Torri M. Houston, Individually and on behalf of all others similarly situated v. Saint Luke’s Health System, Inc.; Saint Luke’s Northland Hospital Corporation

United States Court of Appeals
For the Eighth Circuit
___________________________
No. 22-1862
___________________________
Torri M. Houston, Individually and on behalf of all others similarly situated
Plaintiff - Appellant
v.
Saint Luke’s Health System, Inc.; Saint Luke’s Northland Hospital Corporation
Defendants - Appellees
------------------------------
Secretary of Labor
Amicus on Behalf of Appellant(s)
____________
Appeal from United States District Court
for the Western District of Missouri - Kansas City
____________
Submitted: June 13, 2023
Filed: August 11, 2023
____________
Before GRUENDER, ARNOLD, and KELLY, Circuit Judges.
____________
GRUENDER, Circuit Judge.

-- 1 of 10 --

-2-
Hourly employees of St. Luke’s Health System allege that they were
underpaid over several years due to their employer’s timekeeping policy that
rounded off time at the beginning and end of shifts. The district court granted
summary judgment to St. Luke’s on all claims. We vacate and remand.
I.
St. Luke’s uses an automated timekeeping system. Employees clock in and
out at the beginning and end of their shifts, and the system records the exact time.
The system then applies a rounding policy. Clocked times within six minutes of a
shift’s scheduled start or end get rounded to the scheduled time for compensation
purposes. For example, an employee who clocks in at 8:56 a.m. for a 9:00 a.m. shift
would not be paid for those four minutes. Likewise, an employee who clocks out
early at 4:54 p.m. for a shift ending at 5:00 p.m. would still be paid for those
unworked six minutes.
Torri Houston, a former employee, sued on behalf of herself and similarly
situated employees, claiming that St. Luke’s violated the Fair Labor Standards Act’s
(“FLSA”) overtime provisions by failing to fully compensate employees for work
performed. See 29 U.S.C. § 207(a)(1); 29 C.F.R. § 778.103. She also brought an
unjust-enrichment claim under state law. The district court certified two classes with
different lookback periods: (1) an FLSA collective comprised of employees who
worked for St. Luke’s between September 2016 and September 2018;1 and (2) an
unjust-enrichment class comprised of all employees who worked for St. Luke’s in
Missouri between April 2012 and September 2018. Houston also asserted individual
claims, one under the Missouri Minimum Wage Law, Mo. Rev. Stat. § 290.527, and
one for breach of her employment contract.
1 The FLSA authorizes collective actions, see 29 U.S.C. § 216(b), which are
distinct from class actions governed by Federal Rule of Civil Procedure 23.

-- 2 of 10 --

-3-
After limited yet substantial discovery, St. Luke’s moved for summary
judgment. It stipulated, for the purposes of its motion, that all employee time “on
the clock” was compensable work time and that issues about why employees clocked
in early or clocked out late were immaterial. See Fed. R. Civ. P. 56(c)(1)(A).
Each side submitted expert reports analyzing the rounding policy’s effect on
compensation. The employees’ expert, Scott Baggett, analyzed pay records across
all six years of data on a per-shift basis, a per-workweek basis, a per-employee basis,
and overall. The St. Luke’s expert, Deborah Foster, also analyzed pay records on a
per-shift and per-employee basis but did so separately for three lookback periods.
The parties agree that any minor differences in the data evaluated or in the
conclusions reached by the reports are immaterial for summary judgment.
The reports show that the rounding policy benefited St. Luke’s more often
than not and across different time periods. Baggett reviewed more than 7 million
shifts of 13,000 employees and determined that the rounding policy cut time from
about half of shifts, added time to a little over a third, and had no effect on the rest.
The net loss to employees increased steadily over time. On a per-employee basis,
Bagget concluded that the policy cut time for nearly two thirds of employees and
added time for only a third. Overall, he estimated that the policy favored St. Luke’s
to the tune of 74,000 employee-hours from April 2012 to September 2018. For
damages, Baggett estimated about $140,000 in lost overtime pay for the two-year
FLSA collective period and about $2.2 million in lost earnings for the Missouri
unjust-enrichment class over the corresponding six-year period. As for Houston
individually, from April 2012 to September 2018, Baggett determined that she lost
time on nearly half of her shifts and gained time on only a fifth. Overall, she lost
7.6 hours, amounting to $205.13, or about $32 per year.
Foster’s report tracks Baggett’s. She analyzed three lookback periods:
September 2016 to September 2018; September 2015 to September 2018, and April
2012 to September 2018. For each period, Foster concluded that about half of shifts
had time cut and a little over a third had time added. She also found that across all

