Usual, But Wholly Misunderstood, Effects of Policies on Measures of Racial Disparity Now Being Seen in Ferguson and the UK and Soon to Be Seen in Baltimore
In a February 22, 2016 commentary for The Hill titled “Things DoJ doesn’t know about racial disparities in Ferguson,” I discussed the mistaken premise of the suit the Department of Justice (DOJ) had recently filed against the City of Ferguson, Missouri. I explained that, contrary to the premise of the complaint in the case (and myriad other DOJ actions) that generally reducing adverse criminal justice outcomes will tend to reduce (a) relative racial differences in rates of experiencing the outcomes and (b) the proportion blacks make up of persons experiencing the outcomes, in fact generally reducing the outcomes tends to increase (a) and (b). I had previously attempted to explain the issue with regard to DOJ’s March 2015 report on Ferguson’s police and court practices in a March 9, 2015 letter to the agency and Ferguson officials, and I would later attempt to do so in a submission to the Eastern District of Missouri regarding a proposed consent decree that would ultimately be entered on April 19, 2016 – to no avail in either case.
I have explained this issue here many times between May 6, 2016, and September 12, 2019, in some cases, as in the recent item, discussing the way that the government promotes racial mistrust by leading the public to believe that policies will tend to reduce measures of racial disparity when the policies in fact will tend to increase those measures. For, if measure of racial disparity increase in the face of policies that are supposed to reduce the measures, observers who believe that racial bias is an important cause of racial differences in various adverse outcomes will tend to think that bias must be increasing. And I have repeatedly discussed the DOJ’s obligation to understand the matter and explain it to the public and the other entities it has misled over the years. Such entities include the courts handling the City of Ferguson and Baltimore Police consent decrees (and the citizens of each locale). They also include the government of the United Kingdom, as I pointed out in a November 7, 2017 post here discussed further several paragraphs below.
The consequences of the policies that the DOJ and others mistakenly believe will tend to reduce the aforementioned (a) and (b) are beginning to be seen in two of these three places.
Ferguson and Missouri Disparities. One practice of the City of Ferguson of which DOJ was especially critical in its March 2015 report was the heavy reliance on traffic and court fines for municipal revenue. Some might think the extent to which a jurisdiction relies on traffic and court fines to finance itself was none of DOJ’s business. But the agency regarded the matter within its civil rights enforcement purview because of the fact that blacks made up an especially high proportion of persons paying those fines – though, again, while the agency failed to understand that limiting the use of fines would tend to increase, not reduce, that proportion.
Even before DOJ brought its suit against Ferguson, the Missouri legislature, in 2015, attempted to respond to the Ferguson riots by limiting the proportion of a city’s operating revenues derived from fines, forfeiture, or court fees related to minor traffic violations. Senate Bill 5 (SB5) limited that proportion to 20 percent – except in St. Louis County, where Ferguson is located, where the proportion was limited to 12.5 percent. In establishing these limits, the legislature no doubt acted pursuant the belief that by generally reducing traffic stops, the legislation would reduce the proportion blacks made up persons penalized for traffic violations, especially in St. Louis County. In fact, however, the legislation would tend to increase those proportions, especially in St. Louis County. Paragraph 324 of the Ferguson consent decree then required that Ferguson comply with SB5 and provide DOJ with records reflecting that compliance.
Recently, a number of articles have reported on changes in measures of racial disparity that have followed the general reductions in traffic stops resulting from SB 5 (and presumably also from the Ferguson consent decree). An August 7, 2019 New York Magazine article titled “5 Years After Ferguson, Racial Disparities in Traffic Stops Have Gotten Worse” reported that, while traffic stop revenues had been dramatically reduced in recent years, measures of racial disparity that are functions of the proportion blacks make up of persons stopped had increased in Missouri and are especially large in Ferguson. An August 6, 2019 New York Times article on which the New York Magazine article relied specifically found that the disparities in Ferguson had increased. Both articles discussed the troubled reactions to these patterns on the part of members of the public and civil right organizations. Neither the authors nor the persons interviewed questioned the mistaken belief that the reductions in stops should have reduced the proportion blacks make up of persons stopped. The same misunderstanding is reflected in a June 13, 2019 NPR article, which casts the matter in terms of the increase in the ratio of the black stop rate to the white stop rate.
Thus, due to pervasive innumeracy, especially on the part of the DOJ and social scientists who ought to know better, actions aimed at reducing racial tensions in Ferguson and elsewhere in Missouri instead will exacerbate those tensions.
United Kingdom Disparities. In a November 7, 2017 post here titled “United States Exports Its Most Profound Ignorance About Racial Disparities to the United Kingdom,” I discussed the influence of the U.S. DOJ on the UK government’s September 2017 report titled “Lammy Review: An independent review into the treatment of, and outcomes for, Black, Asian, and Minority Ethnic Individuals in the Criminal Justice System.” As with many of the DOJ’s own activities regarding criminal justice and various other matters, the Lammy Review was premised on the mistaken belief that generally reducing adverse criminal justice outcomes would tend to reduce, rather than increase, (a) relative racial differences in rates of experiencing the outcomes and (b) the proportion racial minorities made up of persons experiencing the outcomes.
