Bos. Chapter, Naacp, Inc. v. Beecher , 295 F. Supp. 3d 26 ( 2018 )

SARIS, Chief U.S.D.J.

*28INTRODUCTION

This case involves two consent decrees entered into more than four decades ago governing the hiring procedures for police and fire departments in certain cities in Massachusetts. Plaintiffs and Defendant Massachusetts Human Resources Division ("HRD") have jointly moved to modify the consent decrees controlling entry-level police officers in eight communities¹ and entry-level firefighters in four communities² pursuant to Fed. R. Civ. P. 60(b)(5). While agreeing on several key modifications, the parties disagree on the factors to be considered in defining the "qualified labor pool" and the appropriate end point for the decrees. After a hearing, the Court ALLOWS Plaintiffs' motion to modify in part and ALLOWS Defendants' motion to modify in part. (Dkt. Nos. 44, 46).

BACKGROUND

I. The History

The two consent decrees that have governed the hiring of entry-level police officers and firefighters were entered in the 1970s in two separate cases. In both Beecher (involving firefighters in Springfield) and Castro (involving police officers in Boston), the courts found that the entrance examinations administered by HRD had a racially discriminatory effect and imposed ratios for hiring Black and Hispanic candidates in most cities as the remedy.³

The First Circuit ordered HRD to create hiring certification lists using fixed ratios of minorities (Black and Hispanic) and other applicants, "perhaps one from the priority pool to every one, two or three from the second pool, until the priority pool has been exhausted." Castro v. Beecher, 459 F.2d 725, 737 (1st Cir. 1972). It later described the prescribed procedure that applied to Boston police officers as follows:

The Castro consent decree required HRD to prepare certification lists by creating two groups. The first, "Group A," would "consist of all Black and Spanish-surnamed applicants who pass a future police entrance examination and are otherwise qualified for appointment on the basis of existing requirements." The second, "Group B," would "consist of all other persons who pass a future police entrance examination and are otherwise qualified for appointment on the basis of existing requirements." When an appointing authority, such as the City [of Boston], sought to hire police officers, it would send a request to HRD, which would send the appointing authority a certification list ordered "on the basis of one candidate from Group A for every candidate certified from Group B." In making its hiring decisions, if the appointing authority chose to reject a candidate *29in favor of another candidate who appeared lower on the HRD list, HRD would not approve the appointment unless the appointing authority "furnished [HRD] with a written statement of [its] reasons" for doing so; HRD would then provide a "written statement of those reasons to ... the candidate upon written request."

The Castro decree was to remain in effect for a given city until that city's police department "achieves a complement of minorities commensurate with the percentage of minorities within the community."

Sullivan v. City of Springfield, 561 F.3d 7, 10-11 (1st Cir. 2009). In 1992, the First Circuit rejected the argument that even if the goals of the Beecher decree had not yet been accomplished, the decree was "constitutionally infirm because it [swept] too broadly." Mackin v. City of Boston, 969 F.2d 1273, 1277 (1st Cir. 1992). The First Circuit has thrice reaffirmed-in Mackin in 1992, Quinn v. City of Boston in 2003,⁴ and Sullivan in 2009-that a city would not be released from the consent decrees until "rough parity" was reached.⁵ Id.

Eventually, the consent decrees applied to hiring of entry-level police officers and firefighters in over 100 municipalities in Massachusetts with a minority population of at least one percent, most of which have subsequently been released. The Court released Boston, which was subject to a one-to-one ratio, in 2004. See DeLeo v. City of Boston, No. 03-12538-PBS, 2004 WL 5740819, at *10-11 (D. Mass. Nov. 23, 2004). As of January 2018, the police consent decree remains binding on eight municipalities and the firefighter consent decree on four. All but one are required to use a certification list based on a one-to-three ratio. Springfield is subject to a one-to-one ratio.

