Trump’s Labor Dept. Has Declared War on Tipped Workers

Heidi Shierholz and David Cooper December 3, 2019

A new rule under Trump’s Dept. of Labor could cost tipped workers more than $700 million. (Shutterstock)

In Octo­ber, the Trump admin­is­tra­tion pub­lished a pro­posed rule regard­ing tips which, if final­ized, will cost work­ers more than $700 mil­lion annu­al­ly. It is yet anoth­er exam­ple of the Trump admin­is­tra­tion using the fine print of a pro­pos­al to attempt to push through a change that will trans­fer large amounts of mon­ey from work­ers to their employ­ers. We also find that as employ­ers ask tipped work­ers to do more non­tipped work as a result of this rule, employ­ment in non­tipped food ser­vice occu­pa­tions will decline by 5.3% and employ­ment in tipped occu­pa­tions will increase by 12.2%, result­ing in 243,000 jobs shift­ing from being non­tipped to being tipped. Giv­en that back-of-the-house, non­tipped jobs in restau­rants are more like­ly to be held by peo­ple of col­or while tipped occu­pa­tions are more like­ly to be held by white work­ers, this could reduce job oppor­tu­ni­ties for peo­ple of color.

Employ­ers are not allowed to pock­et work­ers’ tips — tips must remain with work­ers. But employ­ers can legal­ly cap­ture” some of work­ers’ tips by pay­ing tipped work­ers less in base wages than their oth­er work­ers. For exam­ple, the fed­er­al min­i­mum wage is $7.25 an hour, but employ­ers can pay tipped work­ers a tipped min­i­mum wage” of $2.13 an hour as long as employ­ees’ base wage and the tips they receive over the course of a week are the equiv­a­lent of at least $7.25 per hour. All but sev­en states have a sub­min­i­mum wage for tipped work­ers.

In a sys­tem like this, the more non­tipped work that is done by tipped work­ers earn­ing the sub­min­i­mum wage, the more employ­ers ben­e­fit. This is best illus­trat­ed with a sim­ple exam­ple. Say a restau­rant has two work­ers, one doing tipped work and one doing non­tipped work, who both work 40 hours a week. The tipped work­er is paid $2.50 an hour in base wages, but gets $10 an hour in tips on aver­age, for a total of $12.50 an hour in total earn­ings. The non­tipped work­er is paid $7.50 an hour. In this sce­nario, the restau­rant pays their work­ers a total of ($2.50+$7.50)*40 = $400 per week, and the work­ers take home a total of ($12.50+$7.50)*40 = $800 (with $400 of that com­ing from tips).

But sup­pose the restau­rant makes both those work­ers tipped work­ers, with each doing half tipped work and half non­tipped work. Then the restau­rant pays them both $2.50 an hour, and they will each get $5 an hour in tips on aver­age (since now they each spend half their time on non­tipped work) for a total of $7.50 an hour in total earn­ings. In this sce­nario, the restau­rant pays out a total of ($2.50+$2.50)*40 = $200 per week, and the work­ers take home a total of ($7.50 + $7.50)*40 = $600. The restaurant’s gain of $200 is the work­ers’ loss of $200, sim­ply by hav­ing tipped work­ers spend time doing non­tipped work.

To lim­it the amount of tips employ­ers can cap­ture in this way, the Depart­ment of Labor has always restrict­ed the amount of time tipped work­ers can spend doing non­tipped work if the employ­er is pay­ing the sub­min­i­mum wage. In par­tic­u­lar, the depart­ment has said that if an employ­er pays the sub­min­i­mum wage, work­ers can spend at most 20 % of their time doing non­tipped work. This is known as the 8020 rule: employ­ers can only claim a tip cred­it” — i.e., pay tipped work­ers a base wage less than the reg­u­lar min­i­mum wage — if tipped staff spend no more than 20 % of their time per­form­ing non­tipped func­tions; at least 80 % of their time must be spent in tip-receiv­ing activities.

The pro­tec­tion pro­vid­ed by this rule is crit­i­cal for tipped work­er. For exam­ple, in a restau­rant, the 8020 rule pre­vents employ­ers from expect­ing servers to spend hours wash­ing dish­es at the end of the night, or prep­ping ingre­di­ents for hours before the restau­rant opens. Occa­sion­al­ly, a serv­er might play the role of the host, seat­ing guests when a line has formed, or fill­ing salt and pep­per shak­ers when din­ing ser­vice has end­ed — but such activ­i­ties can­not take up more than 20 % of their time with­out employ­ers pay­ing them the full min­i­mum wage, regard­less of tips.

