Created with Sketch.
Research / Reports / Arizona / Criminal Justice

Arizona’s Imprisonment Crisis:

The Cost to Communities


After decades of sustained prison growth, today Arizona has the fourth highest imprisonment rate in the country and the prison system costs taxpayers over $1 billion each year. This prison growth cannot be explained by more crime or a larger resident population — instead it was driven by policy and practitioner decisions to send more people to prison for first-time and non-violent offenses, and to keep people in prison far beyond the national average.

This imprisonment crisis is not felt equally across the state — some communities bear the burden far more than others. This second report in a three-part series will examine Arizona’s geographical and racial disparities in prison use. It will also take a close look at Maricopa County, and why Arizona’s largest urban area is not following the lead of other similar cities that are safely reducing their reliance on prison.

These disparities come at a high cost. Research has shown that communities with high incarceration rates suffer from lower-than-average wages and diminished labor markets, poor mental and physical health, and weakened family networks.

All three reports in this series were created using individual-level data on people admitted to Arizona prisons. The first part of this series closely examines the choices that led Arizona to this crisis point, and how it has come with a high price to the state’s economy. The third report will explore the dire consequences for Arizona’s women and families. Experts on corrections data cleaned and analyzed prison data in accordance with national standards. See the methodology section for a description of our process and definitions.

County Disparities

Arizona’s criminal laws are set at the state level. However, each county in Arizona has local practitioners — police, sheriffs, county attorneys, defense attorneys, judges, and others — who collectively make decisions about how those laws are implemented. These actors decide who is arrested and convicted of crimes and, ultimately, who goes to prison and how long they stay. These county-level decisions lead to tremendous variation in the use of prison across counties in Arizona.

These county-level choices also have significant consequences. Incarceration not only affects individuals and their families — it also impacts their community. Studies have shown that communities with high incarceration rates have lower incomes, higher rates of disease, and less stable family networks. A recent study found that simply living in a community with a high incarceration rate makes residents more likely to suffer from serious mental health problems, even if they have not had personal contact with the criminal justice system. For these reasons, some researchers have argued that high levels of incarceration destabilize communities to the point of creating more — not less —crime.

Arizona’s most populous counties, Maricopa County, Pima County, and Pinal County, send the greatest number of people to prison. However, when looking at these numbers on a per-capita basis — or how many people are admitted to prison for every 10,000 county residents — rural counties rise to the top. Graham County, a small county with fewer than 40,000 residents, has the state’s highest prison admissions rate, followed by Gila County, Greenlee County, Cochise County, and Yavapai County. This is true even when looking solely at admissions rates for non-violent offenses — or the number of people sent to prison for non-violent offenses for every 10,000 people in the county. In 2017, Graham sent 37 people convicted of non-violent crimes to prison for every 10,000 residents, compared to only 14 people in urban Pima.

Whether someone is sent to prison is not only a function of local decision-makers — it is also determined in part by what options are available in lieu of prison. Eight of the state’s 15 counties have deferred prosecution programs that allow adults charged with a felony to maintain a clean record if they participate in programming. Many counties have adult drug courts for people charged with drug or drug-related crimes. In other areas there are few alternatives beyond probation, which is available in every county.

However, the available data indicates that whether a given county has a high prison admissions rate is not closely connected to the number of prison alternatives. For example, of the five counties with the highest non-violent admissions rates, three (Cochise, Gila, and Graham) have both deferred prosecution programs and drug courts. This is likely due to the fact that many of these alternatives divert a relatively small share of convictions. For example, in Maricopa, where the largest number of convictions occur, 3,641 people were referred to the county’s two adult felony deferred prosecution programs in the last three years while almost 30,000 people were admitted to prison over that period.