II. The Cap

In August 2016, the parties discovered that the consent decrees were actually having the adverse and unintended effect of "capping" minority representation in certain remaining consent-decree cities. The one-to-three ratio effectively created a cap on minority hiring in cities that had a minority population that exceeded 25 percent. As the minority population in these cities increased after the 1970s, the formula, which was supposed to remedy the effects of the racially discriminatory examinations, unfortunately turned out to impede minority hiring.

HRD explains: "Despite census data reflecting Black and Hispanic citizenry exceeding 40% of the total population in all five remaining cities with populations below 100,000, minority exam-passers (while often a majority on these cities' hiring lists) are effectively limited by the decrees to only 25% of certification list slots until one of the racial pools is exhausted." Dkt. No. 32 at 3-4. Due to the "changing demographics" of the consent decree cities and the fact that many more minority candidates are taking and passing the exams, the HRD states that application of the certification ratios contributed to municipalities' inability to make progress towards parity and therefore procure release from the consent decrees. Id. at 4.

To remedy this problem, the parties jointly agree that this Court must modify the decrees to suspend the application of their ratio requirements when they would *30lower the rank order of minority candidates on certification lists. This Court allowed the joint motion to modify the consent decrees on September 14, 2016. Dkt. No. 10. This so-called "safety valve" modification required HRD to suspend the application of the certification ratios otherwise required under the consent decrees when such application would reduce the number of qualified minority candidates who would otherwise appear on a certification list or lower the rank of qualified Black or Hispanic candidates. Since the modifications went into effect, 15 of the 18 certification lists were issued in strict rank order in accordance with the safety valve modifications.

During a September 2016 hearing, the Court ordered the parties to address three subjects: (1) potential remedies for minority candidates who had been passed over due to the terms of the consent decree "cap"; (2) steps to be taken to help the cities reach the goal of parity identified in the consent decrees; and (3) given the age of the consent decrees, whether they should be terminated before "parity" was reached in the remaining consent-decree cities.

LEGAL STANDARD

Fed. R. Civ. P. 60(b) provides "on motion and just terms, the court may relieve a party or its legal representative from a final judgment, order, or proceeding" if "the judgment has been satisfied, released or discharged ... or applying it prospectively is no longer equitable." More broadly, a court may relieve a party from an order for "any other reason that justifies relief." Fed. R. Civ. P. 60(b)(6).

The "flexible standard" in Rule 60(b) applies to consent decrees. Rufo v. Inmates of Suffolk Cty. Jail, 502 U.S. 367, 381, 112 S.Ct. 748, 116 L.Ed.2d 867 (1992). "A party seeking modification of a consent decree may meet its initial burden by showing either a significant change either in factual conditions or in law." Id. at 384, 112 S.Ct. 748. Modifications to the decree are permissible when "changed factual conditions make compliance with the decree[s] substantially more onerous," "unworkable because of unforeseen obstacles, or when enforcement of the decree[s] without modification would be detrimental to the public interest." Id. Consent decrees must be modified if the "obligation placed upon the parties has become impermissible under federal law." Id. at 388, 112 S.Ct. 748.

If the moving party establishes changed circumstances, a court must make modifications "suitably tailored to the changed circumstance." Id. at 391, 112 S.Ct. 748. A court should defer to government administrators who have the primary responsibility for solving the problem of institutional reform when resolving the intricacies of implementing a consent decree modification. Id. at 392, 112 S.Ct. 748 (citations omitted).

