Saturday, February 09, 2013

I got your Big Government right here: Growth in TX convictions unhinged from crime levels

One purpose of this blog is to hone arguments in a more informal arena - using readers as a sort of raucous, unmannerly focus group to vet and refine them - before they're presented in more formal settings such as legislative hearings, public policy reports, and the like. So I was pleased when a combative commenter came forward in a recent post to criticize how I'd portrayed crime rates, which Grits said were declining. While that's true, Texas population increased about 20% over the last decade, meaning that even though crime rates per 100,000 dropped 25%, overall numbers of reported crime either remained steady or dropped only slightly throughout the decade. Here's how my intrepid critic suggested we should think about index crimes in Texas since the turn of the century, expressed as a fraction of the 2001 reported index-crime total:

2001 - 1.00
2002 - 1.03
2003 - 1.04
2004 - 1.03
2005 - 1.01
2006 - 0.98
2007 - 1.01
2008 - 1.00
2009 - 1.02
2010 - 0.97
2011 - 0.91

His point was that, rather than crime declining, with Texas' growing population, the actual total number of reported index crimes remained mostly steady, dropping significantly only recently. Fair enough. And a clever way to present the data. I agree with my anonymous critic it allows for a better apples-to-apples comparison to the other data discussed. So now let's compare reported crime figures to a similar ratio calculating the total number of felony convictions and deferred adjudication verdicts secured by Texas prosecutors over the same period:

2001: 1.00
2002: 1.05
2003: 1.18
2004: 1.20
2005: 1.28
2006: 1.30
2007: 1.42
2008: 1.37
2009: 1.38
2010: 1.36
2011: 1.42

To me, that says that Texas prosecutors no longer need more crime to secure more convictions, for reasons Glenn Reynolds has articulated. Prosecution is a growth industry. Here are the two ratios displayed together graphically, along with a similar metric for growth in arrests:

Convictions, Arrests and Reported Index Crime as a Fraction of 2001 levels, through 2011

Don't like Big Government? There's your Big Government. Find the underlying data, from various sources, in this chart.

See how increases in felony convictions and deferred adjudications have become disconnected from the amount of reported crime or even the number of arrests, rising at far higher rates? If the Legislature wants to close more prison units they must reduce upward pressure on prison admissions over the long term, and this chart shows where that pressure is coming from. The state could chop that top ratio down to size quickly by adjusting drug possession offenses down one notch and/or indexing property crime category thresholds to inflation, but will the Legislature, particularly brand spanking new Criminal Jurisprudence and Corrections Committees in the House with rookie chairmen, be willing to take such bold steps with no (public) plan on the table and little or no time to prepare? That remains to be seen. Like Fox Mulder, I want to believe.



Anonymous said...

It would be interesting to know what the average wait time is from arrest date to actual conviction date. Wonder what the average length of time is from arrest date to conviction date?

Gritsforbreakfast said...

You can get a sense of those dynamics, if not those specific datapoints, from the Office of Court Administration Annual reports. E.g., see the second page of this pdf from FY 2012 data. Their past years' reports are here. Go to District Courts and "Activity Detail."

Anonymous said...

Get rid of, or reduce the penalties associated, w drug free zones. Ever look at a map of your municipality overlayed with drug free zones? Anywhere, and everywhere, are now drug free zones. Its a stealth enhancement.

sunray's wench said...

If reported crime is going down, but convictions are going up, could it be that older crimes are finally being solved?

Gritsforbreakfast said...

I'd be hard pressed to believe, SW, that cold cases account for the increase. They're few and far beween.

Anonymous said...

RE: the 2012 Data sheet:

It would appear that close to %40 of dispositions in a given year are from arrests 12-18 months prior.

Is it me or does this report fail to clearly indicate the number of cases plea bargained. I am assuming this could be under "agreed" dispositions. If so that seems alfully low number.

Gritsforbreakfast said...

8:12, go to page two of the 2012 pdf and look under "Dispositions." They're broken out by "Guilty Plea or Nolo," "By the Court," and "By the Jury." Take the figure from "Guilty Plea or Nolo," add it to the figure for deferred adjudications (just below on the chart), and that gives you the total pleas. The plea bargain rate was 96.5% among 2012 felony convictions, by my calculation.

Anonymous said...

I agree, 4:59:00. Or at least restrict drug-free zone enhancements based on what time of day it is and whether anyone is present, rather than charging someone with an offense in a drug-free zone because they happened to get pulled over in front of a school at 2:00 a.m.

Anonymous said...

Dr. Tony Fabelo doesn't have an analysis of the rising prison vs. declining arrest rates. Would also like to heare his take on what data is missing to give a better picture of where we need to go in terms of sentencing policy.

Anonymous said...

After a person is convicted of a felony, whether guilty of not. especially in Harris, Co, when released that person is still punished and not allowed to get a good job and make a living.

There are many very educated men and women who are not guilty and only convicted because the ADA simply wants to win a case and chalk up that win; they are usually learning to prosecute with someone from the DA's office, then go merrily on to their career.Lies abound! All to be able to chalk up a win.

Why can't the rule be made, after completetion of the sentence, the felony is removed from that person's record. My son is highly educated in computer programming, the conviction in Harris Co. was false and yet, he cannot find a programming job because of a false record. Just how long will this stupidity be allowed to continue, until we run out of qualified people?

locutia7 said...

I wonder if the relative ease of securing a guilty plea (at least 98 percent of all convictions, btw) implicates everyone in the judicial system--prosecutors, defense lawyers, defendants, cops...This is really bad news for all of us.

Anonymous said...

"Combative" here.

I'm happy to be helpful in honing your analysis. But there is still quite a bit more work to do.

