to actual pay, then the calculation of cards/hour will reveal the number of cards that this worker would have had to process in the week in question in order to earn the minimum wage.
(3) Calculation of Number of Cards/Hour Below Which a Minimum Wage Violation Exists
Assuming that the following information is available for the hypothetical employee,
Date Wage Rec'd. Lstd. Rate Pay
1/20/80 $ 3.10 1359 416 0.20 $ 83.20
the inquiry here is what number of cards/hour would this worker have had to process in order for actual pay received ($ 83.20) from DialAmerica to have sufficiently provided the minimum wage of $ 3.10 in 1980?
Devins suggests that the proper formula is
cards/hour = minimum wage x cards processed/gross pay
Inserting the known sample data,
cards/hour = $ 3.10 x 1359/$ 83.20 = 51.
This solution demonstrates that this employee would have had to process 51 cards/hour in order for $ 83.20 to provide the appropriate minimum wage.
With a production rate of 50 cards/hour, this employee would be one card short of the number necessary in order for her actual pay to provide the minimum wage.
Appendix A further displays Devins' mathematical analysis by providing a sample work week for each testifying employee (excluding distributees), and indicating the rate of production (i.e., cards/hour) that each worker would have had to achieve in the given week in order for actual salary received to reach the equivalent of the minimum wage.
Clearly, only three of the 33 witnesses accounted for in Appendix A achieved, or conceivably could have achieved, the rates of production required to render their given salaries consistent with the appropriate minimum wage.
As for the remaining thirty witnesses, the numbers in the far righthand column ("Minimum Wage Viol. Below Indicated Cds/Hr.") establish obvious violations in the sample weeks. For instance, M. Sullivan, in receipt of 545 cards, would have had to, based upon Devins' formula, process cards at the impossible rate of 98/hour in order for her actual pay to provide the minimum wage. However, most employees testified to rates falling only between 30 and 60 cards/hour.
A pattern of violations with regard to those not testifying can be inferred based on the representative testimony of these witnesses. Appendix B demonstrates similar unattainable rates of production for a sampling of 20 non-testifying employees. In fact, only two non-testifying witnesses in this random sample could have achieved the rate of production required to render her actual pay equal to the minimum wage (F. Hecht, at 45 cards/hour and M. Paolacci, at 25 cards/hour).
(4) Calculation of Back Wages Due
Continuing with the prior hypothetical home researcher, the next variable to be determined is the amount of back wages due. Once again, with an assumed rate of 50 cards/hour, Devins reasons that
Back wages due =(cards received/cards/hour x minimum wage) - actual pay
=1359/50 x $ 3.10 - $ 83.20
=27.18 x $ 3.10 - $ 83.20
=$ 84.26 - $ 83.20 = $ 1.06
Therefore, the hypothetical employee should have received $ 84.26 at the minimum wage. Since the actual pay was only $ 83.20, a minimum wage violation of $ 1.06 occurred. Devins excluded this figure ("wage violation") from his computerized summary due to the missing data concerning the number of hours worked by each employee. He testified that once a number of cards per hour is determined for an employee or a group of employees, that number would then be inserted into the computer in the cards/hour slot for a calculation of the back wages due in each work week.
(5) Calculation of Cards Received
Another variable that confronted the Department of Labor was the determination of the total number of cards received when the only information available was the number of listed cards (i.e., telephone numbers found) and the rate paid for these listed cards. To make this determination, Devins used data concerning the "find rates" for various rates of pay per card. For instance, 50-cent cards were found at a rate of 17.43%, 5-cent cards were found at a rate of 57.67%, and 10-cent cards were found at a rate of 37.18%. Armed with this data, Devins postulated that the number of cards received by a given researcher was algebraically ascertainable. Thus, he proposed the following formula:
100%/find rate = x/listed
For the hypothetical researcher:
100%/37.18% = x/listed
100%/37.18% = x/416
x = 100%/37.18% x 416
= 2.6 x 416 = 1119
Therefore, the hypothetical researcher who has found listings for 416 cards payable at the 10-cent rate would have received 1119 cards.
(6) Milo's Records
To further complicate matters, in the case of a distributorship, the number of cards allotted to the distributor and his or her distributees may be the unknown. The records of distributor Thomas Milo, for instance, set forth the total number of cards received from DialAmerica and the amount paid to each of his distributees, but the number of cards worked on by each distributee is unknown. To solve for the number of cards received by each distributee as well as by the distributor, Devins proposes
distributor's gross/total gross = x/cards rec'd
For example, if Milo received a total of 1000 cards from DialAmerica, paid $ 12.45 to one distributee and $ 10.40 to himself, and the total gross pay was $ 23.35, then the formula would apply as follows:
10.40/23.35 = x/1000
x = 10.40/23.35 x 1000
=.445 x 1000 = 445
The formula establishes that Milo received 445 cards out of the total 1000.