The Depart­ment of Labor (DOL), under the Trump admin­is­tra­tion, has pro­posed to do away with the 8020 rule. Work­ers would be left with a tooth­less pro­tec­tion in which employ­ers would be allowed to take a tip cred­it for any amount of time that an employ­ee per­forms relat­ed, non­tipped duties con­tem­po­ra­ne­ous­ly with his or her tipped duties, or for a rea­son­able time imme­di­ate­ly before or after per­form­ing the tipped duties” (see page 53957 of the pro­posed rule).

With no mean­ing­ful lim­it on the amount of time tipped work­ers may per­form non­tipped work, employ­ers could cap­ture more of work­ers’ tips. It is not hard to imag­ine how employ­ers of tipped work­ers might exploit this change in the regulation.

Con­sid­er a restau­rant that employs a clean­ing ser­vice to clean the restau­rant each night: vac­u­um­ing car­pets, dust­ing, etc. Why con­tin­ue to pay for such a ser­vice, for which the clean­ing staff would need to be paid at least the fed­er­al min­i­mum wage of $7.25 per hour, when you could sim­ply require servers to spend an extra hour or two per­form­ing such work and only pay them the tipped min­i­mum wage of $2.13 per hour? Or, a restau­rant that cur­rent­ly employs three dish­wash­ers at a time might decide they can man­age the dish load with only one ded­i­cat­ed dish­wash­er if they hire a cou­ple extra servers and require all servers to wash dish­es peri­od­i­cal­ly over the course of their shifts. Employ­ers could pay servers less than the min­i­mum wage for hours of dish­wash­ing so long as they per­form some tipped work right before or after wash­ing dishes.

The depart­ment rec­og­nizes that work­ers will lose out under this change, stat­ing that tipped work­ers might lose tipped income by spend­ing more of their time per­form­ing duties where they are not earn­ing tips, while still receiv­ing cash wages of less than min­i­mum wage” (see page 53972 of the pro­posed rule). Telling­ly, DOL did not pro­vide an esti­mate of the amount that work­ers will lose — even though it is legal­ly required, as a part of the rule­mak­ing process, to assess all quan­tifi­able costs and ben­e­fits to the fullest extent that these can be use­ful­ly esti­mat­ed” (see Cost-Ben­e­fit and Oth­er Analy­sis Require­ments in the Rule­mak­ing Process).

The depart­ment claims they lack data to quan­ti­fy this poten­tial reduc­tion in tips.” How­ev­er, EPI eas­i­ly pro­duced a rea­son­able esti­mate using a method­ol­o­gy that is very much in the spir­it of esti­mates the Depart­ment of Labor reg­u­lar­ly pro­duces; DOL obvi­ous­ly could have pro­duced an esti­mate. But DOL couldn’t both pro­duce a good faith esti­mate and main­tain the fic­tion that get­ting rid of the 8020 rule is about some­thing oth­er than employ­ers being able to cap­ture more of work­ers’ tips, so they opt­ed to ignore this legal­ly required step in the rule­mak­ing process.

Below we describe the method­ol­o­gy for our esti­mate. The sim­plic­i­ty and rea­son­able­ness of this approach under­scores that by not pro­duc­ing an esti­mate, the admin­is­tra­tion appears to sim­ply be try­ing to hide its anti-work­er agen­da by claim­ing to not be able to quan­ti­fy results.

Method­ol­o­gy for esti­mat­ing tips cap­tured by employers

The remain­der of this piece describes the method­ol­o­gy for esti­mat­ing the total pay trans­ferred from work­ers to employ­ers as a result of this rule described above. To eval­u­ate how this rule change would affect pay, we use data from the Cur­rent Pop­u­la­tion Sur­vey (CPS), restrict­ed to states with a tip cred­it (i.e., that allow employ­ers to pay a sub­min­i­mum wage to tipped work­ers), to esti­mate how much employ­ers might shift work from tra­di­tion­al­ly non­tipped to tipped staff. Doing so would allow them to spread out the total pool of tips received over more peo­ple for whom employ­ers can pay less than the min­i­mum wage, there­by reduc­ing employ­ers’ wage respon­si­bil­i­ty. We then esti­mate the change in total earn­ings that would occur for food ser­vice work­ers if that shift in employ­ment took place.