-- 3 of 10 --

-4-
lookback periods, nearly two thirds of employees had net time lost. On average,
those employees lost 6.5 hours in the first lookback, 8.5 hours in the second
lookback, and 11.5 hours in the third lookback. For employees who had total net
time added, the added time was about half the amount of time that their less fortunate
colleagues lost. Of course, on a per-shift basis, any time lost was small because the
rounding policy operates within a six-minute window.
The district court granted summary judgment to St. Luke’s on all claims. For
the FLSA collective, the court concluded that the rounding policy was neutral as
applied because (1) the time lost per shift was not that much; (2) even if more
employees lost time than gained, the rounding policy both added and subtracted time
during the period, meaning that other time periods might show different results; and
(3) on a per-shift basis, the rounding policy took time from about half of shifts while
it added to or left neutral the other half. For Houston’s Missouri wage claim, all
agreed that it called for the same analysis, so it likewise failed. See Mo. Rev. Stat.
§ 290.505.4.
As for the Missouri unjust-enrichment claim, the court concluded that
allowing St. Luke’s to retain the benefit of the uncompensated work would not be
inequitable because it resulted from a lawful and neutrally applied rounding policy.
Similarly, the court concluded that St. Luke’s did not breach Houston’s employment
contract because she was paid according to a lawful rounding policy to which she
presumably assented under the contract.
Houston appeals. She argues that the district court erred by concluding that
the rounding policy was neutral as applied. She maintains that she has presented
sufficient evidence to raise a genuine dispute that the policy results in systematic
undercompensation over time. We agree.

-- 4 of 10 --

-5-
II.
We review a grant of summary judgment de novo. Lyons v. Conagra Foods
Packaged Foods LLC, 899 F.3d 567, 582 (8th Cir. 2018). We will affirm if the
evidence, viewed in the light most favorable to the non-moving party, shows that no
dispute of material fact exists and that the moving party is entitled to judgment as a
matter of law. Id.
A.
The FLSA requires employers to pay overtime compensation when employees
work more than forty hours in a week. 29 U.S.C. § 207(a)(1). Longstanding
regulations permit employers to “round” an employee’s clocked start and end times
for ease in calculating time worked. 29 C.F.R. § 785.48(b). Time clocks are not
required, but
[i]n those cases where time clocks are used, employees who voluntarily
come in before their regular starting time or remain after their closing
time, do not have to be paid for such periods provided, of course, that
they do not engage in any work. Their early or late clock punching may
be disregarded. Minor differences between the clock records and actual
hours worked cannot ordinarily be avoided, but major discrepancies
should be discouraged since they raise a doubt as to the accuracy of the
records of the hours actually worked.
Id. § 785.48(a). Here, the parties stipulated that the employees engaged in
compensable work at all times “on the clock.” The regulation continues:
It has been found that in some industries, particularly where time clocks
are used, there has been the practice for many years of recording the
employees’ starting time and stopping time to the nearest 5 minutes, or
to the nearest one-tenth or quarter of an hour. Presumably, this
arrangement averages out so that the employees are fully compensated
for all the time they actually work. For enforcement purposes this
practice of computing working time will be accepted, provided that it