It once again becomes necessary to solve for the rate of cards per hour at which the minimum wage would be due to determine if a wage violation occurred. Devin proposes:
cards/hour = minimum wage x cards rec'd/minimum wage due
= 3.10 x 445/10.40 = 133
Milo would, therefore, have to process cards at a rate of 133/hour for his salary of $ 10.40 to reach the equivalent of minimum wage in the week in question.
(7) Poss' Records -- Calculation of Rate of Pay Per Card
In the case of the records maintained by distributor Christine Poss, Devins found that the rate of pay per card was not recorded. With actual pay received and the number of cards completed, actual pay must be divided by the number of cards completed to solve for the rate per card, that is:
rate per card = actual pay/number of cards completed
Therefore, in a week in which a distributee is paid $ 54.90 for the completion of 366 cards,
rate per card = 54.90/366 = 0.15
Here, in this example, the distributee was processing cards payable at the 15-cent rate.
(8) Poss' Records -- Number of Cards Received by Distributees
Several of the records maintained by Poss indicate the rate of pay and exclude the number of cards received by a given distributee. To solve for cards received, the formula suggested is
cards received = listed cards/find rate
Therefore, if a distributee working on cards payable at a rate of 8 cents (as indicated by the records of Poss) found 195 numbers (i.e., "listed" cards), and the find rate for 8-cent cards is 39.35%, then
cards received = 195/.3935 = 495
this employee would have received 495 cards.
A persistent and nagging question arose concerning the definition of a card "processed." It is the Department's position that the task of "processing" cards takes into account not simply time spent on the telephone, but includes the performance of whatever incidental tasks are necessary, such as unpacking and sorting cards by state.
Furthermore, Devins stated that a home researcher, in calculating the number of cards per hour that he or she is able to complete, would include cards passed over because the task of "processing" involves removing the cards from a package, and the determination that certain ones should not be researched.
Thus, the Department asserts that a card received is a card processed. If, however, it is found by the Court that hours worked includes only the time spent on the telephone, as DialAmerica contends, then Devins admits that his formula would have to be adjusted accordingly.
Devins further testified that if a particular home researcher received 1,000 cards in a given week and neglected to so much as look at 900 of them, the Department would only consider 100 as the number of cards received or processed.
However, he added that if a worker at least looked through all of the remaining cards, the time so spent would be included in the calculation of "hours worked."
Upon a thorough review of the deposition testimony of various home researchers, Devins noticed an obvious pattern of minimum wage violations across the board.
At a production rate of 60 cards per hour, he found that 91% of the 4,922 person-work weeks in 1982 would have resulted in minimum wage violations. At a production rate of 90 cards per hour, 53% of the person-work weeks resulted in minimum wage violations. Finally, at the rate of 50 cards per hour, every employee was subjected to a minimum wage violation.
From his investigation Devins concluded that the lowest rate of cards per hour at which the minimum wage violation would be eliminated for every employee would be 107.
Finally, Devins concluded that there existed a pattern as to how the home researchers performed their work. Workers generally first sorted cards, then screened cards to determine which cards to call, made phone calls, and finally initialed each card. Although some worked 30 minutes at a time and some four hours at a time, Devins is convinced that once the researchers sat down to work, each one performed the same basic tasks in precisely the same manner.
C. Deposition Telephone Tests
DialAmerica offered evidence of telephone tests initiated in conjunction with the depositions of several home researchers for this Court's consideration in determining whether there is a reasonable basis for finding a production rate for the testifying home researchers.
These tests, conducted by DialAmerica, required the deponents to research 20 cards by calling Directory Assistance and requesting telephone numbers as they had previously done in their homes. The deponents utilized a touch tone telephone set up at the offices of DialAmerica, and were required to dial "9" for an outside line, and "1" to begin dialing Directory Assistance outside of New Jersey.
The DialAmerica attorneys timed each deponent and subsequently made inquiries as to whether the test was representative of the manner in which he or she performed the work at home, whether the use of the automated response system today made the work go faster, and whether that system made it difficult to obtain more than one listing per directory assistance call.
