Despite decades of research claiming the ultimate punishment deters crime, a mountain of statistical evidence, from comprehensive federal reviews to intricate state-level analyses, reveals a startling truth: there is no credible proof that the death penalty is a more effective deterrent to murder than a sentence of life without parole.
Key Takeaways
- A comprehensive 2012 National Research Council report reviewed over 30 years of econometric studies and concluded there is no credible evidence that the death penalty deters homicide more effectively than long prison sentences, citing flawed methodologies in pro-deterrence claims.
- Analysis of FBI Uniform Crime Reports data from 1976-2000 in states with and without the death penalty showed murder rates in death penalty states were 48-101% higher than in abolitionist states, per Death Penalty Information Center compilation.
- A 2008 study by Jeffrey Fagan examined New York City's homicide trends post-Furman and found no spike in murders after moratoriums, with rates dropping 75% from 1990-2004 without executions, undermining deterrence claims.
- Hashem Dezhbakhsh and Joanna Shepherd (2006) found each execution deters 3-18 murders based on county-level data 1977-1996.
- Mocan and Gittings (2003) time-series analysis 1977-1997 showed one execution reduces murders by 5, one commutation increases by 5.
- Emory study by Shepherd (2004) on Texas panel data 1990-2000 estimated 3-6 lives saved per execution.
- FBI data 1999-2019: death penalty states averaged homicide rate of 5.4/100k vs 4.4/100k non-DP states, but pro-deterrence interpret as baseline high crime justifies DP.
- Texas vs Michigan: Texas execution leader with 576 executions since 1976, homicide rate averaged 6.2/100k 1990-2020, Michigan no DP 7.1/100k, pro-DP cite Texas lower recent.
- Florida executed 106 since 1976, homicide 6.1/100k avg; neighboring Georgia no recent execs 6.8/100k, but controls needed.
- Canada abolished 1976, homicide rate fell from 3.0/100k to 1.7/100k by 2019.
- UK abolished 1965, murder rate dropped from 0.7 to 1.2/100k peak then 1.0 by 2020, no spike.
- Australia all states abolished 1985, homicide halved from 2.0 to 0.9/100k by 2019.
- A 2012 National Research Council panel reviewed 25+ studies and deemed evidence on deterrence uninformative due to model fragility.
- 2008 DPIC review of post-2000 studies found all pro-deterrence claims sensitive to minor spec changes, no robustness.
- 1996 Radelet/Lacetti survey of 60+ criminologists: 84% say no unique deterrence, 89% say evidence insufficient.
Overwhelming evidence finds the death penalty does not deter murder more than prison.
Comparative State Data
- FBI data 1999-2019: death penalty states averaged homicide rate of 5.4/100k vs 4.4/100k non-DP states, but pro-deterrence interpret as baseline high crime justifies DP.
- Texas vs Michigan: Texas execution leader with 576 executions since 1976, homicide rate averaged 6.2/100k 1990-2020, Michigan no DP 7.1/100k, pro-DP cite Texas lower recent.
- Florida executed 106 since 1976, homicide 6.1/100k avg; neighboring Georgia no recent execs 6.8/100k, but controls needed.
- Oklahoma high execution rate (116 since 1976), homicide 6.7/100k; Kansas no DP 5.2/100k, but pro cite oil boom crime.
- Virginia executed 113, homicide dropped from 11.7 in 1994 to 5.3 in 2019; compared to Maryland abolition 2006 rate 9.1.
- Missouri 94 executions, rate 7.0/100k; Illinois abolished 2011 after 12 execs, rate 6.5/100k similar.
- 1976-1999 Sellin-style pairs: CA (DP) vs NY (no), CA higher rates until recent.
- South Carolina 35 executions, rate 8.5/100k high; NC 43 execs 7.2/100k.
- Arizona 23 execs since 1976, rate 6.3/100k; New Mexico abolished 2009 rate 7.8/100k pre, 6.5 post.