-- 5 of 10 --

-6-
is used in such a manner that it will not result, over a period of time, in
failure to compensate the employees properly for all the time they have
actually worked.
29 C.F.R. § 785.48(b). The emphasis on ensuring that employees are “fully
compensated” aligns with the FLSA principle that “[w]ork not requested but
suffered or permitted is work time” that must be compensated. See 29 C.F.R.
§ 785.11. The reason that an employee works outside a scheduled shift time is
immaterial. All that matters is that the employer “knows or has reason to believe
that [the employee] is continuing to work.” Id. If so, “the time is working time,”
id., and a valid rounding policy must be “neutral, both facially and as applied,” so
that the employee is compensated for it, Aguilar v. Mgmt. & Training Corp., 948
F.3d 1270, 1288 (10th Cir. 2020); see Corbin v. Time Warner Entertainment-
Advance/Newhouse P’ship, 821 F.3d 1069, 1075 (9th Cir. 2016).
Here, the parties agree that the rounding policy is facially neutral, but they
disagree on whether, as applied, it fails “over a period of time” to compensate
employees “for all the time they have actually worked.” See 29 C.F.R. § 785.48(b).
As an initial matter, St. Luke’s argues that a facially neutral rounding policy can be
neutral in application even if some employees are negatively impacted over a given
period. The employees do not quarrel with that characterization but argue that the
proper focus is on what happens to employees—whether some or as a whole—in the
long term.
We have not confronted what it means for a rounding policy to “average[]
out” such that employees are compensated properly. See id. By referring to averages
“over a period of time,” the regulation clearly does not require that the rounded time
equal the actual worked time for every shift or workweek or payroll period. It will
often be that over some time periods employees are undercompensated, while over
other periods they are overcompensated or neutrally compensated. That is just the
nature of dealing with averages. So, a single cherry-picked period is likely
insufficient to demonstrate undercompensation in the long run. We must therefore
distinguish statistical anomaly from discernable pattern.

-- 6 of 10 --

-7-
Two cases are instructive. In Aguilar v. Management & Training Corp.,
correctional officers presented unrebutted evidence that their employer’s ten-minute
rounding policy routinely resulted in undercompensation. 948 F.3d at 1274, 1287-
89. The “representative [data] sample” showed that in almost all shifts, the officers
were on the clock for longer than their scheduled shifts. Id. at 1288. Over a four-
year period, 122 officers lost $350,000, or about $717 per officer per year. Id. at
1285. The Tenth Circuit emphasized that if the policy “routinely rounds off . . .
compensable overtime, as the officers’ evidence suggests, then the officers’
rounding theory remains viable.” Id. at 1289. The court reversed summary
judgment to the employer.
In Corbin v. Time Warner Enterprise-Advance/Newhouse Partnership, an
employee argued that a rounding policy violates the FLSA “unless every employee
gains or breaks even over every pay period or set of pay periods analyzed.” 821 F.3d
at 1077. The Ninth Circuit disagreed. It emphasized that the rounding policy need
only “average out in the long-term.” Id. The employee’s pay records showed that
he sometimes gained and sometimes lost minutes and compensation, with a net
deviation of only three minutes and fifteen dollars over the course of a year. Id. at
1079. The court concluded that this “fluctuat[ion] from pay period to pay period”
was not enough to raise a genuine dispute of material fact as to whether the
employer’s rounding policy averaged out over time. Id. Indeed, the court observed
that “a few more pay periods of employment may have tilted the total
time/compensation tally in the other direction.” Id.
We find these cases sufficiently illustrative of the concept for our purposes
here. Houston has raised a genuine dispute that the rounding policy does not average
out over time. No matter how one slices the data, most employees and the employees
as a whole fared worse under the rounding policy than had they been paid according
to their exact time worked. Thus, we need not resolve whether an employer runs
afoul of the rounding regulation whenever it undercompensates any individual
employees over a period of time or only when it undercompensates employees as a
whole, including those who were overcompensated or neutrally compensated. Here,

