Dole v. Haulaway Inc.

DEBEVOISE, District Judge. The Secretary of Labor (sometimes referred to herein as the “government”) filed this action on July 22, 1987 alleging that defendants had since at least July 1984 willfully violated and continued to violate the overtime and record keeping provisions of the Fair Labor Standards Act, 29 U.S.C. § 201 et seq. (hereinafter “the Act”). The complaint seeks to enjoin the defendants prospectively from violating these statutory provisions and restrain defendants from withholding any unpaid overtime compensation found due since July 22, 1984 through the entry of judgment or compliance with the Act, and seeks an equivalent amount of liquidated damages. Defendant, Haulaway Incorporated, (hereinafter “Haulaway”) was incorporated in New Jersey in May 1970 and during the relevant period has been engaged in the trash removal business primarily in the State of New Jersey. Defendant, Joseph Scugoza, has, at all times relevant hereto, been the president of Haulaway; held 100% of the stock of Haulaway; and exercised day-to-day control over the business affairs of the corporate defendant. In this opinion when I refer to the “defendant” I shall mean the corporate defendant unless otherwise indicated. A bench trial commenced on July 18, 1989 and (with a six day interruption) proceeded until July 27 at which time, for *276 reasons which will be described below, it recessed. On August 21 trial resumed and was completed on August 24. This opinion sets forth my findings of fact and conclusions of law. Most of the findings are contained in the critique of the manner in which the court managed the case. It is hoped that the critique may assist other judges avoid the serious mistakes which I made administering this ease. A. Case Management Deficiencies The most critical issue in this and probably all other FLSA overtime cases is the amount of unpaid overtime hours worked by a defendant’s employees. Where, as in this case, the employer has failed to keep adequate records of hours worked, the government of necessity, must devise some reasonable method of ascertaining the overtime hours of affected employees. At no time prior to trial did the government specify the amount of overtime pay it contended was due or its method of calculating that amount. Rather, it prevailed upon a gullible court to include the following provision in the final pretrial order: Many employees will testify to the number of hours worked and the court will make a finding as to their hours. The amount of unpaid overtime compensation due those employees who do not testify will be based on the average number of hours worked found due by the court for the testifying employees. The [Secretary’s] compliance officer will testify to application of the statutory formula for overtime compensation and the method for calculation of overtime due, depending on the court’s finding of hours worked. Thus, the record developed at trial will permit the court to find hours worked and plaintiff will then submit an exhibit listing the amount due each employee for incorporation into the Final Judgment. This sounds most simple. The court merely hears employees’ testimony, decides how many excess hours each worked, averages those results for non-testifying employees, and the Secretary will do all the rest. I first sensed trouble when, immediately prior to trial, I attempted to find a basis for settlement. The government was unable to state a figure which it claimed would be due if its full claim was allowed. It had no figure derived from the evidence it had assembled which it would advance as total overtime pay due. It had no statement of the methodology it intended to apply to determine the number of overtime hours it claimed each employee worked. That ended settlement discussions. The trial commenced on July 18, 1989 and the government called a number of employees to the stand to present evidence on the basis of which overtime hours were to be computed. I shall list some of the factors which, it developed, affected the determination of overtime hours: 1. During most or all of the 1984-1989 period defendant, a trash remover, served the communities of Upper Saddle River, Belleville, Butler and Parsippany-Troy Hills and it also had commercial routes identified as Paterson, Hackensack, Hobo-ken 1 and Hoboken 2. Each route takes a different time to complete; different days on the same route require different times to complete. Most of defendant’s employees were assigned to the same route each day, but some worked on different routes in any given week or in different weeks. This was particularly true with respect to employees who worked in Butler, which had collection service only four days each week, and the employees assigned there, had to be assigned to other routes on the remaining two days in each week. 2. Most employees, both drivers and helpers, usually worked the full six day week — Saturday through the following Friday, with Sundays off. However, some employees worked less than a six day week and, of course, all employees had absences from time to time. 3. Depending on which route was being served, employees reported at defendant’s Hoboken yard at different times in the morning. The reportimg times for the various crews was generally testified to as follows: Upper Saddle River — 4:30-5:30 a.m.; Belleville — 4:00-4:30 a.m.; Butler— *277 4:30-5:00 a.m.; Parsippany-Troy Hills— 4:45-5:00 a.m.; the commercial routes— 5:00 a.m. 4. After collecting all the trash on a given route, it was necessary to take the trash to a landfill, or during later periods to a transfer station. The landfills or transfer stations serving the residential communities were: Upper Saddle River-Bergen County Landfill; Belleville-North Arlington or Kearny Landfill/Newark Transfer Station; Butler-Edgeboro Landfill; Parsippany-Troy Hills-Edgeboro Landfill. Each of the landfills and transfer stations was at a different distance from the communities served and from Hoboken, where the employees returned with the trucks at the end of their work day. There were different waiting times at different landfills and transfer stations; there were different waiting times at the same landfills and transfer stations on different days. 5. On some occasions after all the trash on a route was collected a truck and its crew did not go to the landfill or transfer station but took the load to the Hoboken yard where it remained overnight. 6. Belleville was served by three trucks with their drivers and helpers for a part of the period in question. Thereafter the same territory was served by four trucks and their helpers, thus reducing the time required to complete collections. 