Minimum Wage for Tipped Employees Data, Theory, and Calibrations
Oz Shy
1/31
Observations, motivation, and research goals A tip is a percentage fee (tax) that consumers impose on themselves. The proceeds go directly to servers instead of business owners. Hence, servers derive income from two sources: (i) Customers (via tips) and (ii) Employers (as hourly wages). Although tipping is voluntary, the law permits employers to pay lower minimum wage (referred to as “tip credit”). I investigate a policy whereby servers’ income from tips is replaced by a higher wage. Such a raise could take the form of closing the gap between tipped minimum wage and non-tipped minimum wage.
Introduction
2/31
Minimum wage for tipped employees Federal level: Wage-related rules for tipped employees are governed by the Fair labor Standard Act (FLSA). Tipped employees are those who collect tips more than $30/month. Tips are the sole property of servers (some pooling is allowed). Employers are obligated to pay the regular minimum wage, however they can factor tips into their wage obligations (“tip credit”). Example: Employers must pay the hourly minimum wage $2.13 as long as workers collect enough tips to earn at least $7.25. May be hard to implement in practice. Employers can also deduct their 2-3% credit card fee (if tip is left on a credit card). Several states impose higher minimum wages than the Federal levels. 15 states either adhere to the Federal levels or do not impose any minimum wage, in which case the $2.13 and $7.25 levels apply. Introduction
3/31
Tipping and service quality Is there anyone in the audience thinking that by giving a higher tip she/he gets a higher-quality service? Yes/No? Lynn & McCall (2000), Azar (2007), and Lynn et al. (2012) present evidence for a very weak relationship. 13 studies of over 2500 customers at 20 different restaurants showed mostly near-zero correlation between service evaluation and tip size. This relationship may be statistically significant by very small (less than 2% of the variability in tip percentages). Some waiters documented that they receive higher tips from customers for whom they accidentally spilled water over their shirts. People leave tips even if they receive bad service, and travelers leave tips even if they have no intention to return, Kahhneman et al. (1986). In contrast, Kowrtnik et al. (2009) correlate higher service ratings with tips in the leisure cruise industry. Introduction
4/31
Tipping and service quality (continued) What is the definition of “service?”
Lynn (2009) surveyed female waitresses and found out that tipping rates increased with their breast sizes and decreases with their ages, waist-to-hip ratios, and body sizes.
ROCU (2014) reports that women living off tips in states with a $2.13 hourly tipped minimum wage are twice as likely to experience sexual harassment than woman in states that pay the full minimum wage to all workers.
Introduction
5/31
Observations and motivation It is hard to avoid paying tips in America To avoid paying tips, one must 1. refrain from sit-down/family restaurants, 2. take public transportation, 3. stick to “do it yourself” repairs, 4. turn haircuts into family entertainment.
Tipping in U.S. sit-down restaurants On an average day, 10% of the population eats at sit-down/family restaurants. 58% do so in a month. Tips amount to over $46.6 billion a year.
Social-norm assumption The paper assumes no relationship between tip-giving and service quality. Consumers leave a certain percentage because of a social norm, see a few papers by Azar references therein. Introduction
6/31
History of tipping Where did this “social norm” come from? Hardly existed in the U.S. before the Civil War. May be attributed to emancipation of the large number of ex-slaves who entered the labor market. Evidence of very low wages in passenger trains (expected high tips). Low wages (even negative) in service industries became common. Originated in Europe: Some evidence that British homes in the 17th Century expected guests to tip the servants. American guests were blamed for increasing the size of the tips given to servants in British homes. Please be reminded: Do NOT tip (or feed) this presenter! The jar is on the table...you can’t miss it
Introduction
7/31
Tipping rates in the U.S. Consumer kept raising their tipping rates over the years: 10% was common from the end of the 19th Century until after the Depression (minimum tip size on smaller bills was also common). 15% during the 1980s. 18% from the late 1990s. Now, 20% and even 25% are observed. The higher tipping rates provide additional evidence on the weak relationship between tipping and service quality.
Introduction
8/31
Literature on tipping and market performance
Schwartz (1997) applies a demand-supply model to investigate the effect of tips on profits. Azar (2012) constructs a monopoly-monopsony model to investigate the effects of tipping (vs. service charge) on waiters’ efforts. Shy (2015) analyzes a variety of market structures and shows that an increase in the tipping rate does not translate into the same rate of increase in servers’ tip-inclusive income.
Introduction
9/31
Goals and outline Goals 1. To investigate a policy whereby income collected from tips is replaced by a higher hourly minimum wage for tipped employees. 2. Whether such a raise could take the form of closing the gap between the tipped and non-tipped minimum wages.
