Audit of the Pittsburg Police Department, Part 3

by Tom Flaterty

PERFORMANCE AUDIT
DEPARTMENT OF PUBLIC SAFETY
OFFICE OF PROFESSIONAL STANDARDS

Part III

Report by the Office of the City Controller

TOM FLAHERTY
CITY CONTROLLER

Anthony Pokora, Deputy Controller

Kevin Forsythe, Esq., Management Auditor

Anabell Kinney, Esq., Assistant Management Auditor

Performance Auditors:
Gloria Novak
Joseph Chigier
Jeff Khadem
Woody Mudd

August, 1996

One possible explanation for this lies in the OPS practice of initially distinguishing between serious and less serious citizen complaints. Complaints deemed to meet an appropriate level of seriousness were designated OPR¡sŒ and assigned to OPS investigators. Complaints deemed less serious were designated DR¡sŒ and sent out to zone commanders, who were then supposed to look into the matter. A certain percentage of these DR¡s were, apparently, left to languish on the desks of Public Safety Management at various links in the chain of command. The longer a complaint remains in this limbo state, the more difficult it may ultimately be to impose discipline, even in those cases where discipline would clearly be appropriate.

In addition to the problem of staleŒ cases, these delays jeopardized a crucial segregation of functions. OPS should have been allowed to complete its role as independent fact finder. Public Safety Management should then perform its role as the dispenser of discipline. The inclusion of representatives of management in the fact-finding stages compromises the integrity of the investigation.

RECOMMENDATION NO. 2:

To avoid unacceptable delays in the disposition of complaints, and to safeguard the integrity of its independent fact finding function, OMI should re-examine the practice of referring less seriousŒ citizen complaints (DR¡s) to Public Safety Management for investigation.

RECOMMENDATION NO. 3:

OMI should review all pending complaints for the years 1991 through 1994 in order to identify those cases in which no valid reasons exist for the pending classification. These should be given a final disposition and referred to management for discipline, where appropriate.

RECOMMENDATION NO. 4:

If OMI decides to retain the practice of referring less serious citizen complaints to Public Safety Management for investigation, strict limits should be imposed on the amount of time a complaint is allowed to stay at any level in the chain of command. We suggest 30 days.

III. Race, Gender, Age/Experience Analysis (Objective No. 3)

III. a. Race

To review, 1,674 complaints were filed by citizens over the 10-year audit period (April, 1986 through March 4, 1996). We were able to identify the race of the complainant in 1,507 (90%) of these cases. We were able to identify the race of the accused officer in 1,529 (91%) of these cases. Reasons for lack of identification included no entry in the database, obviously incorrect entries, and entries for officer¡s name such as unknownŒ, homicideŒ, detective.Œ In the Table below, adding the number of known racial entries to the number of unknown, still does not equal the full total. This is because we also eliminated a very small number of racial entries such as HŒ (Hispanic; one Hispanic complainant and one Hispanic officer in the whole 10 years), IŒ (Indian; one Indian complainant in the whole 10 years), etc.

Of 779 complaints known to have been filed by white citizens, we can identify the race of the accused officer 730 times. Of these, 515 were filed against white officers; 201 were filed against black officers.

Of 698 complaints known to have been filed by black citizens, we can identify the race of the accused officer 637 times. Of these, 313 were filed against white officers; 317 were filed against black officers. This is a noteworthy finding. Black officers (male and female) currently comprise about 26% of the Police Bureau¡s sworn strength. Yet in those cases we can identify, when black citizens filed a complaint with OPS, 50% of the time they were complaining about a black officer. [As the table on the next page shows, this disproportion occurred in all years of the audit period, ranging from a high of 63% in 1991 to a low of 37% in 1995].

We next repeated the above analysis for each of the years in the audit period, to see if these ratios stayed roughly constant over time or if noticeable deviations or trends appeared. The results of that analysis are presented in the table below. W/WŒ means a white citizen complainant against a white police officer; W/BŒ means white citizen, black officer, and so on. Total WhiteŒ means the total, for that year, of all complaints filed by white citizens in cases where the race of the officer is known. Similarly, Total BlackŒ means the total, for that year, of all complaints filed by black citizens in cases where the race of the officer is known.

Looking first at the average for complaints filed by white citizens (72% against white officers vs. 28% against black officers, over the entire audit period) we note deviations from this rate falling between 55% vs. 45% in 1993, and 83% vs. 17% in 1995. (It must be kept in mind that the 100% vs. 0% for 1996 is just the first two months of that year). The average for complaints filed by black citizens was 50% against white officers vs. 50% against black officers. This average rate deviated from 37% vs. 63% in 1991, to 63% vs. 37% (a complete reversal of the rate) in 1995.

With the exception of 1990, for each year from 1986 through 1992, white citizens filed more complaints. Then in 1993, complaints by black citizens outstripped those filed by whites 73 to 49. This trend continued in the next three years: 116 to 72 in 1994; 96 to 78 in 1995; and 10 to 7 in (the first two months of) 1996.