These county variations are even more acute when looking only at drug offenses. Five counties — Cochise, Graham, Gila, Yavapai, and Pinal — make up 22 percent of drug admissions to state prison but only 12 percent of the state population. Since 2000, admissions to prison for drug offenses from these five counties have quadrupled, from 258 in 2000 to 1,078 in 2017. A significant portion of the growth was people admitted to state prison for marijuana-related offenses (from 78 people in 2000 to 276 in 2017). In 2017, Cochise alone sent 157 people to prison for marijuana crimes.

However, because of its size, Maricopa still drove the overall growth in the number of people entering prison for drug crimes in Arizona. As noted in the first report of this series, Arizona’s Imprisonment Crisis: The High Price of Prison Growth, the number of people sent to prison in Arizona for drug crimes has nearly doubled since 2000. Almost half of this increase came from Maricopa alone.

Average sentence lengths also differ widely from county to county. Maricopa, for example, sends people to prison with the longest average sentence — 57.4 months, or nearly 5 years. La Paz County, on the other hand, sends people to prison with an average sentence of only 32.9 months, or 2.7 years. As outlined in more detail in the first report of this series, research shows that long sentences are ineffective as a crime control measure, and that additional weeks or months of prison time does not lead to less crime or lower recidivism rates.

Longer sentences in some counties could be driven in part by whether local decision-makers are using sentencing enhancements, including the “repetitive offender” enhancement. For example, as noted in the graph above, Maricopa sends people to prison with the longest average sentence. Maricopa also uses the “repetitive offender” enhancement far more than the rest of the state. Maricopa alone accounted for nearly 76 percent of all the people sent to prison in Arizona with the enhancement, while it only accounted for 61 percent of the state population, and 62 percent of all people sent to prison.

Disparities in sentence length can also be driven by differences in who is charged with a crime and sent to prison in each county (e.g., counties that do not sentence as many people to prison for low-level crimes could have longer sentence lengths due to the composition of their prison admissions). But we still see significant differences when looking at non-violent crimes only. In Maricopa, people sentenced to prison for non-violent crimes receive an average sentence of 43 months, compared to 25 months in La Paz.

The difference is even starker looking only at drug offenses. People sentenced for these offenses in Yavapai, Graham, or Mohave will serve close to four years in prison, compared to two years for those sentenced in Santa Cruz or La Paz.

Tale of two counties: Maricopa and Miami-Dade

With 4.3 million people, Maricopa is home to 61 percent of all Arizona residents. It is over four times the size of Pima, the next largest county in the state. Because of its size, comparing Maricopa to other counties in Arizona or statewide averages may be misleading since the county disproportionately drives the state numbers.

Other large urban counties with similar crime rates and demographics may provide better comparisons. Miami-Dade County has 2.75 million residents and, like Maricopa, encompasses multiple municipalities. The two Metropolitan Statistical Areas have a similar crime rate (approximately 3,500 reported crimes per 100,000 people in Miami-Dade compared to 3,300 for Maricopa). The counties also have a similar age distribution (20 percent under age 18 and 16 percent over age 65 compared to 24 percent and 15 percent respectively) and poverty level (18 percent living under the poverty line compared to 15 percent). Miami-Dade has a higher foreign-born population and a higher Hispanic population than Maricopa but both are well over the national average.

Critically for the purposes of this report, Florida and Arizona have similar sentencing and corrections systems. They are two of only three states in the country where nearly every person sentenced to prison has to serve at least 85 percent of their sentence behind bars.

Even though Miami-Dade and Maricopa are similar in many ways, they are very different in how they use state prison and how that use has changed over time. Since 2000, Maricopa has increased the number of people it sends to prison by 33 percent, while Miami-Dade has decreased the number of people it sends to prison by 46 percent. Today, Miami-Dade, a county almost two-thirds the size of Maricopa, sends only 16 percent as many people to state prison (1,400 compared to 8,929).

Another way to examine prison use in Maricopa and Miami-Dade is through the admissions rate. In 2017, Maricopa sent 20.7 people to state prison for every 10,000 residents, compared to only 5.1 in Miami-Dade, or more than four times as many people per capita.