Outline Explore state-level data on minimum wages. Construct a model of tipping analyzing two market structures: imperfectly competitive service providers and local monopolies. Compare the equilibria with and without tipping (w.r.t. servers’ income, profit of business owners, and consumer prices). State-level calibrations of hourly tip income and comparing those to the gap between tipped and non-tipped minimum wages. Introduction
10/31
Washington Massachusetts California Vermont Arizona New_York Oregon Colorado Hawaii Connecticut Rhode_Island Maine Alaska Minnesota Michigan Maryland Nebraska South_Dakota
Contiguous United States (48 states)
West_Virginia New_Jersey Arkansas Ohio Montana Nevada Illinois Florida Delaware Missouri New_Mexico Wyoming Wisconsin Virginia Utah Texas Tennessee South_Carolina Pennsylvania Oklahoma North_Dakota Min wage
North_Carolina
Min wage tipped
New_Hampshire Mississippi
Difference
Louisiana Kentucky Kansas Iowa Indiana Idaho Georgia Alabama
Data
0
1
2
3
4
5
6
Dollars per hour
7
8
9
10
11
12 11/31
11
11
10
10
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
Min wage (tipped)
Data
Difference (gap)
Tipped minimum wage Median=$3.30, Mean=$4.20 IQR=$2.13–$4.99 Min=$2.13, Max=$11.50 Difference (gap) Median=$5.12, Mean=$4.29 IQR=$3.40–$5.12 Min=$0, Max=$7.50 Non-tipped minimum wage Median=$8.25, Mean=$8.49 IQR=$7.25–$9.79 Min=$7.25, Max=$11.50
Min wage (non-tipped)
12/31
11.5
WA
11.0
10.5
CA OR
10.0 AK
9.5 HI
9.0
8.5
Minimum wage tipped (dollars per hour)
NV
8.0 MN
7.5
AZ CO
7.0
6.5 CT
6.0
5.5 VT
FL
5.0
ME
IL
ND
4.5 IA
SD
4.0
OH MT
MO
RI MA
MD MI
3.5 ID NH
3.0
NY
PA AR
2.5 WI
2.0
Data
WY SC UT VA TN TX OK NC MS KY KS LA IN GA AL
DE
NM
7.5
8.0
WV
NJ
8.5
NE
9.0
(Tipped min wage = 0.517 non-tipped min wage)
9.5
10.0
Minimum wage (dollars per hour)
10.5
11.0
11.5
13/31
The service market Notation
2 restaurants, i = A, B, serving q A and q B meals, respectively. p i = price of a meal by restaurant i, i = A, B. τ = tipping rate. e.g., τ = 0.2 =⇒ 20% rate =⇒ $τ p i total tip. (1 + τ )p i = a buyer’s tip-inclusive price, i = A, B. Restaurants are located on the opposite ends of the unit street [0, 1]. Restaurant A at point 0, restaurant B at point 1. n consumers uniformly distributed on the unit interval [0, 1].
The Model
14/31
The service market Consumer preferences
Utility of a consumer located at x A if chooses restaurant A v − (1 + τ )p − δx B Ux = v − (1 + τ )p − δ(1 − x) if chooses restaurant B v¯ = 0 otherwise, Choose restaurant A
0
x
Neither A nor B
xB
Choose restaurant A
0
Choose restaurant B-
A
Choose restaurant B A
x = xˆ = x
B
1 -
1
x x
Definition: The restaurant market is said to be: consisting of local monopolies if x A < x B , and (imperfectly) competitive if x A = xˆ = x B . The Model
15/31
The service market Restaurants, employment, and profits
Choose restaurant A
0
x
Neither A nor B
Choose restaurant B-
A
xB
Choose restaurant A
0
Choose restaurant B A
x = xˆ = x
1 -
1
B
x x
Restaurants’ output levels: q A = nx A and q B = n(1 − x B ). `A and `B employment levels measured in labor hours. Production of meals: q A = k`A and q B = k`B . Hence, `A = q A /k and `B = q B /k. w = hourly wage (or minimum wage). Profits: Each restaurant solves: max π A = p A q A − |{z} w `A | {z } pA revenue
The Model
labor cost
and
max π B = p B q B − w `B . pB
16/31
Competitive market with tipping (superscript “1”) Equilibrium
Competitive: x A = x B = xˆ. Let w1 = min wage for tipped employees. p1i =
kδ + (1 + τ )w1 , (1 + τ )k
`i1 =
n , 2k
q1i =
n , 2
π1i =
nδ . 2(1 + τ )
Hourly tip income: tip h1 =
τ p1i q1i τ · revenue . = i labor hours `1
Effective hourly wage (tip-inclusive): effective hourly wage
z
}| { τ [kδ + w (1 + τ )] kδτ + (1 + τ )2 w1 1 w1e = w1 + tip h1 = w1 + = . 1 {z +τ 1+τ | } tips per labor hour
The Model
17/31
Competitive market with tipping (superscript “1”) Results
Holding wage w1 constant, 1(a) Equilibrium price and profit of each restaurant i = A, B decreases with the tipping rate. Formally, ∂p1i /∂τ < 0 and ∂π1i /∂τ < 0. 1(b) Consumers’ tip-inclusive price increases with the tipping rate. Formally, ∂[(1 + τ )p1i ]/∂τ > 0. 1(c) Effective hourly wage increases with the tipping rate. Formally, ∂w1e /∂τ > 0. Result 1(a) implies that restaurant owners’ price and profit decrease with the tipping rate. Why is that? Because higher tips weaken price competition among restaurants!
