenth, and it gives 8.5% (95% CI 7.3% to 9.7%). No single value shared by every tenth puts the share above that. If every homicide in the most deprived tenth took that tenth's whole life expectancy and none in the least deprived tenth took any, the share would reach 9.8% (95% CI 8.6% to 11.0%).

Figure 3

Homicide accounts for 3.4% of the gap if each takes 30 years of life and 6.3% if each takes 55, and for at most 8.5% when one value applies to every tenth

Share of the gap in life expectancy between the least and the most deprived tenths of tracts attributed to homicide, by the years of life assumed lost per homicide

Share of the gap attributed to homicide (%)

Show data

Data for Homicide accounts for 3.4% of the gap if each takes 30 years of life and 6.3% if each takes 55, and for at most 8.5% when one value applies to every tenth
conf_high	conf_low	estimate	years_lost
n/a	n/a	0	0
0.131	0.098	0.115	1
0.262	0.197	0.230	2
0.394	0.295	0.345	3
0.525	0.393	0.460	4
0.656	0.491	0.574	5
0.787	0.590	0.689	6
0.919	0.688	0.804	7
1.05	0.786	0.919	8
1.18	0.885	1.03	9
1.31	0.983	1.15	10
1.44	1.08	1.26	11
1.57	1.18	1.38	12
1.71	1.28	1.49	13
1.84	1.38	1.61	14
1.97	1.47	1.72	15
2.10	1.57	1.84	16
2.23	1.67	1.95	17
2.36	1.77	2.07	18
2.49	1.87	2.18	19
2.62	1.97	2.30	20
2.76	2.06	2.41	21
2.89	2.16	2.53	22
3.02	2.26	2.64	23
3.15	2.36	2.76	24
3.28	2.46	2.87	25
3.41	2.56	2.99	26
3.54	2.65	3.10	27
3.67	2.75	3.22	28
3.81	2.85	3.33	29
3.94	2.95	3.45	30
4.07	3.05	3.56	31
4.20	3.15	3.68	32
4.33	3.24	3.79	33
4.46	3.34	3.91	34
4.59	3.44	4.02	35
4.72	3.54	4.14	36
4.86	3.64	4.25	37
4.99	3.73	4.37	38
5.12	3.83	4.48	39
5.25	3.93	4.60	40
5.38	4.03	4.71	41
5.51	4.13	4.83	42
5.64	4.23	4.94	43
5.77	4.33	5.05	44
5.91	4.42	5.17	45
6.04	4.52	5.29	46
6.17	4.62	5.40	47
6.30	4.72	5.51	48
6.43	4.82	5.63	49
6.56	4.91	5.75	50
6.69	5.01	5.86	51
6.82	5.11	5.97	52
6.96	5.21	6.09	53
7.09	5.31	6.20	54
7.22	5.41	6.32	55
7.35	5.50	6.43	56
7.48	5.60	6.55	57
7.61	5.70	6.66	58
7.74	5.80	6.78	59
7.87	5.90	6.89	60
8.00	6.00	7.01	61
8.14	6.09	7.12	62
8.27	6.19	7.24	63
8.40	6.29	7.35	64
8.53	6.39	7.47	65
8.66	6.49	7.58	66
8.79	6.58	7.70	67
8.92	6.68	7.81	68
9.05	6.78	7.93	69
9.19	6.88	8.04	70
9.32	6.98	8.16	71
9.45	7.08	8.27	72
9.58	7.17	8.39	73
9.71	7.27	8.50	74

The share rises in proportion to the years assumed lost per death. The rule on the right marks the largest value a death could take, the life expectancy of the most deprived tenth (74.0 years), and the other rule marks the published mean of 41.5 years. The line plots the rough accounting with a 95% interval from a bootstrap over counties.Note: The accounting multiplies each tenth's homicide rate for 2022 to 2024 by its life expectancy and by the years lost per death, so the share is proportional to that setting. The largest value that can apply to every tenth is the life expectancy of the most deprived tenth, which would mean that every homicide took a whole life. If every homicide in the most deprived tenth took that tenth's whole life expectancy and none in the least deprived tenth took any, which no single value gives, the share would reach 9.8%.Note: Intervals are 95% intervals from a bootstrap that draws whole counties with replacement (replicates and seed are in the methods note).Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024 (modeled estimates for tract clusters); USALEEP life expectancy, 2010-2015; Area Deprivation Index 2023; years of life lost per death assumed from published estimates; population-weighted