7. Prior to 1987 each driver and helper had to wash his truck after returning to defendant’s Hoboken yard at the end of the day. The job took 15-20 minutes but there were often lengthy waits while other trucks were being washed. In 1987 defendant hired other persons to do the washing, and thereafter the drivers and helpers were able to leave immediately after returning to the yard. 8. Occasionally defendant gave a driver or helper extra pay for an unusually long trip or unusual circumstance. 9. Some helpers did not return to the defendant’s Hoboken yard in the evening, but were dropped off nearer their own homes. On occasion some helpers did not go into the Hoboken yard in the morning but were picked up as their truck proceeded to its assigned route. A number of employees testified about various of the above factors which affected the hours of their work and there was stipulated testimony of an even greater number of employees about these factors. There was some variation in the employees’ testimony about the same matters. During the first three days of trial the mass of undigested data referred to above was introduced into evidence by way of testimony and stipulated testimony. By July 20, the third day of trial, it dawned upon me that the government had no intention of using its expertise in this field to pull together the vast array of data and by means of expert testimony or otherwise compute and present what it had concluded was the total overtime for which compensation had not been paid. As government counsel stated “after the Court [makes] a finding on ... how many hours did people work ... the government would then go back and present a schedule of the specific dollar amount due for each individual.” (Tr. at 420, 421). The government correctly pointed out that such a procedure was contemplated by the pretrial order. However, awakening at last, I concluded that the procedure was “totally unacceptable” (Tr. at 421) and required the government to be prepared to present through a qualified witness or witnesses a schedule admissible in evidence showing exactly what it was that the government claimed was owing with respect to each employee. The Court: I’m directing you to have a witness who will give the bottom line and the reasons for the bottom line and a schedule to back it up so that if, theoretically, if I mastered the testimony I could say thank you at the end of the defendants’ case and end of your rebuttal case and give you a judgment. Pretend I’m a jury. You are sending me out to the jury room. Mr. Kay: I didn’t pretend that you are a jury in preparing this case. The Court: I’m directing you to treat me as if I’m a 12 or six person jury who will have the responsibility, when you *278 finish your closing argument, of retiring to the jury room, deliberating and coming back with a judgment which you will want to enforce, levy on their assets for X amount of dollars. (Tr. at 436). That colloquy took place on Thursday, July 20. I adjourned the trial until Wednesday, July 26 so that the government could prepare the report and correlate the supporting evidence which I had required. Wednesday arrived and not unreasonably the defendants objected to being confronted for the first time with a complex report and testimony concerning it. I assured them that once we had gotten the government’s case in manageable form we could turn to defendants’ case and the manner in which they would be given a reasonable opportunity to meet the government’s evidence and to present their own. The government called to the stand Raymond A. Martoccia, an able and experienced compliance officer of the United States Department of Labor who for many years had, among other things, computed back wages in situations where, as in the present case, firms did not maintain records of hours worked. After testifying about his investigation of defendants, which began in March 1986, Martoccia, as an expert, gave his opinion about the amount of overtime which was unpaid for the period August 1984 through March 1989 and explained the basis for that opinion. He provided a figure for each of defendant’s employees. The total figure (as later corrected) was $583,818.75. The formula which Martoccia utilized was: The Number of Weeks X Uncompensated Overtime Hours X IV2 x Regular Hourly Rate of Pay Equaled Unpaid Overtime Compensation Due. (Exh. P22). This formula was applied in the manner described below to each of the 29 workers who testified or as to whom there was stipulated testimony. (Exh. P33, P33A). With respect to each worker, based on his testimony, Martoccia calculated what his hours were if he worked a full six day week and what his average hours per day were. This was done for the period prior to 1987 when the employees were required to wash their trucks each night and it was done for the period after January 1, 1987 when that task was performed by others. With respect to each worker Martoccia was able to determine what his hourly rate of pay was and how many days he worked from defendant’s ADP records. Attached to this opinion as Appendix A are the two computation sheets prepared in order to compute the overtime pay which the government asserted was owing to Kenneth Lancaster, one of defendant’s employees. Similar sheets were prepared with respect to each employee who testified or as to whom testimony was stipulated. I use the Lancaster sheets simply as an example of the government’s method, emphasizing that it was one of the simpler computations. In the upper left hand corner of the first page of the Lancaster computation there appear (i) the number of hours he worked in each week prior to 1987 (67 hours) (when drivers and helpers had to wash their trucks upon returning to Hoboken at the end of the working day); (ii) the number of hours he worked in each day prior to 1987 (11.16 hours); (iii) the number of hours he worked in each week in and after 1987 (60.5 hours); and (iv) the number of hours he worked in each day in and after 1987 (10.08 hours). These figures were based on Lancaster’s testimony and on the assumption that he worked a full work week and full work day. The balance of the first page of the Lancaster computation constitutes a formula for determining the number of uncompensated overtime hours in weeks when Lancaster worked 6, 5 and 4 days. Overtime is the number of hours over 40 that somebody works times one and one-half their hourly rate of pay. In the present case the computation is complicated somewhat because each employee who worked a full six day week was paid for 40 hours at the regular rate and for 8 hours at double rate, which would be the equivalent of 16 hours at straight time. *279 Defendant did have accurate records of the amount which it paid each employee in each week during the post-July 1, 1984 period. This consisted of the ADP Quarterly Employee