CJA
Research Brief No. 6
NEW YORK CITY CRIMINAL JUSTICE AGENCY, Inc.
August 2004
A series of reports summarizing current research from CJA Executive Director, Jerome E. McElroy Director, Research Dept., Richard R. Peterson, Ph.D. Research Brief Editor, Mary T. Phillips, Ph.D. Graphics & Production, Raymond Caligiure Administrative Associate, Bernice Linen-Reed CJA is a not-for-proďŹ t corporation that provides a variety of criminal justice services under a contract with the City of New York. CJA staff interview defendants arrested in New York City, make recommendations for pretrial release, and notify released defendants of upcoming court dates. Within the Agency, the Research Department conducts studies covering a broad array of criminal justice policy concerns. The Research Brief series summarizes the results of some of these studies. New York City Criminal Justice Agency 52 Duane Street New York, NY 10007 PHONE: 646 213-2500 FAX: 646 213-2650 WEB: www.nycja.org Š 2004 NYC Criminal Justice Agency
Release and Bail Decisions in New York City By Mary T. Phillips, Ph.D. Arraignment judges have considerable discretion in deciding whether to release a defendant, or to ďŹ x bail in any amount. What factors inuence their decisions? This Brief summarizes recent research using courtroom observations, supLENGTH OF RESIDENCE IN COMMUNITY plemented by CJA’s computerFAMILY TIES ized database, to analyze the STRENGTH EMPLOYMENT OF EVIDENCE factors affecting release and bail decisions in two borPROBABLE SENTENCE oughs of New York City. CRIMINAL RECORD Figure 1 illustrates the 2)3+ /& &,)'(4 factors that judges in New York State are required by CHARACTER 0RIOR &4! law to consider in order to MENTAL CONDITION secure the defendant’s future appearance in court. However, as this research shows, the illustration bears only a passing resemblance to 2ELEASE "AIL $ECISION reality.
Figure 1 This Research Brief is adapted from: Factors Inuencing Release and Bail Decisions in New York City (July 2004), Parts 1, 2, and 3. The full reports are available on the CJA web site. www.nycja.org/research/research.htm Address comments to the author at mphillips@nycja.org Research project staff: Raymond Caligiure, Taehyon Kim, Elyse J. Revere, Elizabeth Walton. Systems programming: Wayne Nehwadowich.
CJA
Research Brief Introduction
After arrest in New York, an adult’s first court appearance is the arraignment in Criminal Court. There the defendant is informed of the charges and—unless the case is immediately disposed— the judge decides whether to release on recognizance (ROR), set bail, or jail the person without bail pending the outcome of the case. The court’s decision on release and bail at arraignment is one of the most important decisions affecting a defendant facing prosecution. The individual’s right to liberty before being convicted of any crime must be weighed against the risk that the accused might flee before disposition of the case. New York State law enumerates the factors that are to be considered by the courts in assessing a defendant’s risk of flight (Figure 1, previous page). They include prior failures to appear (FTA) and other aspects of criminal record; the strength of the evidence and likelihood of conviction; the probable sentence, if convicted; the defendant’s character and mental condition; employment and financial resources; and, finally, family ties and length of residence in the community. However, it is left to the judge’s discretion to decide what weight to give each factor and whether a given set of circumstances should result in release or bail; and if bail, in what amount.
The data used in the analyses were collected through observations of approximately 2,000 arraignments of non-disposed cases in Brooklyn and Manhattan between September 2002 and March 2003. A total of 15 judges in Brooklyn and 17 in Manhattan were included in the research. To help judges make informed release decisions, CJA provides a recommendation based on information collected in pre-arraignment interviews. The ROR recommendation system was developed from research identifying objectively measured factors that are strong predictors of FTA. At the time of this research, the CJA recommendation was based solely on community ties, but it has since been revised to include other factors as well. CJA does not make a recommendation as to the amount of bail. In addressing the question of what factors influence release and bail decisions, the research also analyzed the role and relative influence of the CJA recommendation. Prosecutors and defense attorneys attempt to influence the decision as well. Judges routinely hear the prosecutor’s bail request first, followed by the defense attorney’s arguments for ROR or lower bail. Presumably, the factors cited by both attorneys to justify their requests are those that they believe to be relevant to the outcome (Figure 2).