The Model
18/31
Competitive market with no tipping (“2”) Equilibrium
Consider a change in social norm! No tipping, τ = 0. Instead, let’s raise the hourly wage from w1 to w2 ≥ w1e (same as the tip-inclusive effective wage before the change). Servers are no worse off now compared with the tipping equilibrium. p2A = p2B =
The Model
kδ + w1e , k
π2A = π2B =
nδ , 2
tip h2 = 0.
19/31
Competitive market with no tipping (“2”) Main results: Comparing no-tipping with tipping equilibria
p2i − (1 + τ )p1i =
δτ nδτ > 0 π2i − π1i = > 0. 1+τ 2(1 + τ )
When transitioning from a tipping equilibrium to a non-tipping equilibrium while raising servers’ wages to maintain the same effective wage level: 2(a) Customers pay higher prices (p2i > (1 + τ )p1i ), 2(b) Employers earn higher profits (π2i > π1i ), and 2(c) Servers earn the same or higher effective wages (w2 ≥ w1e ), and therefore could be made better. Conclusions: The model predicts that substituting tip income with higher wages could make both, restaurant owners and servers, better off. Consumers end up paying higher prices (more surplus is extracted). The Model
20/31
Local monopolies market with tipping (“1”) Equilibrium and results
Choose restaurant A
0
x
Neither A nor B
xB
Choose restaurant A
0 p1i =
Choose restaurant B-
A
Choose restaurant B A
x = xˆ = x kv + (1 + τ )w1 , 2(1 + τ )k
`i1 = n
1 -
1
B
kv − (1 + τ )w1 , 2k 2 δ
x x
kv − (1 + τ )w1 , 2kδ [kv − (1 + τ )w1 ]2 . π1i = n 4k 2 δ(1 + τ ) q1i = n
Now, with local monopolies, output and labor hours vary with wages and tipping rate: ∂`i1 /∂τ < 0 and ∂q1i /∂τ < 0, i = A, B. Intuition: Holding wage w1 constant, higher tip-inclusive price reduces demand. The Model
21/31
Local monopolies market with tipping (“1”) Results
3(a) Both, employment and output decline with the tipping rate. Formally, ∂`i1 /∂τ < 0 and ∂q1i /∂τ < 0 1(a) Equilibrium price and profit of each restaurant i = A, B decreases with the tipping rate. Formally, ∂p1i /∂τ < 0 and ∂π1i /∂τ < 0. 1(b) Consumers’ tip-inclusive price increases with the tipping rate. Formally, ∂[(1 + τ )p1i ]/∂τ > 0. 1(c) Effective hourly wage increases with the tipping rate. Formally, ∂w1e > 0. Result 1(a,b,c) are the same as under competition. However, under local monopolies, employment and output decline (because demand decreases with higher tip-inclusive income).
The Model
22/31
Local monopolies market with no tipping (“2”) Equilibrium
Consider a change in social norm! No tipping so τ = 0. Instead, let’s raise the hourly wage from w1 to w2 ≥ w1e (same as the tip-inclusive effective wage before the change). Servers are no worse off now compared with the tipping equilibrium. p2i =
The Model
kv + w1e , 2k
`i2 = n
kv − w1e , 2k 2 δ
kv − w1e , 2kδ (kv − w1e )2 π2i = n , 4k 2 δ
q2i = n
tip h2 = 0,
23/31
Local monopolies market with no tipping (“2”) Main results: Comparing no-tipping with tipping equilibria
`i2 −`i1 = nτ
(1 + τ )w1 − kv < 0, 4k 2 δ(1 + τ )
p2i −(1+τ )p1i = π2i − π1i = nτ 2
τ [kv − (1 + τ )w1 ] > 0, 4k(1 + τ )
[kv − (1 + τ )w1 ]2 > 0. 16k 2 δ(1 + τ )2
When transitioning from a tipping equilibrium to a non-tipping equilibrium while raising servers’ wages to maintain the same effective wage level: 4(a) Results in some loss of employment (`i2 < `i1 ), [see also: Wessels (1997), Even & Macpherson (2014), Allegretto & Nadler (2015)], 2(a) Customers pay higher prices (p2i > (1 + τ )p1i ), 2(b) Employers earn higher profits (π2i > π1i ), and 2(c) Servers earn the same or higher effective wages (w2 ≥ w1e ), and therefore could be made better. Remark: Reference to Result 2 means no qualitative differences between the competitive and the local monopolies market structures. The Model
24/31
Goals
Calibrations
(a) Explore the possibility of substituting servers’ tip income with higher hourly wage. (b) In particular: by raising the tipped minimum wage to the non-tipped level (closing the gap).