The other settings move the answer less than $L$ does (Figure 4). Rates for 2022 and 2023 alone, the years that are final, give 5.1% (95% CI 4.4% to 5.9%). Neither file covers 2010 to 2015, the years behind the life expectancy estimates, and the county file reaches back only to 2019. Scaling each tenth's accounting to the 2019 county rates gives 3.9% (95% CI 3.3% to 4.5%), and counting only firearm homicide, the measure in Chapter 6.2, Gun homicide and the causes of early death, gives 4.1% (95% CI 3.4% to 4.7%). These are the two lowest shares in the table, and they cannot be told apart. Adding firearm suicide, at the shorter 29.5 years that Zimmerman and colleagues find for suicide decedents, gives 6.2% (95% CI 5.6% to 6.9%), the highest. The rows that fill withheld rates with the low or the high end of the range CDC publishes for them bound what those rates could add, and both stay inside the range, as do the rows that use national life expectancy in place of each tenth's and count each tract once.

Figure 4

Changing the rate window, the set of deaths counted or the conversion moves the share of the gap between 3.9% and 6.2%

Part of the gap in life expectancy between the least and the most deprived tenths of tracts attributed to homicide, in the main setting and in each variation, at the published years of life lost per death

Setting	Years attributed to homicide	Gap (years)	Share of the gap (%)	Interval, low (%)	Interval, high (%)	Tracts with a rate	Of these, in the end tenths
Main setting: homicide rates for 2022 to 2024, all tracts with a rate	0.40	8.28	4.8	4.1	5.5	72,225	14,089
Rates for 2022 and 2023 only (final data)	0.42	8.28	5.1	4.4	5.9	73,581	14,275
Withheld rates filled with the low end of CDC's count range	0.39	8.28	4.7	4.0	5.3	75,862	14,636
Withheld rates filled with the high end of CDC's count range	0.37	8.28	4.5	3.8	5.1	75,862	14,636
Rates scaled to the 2019 level, from the county file	0.32	8.28	3.9	3.3	4.5	70,374	13,779
Firearm homicide only	0.34	8.28	4.1	3.5	4.7	72,285	14,115
Homicide plus firearm suicide	0.52	8.28	6.2	5.6	6.9	50,213	10,102
National life expectancy in place of each tenth's	0.43	8.28	5.2	4.4	5.9	72,225	14,089
Rates scaled to a common share of residents aged 18 to 64	0.41	8.28	5.0	4.3	5.7	72,225	14,089
Gap taken over the tracts that have a rate	0.40	8.44	4.7	4.0	5.3	72,225	14,089
Each tract counted once, not weighted by residents	0.41	8.35	5.0	4.2	5.7	72,225	14,089
New York City tracts at the county file's rate	0.38	8.28	4.6	4.0	5.3	74,299	15,117
Years lost scaled by each tenth's life expectancy over the national	0.36	8.28	4.4	3.8	5.0	72,225	14,089

Show data

Data for Changing the rate window, the set of deaths counted or the conversion moves the share of the gap between 3.9% and 6.2%
conf_high	conf_low	gap	setting	share	tracts	tracts_ends	years
5.45	4.08	8.28	Main setting: homicide rates for 2022 to 2024, all tracts with a rate	4.77	72225	14089	0.395
5.86	4.37	8.28	Rates for 2022 and 2023 only (final data)	5.13	73581	14275	0.424
5.35	4.01	8.28	Withheld rates filled with the low end of CDC's count range	4.68	75862	14636	0.388
5.11	3.81	8.28	Withheld rates filled with the high end of CDC's count range	4.47	75862	14636	0.370
4.48	3.32	8.28	Rates scaled to the 2019 level, from the county file	3.91	70374	13779	0.324
4.68	3.45	8.28	Firearm homicide only	4.06	72285	14115	0.336
6.92	5.62	8.28	Homicide plus firearm suicide	6.23	50213	10102	0.516
5.87	4.42	8.28	National life expectancy in place of each tenth's	5.16	72225	14089	0.427
5.66	4.26	8.28	Rates scaled to a common share of residents aged 18 to 64	4.96	72225	14089	0.411
5.32	4	8.44	Gap taken over the tracts that have a rate	4.68	72225	14089	0.395
5.67	4.24	8.35	Each tract counted once, not weighted by residents	4.97	72225	14089	0.415
5.29	3.95	8.28	New York City tracts at the county file's rate	4.63	74299	15117	0.383
5.05	3.76	8.28	Years lost scaled by each tenth's life expectancy over the national	4.41	72225	14089	0.365