DISCUSSION

The parties agree that the consent decrees must be modified in the interest of justice. Commendably, they have worked well together to address the adverse effect of the cap and develop a narrowly tailored remedy. While disagreements remain, at the outset it should be noted that the gulf between the parties' positions is not as great as the briefing might suggest. Both parties agree that to cure the harm created by the ratios, the 25 percent cap on Black and Hispanic candidates should be lifted and the consent decrees should retain guaranteed spots for minorities with remaining slots filled without regard to race; that ties on the certification list should go in favor of minority candidates;

*31that a qualified and properly defined labor pool-rather than the city's general population-should be used to measure parity; that ratios in Lawrence, Chelsea, and Holyoke (which have over 50 percent minority populations) should be changed to a one-to-one ratio; that prior to any conditional offers of employment, the appointing authority must replace the name of any minority who fails to respond or declines the position with the next-ranked minority candidate who is willing to accept; and that the parties should appear annually to update the Court on the progress of the decrees. As such, these modifications were adopted without objection by the Court on December 19, 2017. Dkt. No. 57.

The government identified 55 minority individuals whose candidacies likely were negatively affected by the cap between 2012 and 2015 and gave them priority in future certifications. See Dkt. Nos. 21, 24. However, HRD was unable to identify candidates earlier than 2012 because of a change in computer systems.

What continues to be the subject of dispute between the parties is which factors should be considered for defining the qualified labor pool, and whether the Court should modify the endpoint of "parity" in determining when the consent decrees should terminate. The parties also disagree as to whether this Court should release the Springfield Fire Department from the 1974 firefighter consent decree immediately.

I. Qualified Labor Pool

The parties agree that this Court should modify the benchmark for defining parity. The original consent decrees use general minority (Black and Hispanic) population as the benchmark for determining parity, but the parties agree that a qualified labor pool is a better measure. However, they dispute the methodology for determining the qualified labor pool. Courts have the discretion to "decide what the most appropriate labor pool is and whether the expert's method utilized the best and most appropriate available labor pool information." Malave v. Potter, 320 F.3d 321, 327 (2d Cir. 2003). Imperfections in data do not render the uses of such data fatal, although they may affect the analysis' probativeness. See Bazemore v. Friday, 478 U.S. 385, 400, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986).

The parties jointly agreed to hire Dr. Bryan Ricchetti, a well-qualified labor economist,⁶ to calculate the share of the population of the eight remaining consent-decree cities (Brockton, Lawrence, Lowell, Springfield, Worcester, Chelsea, Holyoke, and Randolph) that is Black or Hispanic, and determine the share of that population eligible for entry-level police officer and firefighter jobs. He "was not asked to provide an ultimate opinion on the qualified labor pool"; rather, he was tasked with determining the possible labor pool under different statutory permutations. Dkt. No. 45 at 15. He provided his services pro bono. Dr. Ricchetti uses the term "qualified labor pool" to refer to "the share of the population that possesses the relevant qualifications for entry-level jobs" as police officers and a firefighters using age, education, legal residency in the United States, and prior criminal record. Ricchetti Report at 3, Dkt. No. 45-1. Due to data shortfalls, Dr. Ricchetti notes "there is a fundamental tradeoff in public data sources between the geographic detail at which the data are collected (city, county, state, etc.) and the number of economic *32and demographic variables included in the data." Id. at 4.

Chapter 31, Section 58 of Massachusetts General Laws sets out the qualifications for being hired as a police officer or firefighter. To be eligible for original appointment to these positions, one must have reached the age of 21, but may not be older than 32.⁷ Mass. Gen. Laws ch. 31, § 58. Police officers must also have graduated from high school, received a high school graduation equivalency certificate, or served in the military for at least three years leaving under honorable conditions. There are no educational prerequisites for appointment to the fire force. Id.

Chapter 140, Section 131(d)(i) prevents prohibited persons from applying for licenses to carry firearms, thereby preventing people who meet the criteria of a prohibited person from being appointed as police officers. Police officer applicants are disqualified if they have ever been convicted of (a) a felony; (b) a misdemeanor punishable by imprisonment for more than two years; (c) a violent crime; (d) a drug-related crime; (e) a weapon or ammunition-related crime; or (f) domestic violence. Id. With some exceptions, applicants to be firefighters are only disqualified from an appointment to the position for one year following conviction of any crime. Mass. Gen. Laws ch. 31, § 50.