Available data for each U.S. state (a) Tipped and non-tipped minimum wages. (b) Sales revenue in eating and drinking places. (c) Number of jobs in eating and drinking places.
Unavailable data (a) Demand elasticity for these services and production functions.1 (b) Hours servers work each week. [Simulate 20, 30, 40, 50, 60, 70]. (c) Exact tipping rate. [Simulate τ = 5%, 10%, 15%, 20%] 1
Hence, unable to estimate price increase following wage increase.
Calibrations
25/31
Calibration method Step I: Multiply annual sales revenue (pq) by 4 possible tipping rates (τ = 5%, 10%, 15%, 20%) to obtain yearly income from tips (τ pq). Step II: Divide the result from Step I by the number of employees to find the annual server tip income. Step III: Divide the result from Step II by 52 weeks to obtain a server’s weekly income from tips. Step IV: Divide the result from Step III by 6 possible number of hours a typical server works each week: 20, 30, 40, 50, 60, and 70 hours to obtain hourly tip income. Step V:Average Step IV across the 50 (or 15) states. Step VI: Compare the result from Step V to the actual median wage gap between the tipped minimum wage to the non-tipped minimum wage. Calibrations
26/31
Calibration results (50 states) Recalling the gap between the tipped and non-tipped minimum wages: Median=$5.12, Mean=$4.29, IQR=$3.40–$5.12, Min=$0, Max=$7.50. Compare with the calibrated hourly income from tips: 20 hr
30 hr
40 hr
50 hr
60 hr
70 hr
5% tip
2.59
1.73
1.30
1.04
0.86
0.74
10% tip
5.19
3.46
2.59
2.07
1.73
1.48
15% tip
7.78
5.19
3.89
3.11
2.59
2.22
20% tip
10.37
6.91
5.19
4.15
3.46
2.96
Hourly tip income > $5.12 = wage gap (Servers lose from this transition) Hourly tip income < $5.12 = wage gap (Servers gain from this transition) Calibrations
27/31
Calibration results (15 states) 15 states practice the Federally-mandated minimum wages $2.13 and $7.25: Alabama, Georgia, Indiana, Kansas, Kentucky, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Utah, Virginia, and Wyoming. 20 hr
30 hr
40 hr
50 hr
60 hr
70 hr
5% tip
2.49
1.66
1.24
0.99
0.83
0.71
10% tip
4.97
3.31
2.49
1.99
1.66
1.42
15% tip
7.46
4.97
3.73
2.98
2.49
2.13
20% tip
9.94
6.63
4.97
3.98
3.31
2.84
Hourly tip income > $5.12 = wage gap (Servers lose from this transition) Hourly tip income < $5.12 = wage gap (Servers gain from this transition) Calibrations
28/31
NY
MA
RI
NJ
NE
WV
MI
DE
AR
VT
MD
NM
WY
TX
UT
VA
TN
SC
OK
LA
NC
MS
IN
KY
KS
GA
WI
AL
ME
PA
SD
MT
0
NH
5000
0.0
OH
10000
0.5
MO
15000
1.0
IL
20000
1.5
ID
2.0
FL
25000
CT
30000
2.5
IA
35000
3.0
AZ
40000
3.5
CO
45000
4.0
HI
50000
4.5
ND
55000
5.0
CA
5.5
MN
60000
WA
65000
6.0
AK
70000
6.5
NV
75000
7.0
OR
7.5
Sales revenue per employee (dollars per year)
Gap between tipped and non-tiped minimum wages (dollars per hour)
Do state tip credits correspond to actual tip income?
State code
Do state tip credits correspond to actual tip income?
29/31
Behavioral aspects if tips are abolished Consumers
Danny Meyer did that
1. Price transparency (possible increase). 2. Reduced tension associated with service evaluation and guilt effects. 3. Reduced unneeded service.
Servers 1. Reduce income uncertainty. 2. Reduce pressure to please customers. 3. Reduce harassment by customers.
i 15 NYC restaurants
Business owners (employers) 1. Increased marginal cost leading to price increase. 2. Reduce discrimination against cooks and dishwashers. Conclusion
30/31
Summary 1. With tipping, employers incur “artificially-low” marginal costs. 2. Without tipping, employers incur actual marginal costs, hence 3. they raise prices. 4. Model and calibrations show conditions under which servers and employers may become better off under this transition.
Conclusion
31/31