There is significant irony in any situation where policies are premised on beliefs that are the opposite of reality. A particular irony of the Lammy Review, however, arises from the fact that one of its key themes was the importance of promoting trust in the UK’s criminal justice system within minority communities in order that minority defendants would avail themselves of the advantages of pleading guilty to well-founded charges with the same frequency as white defendants. But by promoting the view that observed racial differences were a result of discrimination and by leading readers to believe that measures of racial disparity were increasing “despite” (rather than “because of”) general reductions in adverse criminal justice outcomes, the review was itself promoting racial mistrust.
As discussed in the November 2017 post, there was also reason to believe that the Lammy Review’s recommendations, while touted as means of reducing (a) and (b), would increase those measures still further. The UK Ministry of Justice uncritically accepted the reasoning of the Lammy Review and has been implementing its recommendations while emphasizing that in measuring racial disparity by the method employed by the U.S. DOJ the Ministry of Justice should become a world leader in the analyses of such issues. But, contrary to the expectations of the Lammy Review and the Ministry of Justice, the measures of racial disparity employed in the Lammy Review and adopted by the Ministry of Justice while implementing the Review’s recommendations have increased.
At a March 26, 2019 hearing of the Parliament’s Select Justice Committee, the Lammy Review’s author, MP David Lammy reported that racial/ethnic disparities in adverse criminal justice outcomes had grown “considerably worse” since the September 2017 issuance of his report. Mr. Lammy noted as a “major, major development” that black and minority ethnic individuals now made up 51 percent of the youth prison population.
The fact that measures of racial disparities are increasing the face of actions that the UK government leads the public to believe should decrease the measures will likely lead to further concerns about racial bias and further efforts to reduce adverse criminal justice outcomes in the belief that doings so is an increasingly crucial means of reducing the measures of racial disparity that Mr. Lammy found to be increasing. Absent some authoritative intervention, we can expect this situation to continue indefinitely. But, as here, in the UK no authoritative entity yet understands the matter.
(I am not yet familiar enough with the situation to know whether mistaken perceptions about the effects of policies on the proportion racial minorities (or Muslims, also a subject of the Lammy Review) make up of incarcerated persons may have influenced the policies that led to the early release of the perpetrator of the November 29 terror attack on London Bridge.)
Baltimore Disparities. I am aware of no data yet available reflecting effects of the Baltimore Police consent decree on the measures of racial disparity employed by the DOJ in its August 2017 report on the city’s police practice. But, at least so far as their interactions with the court reveal, the parties and the decree’s monitor continue to believe that things like diversion programs that generally reduce adverse criminal justice outcomes will tend to reduce, rather than increase, the proportion blacks make up of young people in the criminal justice system.
Consistent with its mistaken premise regarding the effects of policies on measures of racial disparity, the decree required the City of Baltimore to comprehensively assess the city’s efforts to decrease youth involvement with the criminal justice system by diversions programs and other means. In October 2018 the city chose the Center for Children’s Law and Policy (CLLP), a Washington, DC nonprofit that is one of many organizations promoting the mistaken belief that diversion programs will tend to reduce (a) and (b), to perform this assessment. In April 2019, the CCLP issued a 55-page report touting the benefits of diversion programs both generally and as a critical means of reducing racial disparities in incarceration and urging further expansion of such programs. The report does not discuss effects of the programs so far implemented in Baltimore on measures on racial disparity.
But, pursuant to the decree’s many requirements regarding the collection and analysis of data on racial differences, the decree’s monitor and the parties should be starting to see some of the decree’s effects on the measures of racial disparity in the DOJ’s underlying report. And, unless something quite unusual occurred in Baltimore, those examining the data should be observing that those measures are changing in the opposite direction of what the parties have led the court and the citizens of Baltimore to believe would occur.
In my September 12, 2019 post here, I discussed the prospects that in the government’s brief in the court of appeals in the case of Council of Parent Attorneys and Advocates, Inc. (COPAA) v. DeVos, then due on September 18, 2019, the government would acknowledge that, contrary to arguments it made in the district court, reducing an adverse outcome tends to increase relative racial differences in rates of experiencing the outcomes. But, also on September 12, the government, which in its earlier motion postponing the filing until September 18 had cited a need to confer with the Solicitor General on the matter, moved to dismiss its appeal. The court granted the motion several days later.
I don’t know whether a recognition by DOJ attorneys that reducing an outcome tends to increase relative differences in rates of experiencing the outcome played into the decision to dismiss the COPAA appeal. Even if it did, however, that hardly means that the understanding will then spread throughout the many parts of the agency where policies and litigation decisions are being informed by beliefs about effects of policies on measures of racial differences that are the opposite of reality. That understanding may have to await such development as having a court handling a decree is a place like Ferguson or Baltimore demand to know why the agency has been leading the court and the public to believe that policies would tend to reduce measures of racial disparity when the policies actually would tend to increase those measures.