1995 warrants particular attention. In that year, complaints by black citizens against white officers reached both their highest absolute number for the whole audit period, 60, and their highest deviation from the 50%-50% average, i.e., 63% against white officers vs. 37% against black officers. A similar jump occurred in complaints filed by white citizens. Their complaints against white officers reached their second highest absolute number, 65 (surpassed only by 1987¡s 75) and their highest deviation from the 72%-28% overall average, i.e., 83% against white officers vs. 13% against black officers. These 1995 figures show a year in which black citizens filed more complaints against the police than white citizens (for the third consecutive year) while at the same time complaints by both black and white citizens against white police officers rose (absolutely compared to white officers in the prior year and relatively compared to complaints against black officers).

Thus, when viewed in summary fashion, against the backdrop of the audit period as a whole, the finding that black citizens filed slightly over 50% of their complaints against black police officers, tends to negate a systemic racial interpretation. Nevertheless, the data for the last three full years, and in particular 1995, could lend support to the anecdotal perception of a recent deterioration in the relationship between the citizenry (black and white) and white police officers.

III. b. Gender

Of the 1,674 complaints filed by citizens over the audit period, we were able to identify the gender of both the complainant and the officer in 1,479 cases. This was 133 more cases than we were able to use in the racial analysis above. Reasons for nonidentification parallel those listed above. Of these 1,479 cases, 923 (62%) were filed by males and 556 (38%) by females. These breakdown as follows:

As we did in the racial analysis, we next broke down the gender data for each year of the audit period to identify trends. In the following table, F/F means female citizen filing a complaint against a female officer, and so on.

The Police Bureau¡s 1995 sworn strength was 60% white male and 15% minority male, for a total of 75% male; 14% white female and 11% minority female, for a total of 25% female. This compares to 1991 figures of 65.5% white male, 13.9% minority male, 9.6% white female, and 10.5% minority female.* Given these numbers, and assuming deployment policies that would increase the incidence of arrests by male officers, the numbers in the tables above do not suggest any patterns of gender-based police misconduct. Yet, in reviewing them for this purpose, we inadvertently discovered a pronounced trend in another area.

From 1991 to 1992, male complaints against male officers jumped from 78% to 91% (as a percentage of total complaints filed by males). This rose again to 93% in 1993, then to a striking 97% in both 1994 and 1995. In each of the years 1992-1995, the absolute number of male complaints against female officers dropped. In both 1994 and 1995 this number was only four complaints, an all-time low which amounted to only 3% of total complaints by males.

Between 1993 and 1994 the number of complaints filed by all citizens increased from 196 to 280, then rose slightly to 292 in 1995. The respective figures for this gender analysis (due to incomplete¡s, unknown¡s, etc.) are 1993 - 129, 1994 - 205, and 1995 - 205. The figures in the table above show that this substantial rise in citizen complaints, starting in 1994, occurred with twice the frequency among male citizens, and that it can be attributed, exclusively, to male police officers. When read in conjunction with the remarks on page 25 above, the evidence mounts for the assertion that a growing number and increasing percentage of these male police officers, were, particularly in 1995, white, male police officers. This statement, in turn, must be tempered by the finding that this increase in complaints against white, male officers came from both white, male complainants and black, male complainants, in roughly equal numbers.

III. c. Age of Citizen Complainant

Of the 1,674 citizen complaints in our modified database, 423 did not have an entry for age of complainant, leaving 1,251 for this portion of the analysis. The following table gives the year-by-year average age for citizens who filed complaints against police officers. We also include the number of citizen complaints in that year which did have an entry for age of complainant; and the total number of citizen complaints for each year.

Again, it should be noted that 1996 figures are for the first two months only. Aside from 1996, the average age of citizen complainants remains fairly constant over the audit period. We ran this particular report to see if a lowering trend appeared in the most recent three or four years. This might suggest an increase in measures taken by police officers to address gang-related crime, leading to a higher percentage of younger arrestees. The table above provides no support for this hypothesis.

Yet, even if trends did appear, any conclusions drawn from this data would have to be qualified, due to the disappointingly high number of incomplete entries in the age of complainantŒ field. This problem, evident in every year, became exceptionally pronounced in 1995. The year before, 236 citizen complaints were filed but 54 (23%)them did not have a usable entry for age of complainant. The same number of citizen complaints, 236, were filed in 1995, and this time 89 (38%) of them were incomplete in this category. These are unacceptably high levels of data entry omissions or errors. In the racial and gender analysis above, satisfactory explanations existed for unusable records, namely that the police officer was frequently listed as homicideŒ, detectiveŒ, or simply unknownŒ by the complainant. The same excuse cannot be made for age of complainant.

RECOMMENDATION NO. 5:

OMI must strive for complete entries and higher accuracy levels at both the intake and data entry stages, in the field of age of complainantŒ and all other fields.