Earnings Record. (Exh. Gl). From those records it was possible to determine the employee’s rate of pay and the number of days he worked in each week. The number of days worked in each week could be ascertained by dividing the hours shown as having been worked in any week by 8, since no matter how many hours were actually worked, these records never recorded more than an 8 hour day or 48 hour week. In Lancaster’s case the ADP records showed that he was paid $46.43 for 8 hours, resulting in an hourly pay rate of $5.8036. In a week in which Lancaster worked six days, he was shown on the ADP records as having worked only 48 hours and was paid accordingly — 40 hours at the regular rate of pay ($5.8036) and 8 hours at double rate for a total of $325. In fact, however, if Lancaster worked a six day week, according to his testimony he actually worked 67 hours prior to 1987 and 60.5 hours after January 1, 1987. As noted above, if a person worked a full six day week, the sixth day (deemed by defendants’ records to be an 8 hour day) was paid at double the regular rate. Sometimes if an employee worked five days, the fifth day would be paid at double time rates, and sometimes if an employee worked four days, one of the days would be paid at double time. All this had to be taken into account when computing overtime due to each employee. Returning to the first page of the Lancaster overtime computation, that page is essentially a breakdown of different length work weeks, i.e., 6 days, 5 days (with and without overtime paid on the last day) and 4 days (with and without overtime paid on the last day). With respect to the work week of each length, a credit for double time paid was applied and there was computed the difference between the time which was properly paid and uncompensated time. For a six day week in Lancaster’s case the table showed in its first column the Actual Hours Paid.. That consisted of 40 hours of straight time plus 8 hours at double time, which was the equivalent of 56 total hours at straight time (40 + 16). The next column consisted of the Hours Properly Paid, i.e., the number of hours Lancaster could have worked in a six day week without defendant having incurred any obligation for overtime. Since Lancaster was paid double time for 8 hours of the first 48 hours, he could work for more than 48 hours without defendant being in violation of the requirement that it pay one and one-half the regular hourly rate for hours in excess of 40 each week. The Hours Properly Paid column computes the amount of hours Lancaster could work without the defendant becoming liable for overtime. In a six day week that number was 50.66 hours. In a five day week it was 45.33 hours if double time had been paid and 40 hours if such double time had not been paid. In a four day week the properly paid hours were 40 and 32, depending on the payment or non-payment of double time. The next and last column on the first page of the Lancaster computation set forth Uncompensated Overtime Hours for six day weeks, five day weeks with and without double time and four day weeks with and without double time. The uncompensated overtime was derived simply by subtracting from Lancaster’s actual hours worked in each of those categories of weeks the hours properly paid. For example, prior to 1987 Lancaster actually worked 67 hours in each six day week, according to his testimony. The number of hours for which he was properly paid was 50.66. Consequently for each six day week he worked prior to 1987 his uncompensated overtime amounted to 16.34 hours. Thus on page 1 of the computation sheet there was provided the data which could be applied to Lancaster’s actual work record. Page 2 of the Lancaster computation applies the page 1 data. It shows that during Lancaster’s first period of employment (3/14/85 to 9/5/85) he worked 17 six day weeks, 2 five day weeks at double time, 2 *280 five day weeks with no double time, and a four day week with double time. Similar data is given with respect to the second period during which Lancaster worked for defendant (7/30/87 to 8/27/87). Total overtime owing to Lancaster amounted to $1,095.44. Similar formulas and overtime computation sheets were prepared for each employee who testified or as to whom testimony was stipulated. Exhibit P33 is the assemblage of these individual computations, and Exhibit P33A contains corrections of certain of the computations. In individual cases other adjustments were entered into the overtime computation to reflect the witness’ testimony. For example, Eugene Bryant testified that during the period when three trucks were assigned to Belle-ville he worked 59 hours per week, and after four trucks were assigned to the town, he worked 53 hours per week. For him, the actual hours worked figure was averaged out at 56 hours, all of which were after January 1, 1987 when drivers and helpers no longer had to wash their trucks at the end of the day. Having described how he computed the unpaid overtime for defendants’ employees who testified or as to whom stipulated testimony was offered, the government’s expert Martoccia explained how overtime for non-testifying employees was computed. This is demonstrated in Exhibit P34 (later corrected in P34A) which averages the overtime established for the employees for whom there was testimony or stipulated testimony. There were excluded from the figures entering into the average the hours to which Jasper Smith and Albert Green testified since they were strikingly above the hours of comparable employees. The averages were the pre- and post-January 1, 1987 periods. Certain adjustments were made to account for different conditions pertaining to the individual employees at different times in their employment careers. Using this approach Martoccia arrived at an average of 66V.1 hours per week for the period prior to 1987 and an average of 60 hours per week for the period after January 1987. These average figures were then applied to each of defendants’ employees who did not testify or as to whom there was no stipulated testimony. Exhibits P35 and P35A contain the computations for all such non-testifying employees. Since the same hours applied to all of them a single sheet could be used containing the formula for determining the number of uncompensated overtime hours in 6, 5 and 4 day workweeks. A copy of that sheet and of the computations for A1 Armour, one of the approximately 170 non-testifying employees is attached as Appendix B. It will be noted that the methodology is the same as that used to compute uncompensated overtime of testifying employees. I should note that Exhibit P35 was prepared during the trial recess that extended from July 20 to July 26, that its preparation required the services of ten investigators