The government urges the Court to adopt all of Dr. Ricchetti's calculations, but Plaintiffs challenge his calculations with respect to the qualified labor pool for both police officers and firefighters.

A. Criminal Record

Dr. Ricchetti adjusted the qualified labor pool population to reflect the statutory disqualification based on a criminal record. To determine the qualified labor pool population for police officers, Dr. Ricchetti relies on the 2009 National Longitudinal Survey of Youth ("NLSY") which calculates "the share of the population that has ever been incarcerated by race, age, gender, and education" and compares this nationwide data with Bureau of Justice Statistics inmate surveys from 1997 and 2003 for purposes of robustness. Ricchetti Report at 9. Dr. Ricchetti opines that "incarceration is typically associated with more serious crimes, and is, thus, a reasonable proxy" for someone being disqualified from the labor pool. Id. at 10. He acknowledges that incarceration rates are an "imperfect proxy" because they can be both over-and underinclusive. Id."[I]t is possible that an applicant who was previously incarcerated for a relatively minor offense would not be disqualified from the labor pool. On the flip side, it is also possible that some applicants might have never been incarcerated in their life despite having committed one of the crimes that disqualifies them from the job." Id. Notwithstanding these data limitations, I agree with Dr. Ricchetti that incarceration is typically associated with more serious crimes and is, thus, a reasonable proxy for the likelihood of a person having committed a serious crime that would disqualify her from being a police officer. See Ricchetti Report at 10.

*33Plaintiffs challenge the data on the ground that the incarceration rates are based on national statistics. Dr. Ricchetti relies on the NLSY data, but he concedes that consent-decree-city incarceration rates may differ from national rates. However, he was not "aware of data that could inform the direction of the potential bias, if any." Id. Based on the statistics from the 2016 Massachusetts Sentencing Commission on report, cited by Defendants, the Court is not persuaded that the national incarceration rates are a reasonable basis for evaluating the qualified pool for police officers. The Commission's statistics indicate that 655 Blacks are incarcerated per 100,000 in the population in Massachusetts, compared to 1,627 Blacks per 100,000 in the nation. However, 401 Hispanics are incarcerated per 100,000 in Massachusetts, as opposed to 362 per 100,000 in the nation. Mass. Sent'g Comm'n, Selected Race Statistics 1, 2 (2016), available at http://www.mass.gov/courts/docs/sentencing-commission/selected-race-statistics.pdf. Thus, the combined statewide rate of incarceration of minorities appears to be lower than the nationwide rate. The Court does not understand the precise impact of this Sentencing Commission data on the qualified labor pool, but a supplemental report is necessary to reflect available statewide data or to explain why that data is inapplicable or unreliable. Plaintiffs also contend that Dr. Ricchetti's incarceration analysis is inaccurate because he does not use city-level incarceration data. The Court disagrees that such localized data are necessary because there is no evidence that the data vary from city to city.

Plaintiffs point out the criminal-record disqualification for entry-level firefighters is much narrower than the factors that lead to disqualification of entry-level police officers. Individuals are only prohibited from becoming firefighters if convicted of "any crime" within the last year (with certain exceptions).⁸ Mass. Gen. L. ch. 31, § 50. I agree with Plaintiffs that national data concerning the population of minorities incarcerated is not a fair proxy for the population convicted of any crime in the previous year. It is unclear if the Massachusetts Sentencing Commission has annual conviction statistics that would be helpful. If so, the report may be supplemented.

B. Legal Residency Data

Dr. Ricchetti used publicly available data to calculate people who may "not be U.S. citizens but [are] ... legally eligible to work in the U.S." Ricchetti Report at 10. Plaintiffs argue that legal-residency data should be excluded because data used by Dr. Ricchetti relied on state-level statistics, and because region of birth was improperly used as a proxy for race. For example, Dr. Ricchetti assumes immigrants from Mexico, Central America, and South America are Hispanic. Although imperfect, statistics regarding region of birth are a "reasonable proxy" for race and can be used to determine what share of these *34populations maintain legal residency and work authorization. See Segar v. Smith, 738 F.2d 1249, 1299 (D.C. Cir. 1984).