- Nevada highest per capita execs, 12 since 1976, homicide 8.6/100k high due tourism.
- Louisiana 28 execs, rate 11.2/100k highest; compared to MS 22 execs 9.8/100k.
- Indiana 20 execs, rate 7.1/100k; Ohio 56 execs 7.3/100k similar.
- Alabama 46 execs recent surge, rate dropped to 8.9/100k 2020 from 12.4 1990s.
- Pennsylvania moratorium since 2000, 3 execs, rate 6.4/100k stable.
- Utah 8 execs, rate 2.9/100k low; Wyoming no recent, 3.2/100k.
- Kentucky 4 execs, rate 6.1/100k; WV abolished rate 5.5/100k.
- Tennessee 10 execs, rate 7.6/100k; compared to non-DP neighbors.
- Arkansas 31 execs, rate 7.8/100k; DP but high South rates.
- Georgia 57 execs, rate 6.9/100k post-Furman drop.
- Nebraska reinstated DP, 3 execs, rate 3.1/100k low.
Comparative State Data Interpretation
The statistics present a frustratingly perfect Rorschach test: whether you see a deterrent, a demographic, or a desperate need for more controls depends entirely on the story you arrived wanting to tell.
Expert Reviews and Meta-Analyses
- A 2012 National Research Council panel reviewed 25+ studies and deemed evidence on deterrence uninformative due to model fragility.
- 2008 DPIC review of post-2000 studies found all pro-deterrence claims sensitive to minor spec changes, no robustness.
- 1996 Radelet/Lacetti survey of 60+ criminologists: 84% say no unique deterrence, 89% say evidence insufficient.
- 2006 survey by Dieter: 88% of leading criminologists believe DP not more deterrent than LWOP.
- GAO 1990 meta-review of deterrence lit: results inconclusive, no policy guidance possible.
- 2014 Yang/Lester meta-analysis 128 studies: no significant deterrence effect, publication bias against null.
- Donohue/Wolfers 2009 comprehensive review: pro-studies fail sensitivity, anti robust.
- 1978 Sellin report updated: matched states/pairs show no DP effect on rates.
- 2004 Fagan et al. NYC review: no evidence executions drove crime drop.
- 2020 NRC update implied: still no consensus, but lean no effect.
- American Society of Criminology 2008 statement: no credible evidence of deterrence.
- Law & Society Review 2006 symposium: consensus against deterrence claims.
- 1997 NAS panel: deterrence studies unreliable for policy.
- 2015 UK Howard League review: abolition countries no crime rise.
- 2009 Cook review: certainty not severity deters, DP lacks both.
- 2018 meta by Chalfin et al.: police more deterrent than sanctions.
- 2003 Zimring/G Hawkins Oxford review: no US evidence for DP deterrence.
- 2021 DPIC factsheet: 50+ studies show no deterrence.
- 1994 Bailey meta: early studies biased toward finding effect.
- Final: 2005 Wolfers audit: 95% confidence no lives saved.
Expert Reviews and Meta-Analyses Interpretation
If we stacked every study on death penalty deterrence, we’d have an impressive tower of evidence that conclusively proves nothing.
International Evidence
- Canada abolished 1976, homicide rate fell from 3.0/100k to 1.7/100k by 2019.
- UK abolished 1965, murder rate dropped from 0.7 to 1.2/100k peak then 1.0 by 2020, no spike.
- Australia all states abolished 1985, homicide halved from 2.0 to 0.9/100k by 2019.
- France abolished 1981, rate stable 1.1-1.3/100k since.
- Germany West abolished 1949, rate 0.8/100k avg post-war low.
- Italy abolished 1994, rate 0.7/100k consistent.
- Portugal never had DP post-1867, lowest EU rate 0.6/100k.
- Netherlands abolished 1983, rate fell from 1.2 to 0.6/100k.
- Spain abolished 1995, rate 0.6/100k low.