Figure 2 Prosecution Arguments
Defense Arguments
Percent of cases in which the prosecutor used each argument
Percent of cases in which the defense used each argument
Strength of case
Criminal record Effect on victim
Flight risk Weapon
Seriousness of case Prior felony conviction Open case
Employed/in school Noninjury to victim Brooklyn N=666 Manhattan N=728
0% 10% 20% 30% 40% 50% 60% 70% 80% Percentages are based on non-disposed cases in which the prosecutor requested bail.
2
Nonserious record
Weakness of case
Prior FTA
No flight risk
Poor health
Not facing jail
Cooperated with police
Lack of resources
Brooklyn N=521 Manhattan N=600
0% 10% 20% 30% 40% 50% 60% 70% 80% Percentages are based on non-disposed cases in which the defense requested ROR.
August 2004
CJA As Figure 2 shows, prosecutors justified their bail requests most frequently in terms of the strength of the case, which pertains to the likelihood of conviction; the defendant’s criminal record; and the effect of the crime on the victim. In Manhattan, prosecutors talked less about the victim and more about the defendant’s flight risk and criminal record, especially prior FTA. The presence of a weapon and the seriousness of the case were also sometimes cited. A prior felony conviction and an open case were the specific aspects of criminal record mentioned most often (after prior FTA).
Research Brief
When defense attorneys requested ROR, two thirds of the time they asserted that their clients did not pose a flight risk. Almost as often they argued that the case was weak. They frequently minimized the seriousness of their clients’ criminal records (or pointed out that they had none) or cited their employment or school enrollment. Occasionally the defense argued that the victim was not hurt, or asked for release on the basis of a client’s poor health, cooperation with the police, or inability to make bail. Sometimes the defense also made the point that a defendant was not facing jail time.
Factors Influencing the ROR Decision
However, this research identified something even more consequential to ROR than charge severity: the prosecutor’s bail request. If the prosecutor consented to ROR, the judge nearly always released the defendant (Figure 4). If the prosecutor requested bail, the dollar amount had a strong effect on ROR. The higher the requested bail, the lower the likelihood of ROR. The defendant was released in over half of the cases in which the amount of bail requested was below $2,500. Very few were released when the request was $10,000 or more, and the ROR rate dropped to zero at requests of $50,000 or more. (Data for Brooklyn and Manhattan are combined in Figure 4 because the results were nearly identical, except that ROR rates in Manhattan tended to be higher.) Research Brief #6
Figure 3 ROR by Charge Severity
"ROOKLYN
-ANHATTAN
Percent ROR’d
.ON FELONY
&ELONY
N=657 N=511
N=342 N=488
Charge Severity at Arraignment
Figure 4 ROR by Prosecutor’s Bail Request
Percent ROR’d
Prior research has shown that charge severity is strongly related to the ROR decision. The severity of the charge was not one of the factors mentioned most often by either prosecution or defense—perhaps it was too obvious to mention— but it was acknowledged indirectly by a small proportion of prosecutors (those who cited the seriousness of the case) and defense attorneys (those who pointed out that a client would not go to jail, even if convicted). Figure 3 shows that charge severity did indeed strongly affect the likelihood of ROR: 67% of non-felony defendants in Brooklyn were released at arraignment, compared to only 36% of defendants facing a felony charge. In Manhattan, comparable ROR rates were 77% and 42%.
#ONSENT UNDER TO 2/2 . . . . . . .