other than Martoccia and the services of the area director and assistant area director of the Northern New Jersey office. This was work which the government originally contemplated doing after the trial was over. In addition to Exhibits P35 and P35A the government offered into evidence Exhibit P21, which consisted of computations of overtime of six other persons computed prior to trial. The very detailed and rather complicated material presented through Martoccia was new to defendants. In fairness I had to give them an opportunity to analyze it and to prepare to cross-examine Martoccia. In addition, defendants had an alternative method of calculating overtime which they wished to prepare and offer into evidence through their expert. In order to accomplish this I deferred Martoccia’s cross-examination and permitted defendants to commence their case in chief. I then adjourned the trial from July 27 until August 21 when, presumably defendants would be ready to cross-examine Martoccia and to present their own overtime computations. It was anticipated that these computations and an expert’s report would be made available to the government well in advance of the adjourned trial date so that it *281 would be prepared to cross-examine the expert and offer any rebuttal evidence it deemed appropriate. This was obviously a most inefficient way to proceed, but since I had not required the government and defendants to prepare their experts’ reports well in advance of trial, fairness and the need to develop the necessary factual data to decide the case seemed to require sacrificing efficiency. The trial resumed on August 21. Unfortunately defendants had not been able to assemble as quickly as they had hoped the array of records from the landfills and transfer stations. On the basis of those records defendants expected to establish the actual time when the drivers and helpers were on the job. The government had not been given the two weeks advance notice of defendants’ expert’s conclusions which I had required, and in fact on August 21 defendants were still working on their expert’s presentation. Not unreasonably the government objected to defendants’ proffer. When I describe below the statistical evidence developed by defendants and the conclusions which their expert drew from it, the merits of the government’s position will be better appreciated. Nevertheless I admitted the evidence and the testimony on the condition that after the trial the government be allowed to submit in affidavit form a full critique of the evidence. So we turn now to defendants’ evidence about overtime hours and computations which was introduced during the period August 21-23. During the July 26-August 21 hiatus in the trial defendants assembled certain data on the basis of which they sought to ascertain overtime hours. First, defendants obtained records of the transfer stations and landfills to which defendants’ trash was hauled. From those records they derived the times when defendants’ trucks entered the facility on each day during selected periods of time. Second, defendants engaged persons to clock the distances to and time required to travel from each transfer station or landfill to defendants’ Hoboken plant. Third, for each route defendants posited a daily starting time, based on testimony. This data, developed as to varying time periods for defendants’ different routes, was fed into a computer which was programmed to produce as to each location a schedule showing for each truck in each day in the selected time period: (i) the date, (ii) employees’ starting time, (iii) the time of the truck’s last arrival at the transfer station or landfill on that day, (iv) travel time from the transfer station to Hoboken, (v) the time of arrival at Hoboken (derived by adding travel time to landfill time), (vi) the length of the actual workday (the period which elapsed between the starting time and the time of the truck’s arrival at Hobo-ken), (vii) the time when the standard non-overtime workday would have ended, and (viii) the amount of overtime, if any (the length of time by which the actual workday exceeded the standard workday). Attached hereto as Appendix C is the first page of Exhibit D44(a) showing a portion of the computer printout applicable to the Parsippany-Troy Hills route (which used the Edgeboro Landfill). Defendants called as their expert Steven I. Schneider, a professional engineer. Previously three youths who had been retained by Schneider testified about their following certain of defendants’ drivers in May and June of 1989. They had timed the drivers’ workdays, observed them and reported on their findings. Schneider himself had followed a number of drivers during May 1989. He had sought to analyze the various routes and determine the time it took drivers on each of those routes to complete their daily tasks. The testimony of the three youths was not particularly persuasive. Their demean- or on the stand demonstrated that they did not take their work following garbage trucks very seriously and viewed it somewhat as a joke, a humorous interlude in their usually more socially elevated lives. Schneider was conscientious and serious about his study, but with respect to the period from July 1984 through March 1989 I do not find his observations very useful. *282 Conditions were dramatically different in May and June 1989 from those which prevailed during most of the period to which the government’s computations pertain. Further, Schneider’s study was very limited and episodic — an inadequate basis on which to make general conclusions. In addition to timing drivers, Schneider gave his opinions concerning the use of transfer station/landfill records to establish overtime. His opinions in this regard are contained in Exhibit D61. He had reviewed the computer printouts from the various transfer stations and landfills showing defendant’s dumping times. He had reviewed defendant’s dump tickets at Edgeboro Landfill and at the Avenue A Transfer Station. He reviewed the computer runs incorporating the same data derived from those documents and the time/distance information with respect to the runs from the transfer stations or landfills to Hoboken. In Schneider’s opinion other than actual observations of the routes from portal to portal, “the times indicated on the dump tickets and/or landfill/transfer station printouts are the most reliable source of information that can be used to calculate actual hours worked by Haulaway employees.” Schneider approved the sample used in the study: “The computer runs contained at [Tabs B, C, D, E and F of Exhibit P61] contain dump times for four (4) random typical months of the year representing each of the four seasons. The months selected were January, April, July and