III. d. Experience Level of Police Officers

Of the 1,674 citizen complaints in our modified database, 168 had no entry for complaint date, or starting date of the police officer, or both, leaving 1,506 for this portion of the analysis. The following table gives yearly averages for experience level of the police officers, in years. As in the age of complainant table above, we also include the number of citizen complaints in each year which did have entries for both complaint date and starting date of the officer; and the total number of citizen complaints for each year.

These data show a pronounced drop, beginning in 1993, in the average experience level of police officers who had citizen complaints filed against them. They lend support to the view that the upswing in citizen complaints which began in 1994, was caused, in part, by an influx of newly hired police officers. From January 7, 1991 to February 15, 1993 the City of Pittsburgh did not hire a single police officer. Then, beginning in the closing months of 1993 and continuing over the next two years, over 40% of the police force retired. Thus, younger, new recruits came into the ranks of the police in unprecedented amounts during 1994 and 1995. The factors which led to this are discussed in detail under Objective No. 6 (see page 46). Here it remains to be said that lack of experience is not a satisfactory excuse for police misconduct. As a partial explanation, however, it is less intractable than racism, and one which bodes better for the future.

RECOMMENDATION NO. 6:

Public Safety Management, the Law Department, OMI (or the ultimate successor of OPS with respect to citizen complaints against police officers), and the City Controller¡s Office should design an on-going monitoring process, incorporating the key indicators of this section¡s racial, gender and experience level analysis. The underlying assumption of this monitoring scheme would be that appropriate training, counseling, and disciplinary measures should, over time, result in, 1) lower percentages of complaints filed by all citizens against white, male officers, 2) lower percentages of complaints filed by black citizens against black officers, and 3) a moderating of the sharp drop in average experience level of those officers against whom citizen complaints are filed.

IV. Multiple Citizen Complaints Against an Officer (Objective No. 4)

Of the 1,674 citizen complaints gleaned from the database, 152 had to be eliminated from this portion of the analysis. In these cases the officer¡s name was either left blank, given as unknownŒ, or as homicideŒ, detectiveŒ, or some other unhelpful designation. In the great majority of these cases, the citizen filing the complaint did not know the name of the officer and subsequent efforts by OPS staff did not identify the officer. This left 1,522 citizen complaints. Over the 10 year audit period these 1,522 complaints were filed against 678 individual police officers. The following table gives the breakdown on all 678 police officers:

From virtually every conceivable vantage point, these figures support the assertion that a few bad applesŒ are causing a disproportionate amount of the problem.

First of all, the numbers should be placed in perspective. The size of the police force has fluctuated over the ten year audit period. The Police Bureau¡s 1995 sworn strengthŒ figure is 1,170. But actually a smaller number of police officers would be susceptible to citizen complaints due to disability, desk duty with no interaction with the public, etc. One would have to do a similar analysis for each year. A very conservative yearly average would be 800. In other words, we assume for purposes of discussion that for each year of the audit period there were about 800 police officers whose duties could subject them to the possibility of having citizen complaints filed against them. If each of these 800 officers would have had only one citizen complaint filed against them each year, it would lead to a grand total of 8,000 citizen complaints filed over the course of the audit period. This is almost five times higher than the actual number of 1,674, and higher than the number we are forced to use for this part of the analysis: 1,522.

If we start with our hypothetical 800 police officers in 1986 and make a further assumption of a 10% yearly turnover rate (through retirements, terminations and new hires) we arrive at an eligible pool of 1,600 individual police officers who, during the 10 year audit period, could have had a citizen complaint filed against them. In fact, 678 did; or 42%. That leaves 922 eligibleŒ police officers (58%) without a single citizen complaint against them.

Furthermore, of the 678 officers who did have citizen complaints filed against them, over half (359, or 53%) had only one in the entire 10 years. 140 officers had only two citizen complaints for the entire 10 years. 90 had three; 30 had four; 49 had between five and 10; and 10 police officers had more than 10 citizen complaints filed against them in the 10 year audit period. Three officers had 21, 28, and 34 citizen complaints filed against them respectively. This is a total of 83 or 5% of the working total of 1,522 citizen complaints; filed against just three police officers out of 1,600 possible candidates. The 10 officers with more than 10 complaints each, accounted for 171 out of 1,522 complaints, or 11%. 59 officers garnered five or more complaints in ten years adding up to 493 out of 1,522 citizen complaints or 32%.

Thus, roughly one-third (493 out of 1,522, or 32%) of the citizen complaints for which we can identify an officer¡s name, are attributable to 59 police officers, or a mere 3.6% of the eligible pool of 1,600 officers over the ten year audit period.