Prosecutor’s Bail Request
3
CJA
Research Brief
Multivariate Analyses of ROR Although the prosecutor’s bail request pro- Some of the components of the recommendation— vides the best single clue to a defendant’s chances employment or residence in the NYC area—were for ROR, some other factors also had an additional also significantly associated with ROR, sometimes impact in multivariate analyses. Charge severity is more strongly than the recommendation itself. not among them because its effects were complete►Judicial variability. Some judges in both ly absorbed by the prosecutor’s bail request—itself boroughs were considerably more or less likely than based largely on charge severity. (We will examine others to order ROR, regardless of the particulars shortly what else may be reflected in the prosecu- of the case or the defendant’s criminal history. An tor’s request.) The following case and defendant exception was found among defendants arrested for charactistics made a difference independently of the first time: for them, decisions were more unithe effect of the prosecutor’s request: form and the identity of the judge was not a factor. ►Criminal history. Criminal history, espe►Seriousness of the case. The presence of cially a record of FTA, had the strongest effect on a weapon, a greater number of charges, and the ROR after the prosecutor’s bail request. A prior prosecutor’s argument that the case was serious FTA, another open case, or a prior misdemeanor were all associated with lower ROR rates in one conviction all reduced a defendant’s chances for or more analyses. Together, these findings indicate ROR, whereas defendants who had never been ar- that the seriousness of the case was a consideration rested before could generally expect to be released. for judges beyond the extent to which it was also a Other significant, but less important, factors: basis for the prosecutor’s bail request. ►CJA recommendation. Defendants who ►Demographics. Females were more likely were recommended for ROR were more likely to than males to be ROR’d in Brooklyn, especially be released than those who were not recommended. when it was the defendant’s first arrest. The factors reported here and on the next page under “Multivariate Analyses” summarize the most important findings from many statistical models of ROR and bail amount presented in the full reports. These factors were statistically significant (meaning that the finding was unlikely to be by chance) while simultaneously accounting for the effects of the other factors in the analysis. Not all of the items were signficant in every model.
Factors Influencing Bail Amount
4
Figure 5 Median Bail Amount by Charge Severity
"ROOKLYN -ANHATTAN
Median Bail Amount
When bail was set at arraignment, the dollar amount was strongly influenced by charge severity. In this respect the findings were similar to those for the ROR decision. The median bail amounts for non-felony and felony cases are given in Figure 5, separately for Brooklyn and Manhattan. (The median is the amount with an equal number of cases above and below.) Bail was much higher for felony cases in both boroughs. In Brooklyn, the median bail amount was $4,000 higher for felonies ($5,000) than for non-felonies ($1,000). In Manhattan, the median bail was $3,000 higher for felonies ($3,500) than for non-felonies ($500). Bail was higher in Brooklyn within each severity level, but the borough made less of a difference than did the severity of the charge.
.ON FELONY . .
&ELONY . .
Charge Severity at Arraignment
August 2004
CJA In another similarity to results for the ROR decision, the prosecutor’s bail request also strongly affected bail amount. Figure 6 shows the correspondence between the amount requested by the prosecutor and the amount set by the judge, combining results for Brooklyn and Manhattan. (Only cases with bail set at arraignment are included.) When the prosecutor consented to ROR or requested bail below $1000, the median amount set was $500—if the judge fixed bail at all. (Most in the lowest request categories were ROR’d.) Within
Research Brief
each range of prosecutor’s requests, the median bail set was far less than the lowest request. As the request rose, however, so did the bail amount. When the prosecutor requested between $50,000 and $499,999, the median amount set was $28,000. Prosecutors sometimes requested $500,000 or more, or remand without bail. For those few (16) cases, the median bail amount was $250,000. This category at the extreme high end is not shown in Figure 6 because the scale would dwarf the other bars in the graph.