The CPS is a house­hold sur­vey that asks work­ers about their base wages (exclu­sive of tips) and about their tips earned, if any. One prob­lem with the CPS data, how­ev­er, is that earn­ings from tips are com­bined with both over­time pay and earn­ings from com­mis­sions. Researchers refer to the CPS vari­able that pro­vides the aggre­gate week­ly val­ue of these three sources of earn­ings (over­time, tips, and com­mis­sions) as OTTC.” In order to iso­late tips using this vari­able, we first restrict the sam­ple to hourly work­ers in tipped occu­pa­tions, to help ensure that we are not pick­ing up work­ers who are like­ly to earn commissions.

For hourly work­ers in these tipped occu­pa­tions who work less than or equal to 40 hours in a week, we assume that the entire amount of OTTC earn­ings is tips. For hourly work­ers in tipped occu­pa­tions who work more than 40 hours, we must sub­tract over­time earn­ings. We cal­cu­late over­time earn­ings for these work­ers as 1.5 times their straight-time hourly wage times the num­ber of hours they work beyond 40. For these work­ers, we assume their tipped earn­ings are equal to OTTC minus these over­time earnings.

Some work­ers in tipped occu­pa­tions do not report their tips in the OTTC vari­able; how­ev­er, the CPS also asks work­ers to report their total week­ly earn­ings inclu­sive of tips, and their base wage exclu­sive of tips. For those work­ers in tipped occu­pa­tions with no report­ed val­ue in the OTTC vari­able, but whose total week­ly earn­ings is greater than the sum of their base wage times the hours they worked, we assume the dif­fer­ence is tips.

In oth­er words, for hourly work­ers in tipped occu­pa­tions we cal­cu­late tips in two ways:

1. For those who report a val­ue for OTTC:

Week­ly tips = OTTC for those who work ≤ 40 hours per week, and

Week­ly tips = OTTC − [(base wage) × 1.5 × (hours worked − 40)] for those who work > 40 hours per week.

2. For those who do not report a val­ue for OTTC:

Week­ly tips = Total week­ly earn­ings inclu­sive of tips – (base wage x hours worker).

In cas­es where tips can be cal­cu­lat­ed both ways, we take the larg­er of the two values.

Stan­dard eco­nom­ic log­ic dic­tates that employ­ers will spread out aggre­gate tips over as many work­ers they can — there­by reduc­ing their wage oblig­a­tions and effec­tive­ly cap­tur­ing” tips. They will shift work from non­tipped to tipped work­ers until the result­ing aver­age wage (com­bined base wage plus tips) of their tipped work­ers is at or just above the hourly wage these same work­ers could get in a non­tipped job. For employ­ers of tipped work­ers to get and keep the work­ers they need, tipped work­ers must earn as much as their out­side option,” since, all else being equal (i.e., assum­ing no impor­tant dif­fer­ence in non­wage com­pen­sa­tion and work­ing con­di­tions), if these work­ers could earn more in anoth­er job, they would quit and go to that job. But for employ­ers to keep these work­ers, they do not need to earn any more than they could earn in anoth­er job (again, assum­ing all else is equal), since as long as they are earn­ing what they could earn in anoth­er job, it would not be worth it to these work­ers to quit.

To cal­cu­late the out­side option wage,” we use regres­sion analy­sis to deter­mine the wage each work­er would like­ly earn in a non­tipped job. We regress hourly wage (includ­ing tips) on con­trols for age, edu­ca­tion, gen­der, race, eth­nic­i­ty, cit­i­zen­ship, mar­i­tal sta­tus, and state, and use the results of that regres­sion to pre­dict what each tipped work­er would earn in a non­tipped job. We set a low­er bound on pre­dict­ed hourly wages at the state min­i­mum wage. We refer to the pre­dict­ed val­ue as the out­side option wage — it’s the wage a sim­i­lar work­er in a non­tipped job earns. We assume if a work­er cur­rent­ly earns less than or equal to their out­side option wage, their earn­ings can­not be reduced because if their earn­ings are reduced, they will leave their job and take their out­side option.

How­ev­er, if a work­er cur­rent­ly earns more than their out­side option wage, their earn­ings can be reduced by the amount the work­er earns above the out­side option wage, since as long as their earn­ings are not reduced below their out­side option wage, they will have no rea­son to leave. We also assume that if their base wage is greater than the state min­i­mum wage — i.e. if their employ­er is not tak­ing the tip cred­it — their earn­ings will not be reduced, since the 8020 rule applies only to tipped work­ers who are paid a sub­min­i­mum base wage. We cal­cu­late new aver­age tips earned as the aggre­gate tips of all tipped work­ers minus the aggre­gate amount, just described, by which their earn­ings can be reduced, divid­ed by the total num­ber of tipped workers.