{The following recommendations should not be interpreted as judgments, by the auditors, of any individual police officer¡s culpability in a particular case, nor as comments on any individual police officer¡s fitness to serve.}

RECOMMENDATION NO. 7:

OMI, Public Safety Management and the Law Department should conduct a detailed review of the three police officers with 21, 28, and 34 citizen complaints, as a way of highlighting the police disciplinary system at its most questionable level of performance. The review (like this audit for the bureau as a whole) should focus on the types of misconduct alleged, the dispositions of the complaints by OPS, the chain of command¡s recommended discipline, the implemented discipline, whether anything in the officers¡ deployment schedules could partially explain the high incidence of complaints, whether any counseling or re-training was recommended and/or received, whether any arbitration proceedings were held, and their results. This review should seek to answer the question: How, after 33 citizen complaints, could a police officer possibly be in a position to attract a 34th?Œ

RECOMMENDATION NO. 8:

Public Safety Management and the Law Department should strive, in the next round of negotiations with the police union, to add a provision to the collective bargaining agreement that would address the issue of police officers with multiple citizen complaints. We suggest a red flagŒ approach, under which any police officer who receives a certain number of citizen complaints over a certain time period (for example, two complaints in one year) would be required to take some mix of counseling and training. More frequent complaints would lead to additional measures such as re- assignment or suspension. The emphasis at this stage should not be on determining culpability, but rather on increasing the awareness and effectiveness of the officer. For this reason, these measures should be taken regardless of the disposition of the complaints by OMI.

We next took the 59 police officers, against whom five or more citizen complaints had been filed over the ten year audit period, and examined the disposition of those cases. Out of the 493 cases, 55 did not have a disposition, the majority of these 55 being pending.Œ This left 438 cases. The dispositions for these 59 officers are summarized in the following table:

These rates for known dispositions on these 59 officers correspond closely with the overall rates for all 1,674 citizen complaints: Overall citizen complaint (OCC) sustained, 12% vs. bad appleŒ (BA) sustained, 14%; OCC not sustained, 46% vs. BA not sustained, 50%; OCC closed, 19% vs. BA closed, 14%; OCC exonerated, 11% vs. BA exonerated, 11%. Perhaps most interesting in this regard is that citizen complaints against the bad applesŒ were disposed of as not sustained at a higher rate than citizen complaints as a whole (50% as compared to 46%).

We found that nearly one third of these officers (19 out of 59, 32.2%) did not have a single complaint sustained against them. There is no correlation between the number of complaints filed against an officer and the number ultimately sustained. Officers with the greatest numbers did not have the highest sustained rates. For example, Officer Alpha, with 21 citizen complaints had only three sustained. Officer Beta, with 28 complaints had seven sustained. Officer Gamma, with 34 citizen complaints (the highest amount in the entire audit period) had only five sustained. On the other hand, Officer Delta, with four complaints had three sustained.

In Objective No. 3 we sorted the citizen complaints by race of officer vs. race of complainant (see pages 23-24). We now conducted a similar analysis on these 59 officers. This group included 29 white officers and 29 black officers with the 59th officer alternately identified as white and Hispanic. For 41 (69%) of these officers (19 white and 22 black) the majority of the citizens filing complaints against them were of the same race. This is true of the three officers (one who is white and two who are black) with the highest number of complaints (21, 28, and 34). For each of these three, the majority of their complainants were of the same race as the officer.

As we said with respect to the racial analysis in Objective No. 3, here again the data do not support the notion of systemic, racially- motivated police misconduct. Pittsburgh police, even the biggest offenders, are, apparently, dispensing their misconduct with an even hand. Statistics, however, can tell us nothing about a particular case. It would be equally incorrect to conclude from this data that racial animosity never figured into any of the 1,674 citizen complaints.

The above analysis can present a good overall picture of the multiple-complaint issue, but it is of limited value for assessing any changes over time. This is because various reasons can account for a low number of complaints filed against an officer. It may be because the officer is doing an excellent job. But it might also be because the officer was recently hired. To identify trends in this area over time, we compiled a listing of all officers against whom two or more citizen complaints had been filed in each of the years 1991 - 1995. This information is summarized in the following table. The second column gives the number of officers who had two or more complaints for the given year. The third column lists officers in the second column with two or more complaints filed against them in one of the other years between 1991 and 1995.

The number of police officers against whom multiple citizen complaints were filed, in each year, more than doubled between 1991 and 1994; from 16 to 35.

Focusing on column three above, provides further evidence for the bad appleŒ theory. The total of 44 is, in reality, just 20 police officers. 18 of them had two or more complaints filed against them in two of the years from 1991-1995; and the other two officers had two or more complaints in four out of those five years.

V. Evaluation of Disciplinary Measures (Objective No. 5)

This section of the audit focused on citizen and management- initiated complaints, sustained by OPS for the years 1991-1995. Summary data for these years are reprinted below from the table on page 18. Asterisks indicate that the dispositions do not add up to the total complaints filed for that year, due to pendingŒ cases. The reader should review the tables on pages 17-18 and the accompanying remarks on the differences in disposition rates between citizen complaints and management-initiated complaints. The sustained cases for 1991-1995 add up to 155.

The rest of our findings with respect to disciplinary measures do not lend themselves well to a tabular presentation. Each case almost had to be considered separately. A great number and variety of things could happen once an OPS finding of sustained began to wind its way up each step in the chain of command. Because of this, none of the figures given below balance or tally into recognizable sums. And so, the remainder of this section presents the findings in a narrative style.