While both counties have seen substantial reductions in crime, Miami-Dade did it without sending more people to prison. In fact, Miami-Dade experienced a greater reduction in crime from 2000 to 2016 — 54 percent — compared to a 47 percent decline in Maricopa.

Miami-Dade is not unique in this way. As shown in the first report in this series, 32 states across the country have experienced major declines in both crime and imprisonment in the last 10 years, while Arizona lags behind. Within these states, many large metropolitan areas are taking the lead in rethinking their use of prison and jails, including Harris County, Texas, New Orleans, Louisiana and Charleston County, South Carolina.

The number of people in state prison per 100,000 residents is a measure of how much a jurisdiction uses prison compared to other sentencing options. Unlike admissions figures, it includes many people serving long sentences who may have been sentenced in the past. Even using this more stable measure, Maricopa imprisons almost twice as many people per capita as Miami-Dade.

Racial and ethnic disparities

Communities of color experience imprisonment at higher rates than white people in Arizona. This is a problem for Arizona — and a problem nationally. In Arizona, communities of color are both disproportionately sentenced to prison, and, in some circumstances, spend longer periods behind bars. These disparities cannot be justified by differences in offending rates. Nationally, criminologists have studied this question for decades, and the most recent study found that a large portion — 45 percent — of racial disparities in imprisonment are not the result of racial disparities in arrest rates. Instead, at numerous points within the criminal justice system, from prosecution to conviction, researchers have identified instances of biased policies, practices, and decision-making. Disparities in who is sent to prison for drug crimes are particularly concerning. Drug use is similar across races, but communities of color are imprisoned at far higher rates.

While these racial and ethnic disparities exist statewide, the scale of disparity often varies depending on where people live. The second half of this section will examine how racial disparities operate at the county level.

In 2017, Hispanic people comprised 31 percent of Arizona’s state population but 37 percent of people admitted to prison. Black people were over-represented at an even higher rate, comprising five percent of the state population but 13 percent of prison admissions. According to these state prison figures, the percentage of Native Americans admitted to prison is roughly equivalent with the number of Native Americans in the state general population. However, these numbers likely do not represent the full range of Native American incarceration in Arizona since crimes committed on reservation land would likely be prosecuted and incarcerated under tribal or federal jurisdiction.

These racial disparities exist across most offense types, though to varying degrees. White people are much less likely to be admitted to prison for drug offenses or violent offenses than their share of the population, while Hispanic people are much more likely to be admitted to prison for drug offenses. Native Americans are disproportionately admitted to prison for offenses in the “Other” category, primarily for driving under the influence (DUI). In 2017, Native Americans made up 19.5 percent of prison admissions for DUI although they make up only five percent of the state’s population. Black people make up five percent of the state’s population but 17 percent of the state’s prison admissions for violent crime, driven largely by aggravated assault and armed robbery.

Hispanic and black people are most overrepresented among those in prison for one of the least serious crimes — simple possession of marijuana. Hispanic people comprise 31 percent of Arizona’s resident population and 32 percent of arrests for marijuana possession, but almost 60 percent of the people admitted to prison for that crime. Black people are five percent of the general population, but three times that share — 16 percent — of the people sent to prison for marijuana possession.

This is particularly concerning because research has consistently shown that marijuana use is similar among white people and people of color.

There are also significant racial and ethnic disparities for marijuana distribution offenses. Hispanic people comprise just under one-third of the general population, and 42 percent of arrests for marijuana distribution, but over 80 percent of people sent to prison for marijuana distribution.

In addition to stark disparities in terms of who is sent to prison, there are also differences in how long people spend in prison by race and ethnicity. On average, black people sentenced to prison in Arizona receive the longest sentences (62 months compared to 54 months for white people and 51 months for Hispanics). Even when controlling for gender, offense type, and the number of prior felonies, black people still receive the longest average sentences.