-- 7 of 10 --

-8-
the rounding policy did both. It resulted in lost time for nearly two thirds of
employees, and those employees lost more time than was gained by their coworkers
who benefited from rounding. This remains the case whether we look at a two-year,
three-year, or six-year period. And unlike in Corbin, the losses here accrued with
sufficient regularity to show a clear trend of undercompensation. Consider the data
another way. Across all employees, on average, St. Luke’s benefited one free hour
of labor per year per employee. If we confine ourselves to only those employees
who were net losers, St. Luke’s benefited nearly two hours per year per employee.
At this stage and on this record, we would be hard-pressed to say that the
rounding policy has “average[d] out” over the long run such that it did not result in
undercompensation. See 29 C.F.R. § 785.48(b). St. Luke’s could have rebutted this
reasonable inference by, for example, pointing to different probative lookback
periods in which employees were overcompensated or neutrally compensated. See
Aguiliar, 948 F.3d at 1289. But it has not.
Instead, St. Luke’s argues about policy. It asserts that finding for the
employees would upend decades of labor law and nullify the rounding regulation
altogether. It insists that no employer would ever implement a rounding policy if it
were required to perpetually audit that policy across various lookback periods. In
other words, St. Luke’s argues that the employees’ theory renders the rounding
regulation mere surplusage or else makes it too burdensome to be worthwhile.
These arguments miss their mark. The rounding regulation does not require
rounding; it permits it. That permission comes with conditions: chiefly, that the
rounding “will not result” in systematic or routine underpayment “over a period of
time” for work performed. 29 C.F.R. § 785.48(b); see Aguilar, 948 F.3d at 1289.
Here, with automated, electronic timing and accounting, this is easy to verify
because the system records the exact time that an employee clocks in or out. There
is no administrative hassle. This is not like the old days of punch cards and hand
arithmetic.

-- 8 of 10 --

-9-
Moreover, given this case’s unusual posture, we think St. Luke’s’ fears about
upending labor law are unfounded. By stipulation, St. Luke’s acknowledges that all
clocked time was worked time. And yet it implies otherwise, appealing to “the
practical realities of time clock placement” in a large facility like its own. Practically
speaking, employees might clock in and walk down the hall for a cup of coffee
before actually starting work. Or they might, at the end of the shift, pack up their
belongings and say goodbye to coworkers before clocking out. Maybe some of that
occurred here. And maybe after we vacate the summary-judgment order, the parties
will have to expend significant effort litigating whether all clocked hours were
actually hours worked. But that is not our concern. We hold St. Luke’s to its
stipulation, see 303 Creative LLC v. Elenis, 600 U.S. ---, 143 S. Ct. 2298, 2316-17
(2023), which formed the basis for its summary-judgment position that the rounding
policy itself was lawful, neutral, and sufficient to defeat the employees’ claims.2 We
conclude that the employees have raised a genuine dispute that the rounding policy,
as applied, did not average out over time. The district court therefore erred in
granting summary judgment on the FLSA and Missouri wage claims.3
B.
The rest of the district court’s summary-judgment order followed from its
conclusion that the employees could not prevail on their rounding-policy theory. For
the unjust-enrichment and breach-of-contract claims, the court reasoned that if the
2 We cannot help but find this posture odd. By stipulating away the significant
question of whether all rounded-off time was indeed compensable time, the parties
have plucked the rounding regulation from the broader regulatory scheme and
subjected it to artificial examination unmoored from practical realities.
3 For these claims, the district court did not address whether any
undercompensation from rounding was de minimis according to 29 C.F.R. § 785.47.
We decline to do so in the first instance. See Fergin v. Westrock Co., 955 F.3d 725,
730 n.3 (8th Cir. 2020) (declining to consider issues that were briefed to the district
court but “not passed on”).

-- 9 of 10 --

-10-
rounding policy was lawfully applied then any enrichment was not unjust and no
contract was breached. See Jennings v. SSM Health Care St. Louis, 355 S.W.3d 526,
536 (Mo. Ct. App. 2011) (“Demonstrating unjust retention of the benefit is the most
significant element of unjust enrichment . . . .”). But a genuine dispute remains that
the rounding policy was not lawfully applied. Finally, although the district court
undertook a partial FLSA de minimis analysis for the unjust-enrichment claim, it
never actually resolved that issue (nor the threshold issue of whether the Missouri
common-law doctrine incorporates the FLSA’s de minimis standard). We therefore
decline to resolve these issues at this juncture, and we vacate the judgment as to all
claims and remand for consideration in light of this opinion.
III.
For the foregoing reasons, we vacate the district court’s grant of summary
judgment to St. Luke’s and remand for further proceedings consistent with this
opinion.
______________________________

-- 10 of 10 --