November. It should be noted that November represents a later month for route times since leaf collection is heaviest in that month.” Schneider approved the methodology employed to derive overtime from the computer runs. That method involved first adding all overtime minutes for all trucks sampled at each disposal site, and dividing the sum by the total number of truck shifts sampled. The resulting figure was the average overtime minutes per shift. Appendix D to this opinion is a copy of Rider 1 of Exhibit D44C which sets forth the computation of overtime for each municipality which defendant served. Having determined the average overtime minutes per shift, the next step was to determine annual overtime pay owing for each location for each year. It was assumed that drivers received $10 per hour and helpers $7 per hour. The following formula was applied to each community: Overtime payable = (hourly rate) X (1.5) X (No. of truck shifts per year) x (average overtime per shift). The total of the figures derived for each community constituted total annual overtime, which amounted to $41,849. Assuming equal overtime in each year, the total overtime which defendants computed for the August 1984 to April 1989 period was $200,875. Appendix E to this opinion is a copy of Rider 2 of Exhibit D44C which sets forth defendants’ computation of annual and total overtime. The government objected strenuously to the admission into evidence of all the material involving the transfer station/landfill dumping times and the calculations based thereon. They noted that they were not alerted to defendants’ new method of computing overtime until confronted with it in court, that they had had no chance to review the immense amount of documentation on which it was based, and that they had had no chance to prepare to cross-examine the witnesses who testified on the subject. Giving added credence to the government’s protestations of surprise and prejudice is defendants’ own Summation Brief submitted on September 15, three weeks after the conclusion of the trial. Apparently defendants themselves were surprised and prejudiced by their own expert’s report and exhibits which concluded that the total amount of overtime computed on the basis of the transfer station/landfill data and travel time data was $200,875. The Summation Brief sets forth new assumptions, new corrections and new figures entering into the computations. Defendants submit as exhibits to the Summation Brief new computer printouts in substitution for Exhibits D43 (Edgeboro-Parsippany-Troy Hills), D44a (Morris County Transfer Sta *283 tion — Parsippany-Troy Hills), D45a (Belle-ville Transfer Station), D46 (Bergen County Utilities Authority — Upper Saddle River) and D47 (Hackensack Meadowlands Baler-Commercial). Defendants also submit as an exhibit to the Summation Brief a new Rider 2 which would replace the Rider 2 which is a part of trial Exhibit D44C. On the basis of the new material defendants argue that appropriate adjustments to the data which they placed in evidence require a finding that the actual overtime for the 4 years 8 month period is $47,736, not $200,875 as they stated at trial. Defendants also note a “worst case scenario” which does not take into account days in a week when an employee worked less than eight hours and which assumes that each crew worked at a minimun an 8 hour day. The “worst case scenario” produces total overtime for the entire period of $233,-731.79 as shown in defendants’ new Rider 2. The government’s objections to the computation evidence which the defendants offered at trial were eminently reasonable and under ordinary circumstances I would not have admitted the defendants’ belated proffer. Nevertheless, this was not the ordinary case. The government had failed to prepare its computations until mid-trial and over defense objections I admitted them. To have done otherwise would have resulted in victimized employees not recovering what was due them. In fairness defendant was entitled to the same leniency as was accorded the government. To overcome any unfairness to the government I allowed it to file a post-trial affidavit criticizing the defense overtime computations. The government took advantage of that opportunity by submitting an affidavit of Raymond A. Martoccia. However, this leniency does not extend to defendants’ post-trial efforts to alter the evidence submitted by them and received at the trial. I shall ignore the exhibits which accompanied defendants’ Summation Brief. It should be apparent that the tortured course of the trial of this ease (which had to be tried in three segments during the July 18-August 24 period) was caused by the court’s failure to require both the government and the defendants prior to the pretrial conference to submit detailed reports setting forth their computations of overtime hours and the entire basis for those computations. If that had been done the respective experts could have been deposed, the underlying bases for the computations could have been explored and verified, settlement discussions could have been conducted and failing settlement the trial could have been conducted expeditiously without periodic adjournments and without any party facing any surprise or prejudice. It would appear that the government contemplated putting in the employees’ testimony and other evidence going to overtime hours and then asking the court to make a finding of the number of hours worked by each testifying employee. That finding would have constituted the court’s judgment. It would have provided the government with the “Actual Hours Worked” figure on its chart entitled “Formula For Determining Number of Uncompensated Overtime Hours in 6, 5, and 4 Day Workweeks.” (Appendix A, page 290). I gather that the government intended to send the court a post-trial submission which would have used the hours worked figure found by the court to compute the payments due to the testifying employees for overtime. The post-trial submission would have averaged the hours of the testifying employees, and then used the average hours worked to compute the payments due to the non-testifying employees. In other words after the trial the government would have made the computations which I required be put in evidence at the trial. If that course had been followed there would either have had to have been a second trial on the various steps of the computations or a judgment would have been entered without giving the defendants their constitutional right to a trial on actual factual issues. The actual hours worked by each employee was not the only issue to be tried. The reasonableness of each step of the govern- *284 merit’s computation raised factual questions. For example, the concept of averaging the hours