Most deponents indicated that the process was much quicker at the deposition test, and many suggested that they could not have completed the number of cards per hour at home that were completed during DialAmerica's test.
Finally, the cards provided at the deposition test were from one state, whereas researchers, in reality, were given cards from a variety of states. Since many deponents testified that certain states were faster or slower than others, the critical question to ask is whether these tests provide an accurate representation of the probable production rates achieved at home.
Presumably, DialAmerica submitted evidence of its telephone tests in an effort to show that most of the home researchers had the ability to complete more cards than the records and their testimony indicate. If such a showing were in fact made and accepted, this Court could then determine that the minimum wage violations calculated by compliance officer Devins are grossly overstated and that the "pattern" of minimum wage violations is illusory at best.
This Court refuses, however, to ignore the pattern established by the testimony of so many home researchers and the records submitted in evidence. While it may be true that certain home researchers sporadically could have achieved higher rates of production, an assertion of what they "should" have done begs the question and in no way lessens the wage violations that have been committed by DialAmerica.
D. Bellcorp District Manager
The Department of Labor called Donna Bastien of Bell Communications Research ("Bellcorp") as an expert witness. Bastien maintains the title of district manager for open network architecture and has previously been involved in the management of operator systems and technical studies and consultation in the area of Directory Assistance.
Bastien offered testimony as to the average time spent by an operator on a Directory Assistance call between 1980 and 1982 using microfilm or microfiche Directory Assistance systems. This period, the "average work time," consists of the time that the customer is connected to the particular operator system and the operator is serving that customer.
In other words, its duration is measured from the time the operator system identified the operator to whom it would forward the call until the time that operator disconnected from the call.
Based on documents in evidence, it can be gleaned that in the early 1980s, the average work time for an operator using a computerized system was 30.5 seconds; for an operator using microfilm or microfiche, 33 seconds; and for an operator using a telephone book, 38 seconds in the early 1980's.
In an attempt to demonstrate that the deposition telephone test of 1987 was not conducted under conditions sufficiently similar to those existing in 1980-83, the Department elicited testimony from Bastien as to the rate of growth of computerized Directory Assistance systems. In 1983, 79 computer systems and three microfilm systems were in existence for the Bell operating companies in the United States.
By contrast, in 1980, 35% of Directory Assistance systems were using microfilm or microfiche, 14% were using paper records, and 51% were computerized. During her testimony, Bastien described how a call to Directory Assistance proceeds. Once a caller has dialed the appropriate number of digits (depending on which section of the country one is dialing from, this may be as few as three digits or as many as 10),
the call is routed through the telephone network by various "switching systems" and eventually terminates at an operator system. The system requires a certain amount of time in order to locate an available operator but, once found, the operator is attached to the call. When the caller makes an information request, the operator must search for the number or numbers using one of the search mechanisms (microfilm/microfiche, computerized database, or phone book). The operator then provides the number or numbers to the caller manually or by use of a computerized announcement machine (audio response system),
and then disconnects from the call.
In terms of the amount of time required to complete a Directory Assistance call, Bastien estimated one second per digit dialed using a rotary phone, and at least half of that time for a caller using a touch tone phone.
In addition, 140 milliseconds (.140 seconds) are required for the transmission of each digit dialed.
She further explained that there exists a hierarchy of switching machines (i.e., classes) used in the telephone network, including the caller's local machine. A call could conceivably be routed through eight classes before obtaining a connection with an operator.
Thus, hypothetically, if a caller must dial an area code plus a seven-digit telephone number (i.e., 10 digits in sum), and 140 milliseconds per digit is required, the call could be routed in 5.6 seconds.
Although the witness acknowledged that a shortage of operators or a heavy flow of traffic could cause a caller to wait longer than usual, the normal pre-divestiture procedure required a 5.8 second speed of answer time. In other words, 5.8 seconds was the normal time required to actually get connected to an operator.
Therefore, assuming 10 seconds for dialing (one second per digit for a 10-digit call), 5.6 seconds for the call to be routed, 5.8 seconds to get the operator, and approximately 30 seconds for average working time, the average Directory Assistance call could be completed in 51 seconds.
The average waiting time could, of course, be extended where the caller requested more than one telephone number from the operator. Before the telephone company began charging for Directory Assistance, there was more flexibility in the amount of requests it would tolerate. In light of the subsequent charge for Directory Assistance and the automated response mechanism, an operator could manually monitor the number of requests permitted.