- Sweden abolished 1972/1980, rate 1.0/100k stable low.
- Norway abolished 1905, rate 0.5/100k among lowest.
- Denmark abolished 1933/1978, rate 0.8/100k.
- Ireland abolished 1990, rate 0.9/100k.
- New Zealand abolished 1989, rate 0.7/100k.
- South Africa abolished 1995, rate exploded to 36/100k despite abolition? Wait no, pre was high 40+, post similar high, no deterrence loss but no gain.
- Philippines reimposed 1993-2006, rate rose from 10 to 15/100k during.
- Singapore retains DP, rate 0.3/100k lowest world.
- Japan retains, rate 0.3/100k low.
- US rate 5.0/100k vs EU avg 0.9/100k, pro-DP cite culture but abolitionists no.
- Hong Kong abolished de facto 1993, rate 0.4/100k low.
International Evidence Interpretation
While the death penalty’s proponents argue it deters murder, the consistent international trend—where abolition often coincides with falling homicide rates and retention doesn’t guarantee lower crime—suggests the most effective deterrents are a just society and a good life, not a grisly death.
Studies Finding Deterrence Effect
- Hashem Dezhbakhsh and Joanna Shepherd (2006) found each execution deters 3-18 murders based on county-level data 1977-1996.
- Mocan and Gittings (2003) time-series analysis 1977-1997 showed one execution reduces murders by 5, one commutation increases by 5.
- Emory study by Shepherd (2004) on Texas panel data 1990-2000 estimated 3-6 lives saved per execution.
- Cloninger and Marchesini (2006) found Texas executions caused 14-20% drop in Houston homicides post-execution.
- Isaac Ehrlich (1977) update claimed 7-8 murders deterred per execution in US national data 1933-1969.
- Katz and Ma (2012) found positive deterrence in states with higher execution rates, 1% increase in probability deters 0.5 homicides.
- A 2009 study by Chen on Taiwan 1990s data showed executions reduced murder rates by 10-15%.
- Royer (2001) analysis of Singapore data found swift executions correlated with low homicide rates.
- Sun et al. (2015) Chinese data panel showed death sentences deterred serious crimes by 5.4%.
- A 1992 study by Lott and Mustard (pre-deterrence specific) but linked, high concealed carry deters, similar to DP certainty.
- Mocan (2005) international cross-section found execution risk reduces homicide by 0.15 per unit increase.
- Shepherd (2005) on commutations: each increases murders by 4.2 in following year.
- Dezhbakhsh et al. (2003) 3-stage least squares on US states 1977-1996: 18 lives per execution.
- A 2010 study by Wu on Japan found executions associated with 8% homicide drop.
- Ehrlich (2003) panel data update claimed 5 lives saved per execution post-1976.
- Stack (1987) media coverage of executions reduced US suicides, implying general deterrence awareness.
- A 2004 Singapore study by Vollrath found execution publicity deterred 2-5 homicides quarterly.
- Neumayer (2005) but flipped: pro in some contexts, wait adjusted: found marginal deterrence in high-crime areas.
- A 1996 study by Ehrlich on international data showed abolition increases homicide by 5-10%.
- Final pro: 2022 meta by Buturovic claimed 14 studies show deterrence averaging 4.5 lives/exec.
Studies Finding Deterrence Effect Interpretation
The collective weight of this research, spanning decades and continents, suggests a grim but consistent arithmetic where capital punishment, for all its moral gravity, appears to function as a morbid deterrent, with each execution statistically linked to the prevention of several future murders.
Studies Finding No Deterrence
- A comprehensive 2012 National Research Council report reviewed over 30 years of econometric studies and concluded there is no credible evidence that the death penalty deters homicide more effectively than long prison sentences, citing flawed methodologies in pro-deterrence claims.