V. a. Overview

Of those 155 sustained cases in the OPS database for 1991-1995, we were able to examine the actual OPS hard copyŒ files for 154; 73 were management-initiated and 81 were filed by citizens. Based on our examinations and on discussions with the person who served as the Civilian Coordinator of OPS, we were able to say with certainty that discipline had been given in 62 cases. Of these 62 cases in which discipline had been given, we found 38 oral reprimands and 24 written reprimands.

In addition to those 62 cases, we found these final recommended disciplines in the following cases: Nine one-day suspensions; five three-day suspensions; five suspensions greater than three days; and 13 terminations. These 13 terminations involved only nine officers (because the City tried to terminate two of the officers three times each). Of those nine officers, two were reinstated by arbitrators; three retired or quit; two were finally terminated; and two are terminated but awaiting a decision on their arbitration appeals. Every one of these terminations involved a management- initiated complaint.

The oral reprimands were given in cases with the following offense codes: 23 for Conduct Unbecoming an Officer; three for Use of Force/Arrest; two for Conduct Towards the Public; two for Failure to Perform/Neglect of Duty; one for Firearm Discharge-No Injury/Warning Shots/Mishandling; one for Abuse of Authority- Self Assigned Investigation; one for Lost, Stolen, Seized/Abandoned Property.

The written reprimands were given in cases with the following offense codes: 13 for Conduct Unbecoming an Officer; two for Theft; two for Firearm Discharge-No Injury/Warning Shots/Mishandling; one for Use of Force/Arrest; one for Lack of Sound Judgement-Obedience to Orders/Laws.

The recommended one-day suspensions were given in cases with the following offense codes: five for Conduct Unbecoming an Officer; one for Theft; one for Conduct Towards the Public; one for Use of Force/Arrest; one for Lost, Stolen, Seized/Abandoned Property.

The recommended three-day suspensions were given in cases with the following offense codes: three for Conduct Unbecoming an Officer; one for Abuse of Authority-Self Assigned Investigation. The recommended suspensions greater than three days were given in three cases for Conduct Unbecoming an Officer.

The recommended terminations (usually, a five-day suspension pending termination) were given in cases with the following offense codes: eight for Conduct Unbecoming an Officer; one for Use of Force/Custody; one for Handling of Property-Prisoner/Personal Property; one for Insubordination/Written Directions; one for Firearm Discharge-No Injury/Warning Shots/Mishandling.

Counseling was required, sometimes on its own and sometimes in addition to the other discipline given, in cases with the following offense codes: two for Use of Force/Custody; one for Use of Drugs/Alcohol; one for Lack of Sound Judgement-Obedience to Orders/Laws; one for Racial, Religious/Ethnic Intimidation;

V. b. Chain of Command

We next looked at modifications and reversals within the chain of command. These came about in the following way. After OPS made a finding of sustained, the case was forwarded to the police officer¡s Zone Commander who decided on an appropriate discipline or rejected the OPS finding. This decision then went to the Assistant Chief of Police who approved it, modified it (made it harsher or easier), or reversed it. Then, once again, the process was repeated with the Chief of Police, and then, finally, with the Director of Public Safety.

We found the following in this regard: In seven cases the Assistant Chief increased the recommended discipline; in three cases the Assistant Chief decreased it. In six cases the Chief of Police increased the recommended discipline; in nine cases the Chief decreased it. In four cases the Director of Public Safety increased the recommended discipline; in five cases the Director decreased it.

In 16 cases the Zone Commander rejected and ignored the OPS finding of sustained (11 times changing the OPS finding from sustained to not sustained and five times changing it to unfounded). In two cases the Director rejected an OPS finding of sustained, changing it to not sustained. To its credit, in all instances where a representative of Public Safety Management sought to change findings of sustained, OPS did not change the sustained finding in its database, nor, to our knowledge, did OPS alter its final report.

RECOMMENDATION NO. 9:

In the future, it should be impermissible for a representative of Public Safety Management to change or attempt to change a finding of OMI. Recommending, accepting, rejecting, and modifying discipline is a legitimate role for management. But changing an OMI disposition of a particular case interferes with OMI¡s role as independent fact-finder and compromises the integrity of the investigation.

RECOMMENDATION NO. 10:

In the future all OMI files should have a complete record which includes the findings and disposition of the fact-finder, all recommended discipline through the chain of command, some entry stating that the discipline has or has not been actually given, and, where applicable, the record of any arbitration proceedings. This would greatly aid Management and future auditors in evaluating disciplinary measures.

V. c. Arbitration

We next looked at sample of 17 arbitration proceedings in which Public Safety Management attempted to discipline police officers (seven of these, as a result of sustained OPS complaints). Five of the 17 are pending or awaiting a decision. We examined the remaining 12. One of these was a recurring case in which the officer won the first time but lost the second time, due to a repetition of the proscribed conduct (drug use). So although there were only 12 cases, there were 13 arbitration decisions. The City won six times and lost seven times.