C. Education Requirements for Firefighters

Plaintiffs challenge the consideration of a high-school diploma or its equivalent as a factor in determining the qualified labor pool for entry-level firefighter jobs. Plaintiffs argue that because the statute provides no education requirement, it should not be used as a factor in determining the qualified labor pool.

Defendants respond that because approximately 98 percent of firefighters have attained at least a high-school diploma, the qualified labor pool should include only members of the community who have reached such an educational attainment. Plaintiffs have the more persuasive argument. The statutory requirements should be used in determining the qualified labor pool.

D. Updated Population Data

Defendants argue that it is appropriate to measure parity based on 2010 census data as opposed to more recent data, given the significant rate at which demographics are shifting in some of the consent-decree cities. The demographic shifts in minority communities, Defendants argue, makes parity a moving target. Plaintiffs argue that parity should be calculated based on 2011-2015 community data as a requirement of the original consent decrees. The Court agrees with Plaintiffs that the qualified labor pool should be calculated based on updated population data. Accordingly, the numbers that are adjusted for 2011-2015 are the appropriate numbers to be used for the qualified labor pool. Mackin, 969 F.2d at 1276 (stating the decree refers to "contemporaneous population figures").

II. Termination of the Consent Decrees

The parties dispute whether the consent decrees should be modified by amending the termination provisions. The government argues that the consent decrees have been in place for more than forty years and consent decrees should not go on in perpetuity. In its view, the consent decrees should run for another five years and then sunset, irrespective of whether the goal of parity has been reached. Relying on the doctrine of stare decisis, Plaintiffs argue in favor of retaining parity as the goal, and advocate for annual monitoring and reporting to keep the Court apprised of progress towards meeting the goals of the consent decrees. They oppose any predetermined cutoff date as arbitrary. They also stress that the goal of parity has been directed by the First Circuit in Castro and reaffirmed in Mackin, Quinn, and Sullivan.

In public institutional reform cases, the Supreme Court has warned that consent decrees should not last forever. See Rufo, 502 U.S. at 392, 112 S.Ct. 748. In deciding whether to terminate a consent decree, a court must consider various factors. First, a court must be satisfied that there is relatively little or no likelihood that the original constitutional violation will be continued when the decree is lifted. Id. (citing Board of Educ. v. Dowell, 498 U.S. 237, 247, 249-50, 111 S.Ct. 630, 112 L.Ed.2d 715 (1991) (in a desegregation case, recognizing the important value of local control) ).

Second, the Court must examine whether the goals of the consent decree have been met. Labor/Cmty. Strategy Ctr. v. Los Angeles Cty. Metro. Transp. Auth., 564 F.3d 1115, 1120 (9th Cir. 2009). "[A]n inquiring court should ask whether the goals of the litigation, as incorporated in the outstanding decree, have been completely *35achieved." Mackin, 969 F.2d at 1275. A court should not modify long-standing goals in a consent decree just because the goals have not been achieved. Ensley Branch, NAACP v. Seibels, 31 F.3d 1548, 1563 (11th Cir. 1994).

Third, federalism concerns instruct the courts that once a constitutional "violation has abated and its pernicious effects have been cured," federal intrusion into local affairs must cease. Mackin, 969 F.2d at 1275-76. District courts enforcing public-law decrees have "broad discretion in determining such matters as whether the objectives of the decree have been substantially achieved." Navarro-Ayala v. Hernandez-Colon, 951 F.2d 1325, 1337 (1st Cir. 1991). "Ordinary contract principles, involving determination of the parties' intent when they entered into the stipulation, are the appropriate interpretative guideposts here." Id. at 1339. Consent decrees "should not operate inviolate in perpetuity." In re Pearson, 990 F.2d 653, 658 (1st Cir. 1993).