The first row is the main setting and each other row changes one thing. The lowest shares come from scaling the rates to the 2019 level and from counting only firearm homicide, and the highest from adding firearm suicide. The columns give the part attributed to homicide in years, the gap it is measured against, its share, and a 95% interval for the share from a bootstrap over counties.Note: Firearm suicide enters at the mean years of life lost per suicide, 29.5, and only for tracts with a published rate for both. Rates scaled to the 2019 level use the county file's ratio of the 2019 rate to the 2022 to 2024 rate in each tenth, so they describe counties and not clusters. The two rows that fill withheld rates use the low and the high end of the count range CDC publishes for them, so they bound what the withheld rates could add and are not estimates. The row for New York City gives the city's tracts the county file's rate for their borough. The row that scales the years lost multiplies them by each tenth's life expectancy over the national one. The rows that measure the gap over the tracts that have a rate and that count each tract once have gaps of their own, which the table gives.Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024 (modeled estimates for tract clusters); USALEEP life expectancy, 2010-2015; Area Deprivation Index 2023; years of life lost per death assumed from published estimates; population-weighted

A national gap averages over a wide range of places. Grouping clusters by their own homicide rate shows the other end. The clusters fall into six groups, from those with rates up to 2 to those over 40 deaths per hundred thousand residents a year, and Figure 5 compares the life expectancy of each group with that of the lowest. If the annual errors in a cluster's rate are independent, chance would raise the measured rate of a group chosen for its high rate, so each cluster is placed by two of the three annual rates and the accounting uses the third year's, scaled to the level of the 2022 to 2024 mean. Forming the highest group and measuring it on the same years would put its mean rate at 59.5 instead of the 52.0 (95% CI 46.9 to 56.1) used here.

Life expectancy is 80.2 (95% CI 80.0 to 80.4) years in the lowest-rate group, whose mean rate is 1.8 (95% CI 1.7 to 1.9), and 72.1 (95% CI 71.8 to 72.4) years in the group over 40. That group holds 3.6 million residents, 1.3% of the residents of tracts with a rate. The difference of 8.1 (95% CI 7.7 to 8.4) years is close to the gap of 8.3 years between the least and most deprived tenths. By the accounting, homicide costs this group 1.6 (95% CI 1.4 to 1.7) years of life expectancy, and the part beyond the lowest-rate group's loss, 1.5 (95% CI 1.3 to 1.6) years, is 18.5% (95% CI 16.5% to 20.3%) of the difference. The share is higher in each higher group, from 4.5% (95% CI 3.9% to 5.1%) in the second-lowest to 12.4% (95% CI 11.4% to 13.4%) in the second-highest and 18.5% (95% CI 16.5% to 20.3%) in the highest.

Figure 5

Life expectancy is 8.1 years lower in clusters with homicide rates over 40 per hundred thousand than in clusters with rates up to 2, and homicide accounts for 18.5% of the difference

Years of life expectancy at birth below the lowest-rate group of tract clusters, by group of clusters ranked on the homicide rate for 2022 to 2024: the observed difference split into the part attributed to homicide and a remainder

Years of life expectancy below the lowest-rate group

Homicide (rough accounting)
All other causes (remainder)

Homicide rate of the tract cluster, deaths per 100,000 residents a year, 2022 to 2024