Using this esti­mate of new aver­age tips earned, we can esti­mate how much employ­ers might shift the com­po­si­tion of employ­ment by reduc­ing the num­ber of non­tipped work­ers and adding more tipped ones. We assume that the total amount of tips earned remains the same— it is just spread out over more tipped work­ers (who are now doing more non­tipped work). In par­tic­u­lar, we assume that the new num­ber of tipped work­ers is the num­ber that, when mul­ti­plied by the new aver­age tips earned, is equal to the total aggre­gate tips before the change.

We oper­a­tional­ize this by mul­ti­ply­ing the sam­ple weights of tipped work­ers by total aggre­gate tips divid­ed by the dif­fer­ence between total aggre­gate tips and the aggre­gate amount by which earn­ings can be reduced. We then assume that the num­ber of tipped work­ers added is off­set one-for-one by a reduc­tion in the num­ber of non­tipped work­ers who have food ser­vice occu­pa­tions. We oper­a­tional­ize this by mul­ti­ply­ing the sam­ple weights of non­tipped work­ers by one minus the ratio of the increase in tipped work­ers to the orig­i­nal num­ber of non­tipped work­ers. We find that employ­ment in non­tipped food ser­vice occu­pa­tions will decline by 5.3% and employ­ment in tipped occu­pa­tions will increase by 12.1%, result­ing in 243,000 jobs shift­ing from being non­tipped to being tipped as a result of this rule. The work that had been done by those non­tipped work­ers will now be done by tipped work­ers, with tipped work­ers spend­ing less time doing work for which they receive tips.

The loss in pay is cal­cu­lat­ed as the dif­fer­ence between cur­rent aggre­gate food ser­vice tips and new aggre­gate food ser­vice tips using the new employ­ment weights just described for tipped and non­tipped work­ers and the new aver­age wages for tipped work­ers. We assume aver­age wages for non­tipped work­ers do not change. We esti­mate that there will be a trans­fer of $705 mil­lion from work­ers to employ­ers if this rule is finalized.

Final­ly, it should be not­ed that our esti­mate of the trans­fer from work­ers to employ­ers is like­ly a vast under­es­ti­mate for three rea­sons. First, tips are wide­ly known to be sub­stan­tial­ly under­es­ti­mat­ed in CPS data, thus it is high­ly like­ly that we are under­es­ti­mat­ing the amount of tips employ­ers would cap­ture as a result of this rule change. For exam­ple, we find that 47.6% of work­ers in tipped occu­pa­tions do not report receiv­ing tips. Sim­i­lar­ly, using rev­enue data from the full-ser­vice restau­rant indus­try and updat­ing the method­ol­o­gy from Table 1 here to 2018, we find that tips in full-ser­vice restau­rants are $30.5 bil­lion, which is rough­ly twice the amount of tips report­ed in food ser­vice in the CPS. This means the amount employ­ers will real­ly cap­ture is like­ly rough­ly twice as large as our estimate.

Sec­ond, we only esti­mat­ed loss­es in food ser­vice. How­ev­er, about 26.0 % of tips earned in the econ­o­my are not earned in restau­rants or food ser­vice occu­pa­tions. Com­bin­ing these two fac­tors togeth­er means what employ­ers will real­ly cap­ture may be 2.5 times as large as our esti­mate. Third, our esti­mates assume that get­ting rid of the 8020 rule will only have an effect if the employ­er is already tak­ing a tip cred­it. This ignores the fact that some employ­ers may be incen­tivized to start using the tip cred­it if the 8020 rule is abol­ished, know­ing that with­out the rule they will be able to cap­ture more tips. Account­ing for this fac­tor would increase our esti­mate further.

The piece was also pub­lished at the Eco­nom­ic Pol­i­cy Insti­tute’s Work­ing Eco­nom­ics Blog.

Hei­di Shier­holz is the Senior Econ­o­mist and Direc­tor of Pol­i­cy at the Eco­nom­ic Pol­i­cy Insti­tute. David Coop­er is a Senior Eco­nom­ic Analyst.
Limited Time:

SUBSCRIBE TO IN THESE TIMES MAGAZINE FOR JUST $1 A MONTH