of the testifying employees to find an hours figure for each non-testifying employee raises a number of factual issues. Early in this opinion I noted some of the many variables which affect the hours which employees worked. Each of those variables presents factual questions. Further, assuming that each testifying employee accurately stated the total hours he worked each week, one could still question whether the sample of employees whose hours entered into the average was representative of all the employees, whether each of the critical variables was accounted for in the sample, etc. I have no doubt a statistician would have a field day analyzing simply the averaging aspect of the government’s methodology. Unless the government is required to prove the validity of its computations in toto at a trial, a defendant would not have an opportunity to challenge them and thus would be deprived of due process. The computations are not mechanical matters which the trial court can refer to the government or even a special master. The pretrial order which the court entered contemplated such a course and that is why in mid-trial I corrected the errors of my ways. B. Overtime Computation Both the government’s and defendants’ methods of computing overtime hours are imperfect devices necessitated by defendants’ inexcusable failure to keep adequate records. Neither method produces a perfect result, but I conclude that of the two the government’s is the more likely starting point. Substantially for the reasons set forth in Martoccia’s rebuttal testimony and in his post-trial affidavit (which will be received in evidence as Exhibit P45) I conclude that defendants’ approach is fatally flawed. These reasons can be summarized as follows: 1.Defendants’ calculations took no account of the considerable time employees spent waiting to wash their trucks and washing their trucks prior to 1987. 2. The calculations used an erroneous 5:15 a.m. starting time. The starting time of employees working in Belleville was 4:30. The starting time of employees working in some of the other communities was earlier than 5:15 a.m. 3. The computer runs for Belleville and Parsippany-Troy Hills erroneously included respectively 180 and 160 lines where no transfer station or landfill times were noted, resulting in a smaller average overtime figure. 4. There were numerous instances where a trip to the dump should not have been included because the same truck was counted again later in the same day. By counting the same truck twice in a single day, the reports inflated the entry columns labeled “# shifts sampled per exhibit” on Rider 1 of D-61. Since the “# shifts sampled per exhibit” was divided into the total overtime minutes, a lower divisor would increase the average overtime minutes per shift. 5. The defendants’ report does not reflect the time of trucks which continued to work after the time they were at the transfer station and returned to the Hoboken yard with partial loads which would be unloaded the next day. 6. Defendants’ calculations are based on data from limited periods of time which are not representative of the entire August 1984-March 1989 period. The periods for which the data was taken were generally in more recent months, while, as I shall note below, the more serious overtime violations occurred in the earlier years. 7. The report does not take into account that in earlier years there were only three trucks in Belleville and only three pickup trucks in Saddle River. 8. There were days when defendant provided services, yet no trucks were listed in defendants’ computations. 9. Defendants’ calculations failed to take into account the waiting time at the transfer stations and landfills after the trucks were weighed in. That time could be considerable. *285 10. The formula used in Rider 2 of Exhibit D44C is flawed and results in an understatement of overtime. The reasons for this are explained in paragraphs 15-18 of Martoccia’s affidavit. For these reasons I conclude that overtime cannot be derived using defendants’ methodology. There are distortions in the government’s approach, but I conclude it is basically sound and it was appropriate to use it under the circumstances. I will adopt that approach but discount the overtime figure derived from it in an attempt to adjust for the distortions. I reject two of defendants’ grounds for objecting to the government’s methodology —(i) the asserted failure to give defendants credit for breaks which the employees took during the working day and (ii) the asserted use by employees of work time to perform personal tasks. The testimony of the employees called by the government was consistently to the effect that employees stopped to purchase food or soft drinks when the occasion presented itself and ate on the run, often taking their food and drinks to the trucks and eating while on the road. Some of defendants’ witnesses described longer stops and regular practices of having breakfast at certain locations or receiving sandwiches from customers or gathering for a while and eating and resting. “To qualify as a bona fide noncompensable break, the respite must be uninterrupted and at least 30 minutes in duration, and the employee must be completely relieved from duty. See 29 C.F.R. subsection 785.-19.” Donovan v. Bel-Loc Diner, Inc., 780 F.2d 1113 (4th Cir.1985). Since 1940 the Administrator of the Wage and Hour Division of the U.S. Department of Labor, whose duty is to administer and interpret the Act, has stated that short breaks are compensable: Rest periods of a short duration, running from 5 to about 20 minutes, are common in industry. They promote the efficiency of the employee and are customarily paid for as working time. They must be counted as hours worked. Defendants had no announced policy with respect to breaks. Notice was not given to employees that during the working day they were entitled to free time of any specified duration in order to eat or rest. The haphazard breaks taken to grab a sandwich, coffee or a soft drink did not qualify as an interruption of the workday sufficient to constitute noncompensable time. Defendants presented evidence that on occasion after a route had been completed a truck and its crew would proceed to perform a private job for a resident. Defendants also presented evidence that on occasion an employee did not return directly to Hoboken after completing his route and instead stopped off at a motel or tavern. It is quite likely that events such as these which defendants’ witnesses described took place from time to time, but the evidence is anecdotal. The evidence does not support a finding that these events took place with sufficient frequency to affect the government’s calculations of overtime hours. If the events had occurred