Appendix – Innumeracy of the Disparities Industry
The Center for Children’s Law and Policy (CCLP) is part of a vast and growing disparities industry and the organization’s activities strikingly illustrate the anomaly of an industry that emphasizes the importance of understanding data in addressing disparities issues at the same time that it leads the public and policymakers to believe that policies will tend to reduce measures of racial disparity when in fact the policies will tend to increase those measures. Often those being misled on this matter, including many officials at all levels of government, pay substantial fees to be misinformed, as in the $3000 per person program CCLP conducts in partnership with Georgetown University’s Center for Juvenile Justice Reform.
One peculiar aspect of such programs, and like programs promoting restorative justice as a means of reducing racial differences in public school suspensions, is that by promoting policies that actually increase the measures of disparity on which the programs encourage participants to rely the programs increase the demand for the programs. This might be considered a remarkable fraud if the groups marketing their expertise in data analysis were actually aware that the policies they promote increased those measures.
CCLP also produces a Racial and Ethnic Disparities Reduction Manual based on work funded by, along with various foundations, the DOJ’s Office of Juvenile Justice and Delinquency Prevention. The manual stresses both the need to understand data in order to reduce disparities and the importance of reducing adverse criminal justice outcomes by things like diversion programs in order to reduce the ratio of the black rate of experiencing the outcomes to the white rate of experiencing the outcomes (the main measure of racial disparity in the manual, which, like the Lammy Review and the UK Ministry of Justice, the manual borrowed from the DOJ).
The manual’s second chapter, titled “Using Data Strategically to Reduce Racial and Ethnic Disparities,” notes encouragingly (at 7) that as a result of certain program in Sedwick County, Kansas, shoplifting thefts by young people were reduced by 27 percent in one year, “including 26% for black youth.” But the fact that the black rate was reduced less than the overall rate means that the ratio of the black rate to the overall rate increased. A chart on page 20 then shows that the white rate decline by 31.5 percent, which means there was an even more pronounced increased in the ratio of the black rate to the white rate.
Referring to the same program, the chapter (at 21) notes, also encouragingly, that juvenile arrests declined almost 20 percent for black youth. Here one must go to the cited reference to learn (at 3) that that the white rate declined by 21.7 percent, again meaning the ratio of the black rate to white rate increased.
In the school discipline context, another aspect of the disparities industry’s innumeracy is illustrated by the fact that in the occasional situations where reportage or studies indicate or suggest that a general reduction in discipline rates was accompanied by a decrease, rather than an increase, in the ratio of the black rate to the white rate, the data show that the ratio in fact increased (as in treatments of disparities issues concerning Oakland (CA), Denver (CO), Allegheny County (PA), and Massachusetts). The school district in Oakland, in particular, has received much acclaim for the way that its restorative justice programs reduced the ratio of the black suspension rate to the white suspensions, even though, unbeknownst to those seeking to learn from Oakland’s experience and Oakland officials themselves, the ratio actually increased.
Materials produced by the disparities industry commonly promote the view that bias plays a large role in observed racial differences. Sometimes, especially with regard to school discipline, the materials make the facially implausible claim there are no racial differences in student behavior (even though blacks are disproportionately represented among the groups defined by background characteristics and academic achievement levels where behavior problems are most pronounced). But always there is a tendency to suggest that behavior differences are minor and, in event, are insufficient to explain differences in outcomes.
Part of the problem in such discussions is the general failure to understand the issues I addressed here in “The Misunderstood Relationship Between Racial Differences in Conduct and Racial Differences in School Discipline and Criminal Justice Outcomes” (Dec. 20, 2017). That is, the same failure of understanding of features of most risk distributions that underlies the mistaken belief that reducing an outcome tends to reduce relative differences in rates of experiencing the outcome underlies the failure to understand that a certain relative difference in behavior could explain a much larger relative difference in sanctions for more serious manifestations of that behavior. A good example of the point may be found in Table 1 (at 5) of my August 9, 2019 letter to the DOJ attorney handling the COPAA v. DeVos appeal. It shows that if students are sanctioned for any missing of class, the ratio of the black rate of one or more sanctions to the white rate would be 1.08. But as sanctions are limited to more extreme cases, the ratio would consistently rise, reaching 2.67 when sanctions are limited to students who miss class more than weekly.
But the CCLP manual provides an extreme illustration of the disparities industry’s approach to minimizing the role of racial differences in behavior on racial differences in outcomes. The manual’s Chapter 3 (at 6) states that that “marijuana use is only slightly higher among black (28.9%) … students than it is for white (20.4%) students” and that such fact cannot explain why “[b]lack youth are almost 40% more likely to be arrested for a drug violation than their white counterparts.” But 28.9 percent is 42 percent greater than 20.4 percent. Thus, in this instance the manual is saying that a 42 percent greater use of marijuana by blacks than whites is insufficient to explain a 40 percent greater black arrest rate. This, however, is less likely an example of the authors’ taking advantage of the limited numeracy of the manual’s readership than it as illustration of the authors’ own limited numeracy.