Assuming a slight overload of requests for Directory Assistance, the speed of answer would naturally be longer, according to Bastien. A 5% overload, for instance, would cause a 10-second delay at most. However, there is no direct proportion among these figures, as a 30% overload would not necessarily indicate a 60-second delay.
Finally, Bastien concluded that the use of the automated response mechanism today requires a caller to spend more time to acquire two listings than the caller would have spent in the early 1980s. Thus, while automation may be less expensive to the telephone company, it is not necessarily quicker.
This Court finds Bastien's testimony credible. Accounting for the dialing process, the routing process, the speed of answer time, and average work time, the Court is satisfied that 51 seconds to complete a Directory Assistance call is a reasonable approximation and entirely compatible with the Secretary's finding of 53 cards/hour as the average number processed by the typical home researcher.
A. Fair Labor Standards Act (FLSA)
Pursuant to § 11(c) of the FLSA, 29 U.S.C. § 211(c), every employer is required to make, keep and preserve records of persons employed by it and of the wages, hours and other conditions and practices of employment maintained by it. This duty is imposed upon the employer as employees rarely maintain such records, and it is the employer who is in the position to know and produce the most probative facts regarding the nature and amount of work performed. Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680, 687, 66 S. Ct. 1187, 90 L. Ed. 1515 (1946).
The production records maintained by DialAmerica include name, date, number of cards received, number of cards for which listings were found, and amount paid. However, there is no record of the number of hours worked by the home researchers, and DialAmerica never required its home researchers to keep track of the number of hours worked. A violation of the record-keeping provisions of the FLSA is, therefore, patently obvious, as the number of hours worked is critical to a determination of the minimum wages due.
Section 6(a) of the Act requires an employer to pay each of its employees an amount not less than the applicable minimum wage times the number of hours worked. 29 U.S.C. § 206(a).
This work performance payment must be made on a weekly basis. In other words, each work week stands on its own and payments made in productive and unproductive weeks cannot be averaged together for the purpose of a minimum wage determination.
Impliedly, therefore, the speed at which one works is not relevant to the calculation of minimum wage.
The Department contends that the unknown number of hours actually worked by each home researcher can be reconstructed through a formula utilizing such known factors as minimum wage, cards received, and amount actually paid. The Department posits that the application of its theory of proof will permit this Court to ascertain a minimum wage violation for each work week by comparing the home researcher's rate of processing cards to the rate at which the researchers would have been required to process cards in order to earn the minimum wage.
An employee who brings suit under § 16(b) of the Act, 29 U.S.C. § 216(b), for the recovery of unpaid minimum wages must prove that he or she has performed work for which he or she was not properly compensated. Anderson, 328 U.S. at 687. "The remedial nature of this statute and the great public policy which it embodies, however, militate against making the burden an impossible hurdle for the employee." Id.
The Anderson case and its progeny establish that when the employer has complied with the Act and has maintained proper and adequate records, the burden of the employee is easily discharged by securing production of those records. Id. However, where, as in the case at bar, the employer's records are inaccurate or inadequate, and the employee alone is unable to offer convincing proof, a serious problem arises. The solution, however, is not to penalize the employee by denying recovery based on an inability to prove the extent of the undercompensated work. Such a result "would place a premium on an employer's failure to keep proper records in conformity with his statutory duty; it would allow the employer to keep the benefits of an employee's labors without paying due compensation as contemplated by the Fair Labor Standards Act." Id. The employee in a case such as this meets the required burden if he or she can prove that work was performed for which the worker was not properly compensated and if he or she produces sufficient evidence to show the "amount and extent of that work as a matter of just and reasonable inference." Id. (emphasis added); see Brock v. Seto, 790 F.2d 1446, 1448 (9th Cir. 1986); Donovan v. New Floridian Hotel, Inc., 676 F.2d 468, 472 (11th Cir. 1982); Marshall v. Van Matre, 634 F.2d 1115, 1119 (8th Cir. 1980).
Once the employee has satisfied this standard, the burden shifts to the employer to produce evidence of work performed or to negate the reasonableness of the inference to be drawn from the employee's evidence. Anderson, 328 U.S. at 687-88. If the employer is unable to meet this burden, the Court may determine and award damages to the employee. Id. As the Anderson Court held:
The employer cannot be heard to complain that the damages lack the exactness and precision of measurement that would be possible had he kept records in accordance with the requirements of § 11(c) of the Act. . . . The employer, having received the benefits of such work, cannot object to the payment for the work on the most accurate basis possible under the circumstances.