Show data

Data for Life expectancy is 8.1 years lower in clusters with homicide rates over 40 per hundred thousand than in clusters with rates up to 2, and homicide accounts for 18.5% of the difference
band	component	conf_high	conf_low	years
Over 2 to 5	Homicide (rough accounting)	0.055	0.046	0.051
Over 2 to 5	All other causes (remainder)	1.25	0.940	1.09
Over 5 to 10	Homicide (rough accounting)	0.140	0.127	0.134
Over 5 to 10	All other causes (remainder)	2.42	2.03	2.22
Over 10 to 20	Homicide (rough accounting)	0.334	0.305	0.319
Over 10 to 20	All other causes (remainder)	3.90	3.43	3.67
Over 20 to 40	Homicide (rough accounting)	0.755	0.675	0.718
Over 20 to 40	All other causes (remainder)	5.42	4.74	5.09
Over 40	Homicide (rough accounting)	1.62	1.34	1.50
Over 40	All other causes (remainder)	6.98	6.17	6.58

The difference grows with the homicide rate, and the homicide part grows faster, so it is a larger share of the difference in each higher group. Each bar is the observed difference in life expectancy for deaths in 2010 to 2015 from the lowest-rate group. The lower segment is the part attributed to homicide by rough accounting and the upper segment is the remainder.Note: Clusters are ranked on the mean of two of the three annual rates, and the rate used in the accounting is the third year's, put on the level of the three-year mean by one national factor. The figure averages the three ways of leaving out a year. If the annual errors in a cluster's rate are independent, this keeps chance from inflating the rate of a group chosen for its high rate: the highest group's rate is 52.0 this way and would be 59.5 if the same years chose the group and measured it. About 62% of the highest group's residents have a 2024 rate that is a modeled, provisional estimate.Note: The groups differ between the three ways of leaving out a year. The highest group holds 1,165 to 1,550 tracts, 194 to 264 clusters and 54 to 67 counties, and the 10 counties with most of its residents hold 52.7% to 57.1% of them, so its interval rests on few counties. Using only 2022 and 2023, one year placing a cluster and the other measuring it, the highest group's share is 20.0%.Note: Groups hold the residents of tracts outside New York City whose cluster has a rate in every year. The lowest group, rates up to 2, has a mean rate of 1.8 and a life expectancy of 80.2 years.Note: Life expectancy is the USALEEP estimate for deaths in 2010 to 2015, area-weighted from 2010 onto 2020 tracts; tracts where group quarters hold a large share of residents are left out. Tenths hold equal numbers of tracts and means are weighted by residents.Note: Intervals are 95% intervals from a bootstrap that draws whole counties with replacement (replicates and seed are in the methods note).Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024 (modeled estimates for tract clusters); USALEEP life expectancy, 2010-2015; Area Deprivation Index 2023; years of life lost per death assumed from published estimates; population-weighted

The highest group is a narrow slice and its membership shifts with the year left out. It holds between 1,165 and 1,550 tracts, in between 194 and 264 clusters and between 54 and 67 counties, depending on the split. The 10 counties with most of its residents hold between 52.7% and 57.1% of them, and Cook County alone between 12.2% and 15.1%. The interval for its share is a bootstrap over those counties. About 62% of its residents have a 2024 rate that is a modeled estimate and provisional. Using 2022 and 2023 alone, with one year placing a cluster and the other measuring its rate, gives 20.0% (95% CI 17.9% to 21.9%) for the highest group.

An age-specific decomposition from one place gives a figure for scale. In Washington, D.C., Roberts, Reither and Lim attribute 2.43 of the 17.23 years by which Black men's life expectancy trailed white men's for deaths in 2015 to 2017 to homicide, or 14.1%.[3] That is a gap between groups of men in one city, not a gap between kinds of tracts. Light and Vachuska find that homicide contributed more than any other cause of death to the national gap between Black and white men's life expectancy in 2020 and 2021.[4] Chapter 6.2, Gun homicide and the causes of early death shows the same pairing of high rates and short lives inside Cook County.

Within a deprivation tenth, life expectancy is lower where homicide rates are higher

Deprivation and homicide move together, so the comparison of tenths mixes them. Splitting each tenth at its median homicide rate, among its tracts with a rate, holds the deprivation rank nearly fixed. In every tenth, life expectancy is lower in the group with the higher rate (Figure 6), by 0.8 (95% CI 0.5 to 1.1) years in the least deprived tenth and 2.1 (95% CI 1.7 to 2.6) in the most deprived, where the two groups' mean rates are 5.0 (95% CI 4.7 to 5.4) and 24.6 (95% CI 21.7 to 27.3) per hundred thousand. The accounting attributes 0.6 (95% CI 0.5 to 0.7) years of the most deprived tenth's difference to homicide, 27.5% (95% CI 22.4% to 33.8%), and 6.3% (95% CI 4.9% to 8.4%) of the least deprived tenth's.