Fourth, any interpretation and application of a judicial decree affording race-based relief must survive strict scrutiny. Quinn, 325 F.3d at 28. However, the First Circuit is "sympathetic" to the argument that "communities place more trust in a diverse police force and that the resulting trust reduces crime rates and improves policing." Cotter v. City of Boston, 323 F.3d 160, 172 n.10 (1st Cir. 2003) ("A law enforcement body's need to carry out its mission effectively, with a workforce that appears unbiased, is able to communicate with the public and is respected by the community it serves, may constitute a compelling state interest." (quoting Patrolmen's Benevolent Ass'n of N.Y., Inc. v. City of N.Y., 310 F.3d 43, 52 (2d Cir. 2002) ).

These consent decrees setting racial quotas are among the oldest in the nation, perhaps the oldest. There is no evidence of any lingering effects of the unconstitutional examinations conducted in 1968, 1969, and 1970. The HRD has been operating in good faith, so there is little likelihood that the original constitutional violation will continue. In a vacuum, these factors might suggest the modification and termination of the consent decrees even though parity has not been reached. Over the decades, however, it was not foreseen that due to demographic surges, a strict ratio requirement would prevent cities from reaching the core goal of the consent decrees: parity. In other words, the Court-imposed formula turned out to be not a remedy for the past discrimination but a cause of new discrimination, which does have lingering effects into the present. The Court must decide what modifications are necessary in light of these unforeseen consequences of the Court-imposed ratio.

The party seeking release from a consent decree bears the burden of proof. DeLeo, 2004 WL 5740819, at *6. The government points out that the census data from 2010 indicates that it is unlikely the "cap" harmed certain consent-decree cities (Springfield, Lowell, and Worcester) if the qualified labor pool calculations in Dr. Ricchetti's report are used. This argument is unpersuasive because the benchmark as provided in the consent decrees in effect at the time is general population rather than the qualified labor pool analysis in the proposed modified consent decree. The parties provided these helpful statistics from the census in 2000 and 2010.

*36Brockton Lawrence Lowell Springfield Worcester Chelsea Holyoke Randolph Black and 39.8% 76.1% 23.2% 58.4% 31.1% 68.8%% 50.8% 43.5% Hispanic Population - 2010 Census Data Source: United States Census, 2010 - Percentage of "Black or African American" and "Hispanic or Latino" population in Consent Decree cities. Dkt 45-1 at Exhibit 4.C, line 1. Brockton Lawrence Lowell Springfield Worcester Chelsea Holyoke Randolph Black and 25.73% 64.54% 18.22% 48.22% 22.04% 55.66% 45.1% 24.1% Hispanic Population - 2000 Census Data Source: United States Census, 2000 - Percentage of "Black or African American" and "Hispanic or Latino" population in Consent Decree cities; https://www.census.gov//census2000/states/ma.html.

Based on these census numbers, the Court finds the decrees likely substantially undermined the efforts towards parity in at least two communities, dating back to at least 2000. Lawrence and Chelsea had an overall population that exceeded 50 percent Black and/or Hispanic in the 2000 federal census. It is likely that unidentifiable minority candidates for positions in these cities were treated unfairly as far back as 2000. Sadly, the government states: "In Chelsea, the problem had become so acute a number of years ago that Hispanic candidates consciously chose to identify as Caucasian so as to boost their prospects for a job offer." Dkt. No. 59 at 8.

In Brockton, Holyoke, and Randolph,⁹ the 2010 census shows that the cap likely adversely affected minorities under the consent decrees. In the 2000 census, however, the minority shares of the overall population in Worcester (22.04 percent) and Randolph (24.1 percent) dipped just below 25 percent, but it is unclear when the minority population exceeded 25 percent. In Springfield, which has the one-to-one ratio, the cap adversely affected minorities as of the 2010 census, and likely much earlier, as there was a 48.2 percent minority population in 2000. It is unlikely that Lowell has been harmed at all by the cap since its minority population stayed below 25 percent in both the 2000 and 2010 censuses.