The City¡s Arbitration Losses

Of the City¡s seven losses, four involved attempts to terminate the officer; two for drug use, one for over-time irregularities, and one for what the City argued was rape, but what the arbitrator (in an uncanny written opinion which tarnishes the reputation of the American Arbitration Association) chose to view as sex on the job. All four police officers were ordered reinstated (full reinstatement with back pay for the over-time case; the drug users received no back pay and had to undergo counseling and random drug testing; the officer in the rape case received no back pay and a probationary period).

Two of the City¡s seven losses involved disputes over whether the officer should be required to return the over-time pay they had received in the alleged infraction. The arbitrators felt the City had failed to produce adequate evidence and ordered that the officers be allowed to keep the money.

The last of the City¡s seven losses involved an attempted three day suspension for excessive force. The arbitrators set aside the suspension and directed that the police officer undergo counseling instead.

The City¡s Arbitration Victories

Of the City¡s six victories, five involved suspensions; one for drug use (unspecified number of days), three for unlawful strip searches (all for one day), and one for over-time irregularities (two days). The last of the six victories was a termination in the recurring case mentioned above (the officer won the first time but was terminated when another instance of drug use occurred).

Victories and losses aside, a number of things can be said about the City¡s arbitration history in regard to this performance audit. The first is that arbitration was infrequently used to resolve matters which began as citizen complaints. Only four out of the 12 arbitration cases we examined (one excessive force and three unlawful strip searches) would have begun as citizen complaints. The rest were management-initiated investigations. This finding parallels our finding on the marked differences in disposition rates between citizen complaints and management-initiated complaints (see pages 17-18). It also makes it difficult to argue that any perceived short comings in the dispensing of discipline, for citizen complaints, by Public Safety Management, are caused by the arbitration system. There simply are not enough arbitration cases to make that argument credible.

The second point we note is that the City tried four times to discipline police officers who were using drugs. Two out of those four times it attempted to terminate the officer and was forced by arbitrators to reinstate the drug user with counselingŒ and random drug testing. One of these officers, as a result of a these random tests, was caught again and successfully fired, which seems to vindicate management¡s initial determination that firing this police officer was appropriate. Once again, we are very concerned that an organization as prestigious as the American Arbitration Association could have recommended to the City and the police union, individuals to serve as arbitrators, who have such a cavalier attitude regarding illegal drug use by police officers sworn to enforce the laws of this City and Commonwealth. Of the 2,494 total OPS complaints in the database for the 10-year audit period, 74 had offense code 26, use of drugs/alcoholŒ (this number includes on- duty and off-duty incidents). This is a serious offense and a serious problem for Public Safety Management.

VI. Additional Statistical Analysis (Objective No. 6)

In this last Objective, we recap some of the more significant findings of this performance audit, then present them against the backdrop of crime statistics and other data on police activity during the audit period.

VI. a. Recap of Significant Findings

I. Out of the total 2,495 complaints filed with OPS from April, 1986 through March 4, 1996, 1,980 were filed against police officers. 306 of these were management-initiated complaints. This leaves a total of 1,674 complaints filed against police officers by citizens.

II. Striking differences appear in disposition rates between citizen and management complaints. Overall, the sustained rate for citizen complaints was 12%, while for management complaints it was 52%. The rate for not sustained citizen complaints was 46%, compared to only 6% not sustained for management complaints. 11% of citizen complaints were exonerated. Less than 1% of management complaints were exonerated.

III. Almost all of the 1996 citizen complaints were still pending at the close of the audit period, March 4, 1996, and that is not unusual. The number of pending 1995 citizen complaints is 82 out of 292 (28%) and this does not seem an unreasonably high number. But the further back one goes the less acceptable the pending cases become: 38 in 1994; 31 in 1993; 26 in 1992; and 10 in 1991.

IV. Of 779 complaints known to have been filed by white citizens, we can identify the race of the accused officer 730 times. Of these, 515 (72%) were filed against white officers; 201 (28%) were filed against black officers. Of 698 complaints known to have been filed by black citizens, we can identify the race of the accused officer 637 times. Of these, 313 (50%) were filed against white officers; 317 (50%) were filed against black officers.

V. With the exception of 1990, for each year from 1986 through 1992, white citizens filed more complaints. Then in 1993, complaints by black citizens outstripped those filed by whites 73 to 49. This trend continued in the next three years: 116 to 72 in 1994; 96 to 78 in 1995; and 10 to 7 in (the first two months of) 1996.

VI. In 1995 complaints by black citizens against white officers reached both their highest absolute number for the whole audit period, 60, and their highest deviation from the 50%-50% average, i.e., 63% against white officers vs. 37% against black officers. A similar jump occurred in complaints filed by white citizens, whose complaints against white officers reached their second highest absolute number, 65 (surpassed only by 1987¡s 75) and their highest deviation from the 72%-28% overall average, i.e., 83% against white officers vs. 13% against black officers.