with such frequency, surely defendants would have learned of them and stopped them. However, in one respect defendants’ critique of the government’s calculations has some (though limited) merit. The government’s calculations depend upon the accuracy of the employees’ testimony about the number of hours they worked. One would expect that there would be differences in such testimony considering that a long period of time was involved, some employees worked the same route as others but at times when conditions were different, and human memories are fallible. Although certain variations in testimony are to be expected, I conclude that the variations in testimony in the present case went beyond the expectable. There were exaggerations of starting times and ending times which had the effect of unduly increasing the hours of the working day, skewing in an upward manner the hours testified to and the hours found by averaging the hours testified to. Upward skewing also arises from the fact that it appears that the uncompensat *286 ed overtime diminished towards the end of the computation period. In fact the testimony of defendants’ managers, and the testimony of those who observed the operation of the routes in May and June 1989, suggests that by then defendants had brought themselves into substantial compliance with overtime requirements. This is not a surprising result since trial of this case was approaching and defendants could have expected closer scrutiny. Further, efficiencies had been introduced into the operations; the new transfer stations permitted more rapid dumping of loaded trucks; and recycling regulations in some municipalities reduced the loads which had to be picked up. Thus to apply the same average overtime hours for the entire period from January 1, 1987 (when drivers and helpers no longer had to wash their trucks) to March 31, 1989 most likely inflated the total overtime figures. There are probably a number of ways in which these two inflating tendencies could be dealt with, such as adjusting the hours which went into the average hours and recomputing the amount owed each of the 185 employees. Given the imprecision which necessarily accompanies any method, I conclude that a reasonable and practical means of adjusting for the two skewing tendencies which I noted is to discount the government’s total overtime figure by a reasonable percentage and reallocate the resulting amount in an appropriate manner among the employees. Defendants cannot complain about imperfections in the government’s methodology nor about my imprecise means of adjusting the government’s figures. These steps are necessitated by defendants’ failure to keep proper records over a very long period of time and are permissible. Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680, 66 S.Ct. 1187, 90 L.Ed. 1515 (1946). The government’s corrected total overtime figure is set forth in Martoccia’s post-trial affidavit. (Exhibit P45). It amounts to $583,816.74. It will be adjusted to account for the two inflating factors I noted in two ways: (i) The $583,816.74 figure will be reduced by 25% to $437,862.56 and (ii) Defendants will be deemed to have complied with all overtime requirements for the period April 1, 1989 to the date of judgment. This will not bar any employee or employees in a new action from asserting a claim that he or they were not paid proper overtime pay during that period and seeking to prove that claim by means of evidence applicable to that limited period. The awards payable to non-testifying employees as shown in Exhibit P45 will each be reduced by 25%. In the case of testifying employees, at the government’s discretion, either the individual awards may each be reduced by 25% or else the individual awards may be reduced in a manner devised by the government which will equitably distribute the total amount payable among the individual members of that group (such as applying a uniform average hours figure to all these employees so as not to penalize the employees who testified to more conservative and perhaps more accurate figures). C. Conclusions of Law Section 17 of the Fair Labor Standards Act, 29 U.S.C. § 217 and 28 U.S.C. §§ 1331 and 1345 confer subject matter jurisdiction on the court. Defendant Haulaway is an employer which has employees who are employed in an enterprise engaged in commerce or in the production of goods for commerce. 29 U.S.C. § 203(s). The evidence established that Haulaway’s employees used trucks, truck parts, containers, fuel and other materials that moved in commerce. Employees of commercial trash collectors are within the coverage of the Act, e.g., Brennan v. Metropolitan Trash, Inc., 513 F.2d 1324 (10th Cir.), cert. denied, 422 U.S. 1042, 95 S.Ct. 2656, 45 L.Ed.2d 693 (1975). The Act’s overtime provisions prohibit any “employer” from failing to pay overtime compensation. The Supreme Court has characterized the Act’s statutory definition of an “employer”, 29 U.S.C. § 203(d), as the “broadest ... that has ever been included in any one act”, United States v. Rosenwasser, 323 U.S. 360, 363 n. 3, 65 S.Ct. 295, 296 n. 3, 89 L.Ed. 301 (1945). A corporate officer with operational control is *287 an “employer”, along with the corporation, jointly and severally liable under the Act for unpaid wages. Donovan v. Agnew, 712 F.2d 1509, 1511 (1st Cir.1983). Further, any such corporate officer is liable in his individual, not representative, capacity. Donovan v. Grim Hotel Company, 747 F.2d 966, 971 (5th Cir.1984). Brock v. Superior Care, Inc., 105 CCH Lab. Cases H 34,839 (E.D.N.Y.1986) (holding two corporate officers individually liable as employers), af f'd on other issues, 840 F.2d 1054 (2d Cir.1988). The evidence established that Mr. Scugoza was the president and sole shareholder of the corporate defendant. In addition to holding corporate authority, Mr. Scugoza actually exercised day to day control over the affairs of the corporate defendant including approving promotions and raises, hiring and discharging employees, and establishing other terms and conditions of employment. Section 7(a)(1) of the Act requires employers to pay employees that work in excess of forty hours per week overtime compensation equivalent to one and one-half times the employees’ regular hourly rate. 29 U.S.C. § 207(a)(1). I have set forth above my findings that defendants’ employees were not paid for overtime as required by the Act and that the amount of the nonpayment was $437,862.56. Section 11(c) of the Act, 29 U.S.C. § 211(c), requires that an employer “shall make, keep, and preserve such records of the persons employed by him and of the wages, hours, and other conditions and practices of employment maintained by him____” 29 U.S.C. § 211(c). Section 15(a)(5) of the Act, 29 U.S.C. § 215(a)(5), makes it unlawful to violate § 11(c). The Secretary’s implementing regulations are found at 29 C.F.R. Part 516 and require, inter alia, that employers preserve, for three years, records of the (1) total daily and weekly hours employees work; (2) employees’ regular hourly rates of pay for each week that overtime is worked; (3) total daily or weekly straight time earnings’ and (4) total weekly premium pay for overtime hours. 