Figure 6 Median Bail Amount by Prosecutor’s Bail Request Median Bail Amount
$30,000
$28,000
$25,000 $20,000 $15,000 $10,000
$10,000 $5,000
$5,000
$2,500
$500
$500
$750
$1,000
Consent to ROR
under $1,000
$1,000$2,499
$2,500$4,999
$5,000$7,499
N=75
N=164
N=81
N=92
$3,000
$3,250
0 N=4
$7,500- $10,000- $15,000- $25,000- $50,000$9,999 $14,999 $24,999 $49,999 $499,999
N=73
N=92
N=83
N=31
N=60
Prosecutor’s Bail Request
Multivariate Analyses of Bail Amount The prosecutor’s bail request overshadowed affect this aspect of the decision because CJA does everything else in multivariate analyses of bail not make any recommendation for bail amount. amount. Charge severity was not significant in most There were a couple of additional factors that multivariate analyses because—as mentioned in the independently affected bail amount: ►Judicial variability. Some judges in each discussion of the ROR decision—it overlapped with the prosecutor’s bail request too much to have an borough set bail much higher or lower than other independent effect. When the prosecutor’s request judges, for the same types of cases, regardless of was omitted from analyses, charge severity appeared the prosecutor’s bail request. The most variation to be the most powerful predictor of bail amount. It among judges was found when the defendant had a was not an adequate substitute for the prosecutor’s criminal record. request, though. A much better prediction of the bail ►Seriousness of the case. The greater the amount was possible from the prosecutor’s request number of Violent Felony Offense (VFO) arrest than from the severity of the charge. charges, the higher the bail in Manhattan. The prosThe amount of bail was not also influenced by ecutor’s mention of a weapon or the seriousness of other factors that had an effect on ROR: criminal the case were also factors associated with higher history, the CJA recommendation, or demograph- bail, primarily in Manhattan. These had a weak but ics. The CJA recommendation was not expected to statistically significant impact. Research Brief #6
5
CJA
Research Brief The Prosecutor’s Bail Request
Figure 7 Type of Bail Request Request remand 1% Consent to ROR 21%
Request bail 78%
Because the prosecutor’s bail request was found to have the strongest influence on both the ROR decision and the amount of bail set at arraignment, further analysis was done to discover what influences the prosecutor. Those factors have an indirect effect on the judge’s decisions as well. Figure 7 shows that prosecutors requested bail in the majority of cases (78%), consented to ROR 21% of the time, and in a few cases requested that the defendant be remanded without bail (1%). These proportions were virtually identical in Brooklyn and Manhattan, so the boroughs were combined in Figure 7.
N=1,796
M u l t i v a r i a t e A n a l y s e s o f P r o s e c u t o r ’s B a i l R e q u e s t In general, the same considerations influenced whether the prosecutor consented to ROR and—when there was no consent to ROR—the amount of bail requested. The factors that were most important were the same in both boroughs; these are listed below. (A few additional factors that were significant but weak predictors in only one borough are not shown.) The research was not as successful in explaining the prosecutor’s requests as in explaining judicial decisions, especially consent to ROR. More than half of the variance in the prosecutor’s consent to ROR was unaccounted for by our analyses. ►Charge severity. The more severe the charge, the less likely was consent to ROR, and the higher the dollar amount of bail requested. This was of overwhelming importance to bail amount; it was one of several factors important to ROR. ►Criminal history. A prior FTA greatly decreased the likelihood of consent to ROR, and
this was even more important than charge severity in Brooklyn. Having an open case also had a significant effect on consent to ROR. A prior felony conviction raised the amount of bail requested. ►Violent offense. A violent offense or weapon lessened the likelihood of consent to ROR and raised the bail request significantly.
Agreement Between Prosecutor, CJA Recommendation, and Judge Figure 8 illustrates the threeshould not be ROR’d (34%), the Figure 8 way agreement between the prosjudge ordered ROR anyway for ROR Rate by ecutor (assistant district attorney, 38% of them. ADA-CJA Agreement or ADA), CJA recommendation, When CJA and the ADA and the judge’s ROR decision. disagreed, the judge tended to ADA+ “ADA+” indicates that the side with the prosecution, but it ADA– CJA+ CJA– prosecutor consented to ROR; made a big difference what the 14% 34% “CJA+” indicates that CJA recADA requested. If the ADA conADA+ 99% ommended ROR. A negative sented to ROR, the judge usually CJA– 38% ROR’d 7% ROR’d sign indicates the opposite. went along with it against CJA’s 91% d R’ RO The ADA and CJA agreed recommendation (91% ROR’d). that a defendant should be ROR’d If the ADA did not consent to ADA– 47% CJA+ only 14% of the time. When that ROR (and CJA recommended it), ROR’d 45% happened, the judge nearly aljudges sided with the ADA just ways ordered ROR (in 99% of over half of the time (47% were those cases). When the ADA and ROR’d). This was the most comN=1,723 CJA agreed that the defendant mon scenario (45% of cases). 6 August 2004
CJA
Research Brief