VII. The fact that black citizens filed slightly over 50% of their complaints against black police officers, tends to negate a systemic racial interpretation. Nevertheless, the data for the last three full years, and in particular 1995, could lend support to the anecdotal perception of a recent deterioration in the relationship between the citizenry (black and white) and white police officers.

VIII. From 1991 to 1992, male complaints against male officers jumped from 78% to 91% (as a percentage of total complaints filed by males). This rose again to 93% in 1993, then to a striking 97% in both 1994 and 1995. In each of the years 1992-1995, the absolute number of male complaints against female officers dropped. In both 1994 and 1995 this number was only four complaints, an all-time low which amounted to only 3% of total complaints by males.

IX. Between 1993 and 1994 the number of complaints filed by all citizens increased from 196 to 280, then rose slightly to 292 in 1995. This substantial rise in citizen complaints, starting in 1994, occurred with twice the frequency among male citizens, and can be attributed, exclusively, to male police officers. When read in conjunction with the racial analysis, the evidence mounts for the assertion that a growing number and increasing percentage of these male police officers, were, particularly in 1995, white, male police officers. This statement, in turn, must be tempered by the finding that this increase in complaints against white, male officers came from both white, male complainants and black, male complainants, in roughly equal numbers.

X. The average experience level of police officers who had citizen complaints filed against them reached a 10-year high of 13.77 years¡ experience in 1988. It then dropped significantly from 1993¡s 9.20 years¡ experience down to 5.79 in 1995 and 4.10 for the first two months of 1996. This lends support to the view that the upswing in citizen complaints which began in 1994, was caused, in part, by an influx of newly hired police officers.

XI. From virtually every conceivable vantage point, our multiple complaint analysis supports the assertion that a few bad applesŒ are causing a disproportionate amount of citizen dissatisfaction with the police. Of the 678 officers who did have citizen complaints filed against them, over half (359, or 53%) had only one in the entire 10 years. 140 officers had only two citizen complaints for the entire 10 years. 90 had three; 30 had four; 49 had between five and 10; and 10 police officers had more than 10 citizen complaints filed against them in the 10 year audit period.

XII. Three officers had 21, 28, and 34 citizen complaints filed against them respectively. This is a total of 83 or 5% of the working total of 1,522 citizen complaints, filed against just three police officers. The 10 officers with more than 10 complaints each, accounted for 171 out of 1,522 complaints, or 11%. 59 officers garnered five or more complaints in 10 years adding up to 493 out of 1,522 citizen complaints or 32%. Thus, roughly one-third (493 out of 1,522, or 32%) of the citizen complaints for which we can identify an officer¡s name, are attributable to 59 police officers, or a mere 3.6% of the eligible pool of 1,600 officers over the ten year audit period.

XIII. The 59 police officers with five or more citizen complaints over the audit period included 29 white officers and 29 black officers. For 41 (69%) of these officers (19 white and 22 black) the majority of the citizens filing complaints against them were of the same race. This is true of the three officers (one who is white and two who are black) with the highest number of complaints (21, 28, and 34). These data do not support the notion of systemic, racially- motivated police misconduct. It would be equally incorrect, however, to conclude from this data that racial animosity never figured into any of the 1,674 citizen complaints.

XIV. The number of police officers against whom multiple citizen complaints were filed, more than doubled between 1991 and 1994; from 16 to 35.

XV. The 155 cases from 1991-1995 in which OPS made a finding of sustained led to 38 actual oral reprimands; 24 actual written reprimands; 13 attempts to terminate an officer (five successful, six unsuccessful, and two more pending an arbitration decision); and the following recommended disciplines: Nine one-day suspensions; five three-day suspensions; five suspensions greater than three days.

VI. b. Pittsburgh Police Bureau Statistics

Reporting Standard Number 3 of the General Accounting Office¡s Yellow BookŒ (Generally Accepted Governmental Auditing Standards, revised 1994) indicates that a performance audit report should contain a description of any significant noteworthy accomplishmentsŒ of the auditee. The following bar graphs demonstrate such a significant and noteworthy accomplishment of the Department of Public Safety, particularly police officers and Police Bureau Management.

In 1995 Pittsburgh recorded its lowest crime rate since 1966. The number of occurrences for every Part IŒ offense declined compared to 1994: Criminal Homicides down 6%, Forcible Rapes down 7%, Robberies down 16%, Aggravated Assaults down 16%, Burglaries down 17%, Larcenies down 10%, Vehicle Thefts down 37%, Arsons down 22%.

Multiple-year comparisons in certain categories show even more dramatic drops. Vehicle Thefts dropped 58% since 1992, from 8,124 thefts to 3,387. In other words, almost 5,000 less cars were reported stolen in 1995 than in 1992. Robberies dropped 31% since 1992, from 2,989 robberies to 2,077. Aggravated Assaults dropped 51% since 1990, from 2,237 incidents to 1,096.