The Court declines to impose a termination date with respect to the consent-decree cities whose harm pre-dated the 2010 census. Hopefully, the consent decrees as modified will result in reaching the goal of parity within the five-year period. However, the Court imposes a presumptive five-year termination date for Lowell, which has likely never been harmed by the cap, as indicated in both the 2000 and 2010 censuses. If the government proves that parity in Lowell is precluded by nondiscriminatory other factors *37(like veterans' preference), it may move for release of Lowell.

ORDER

The Court ALLOWS the motions to modify the consent decrees so that parity is measured against a qualified labor pool, rather than the general minority population (Dkt. Nos. 44, 46). The parties shall submit a refined expert analysis consistent with this opinion within 30 days. The Court defers ruling on releasing the Springfield Fire Department until the revised statistics are submitted. The parties shall confer and submit a proposed modified consent decree within 30 days.

SO ORDERED.

1 Brockton, Chelsea, Holyoke, Lawrence, Lowell, Randolph, Springfield, and Worcester.

2 Chelsea, Holyoke, Lawrence, and Springfield.

3 See Boston Chapter, NAACP, Inc. v. Beecher, 371 F.Supp. 507 (D. Mass. 1974), aff'd, 504 F.2d 1017 (1st Cir. 1974), cert. denied, 421 U.S. 910, 95 S.Ct. 1561, 43 L.Ed.2d 775 (1975) (" Beecher "); Castro v. Beecher, 334 F.Supp. 930, 934 (D. Mass. 1971), aff'd in part, rev'd in part, 459 F.2d 725 (1st Cir. 1972) (upholding the finding that police entrance exams in 1968, 1969, and 1970 were not rationally related to a capacity to do the job, but remanding because the district court's remedy was too narrow).

4 325 F.3d 18, 18-26 (1st Cir. 2003).

5 The consent decrees have been subject to litigation and modification over the decades. The complex history of the entry of the consent decrees and modification of the decrees can be found in Sullivan and Quinn, 325 F.3d at 18-26.

6 Dr. Ricchetti has a Ph.D. in Economics from Cornell University, and has worked for the United States Census Bureau and in consulting at Cornerstone Research. He has also testified several times in federal court as an expert in labor economics and econometrics.

7 There are statutory exceptions made for veterans, who are "allowed to exceed the maximum age provision of this section by the number of years served on active military duty, but in no case shall said candidate for appointment be credited more than four years of active military duty." Mass. Gen. Laws ch. 31, § 58A. Cities can also elect not to adopt this section and have higher age limits. See Donahue v. City of Boston, 264 F.Supp.2d 74, 81-82 (D. Mass. 2003), aff'd, 371 F.3d 7 (1st Cir. 2004) ("Each city, town or district [is] thus able to determine for itself whether it wishe[s] to exclude applicants age thirty-two and over from the certification list from which police officers and firefighters are hired.").

8 "No person ... shall ... be appointed to or employed in any such position within one year after his conviction of any crime except that the appointing authority may ... appoint or employ within such one-year period a person convicted of any of the following offenses: a violation of any provision of chapter ninety relating to motor vehicles which constitutes a misdemeanor or, any other offense for which the sole punishment imposed was (a) a fine of not more than one hundred dollars, (b) a sentence of imprisonment in a jail or house of correction for less than six months, with or without such fine, or (c) a sentence to any other penal institution under which the actual time served was less than six months, with or without such fine. Violations of statutes, ordinances, rules or regulations regulating the parking of motor vehicles shall not constitute offenses for purposes of this section." Mass. Gen. Laws ch. 31, § 50.

9 Randolph has done little hiring in recent years and so it is more difficult to ascertain the extent of any problems there.