Conclusions and Discussion Results from this research suggest several ways in which the criteria for actual release and bail decisions differ from those listed in New York Criminal Procedure Law, illustrated in Figure 1. For one thing, the statute makes no distinction between ROR and bail amount. In reality, the considerations for the two aspects of the decision were found to be somewhat different. Criminal history and community ties (including the CJA recommendation) affected ROR but not bail amount. In addition, the most important factor affecting both ROR and bail—namely, the prosecutor’s bail request—is not specifically listed in the statute. Our main conclusion is that prosecutors dominate release and bail decisions. The powerful statistical association between prosecutors’ requests and judges’ decisions certainly suggests this conclusion, but it is also possible that the same factors independently influenced both. If this were the case, though, it should have been possible to predict the judges’ decisions equally well without recourse to the prosecutor’s request—but it was not. It is still conceivable that unknown factors, not available to our analysis, independently affected both. Nevertheless, it seems likely that at the very least, prosecutors influence judicial decisions by organizing the welter of case-related details to emphasize some
and de-emphasize others. In addition, the fact that judges rarely made a decision that was harsher than the prosecutor’s request suggests that prosecutors establish boundaries that judges are reluctant to exceed. This research does not contradict previous findings that charge severity strongly affects release and bail decisions. Rather, it contributes a more nuanced understanding of this relationship by showing that a large part of the effect is indirect, coming through the prosecutor’s bail request. The importance of charge severity in release and bail decisions has often been interpreted to mean that public safety, rather than risk of flight, is the paramount concern. The significance in this research of several other measures of the seriousness of the case, including violent offense and presence of a weapon, also suggests that judges (and prosecutors) are concerned with public safety. However, their reasoning may be that the motivation to flee is greater when a defendant is facing a severe sentence. Thus, it is difficult to disentangle risk of flight, which is a legitimate criterion for release under New York law, and public safety, which is not. Finally, it made a difference which judge heard the case. Some judges released more defendants, and some set higher bail, regardless of case and defendant characteristics.
Policy Implications
If the goal of release and bail decisions is to assure the defendant’s appearance in court, then rational decision making would require that factors known to predict FTA (such as the CJA recommendation) should be be given more weight than other factors (such as the prosecutor’s bail request). The revised recommendation system, which takes criminal history into account as well as community ties, may indeed inspire more confidence than the one in use at the time of this research. If the goal is also to protect the public from dangerous criminals, this should be made explicit and criteria known to predict re-arrest (or some other measure of dangerousness) should also be used in making the decision. However, it is not clear from other research that either the prosecutor’s bail Research Brief #6
request or charge severity predicts dangerousness. Making public safety an explicitly recognized goal of release and bail decisions in New York would require legislative action. Such legislation has been introduced in the State Assembly but has not been enacted to date. Proposals include options for supervisory conditions for release and remand without bail for dangerous criminals. A remand option might decrease the reliance on cash bail for the purpose of detention by providing for explicit preventive detention, with procedural safeguards. Equity in the criminal justice system requires that decisions made by different judges be consistent for similarly situated defendants, yet significant disparities were found. This issue could be addressed through training. 7
Research Brief from
CJA
#6 (August 2004): Release and Bail Decisions in New York City Already published in this series: No. 1 (December 2002): Jail Displacement for ATI Programs by Mary T. Phillips, Ph.D. No. 2 (April 2003): The Impact of Felony ATI Programs On Recidivism by Jukka Savolainen, Ph.D. No. 3 (August 2003): The Impact of Quality-Of-Life Policing by Freda F. Solomon, Ph.D. No. 4 (December 2003): Combating Domestic Violence in New York City, 2001 by Richard R. Peterson, Ph.D. No. 5 (April 2004): CJA’s New Release-Recommendation System by Qudsia Siddiqi, Ph.D.
Send us your e-mail address! E-mail breed@nycja.org and tell us if you would like to be notiďŹ ed when new reports become available on our web site. www.nycja.org/research/ research.htm
Forthcoming: No. 7 in the series (December 2004) will compare the processing of domestic violence cases in Manhattan before and after the establishment of a specialized Domestic Violence court part. By Research Department Director Richard R. Peterson, Ph.D.
Research Brief from
CJA
#6 (August 2004): Release and Bail Decisions in New York City The New York City Criminal Justice Agency 52 Duane Street New York, NY 10007
TO: