This commentary originally appeared in Real Clear Policy on August 17, 2016.

In a recent Real Clear Policy article, Sean Kennedy examines Texas’ violent crime rate, questioning the Lone Star State’s policy of improving public safety while reducing incarceration. Unfortunately, by cherry-picking data of questionable quality, Mr. Kennedy undermines his central claims.

Start with the facts. Over the past decade, Texas closed three prisons while cutting its juvenile detainee population from about 4,000 in 2006 to 1,331 earlier this year. These reforms focused on keeping non-violent offenders out of costly lock-ups and used a portion of the dollars saved toward proven treatment, rehabilitation, and reentry programs. Texas did not reduce penalties for violent offenders, let alone murderers. The result: Reduced recidivism, lower costs, and a state-wide crime rate reduced to levels not seen since 1968.

This is the Texas model of reducing both crime and incarceration rates, and it has been successfully implemented in varying degrees in some 40 states as well as informing pending federal criminal justice legislation. So impugning Texas’ criminal justice reform efforts isn’t just messing with Texas, it’s questioning the basis of criminal justice reform work across the nation.

The idea that crime rates and incarceration are joined at the hip has been thoroughly discredited. Incapacitation of criminals via incarceration is a factor in crime rates, to be sure. But most criminals eventually get out of prison, and, when they do, helping them stay out is vitally important.

Moreover, there’s no evidence that recent criminal justice reforms pioneered in Texas have any connection to the increases in violent crime seen in some U.S. cities. To the contrary, over the past several years, crime rates have fallen faster in states that have reduced imprisonment rates than in those where prisons have continued to grow. Many of the cities now experiencing violent crime increases, such as Chicago, are in states that have yet to implement comprehensive sentencing and corrections reforms.

This isn’t to say that current criminal activity, largely concentrated in major urban centers, might not have something to do with the so-called Ferguson Effect or the Mexican drug cartels’ replacement of marijuana smuggling with heroin. Criminal behavior and crime rates are the result of a complex interplay of demographics, policing, sentencing, incarceration, rehabilitation, reentry, and other factors. For that reason, parts of the system can be improved while others fail, resulting in increased overall crime.

So, what, exactly, is the problem with Texas? 

Mr. Kennedy admits that crime plummeted in Texas as prison populations were reduced right through 2014. But he goes on to say, “Now violent crimes — especially homicide — have spiked again in Texas’ biggest cities,” suggesting that Texas’ criminal justice reforms are to blame. But there is no correlation between increased violent crime in Texas’ largest cities and criminal justice reforms largely aimed at nonviolent offenders. In fact, there’s a long history in Texas and a large body of research that shows these well-implemented alternatives not only cost less but work better than the prison cell. Such reforms, when implemented correctly, reduce recidivism and crime rates.  

If we misdiagnose the problem, we won’t find the solution. Kennedy points to two charts showing the number of homicides and violent crimes committed in Texas’ five largest cities for the first six months of 2014, 2015, and 2016. Because FBI statistics are only available through the first half of 2015, Kennedy combines raw data from the police departments themselves as well as from the Major Cities Police Chiefs Association, a lobbying and advocacy group.

The problem? A staffer at the Major Cities Police Chiefs Association warned us that their data are “not scientific” and used only “as a benchmark for the agencies to see where they stand in relation to one another.” Further, the data are not checked for accuracy but simply self-reported by the member agencies, who return a survey sent out by the association. This ought to raise red flags.

How do Kennedy’s data compare to the official FBI data available for two of the three periods on which he reports? Not well.

For instance, Kennedy asserts that the number of homicides in Houston went up 57.8 percent in the first six months of 2014 compared to 2015; the FBI data says 44 percent. Both indicate a large jump, but the difference between 57.8 percent and 44 percent is statistically significant. 

Kennedy’s overall violent crime number for Houston is even more at odds with available data. He claims that the number spiked by almost 25 percent, comparing the first half of 2014 to first half of 2015. The FBI data for the same period doesn’t show such increase, but, rather, a decrease of 1.8 percent. Factoring in population growth, the decline in violent crime in Houston over that period is closer to 3 percent — far lower than Kennedy’s 25 percent.

Since the Major Cities Police Chiefs Association didn’t report Houston’s statistics in their compilation of data, Kennedy likely derived the figure (7,957 violent crimes from January 2014 to June 2014) from the Houston Police Department. But that department reported 10,000 violent crimes for the period while the FBI lists 10,401. Thus Kennedy’s base year for Houston is more than 20 percent lower than it should be, undermining his subsequent calculations. 

Now, let’s take a look at 2013 data. For the first half of that year, Houston reported 10,106 violent crimes. Comparing the first six months of 2013 to 2016, the overall number of violent crimes rose by 14.7 percent in three years. Factoring in population growth and using the proper baseline for 2013, we see that violent crime rate is up about 9 percent in three years and about 13 percent over the last two, more than a third less than Kennedy’s 36 percent jump.

Looking at the 29 major Texas cities comprising 39 percent of the state’s population that reported 2014-2015 data to the FBI, (Austin didn’t report), we see violent crime going up in 15 cities and down in 14, an average increase of 4 percent (not factoring for population growth). For urban centers statewide, violent crime was up 2.4 percent. Factoring in population growth, the violent crime rate in these cities over the first six months of 2015 compared to 2014 increased about 0.6 percent.

Murder spikes are concerning, but the violent crime rate is more telling. Murders are, thankfully, a small proportion of the aggregate of violent crimes in any city, so a movement up or down is a large percentage of that small number. Keep in mind that the Uniform Crime Report, a database maintained by the Federal Bureau of Information on national crime statistics and used in innumerable research efforts, only captures the most serious offense from any single incident. For instance, if a victim is killed as a result of a rape, robbery, or aggravated assault, only the murder is counted. So it would be more alarming if the murder rate and the violent crime rate were rising in tandem (they’re not).

Since violent crime, especially homicide, is relatively rare, property crime rates can provide a better barometer of trends and the effectiveness of criminal justice policies. What happened to the number of property crimes in Texas’ major cities (excluding Austin) reported to the FBI? Down by 5.9 percent from the first half of 2014 to the first half of last year. Converting to a crime rate, property crime in these cities is down almost 8 percent.

The Major Cities Police Chiefs Association report, which is comprised of 61 urban law enforcement agencies (and does not take population growth into account), indicates that violent crime, including homicide and non-fatal shootings, is up nationally by 2.3 percent from the first half of 2015. This means that the violent crime rate in this subset of cities is up on the order of just over 1 percent — not good, but certainly not a massive crime spike. For the full year, 2014 to 2015, the violent crime rate is up a similar 2.2 percent.

So it’s simply not the case that, in aggregate, violent crime in Texas’ major cities “has risen year-on-year for the first time in a generation,” as Kennedy asserts. On the contrary, Texas crime rates are at historic lows and its incarceration rates are heading lower, mainly due to a shift in treatment of non-violent offenders, which leaves more room in state lockups for the violent. These successful criminal justice reforms have resulted in improved public safety and a lower cost to taxpayers. 

The evidence from Texas is clear: Criminal justice reforms have improved public safety, not imperiled it. The Lone Star State is a model for the rest of the nation.

Chuck DeVore is a vice president with the Texas Public Policy Foundation and a former California lawmaker. Randy Petersen is a senior researcher with the Texas Public Policy Foundation’s Right on Crime initiative and a veteran of 21 years of law enforcement as a sworn officer.