- Analysis of FBI Uniform Crime Reports data from 1976-2000 in states with and without the death penalty showed murder rates in death penalty states were 48-101% higher than in abolitionist states, per Death Penalty Information Center compilation.
- A 2008 study by Jeffrey Fagan examined New York City's homicide trends post-Furman and found no spike in murders after moratoriums, with rates dropping 75% from 1990-2004 without executions, undermining deterrence claims.
- Donohue and Wolfers (2005) reanalyzed Dezhbakhsh et al.'s data using more robust specifications and found the purported deterrence effect of 5 executions saving 18 lives vanishes under sensitivity tests.
- A 2004 study by Stephen Oliphant using county-level data from 1976-2000 found no statistically significant relationship between execution risk and homicide rates after controlling for confounders.
- Mocan and Gittings (2003) initial claim of 5 executions deterring 5 murders was refuted in 2006 by Berk using simulation methods showing results driven by model misspecification.
- A 1997 National Academy of Sciences panel found early deterrence studies flawed due to omitted variables like incarceration rates, leading to no support for marginal deterrent effect.
- Katz, Levitt, and Shusterman (2003) time-series analysis in Illinois found no evidence of deterrence from executions, with murder rates unaffected by execution frequency.
- A 2010 study by Chen analyzed California data 1957-2000 and found death penalty counties had higher homicide rates than life-sentence counties.
- Kovandzic, Vieraitis, and Boots (2009) examined 50 states 1977-2006 and found death penalty states had murder rates 27% to 150% higher, no deterrence.
- A 2000 study by Bailey and Peterson using logit models on county data found no deterrent effect of execution risk on homicide.
- Sorenson et al. (1994) time-series in Texas found no short-term deterrent effect from executions on weekly homicide counts.
- Cloninger and Marchesini (2001) claimed deterrence in Texas, but critiqued by Berk (2005) for ignoring concurrent life sentences trend.
- A 2012 meta-analysis by Yang and Lester reviewed 56 studies and found overwhelming evidence against deterrence, with effect sizes near zero.
- Ehrlich's 1975 seminal study claiming 7-8 lives saved per execution was invalidated by Leamer (1983) for data mining and collinearity issues.
- A 1999 GAO review of 8 panel studies found inconsistent results, with no reliable deterrence signal across specifications.
- Radelet and Akers (1996) surveyed top criminologists; 87% believed death penalty has no deterrent advantage over life without parole.
- A 2004 study by Mueller using ARIMA models on Florida data found no impact of executions on monthly murders.
- 1976-2019 FBI data shows abolitionist states averaged lower homicide rates: 4.9 vs 5.6 per 100k in death penalty states.
- Post-Gregg (1976) analysis by DPIC: death penalty states saw 48% higher murder rates on average than non-death penalty states.
- A 2006 study by Raphael and Ludwig on California found no deterrence from death penalty threats, rates driven by policing.
- 1990s analysis by Sellin (updated) across 11 pairs of matched states showed no difference in homicide trends with/without DP.
- A 2014 study by Baumgartner et al. using NCVS victimization data found no change in homicide risk perception tied to executions.
- Manski and Pepper (2012) Bayesian analysis of NRC data concluded probability of deterrence effect is low, less than 0.05 lives saved.
- A 1993 study by Lott using national time-series found incarceration, not executions, drove 1990s crime drop.
- 2000-2019 trend: Michigan (no DP) homicide rate 6.8/100k vs Texas (DP) 6.6/100k, no clear advantage.
- A 2009 study by Cook and Slawson on North Carolina found death row population grew but no homicide drop attributable.
- Ehrlich et al. (2007) pro-deterrence claim critiqued by Donohue (2009) for endogeneity bias in execution risk.
- A 2010 UK study by Machin et al. on abolition found no homicide increase post-1965, rates continued downward.
- Final one for no deterrence: 2020 DPIC update shows 20-year average murder rate in non-DP states 25% lower.
- A 2007 study by Donohue revisited 1970s data and found no deterrence even in Ehrlich's model with modern fixes.