Figure 6

Within each tenth of tracts on the deprivation index, life expectancy is lower in the group with the higher homicide rate, by 0.8 years in the least deprived tenth and 2.1 in the most deprived

Population-weighted mean life expectancy at birth for deaths in 2010 to 2015 in the lower-rate and higher-rate groups of each tenth of tracts on the Area Deprivation Index, split on the homicide rate for 2022 to 2024

Lower-rate group
Higher-rate group

Show data

Data for Within each tenth of tracts on the deprivation index, life expectancy is lower in the group with the higher homicide rate, by 0.8 years in the least deprived tenth and 2.1 in the most deprived
hi1	hi2	higher	lo1	lo2	lower	n1	n2	tenth
82.98	82.28	81.97	82.63	81.73	82.81	14320000	12679000	1 (least deprived)
81.64	80.94	80.57	81.31	80.29	81.47	15296000	15016000	2
80.99	80.19	79.90	80.73	79.64	80.86	16173000	15971000	3
80.42	79.42	79.22	80.09	79.01	80.26	16346000	16076000	4
79.79	78.81	78.61	79.50	78.40	79.64	15852000	15625000	5
79.18	77.95	77.74	78.90	77.55	79.04	15939000	15798000	6
78.45	77.25	77.07	78.16	76.91	78.30	14581000	14508000	7
77.77	76.32	76.15	77.47	75.98	77.62	13908000	13808000	8
76.75	75.34	75.14	76.45	74.94	76.59	12765000	12689000	9
75.45	73.16	72.90	74.71	72.68	75.04	11177000	11140000	10 (most deprived)

In every tenth the filled dot, the higher-rate group, lies left of the open dot, and the difference is widest in the most deprived tenth. Each tenth is split at its median homicide rate, weighted by residents. Dots are population-weighted means with 95% intervals from a bootstrap over counties.Note: The groups are formed by cross-fitting, as in the figure of groups: each tenth is split on the mean of two of the three annual rates and the figure averages the three ways of leaving out a year. The n columns give residents. Each tenth is split among the tracts with a usable homicide rate, so a tenth has fewer tracts here than in the other figures.Note: The split is on the homicide rate and the outcome is life expectancy from another source and another period, so the difference is not built by the sorting. It compares places and does not show what violence does to the people in them.Note: Life expectancy is the USALEEP estimate for deaths in 2010 to 2015, area-weighted from 2010 onto 2020 tracts; tracts where group quarters hold a large share of residents are left out. Tenths hold equal numbers of tracts and means are weighted by residents.Note: Intervals are 95% intervals from a bootstrap that draws whole counties with replacement (replicates and seed are in the methods note).Source: CDC Mapping Injury, Overdose, and Violence, tract file, homicide, 2022-2024 (modeled estimates for tract clusters); USALEEP life expectancy, 2010-2015; Area Deprivation Index 2023; population-weighted

Most of each difference is left to other causes of death. The index is built from census measures of income, education, employment and housing, so tracts of the same rank can differ in conditions it does not measure, and homicide may mark them. The pattern also fits homicide being tied to other deaths, for example among residents who live with it. The data cannot separate the two.

What these data cannot show

The accounting is not an age-specific decomposition. It applies one value of $L$ to every tenth, although victims' ages differ by place, and that value is a national mean for victims, most of them from 17 states, specific to each victim's race, ethnicity and sex and floored at zero. It converts a crude rate to a lifetime risk by assuming a stationary population. The share scales with $L$ , so a reader who believes victims lose more or fewer years should read the answer off Figure 3. The intervals around each share treat $L$ as known, and the sweep is the answer to that uncertainty. Arriaga's method needs rates by age and cause for each group, and neither file provides them below the county.