Sentence length disparities are particularly acute for some offenses. The first report in this series showed that the average prison sentence for people convicted of simple drug possession was 36 months. For black people sentenced to prison for simple drug possession, the average sentence was 56 months, or more than a year and a half longer.

This is almost certainly driven by differences in the use of sentencing enhancements. In 2017, 28 percent of eligible black people sentenced to prison for drug possession received a “repetitive offender” enhancement versus 18 percent of white people. This comports with research showing that black people are more likely than white people to be prosecuted under habitual offender statutes, controlling for criminal history and crime type.

Across all but one county in Arizona, black people are disproportionately admitted to prison, though the issue is greater in some counties than others (Cochise County is the exception). In five of the 15 counties, black people make up more than three times as high a share of prison admissions as they do of the general population. In La Paz, black people account for seven percent of prison admissions but only one percent of the county population. Even in the state’s largest county, Maricopa, black people make up 17 percent of prison admissions from the county but only six percent of the county population.

Hispanic people are disproportionately sent to prison in all but one county — Yuma. The largest disparities are in the predominantly rural Apache, Cochise, Graham, Greenlee, and La Paz counties. Some of these disparities are even larger when looking at specific crimes.

While Hispanic people across Arizona are overrepresented among those sent to prison for drug distribution, there are particularly significant disparities in a handful of counties. In Apache, Cochise, Coconino, Graham, Maricopa, Navajo, and Pinal counties, Hispanic people are more than twice as likely to go to prison for drug distribution. For instance in Apache, Hispanic people make up six percent of the population but 41 percent of people sentenced to prison for drug distribution. In some small counties, disparities may be driven by one-time events or enforcement actions. However, in Maricopa , where the size of the population makes this unlikely, Hispanic people make up 72 percent of people sentenced to prison for drug distribution and only 31 percent of the county population.

This is despite the fact that Hispanic people are overall no more likely to be arrested for distribution in Arizona than any other group (making up 33 percent of distribution arrests but 31 percent of the population).


Arizona has an imprisonment crisis — but it is not experienced uniformly. The problem is concentrated in some communities far more than others. This includes many rural communities that send people to prison at a high rate, as well as Maricopa, which sends people to prison with the longest average sentence of any county in the state. Communities of color are particularly impacted by this crisis. Black and Hispanic people are disproportionately sent to prison across most crime types, and black people receive the longest average sentences in Arizona.

These findings highlight the need to begin a conversation in Arizona around reforming the state’s prison population and, importantly, to ensure that all communities can engage in that discussion. Imprisonment is not only a function of statewide choices and statewide policies — much of it is determined by local practitioners. For Arizona to make meaningful reductions to its prison population, all of its diverse communities need to be consulted.


Unless otherwise cited, the analyses in this report were conducted by using individual-level data acquired from the Arizona Department of Corrections (ADC) covering over 30 years and nearly half a million records. Data was cleaned and coded for analysis by researchers and statisticians with extensive experience working with federal, state, and local corrections agencies across the country. To learn more about our researchers and the data and definitions used, see the full methodology in the first report in this series — Arizona’s Imprisonment Crisis: The High Price of Prison Growth.

In addition to the data definitions provided there, this report includes information on admissions and sentence length by county and race/ethnicity.

County-level information is based on an ADC variable (“ADMIT_FROMLOC”) identifying the jurisdiction an individual “came from” according to the codebook provided. The codebook further notes that this generally means the county of conviction. Individuals coming from jurisdictions outside of Arizona were excluded from this analysis.

Information on the race and ethnicity of individuals admitted to ADC custody comes from the ADC variable “RACE.” The values in this variable included both racial and ethnic groups and does not distinguish by race within ethnic groups (that is, it does not distinguish between “white/Hispanic” and “black/Hispanic” for example). The names of the groups were re-coded to match general population data more closely (“Caucasian” in the ADC variable was coded as “white,” “African-American,” as black, etc.). This variable does not include information on people with multiple ethnicities.

Tell the world; share this article via...