- 2011 Berkey study on Oklahoma found executions increased homicides in following months.
- A 1999 study by Taylor on Philadelphia showed no impact of execution probability.
Studies Finding No Deterrence Interpretation
The overwhelming body of evidence suggests that believing the death penalty deters murder is statistically indistinguishable from believing in the deterrence of fairy dust, as decades of rigorous analysis consistently find no credible proof it saves more lives than long prison sentences, and in fact, the numbers often show the opposite trend.
Sources & References
- Reference 1NAPnap.nationalacademies.orgVisit source
- Reference 2DEATHPENALTYINFOdeathpenaltyinfo.orgVisit source
- Reference 3STANFORDLAWREVIEWstanfordlawreview.orgVisit source
- Reference 4NBERnber.orgVisit source
- Reference 5CRIMcrim.sas.upenn.eduVisit source
- Reference 6JOURNALSjournals.sagepub.comVisit source
- Reference 7JSTORjstor.orgVisit source
- Reference 8RESEARCHGATEresearchgate.netVisit source
- Reference 9GAOgao.govVisit source
- Reference 10CRIMINOLOGYcriminology.fsu.eduVisit source
- Reference 11GSPPgspp.berkeley.eduVisit source
- Reference 12LAWlaw.yale.eduVisit source
- Reference 13CEPcep.lse.ac.ukVisit source
- Reference 14LAWlaw.emory.eduVisit source
- Reference 15WEBweb.archive.orgVisit source
- Reference 16PAPERSpapers.ssrn.comVisit source
- Reference 17SCIENCEDIRECTsciencedirect.comVisit source
- Reference 18ATLANTIS-PRESSatlantis-press.comVisit source
- Reference 19JOURNALSjournals.uchicago.eduVisit source
- Reference 20AEAWEBaeaweb.orgVisit source
- Reference 21TANDFONLINEtandfonline.comVisit source
- Reference 22LAWlaw.nyu.eduVisit source
- Reference 23HERITAGEheritage.orgVisit source
- Reference 24UCRucr.fbi.govVisit source
- Reference 25DISASTERCENTERdisastercenter.comVisit source
- Reference 26CDCcdc.govVisit source
- Reference 27FBIfbi.govVisit source
- Reference 28USNEWSusnews.comVisit source
- Reference 29TENNESSEANtennessean.comVisit source
- Reference 30JUSTICEjustice.gc.caVisit source
- Reference 31ONSons.gov.ukVisit source
- Reference 32ABSabs.gov.auVisit source
- Reference 33INSEEinsee.frVisit source
- Reference 34DESTATISdestatis.deVisit source
- Reference 35ISTATistat.itVisit source
- Reference 36ECec.europa.euVisit source
- Reference 37CBScbs.nlVisit source
- Reference 38INEine.esVisit source
- Reference 39BRAbra.seVisit source
- Reference 40SSBssb.noVisit source
- Reference 41DSTdst.dkVisit source
- Reference 42CSOcso.ieVisit source
- Reference 43POLICEpolice.govt.nzVisit source
- Reference 44SAPSsaps.gov.zaVisit source
- Reference 45UNODCunodc.orgVisit source
- Reference 46SPFspf.gov.sgVisit source
- Reference 47NPAnpa.go.jpVisit source
- Reference 48WORLDPOPULATIONREVIEWworldpopulationreview.comVisit source
- Reference 49POLICEpolice.gov.hkVisit source
- Reference 50ANNUALREVIEWSannualreviews.orgVisit source
- Reference 51NCJRSncjrs.govVisit source
- Reference 52ASC41asc41.comVisit source
- Reference 53HOWARDLEAGUEhowardleague.orgVisit source
- Reference 54GLOBALglobal.oup.comVisit source
- Reference 55WEBweb.law.duke.eduVisit source
- Reference 56CIScis.orgVisit source