The two sources do not describe the same years. Life expectancy is the USALEEP estimate for deaths in 2010 to 2015, area-weighted onto 2020 tracts with standard errors that are good for mapping only, and the homicide rates are for 2022 to 2024. The county file's rates for 2019 are below its rates for 2022 to 2024 in every tenth, and no file here describes 2010 to 2015, so the share for those years could be lower or higher than shown. The rates are CDC's modeled estimates for clusters of neighboring tracts, repeated on each tract, which averages over neighborhoods and lowers measured concentration. That they are crude per-resident rates is inferred from the check against exact counts and not from CDC documentation. The 2024 values are provisional, and the rows that fill withheld rates from CDC's count ranges give bounds and not estimates. The 95% intervals come from a bootstrap that resamples counties, so they describe which counties are in the file and not the error of either model.

The comparisons are of places. A tract's life expectancy and its homicide rate come from different people and different decades, and a difference between a higher-rate and a lower-rate group does not show what homicide does to the people who live there. The accounting counts deaths from homicide and no other consequence of it, so it leaves out any association of violence with other deaths and with health short of death. The Area Deprivation Index understates deprivation in high-cost cities, and New York City, which holds a large share of the least deprived tenth, has published rates near zero that the accounting leaves out. Nothing in this chapter shows what reducing violence would do to life expectancy; Chapter 6.2, Gun homicide and the causes of early death reviews evidence on programs aimed at violence.

Data sources

CDC injury and violence mappingCDC homicide rates for tract clusters, annual 2022, 2023 and 2024 (modeled estimates for 1 to 9 deaths, zero for none, direct rates for 10 or more); county file 2019 and 2022 to 2024 for one checkYears: 2019-2024 for counties, 2022-2024 for tracts (data as of 2026-05-14) · Weights: ACS 2019-2023 residents of each tract · License: Public domain (CDC, U.S. Department of Health and Human Services)
USALEEP life expectancyLife expectancy at birth by tract, deaths in 2010 to 2015, area-weighted onto 2020 tractsYears: 2010-2015 · Weights: 2019-2023 tract population · License: Public domain (NCHS, CDC)
Area Deprivation IndexArea Deprivation Index national rank, 2023, for the tenths of tractsYears: 2023 release (version 4.0.1; built on ACS 2019-2023) · Weights: none (rank) · License: Free for non-commercial use; cite the Neighborhood Atlas and Kind and Buckingham
American Community SurveyTract population (weights) and age shares, 2019-2023Years: 2019-2023 (5-year) · Weights: none (full count) · License: Public domain (U.S. Census Bureau)
Published estimates from cited studiesMean years of potential life lost per homicide victim and per suicide decedent (Zimmerman, Fridel and Trovato 2024) and a decomposition of the Black-white gap in male life expectancy in Washington, D.C. (Roberts, Reither and Lim 2020)Years: as reported in each study · Weights: none (published estimates) · License: Quoted from published research for commentary

All data sources

Methods

This chapter approximates how much of the life-expectancy gap between the most and the least deprived tenths of tracts homicide accounts for. It is a rough accounting and not an age-specific decomposition, because neither source gives age-specific death rates for a tract, and the two cover different periods (life expectancy for deaths in 2010 to 2015, homicide rates for 2022 to 2024) and come from different models. A tract enters only if CDC publishes the homicide rate of its cluster in all three years; New York City is left out, and suicide is not counted. The years of life lost per homicide are an assumption taken from a published study, and the chapter varies it.

Tracts, periods and homicide rates

The sample, tenths, weights and gap are those of the chapter Place and the length of a life: tracts with a USALEEP life expectancy for deaths in 2010 to 2015 (area-weighted onto 2020 tracts), institutional tracts (group quarters at least 25 percent of residents) dropped, an ADI 2023 national rank, tenths of tracts with equal numbers of tracts, means weighted by ACS 2019-2023 tract population.

The homicide rate is CDC's rate for the cluster of about five tracts that holds the tract, repeated on each member tract; a tract carries the mean of its cluster's annual rates for 2022, 2023 and 2024 and is used only if CDC publishes all three, which 72,225 tracts outside New York City satisfy. New York City is left out because the file records almost no deaths there. Withheld rates are not filled in the main setting; two checks fill them from the published count range.