Your revised analysis suffers from two fundamental problems:

First, it is a rather blatant apples-to-oranges analysis, in that it compares index crime data, on the one hand, to overall arrest and conviction/deferred adjudication data, on the other hand. Index crimes are a restricted data set. How restricted depends on the source of the data. If the source is the FBI's annual summary data, then it is the eight crimes of willful homicide, assault, rape, burlary, theft, larceny, motor vehicle theft, and arson. However, if the source is DPS's annual summary data then it is the seven crimes of willful homicide, assault, rape, burlary, theft, larceny, and motor vehicle theft (i.e., arson isn't included in DPS's summary data). In contrast, your numbers for Texas's arrests and convictions/DAs is for all crimes, not specifically index crimes. It is perfectly reasonable to use the index crime rate as a measure of the prevalence of crime. That is standard. But it is then a bit contrived to evaluate the connectness/disconnectedness of the criminal justice system to the crime rate by throwing in lots of other crimes into the mix that are not reflected in the crime rate data you are using.

Second, you approach the analysis as true-believers (in Hoffer's sense) generally approach analysis. You approach it to support a bias rather than to ask a series of meaningful questions to clarify relationships. As a result, you stop as soon as you observe something that fits your bias.

So, if you are going to use index crimes as the measure of the prevalence of crime, then you need to look at arrests and convictions/DAs for index crimes. These data can be extracted fairly easily from the annual DPS UCR reports, and the annual reports for the judiciary.

So what does the relationship between index crimes and index crime arrests look like? Does it support the idea that the criminal justice system is disconnected from the actual occurence of crime? The count data normalized to 2001 counts look like this:

2001 1.00
2002 1.06
2003 1.06
2004 1.13
2005 1.08
2006 1.04
2007 1.12
2008 1.17
2009 1.24
2010 1.19
2011 1.12

Well, that sort of looks like an upward trend overall. So maybe there really is a disconnect between arrests and crime rates. If I were a true believer, I would stop there. But as someone who wants to understand what is going on I wonder about a different relationship: the relationship between index crime arrests each year and the number of reported index crimes in that same year. That relationship would seem to be reflected in the annual index crime arrests expressed as a percentage of annual index crimes reported. So what does that trend data look like? Like this:

2001 13%
2002 14%
2003 14%
2004 15%
2005 14%
2006 14%
2007 15%
2008 16%
2009 16%
2010 16%
2011 16%

Well, there still looks to be an upward trend. In 2001 arrests for index crimes were 13% of the reported index crimes. In 2011 that percentage increased to 16%. But what is glaringly apparent is that 16% is as good as it gets. That seems like a pretty piss-poor percentage to my mind. I'm paying taxes for my police agency to investigate and solve crimes. But for the most serious of crimes the number of arrests in any given year is only 16% of the number of corresponding crimes. Based on these numbers, I as a taxpayer would feel pretty comfortable telling the police that they need to do a better job at investigating these serious crimes. And if they do a better job, what would I expect to happen to these numbers? I'd expect the percentages to gradually rise over time. Which is exactly what they seem to be doing. Which would be a very good thing in my mind. But more imporantly it would reflect a very close connectedness between crime rate and arrests, not the disconnectedness that is claimed by the true-believers.

So, you get the idea. Hopefully.

Gritsforbreakfast said...

11:06, I'm not sure what you think you've demonstrated here. Burrowing down into index crime arrests (one notices you've shifted the conversation, having lost the argument in earlier threads about convictions), you concur that numbers of arrests are disconnected from crime levels, just as I'd demonstrated numbers of convictions are disconnected . So far so good.

Then you conclude that that's okay because, based on your own ideology, you'd prefer more arrests and prosecutions, anyway, so you don't really care. But that normative claim doesn't dispute the mathematical trends presented, nor does the modest increase in arrest percentages described in the second half of your comment explain the numbers. The disconnect is real, obvious and even supported by your own calculations.

Finally, I'll readily grant index crimes numbers aren't direct apples to apples comparisons to total arrests and felony convictions - one can't manufacture directly comparable data out of thin air, and that's what exists. (And of course, when you found more precisely fitting data on index crime arrests they showed the same pattern.) But they can demonstrate general trends, and the trend has been that felony convictions continued to rise dramatically during a period that crime was stagnant or falling. Moreover, the data show that arrests more closely tracked crime rates than convictions, which rose more significantly than arrests and remained high even when arrests declined. Nothing you've said here or in the other posts changes that basic fact, which is the main driver behind upward pressure on TDCJ admissions.

Gritsforbreakfast said...

BTW, 11:06, would you feel more comfortable if, instead of using reported UCR crime, I'd used crime victimization survey data to estimate crime levels? They capture the larger categories of crime comparable to the all-arrests, all-convictions numbers, etc., and those trends show crime declining at similar rates to the UCR vis a vis 2001. It certainly wouldn't change the analysis much.

Anonymous said...

I don't know if this has any impact on the increase, but in the last 2 years all of the counties have been working to get their disposition completeness reports in compliance. That requires 95% of all arrests in year X to be disposed of by the end of year X+2. In the process of doing this, they may have resolved a good number of old cases. Not exactly what SW suggests with her cold cases solved suggestion, but along those lines. If this has anything to do with it, you should start to see a leveling out between arrests and convictions as they stay caught up. Of course, you also have to factor in multiple convictions off the same arrest, so there will always be some discrepancy. Admittedly, I didn't peruse the reports like you did, but thought I'd offer this as a potential factor to consider.

Anonymous said...

"BTW, 11:06, would you feel more comfortable if, instead of using reported UCR crime, I'd used crime victimization survey data to estimate crime levels? "

I would prefer something more than a faux analysis based on selective and biased filtering of information. There is way too much of that already. Opinions are fine. But when you gussie-up opinion mongering as faux analysis, you cheapen both opinion and analysis.