29 C.F.R. §§ 516.2 and 516.5. Defendants were in flagrant violation of these record keeping provisions, and even though shortly before trial defendants installed a time clock for punching in and out, it is not clear that they have yet met the statutory requirements. Under § 6(a) of the Portal-to-Portal Act, 29 U.S.C. § 255(a), the statute of limitations bars wage claims accruing more than two years prior to the filing of the complaint, unless the violation is found to be willful, in which case recovery of back wages can be had for a three year period. The Supreme Court has recently held that a violation of the Act is willful if “... the employer either knew or showed reckless disregard for the matter of whether its conduct was prohibited by the statute.” McLaughlin v. Richland Shoe Co., 486 U.S. 128, 108 S.Ct. 1677, 100 L.Ed.2d 115 (1988). In its decision the court found that willfulness was synonymous with the terms “voluntary”, “deliberate”, and “intentional”. Id. 108 S.Ct. at 1681. Here the evidence establishes willful violation of the Act. Louis Garibaldi, the defendants’ general manager from 1984 to January 1987, was aware of the time and one-half requirement. Eddie Rodriguez, payroll supervisor since 1983, was aware of the requirement. Scugoza authorized entra pay on certain occasions, showing a recognition of overtime obligations. Employees complained to defendants’ supervisory personnel about working overtime without proper pay and were told, in effect, that if they didn’t like it they could leave. Defendants sought to conceal the fact that they were not paying required overtime pay by failing to keep proper records and by demanding that employees sign untruthful statements about the correctness of their pay as a condition to receiving their paychecks. In light of the foregoing findings the three year, rather than the two year statute of limitations is applicable, and defendants are each liable for overtime pay for the period August 1, 1984 through March *288 31, 1989 (the cutoff date of the government’s computations). Section 16(c) of the Act, 29 U.S.C. § 216(c), authorizes the Secretary of Labor to present a claim to recover liquidated damages in an amount equivalent to the unpaid overtime compensation found due. The purpose of liquidated damages is to compensate employees by making them whole for the delay in receiving wages improperly denied them through violations of the Act. Overnight Motor Co. v. Missel, 316 U.S. 572, 582-584, 62 S.Ct. 1216, 1222-1223, 86 L.Ed. 1682 (1942). Brock v. Superior Care, Inc., 840 F.2d 1054, 1064 (2d Cir.1988). Its purpose is not punitive. Marshall v. Brunner, 668 F.2d 748 (3d Cir.1982). Under the principles enunciated by the Third Circuit in Brunner and more recently in Williams v. Tri-County Growers, Inc., 747 F.2d 121 (3d Cir.1984), if the employer proves to the satisfaction of the court that (1) the acts or omissions giving rise to minimum wage or overtime violations were in good faith, and, (2) that the employer had reasonable grounds for believing that the acts or omissions were not in violation of the Act, the court may, in its discretion, decide not to award any liquidated damages, or to reduce the amount of liquidated damages. This is an affirmative defense afforded the employer by 29 U.S.C. § 260. See also Brock v. Wilamowsky, 833 F.2d 11, 19 (2d Cir.1988). Defendants have failed utterly to establish either the good faith or reasonable grounds basis for escaping liquidated damages. The facts which establish the willfulness of their conduct affirmatively negate their good faith and reasonable grounds contentions. Further evidence of bad faith and unreasonableness is defendants’ evasive, dilatory and obstructive conduct after Martoccia commenced his investigation. The investigation began in March 1986. Great difficulty was encountered in meeting with defendants or their attorneys. Records were requested and not provided. When records were sent they were incomplete or not responsive to requests. Representatives of the government met with defendants’ attorneys in December 1986 and pointed out that defendants were not keeping records and that it appeared on the basis of a preliminary examination that employees were working on the average 67V2 hour six day weeks. The defendants were informed how they could get into compliance. This advice was not followed. Quite clearly defendants are liable for liquidated damages for this entire period. Section 17 of the Act, 29 U.S.C. § 217, provides for equitable relief in the form of injunctions to restrain future violations of the Act and to restrain the continued withholding of back wages due. The decision to grant injunctive relief under § 17 lies within the sound discretion of the district court. Mitchell v. Lublin, McGaughy & Associates, 358 U.S. 207, 215, 79 S.Ct. 260, 265-266, 3 L.Ed.2d 243 (1959). The defendants have been guilty of egregious violations of the record keeping requirements of the Act and of the overtime payment requirements. It has not been established that their record keeping now complies with the Act. I have concluded that in recent months defendants have been making the overtime payments required by the Act in that it appears that the work week has been shortened so as not to require overtime. There is no assurance, however, that this eve-of-trial reform will continue. Under all the circumstances described above injunctive relief is appropriate and necessary to ensure future compliance with the Act. D. Conclusion Both defendants have violated the record keeping and overtime payment requirements of the Fair Labor Standards Act. Their conduct was willful, and consequently the three year statute of limitations is applicable. The total amount of overtime payments for which defendants are liable is $437,-862.56. Defendants are obligated to pay liquidated damages on that full amount. The government is entitled to injunctive relief to ensure future compliance with *289 both record keeping and overtime payment provisions of the Act. The government is requested to submit a form of order implementing the foregoing. *290 [[Image here]] *291 [[Image here]] *292 [[Image here]] *293 [[Image here]] *294 [[Image here]] *295 APPENDIX C [[Image here]] APPENDIX D Rider 1 [[Image here]] *296 Rider 1 [[Image here]] APPENDIX E Rider 2 [[Image here]]