Figures explained here: Figure 1 and Figure 6

How the accounting approximates the share of the gap

The chapter is a rough accounting and not an age-specific decomposition. Neither source gives age-specific death rates for a tract, and the two cover different periods and come from different models, so the part of a life-expectancy gap that homicide accounts for is approximated by the product of three terms: the homicide rate, life expectancy at birth, and the mean years of life a homicide takes.

The first two give the lifetime risk of dying of homicide in a stationary population, where the crude death rate times life expectancy equals the lifetime probability of dying of that cause; the third converts the risk to years. The product is a first-order approximation to the life expectancy that would be gained if the cause were removed, and it leaves out competing risks and the age mix of each tract; it also treats a victim's life expectancy at birth minus age at death, the published measure, as the remaining life expectancy at the age of death. Arriaga's method (1984) needs death rates by age and cause for each group, which neither file provides below the county.

Figure explained here: Figure 2

Years of life lost per homicide are an assumption

The years of life lost per homicide are an assumption: 41.5 years, the mean of life expectancy (specific to race, ethnicity and sex) minus age at death over homicide victims in the National Violent Death Reporting System, 2003 to 2019, from Zimmerman, Fridel and Trovato (2024); the chapter varies it from zero to the life expectancy of the most deprived tenth.

Figure explained here: Figure 3

What counts as a violent death

In this chapter a violent death means a homicide by any means as the CDC injury dashboard defines it (its All_Homicide measure). The National Violent Death Reporting System, the source of the years lost, counts suicide among violent deaths and this chapter does not. Suicides and deaths in legal intervention are not included; one check adds firearm suicide.

How uncertainty is measured

Intervals are percentile intervals from a bootstrap that draws whole counties (3,104 counties and Connecticut planning regions) with replacement, 2,000 replicates, seed 12345, shared by every statistic so that a share or a difference has one interval; the gap and its interval equal that chapter's.

Groups formed on the homicide rate and the 2019 check

For groups formed on the homicide rate, each cluster is placed by the mean of two of the three annual rates and the rate in the accounting is the third year's, scaled by one national factor to the level of the three-year mean; the three ways of leaving out a year are averaged. The 2019 check scales each tenth's accounting by the ratio of the county file's 2019 rate to its 2022 to 2024 rate.

Figures explained here: Figure 5 and Figure 4

References

Arriaga, Eduardo E. (1984). Measuring and explaining the change in life expectancies. Demography 21(1): 83-96. doi:10.2307/2061029
Zimmerman, Gregory M.; Fridel, Emma E.; Trovato, Daniel (2024). Disproportionate burden of violence: Explaining racial and ethnic disparities in potential years of life lost among homicide victims, suicide decedents, and homicide-suicide perpetrators. PLOS ONE 19(2): e0297346. doi:10.1371/journal.pone.0297346
Roberts, Max; Reither, Eric N.; Lim, Sojung (2020). Contributors to the black-white life expectancy gap in Washington D.C.. Scientific Reports 10: 13416. doi:10.1038/s41598-020-70046-6
Light, Michael T.; Vachuska, Karl (2024). Increased homicide played a key role in driving Black-White disparities in life expectancy among men during the COVID-19 pandemic. PLOS ONE 19(8): e0308105. doi:10.1371/journal.pone.0308105

How to cite

@misc{singh2026violenceandthelifeexpectancygap,
  author = {Singh, Vishal},
  title = {Violence and the life-expectancy gap},
  year = {2026},
  url = {https://vishalsingh.org/zipcode/violence-and-the-life-expectancy-gap},
  note = {In ZIP Code Destiny. Published 9 October 2026.}
}

Homicide rates rise across deprivation tenths, with the steepest step at the top

A homicide rate becomes years of life through two assumptions

Most of the gap belongs to causes of death other than homicide

The share is proportional to the years of life each homicide takes

In the clusters with the highest homicide rates, homicide accounts for a larger share

Within a deprivation tenth, life expectancy is lower where homicide rates are higher

What these data cannot show

Data sources

Methods

Tracts, periods and homicide rates

How the accounting approximates the share of the gap

Years of life lost per homicide are an assumption

What counts as a violent death

How uncertainty is measured

Groups formed on the homicide rate and the 2019 check

References

How to cite