Of course, several things can combine to produce decreases of this nature, including age patterns, population shifts, changing economic and social conditions, etc. But even after taking these into account, a significant portion of the drop in crime can still be attributed to the activity of police officers, the Bureau of Police, Public Safety Management, and the Administration.

From Public Safety Management we obtained arrest figures for the years 1988 through 1995. In the following table we calculated citizen complaints as a percentage of arrests for each of these years.

By way of preface, we note that over these eight years citizen complaints, as a percentage of arrests, averaged less than 1% (0.84%). However, our primary interest here was to see if the rate of alleged police misconduct remained the same in those years (1994 and 1995) when the absolute number of citizen complaints rose significantly. It did not. In fact, it rose from 0.79% in 1993 to 1.14% in 1994 and dropped slightly to 1.13% in 1995. This data with respect to arrests would not support the claim that citizen complaints rose because the number of arrests rose. Arrests, it is true, did rise in 1994 from 20,299 to 20,688; and again in 1995 to 20,829. But, as the percentages show, this increase was not enough to explain the increase in citizen complaints from 162 to 236.

Also from Public Safety Management we received figures on dispatch callsŒ for the years 1991 through 1995. A dispatch call is a communication with the police bureau which results in a police vehicle being sent to the incident. These, like arrests, are another indicator of the level of police interaction with the public. The following table presents citizen complaints as a percentage of dispatch calls for this period.

Unlike arrests, the absolute number of dispatch calls did not increase from 1993 to 1994; nor in the next year. This, combined with the increase in citizen complaints accounts for the increased rates in the fourth column for 1994 and 1995. As we said above in the case of arrests, this data does not support the assertion that the increase in citizen complaints was due to increased police contact with the citizenry. It does, however, reflect favorably upon the police bureau when one considers that, even at its highest rate, about seven citizen complaints are filed for every 10,000 police vehicle responses to a dispatch call; less than one in 1,000.

In addition to arrests and dispatch calls, this analysis could be done with other indicators of police contact with the public, for example, traffic stops.

VI. c. The Police Hiring Crisis Revisited

No discussion of the work of the Bureau of Police in the last five years would be complete without taking into account the unique challenge faced by the Murphy Administration since its first day in office, January, 1994. In that month the City Controller¡s Office released our Performance Audit of Police Officer Hiring for the years 1991-1993. We quote at length from the Conclusion to that report:

Throughout the audit period, 1991-1993, with the exception of one quarter (third quarter, 1993), the Masloff Administration failed to meet the budgeted figure for police officer positions. The variances ranged from 11 below budget in first quarter 1991 to 89 below budget in fourth quarter 1992.

From January 7, 1991 to February 15, 1993, a period of 25 months, the City did not hire a single police officer. The City did hire 243 police officers in 1993.

The corollary to this 25-month hiring freeze was a 14-month freeze on the 75% early retirement incentive program. Police officers retiring under this program finally started to leave the force in November, 1993, the last two months of the Masloff Administration. As of the close of 1993, approximately 390 officers are still waiting to retire under this program, roughly 32% of the force; and all these must be allowed to leave by December 31, 1995.

In September, 1992, when early retirements could have begun, the opportunity existed to stretch this exodus of seasoned officers (at that time, 453, approximately 40% of the force) over a period of three years, four months (September, 1992 - December, 1995). Now, because of the actions of the Masloff Administration, the entire problem has been shifted to the incoming Murphy Administration, and the time span has been shortened from 40 months to 24 months; 24 months in which nearly 400 experienced police officers (not to mention those reaching mandatory retirement age plus all other ‰leaves¡) will have to be replaced by new officers.Œ [emphasis added].

CONCLUSION

Unquestionably the in-coming Murphy Administration faced a public safety crisis in early 1994, and met the challenge. Furthermore, a credible argument can be made that the new police officers (the majority of whom, have not had a single citizen complaint filed against them) deserve a significant portion of the credit for the recent drop in crime. Yet, this aggressive drive to take back the streetsŒ has coincided with an upswing in citizen discontent with the police and created a new, and perhaps even more difficult, challenge.

The Administration, Public Safety Management, the Law Department, the Fraternal Order of Police, and every sworn police officer must now work together to win back the trust of ordinary citizens. They must, in a sincere and convincing fashion, reject the notion that increased citizen discontent is an inevitable by-product of better crime fighting. On the contrary, they must demonstrate by their actions that the opposite is true; that courtesy, professionalism, cooperation with the community, respect for rights, and seasoned restraint, are indispensable components of any successful effort to guard the public¡s safety. The majority of the police in this City already know this. Many practice it every day. It is the sincere hope of the City Controller¡s Office that this audit will enable them to do their jobs even better. We also hope it will help provide Public Safety Management and the citizens of Pittsburgh with the best possible system to receive, investigate, monitor, render decisions on, mete out discipline for, and reduce the frequency of, citizen complaints against the police.

* See, Commonwealth of Pennsylvania v. Peter Flaherty, 760 F. Supp 472, 478 (W.D. Pa. 1991).