Tipping in Restaurants and Around the Globe: An Interdisciplinary Review Michael Lynn Introduction On an average day, approximately ten percent of the U.S. population eats at sitdown/family restaurants. In an average month, approximately 58% do so (Media Dynamics, 2001).
After completing their meals, almost all of these restaurant diners leave a voluntary gift of money (or tip) for the server who waited on them (Speer, 1997).These tips, which amount to approximately $21 billion a year, are an important source of income for the nation’s two million waiters and waitresses (Lynn, 2003b). In fact, tips sometimes represent 100 percent of waiters and waitresses take home pay, because tax withholding eats up all of their hourly wages (Mason, 2002). Of course, tipping is not confined to restaurant servers or to the United States.
In the U.S., consumers also tip barbers, bartenders, beauticians, bellhops, casino croupiers, chambermaids, concierges, delivery persons, doormen, golf caddies, limousine drivers, maitre-d’s, masseuses, parking attendants, pool attendants, porters, restaurant musicians, washroom attendants, shoeshine boys, taxicab drivers, and tour guides among others (Star, 1988).Although not as common as in the U.S.
, tipping is also practiced in most countries around the world (Putzi, 2002). In fact, national differences in tipping are a source of uncertainty for many international travelers and local tipping practices are a topic covered in most travel guides. Tipping is an interesting economic behavior, not only because it is widespread and practically important, but also because it is an expense that consumers are free to avoid. Although called for by social norms, tips are not legally required.Furthermore, since tips are not given until after services have been rendered, they are not necessary to get good service in establishments that are infrequently patronized. For this reason, many economists regard tipping as “mysterious” or “seemingly irrational” behavior (e.
g., BenZion and Karni, 1977; Frank, 1987; Landsburg, 1993). The present chapter explores this behavior and its implications for economic theory and public policy. The chapter is divided into four sections.The first two sections provide more detail about the phenomenon of tipping by summarizing and discussing the results of empirical research on the determinants and predictors of restaurant tipping and of national differences in tipping customs respectively. Then, economic theories about tipping are reviewed in light of the previously summarized empirical literature.
Finally, the public welfare and policy issues raised by tipping are discussed.Determinants and Predictors of Restaurant Tipping Restaurant tips in the United States vary substantially across dining occasions, dining parties, servers, and restaurants. Numerous studies attempting to explain this variability in restaurant tipping have appeared in the psychology and hospitality management literatures and a few such studies are beginning to appear in the economics literature (e.g.
, Bodvarsson and Gibson, 1994; Bodvarsson, Luksetich and Mcdermott, 2003; Conlin, Lynn and O’Donahue, 2003; Lynn and McCall, 2000a; McCrohan and Pearl, 1991).This research has generally relied upon one or more of the following three methodologies: (1) researchers have stood outside of restaurants and conducted exit surveys of departing patrons about their just completed service encounters and tipping behaviors, (2) researchers have created panels of consumers who agreed to keep diaries of their restaurant dining experiences and tipping behavior, and (3) researchers have recruited restaurant servers to record information about their own behavior, their customers’ characteristics, and the tips those customers leave.Among the variables whose effects on restaurant tipping have been studied using these methodologies are bill size, payment method, dining party size, service quality, server friendliness, server sex, customer sex, customer patronage frequency, customer ethnicity, and various interactions between these variables. The results of this research are briefly reviewed in the paragraphs below.Bill Size Social norms in the United States call for tipping restaurant servers 15 to 20 percent of the bill, so it should not be surprising that dollar tip amounts are positively related to bill size. What may be surprising is how strong this relationship is.
In a quantitative review of 36 studies involving 5, 016 dining parties from over 40 restaurants, Lynn and McCall (2000b) found that 69 percent of the average within-restaurant variability in dollar tip amounts can be explained by bill size alone.This suggests that bill size is twice as powerful as all other factors combined in determining dollar tip amounts within restaurants. Of course, the effects of bill size are not invariant. Research suggests that bill size predicts dollar tip amounts better when the tipper is a regular patron of the restaurant (Lynn and Grassman, 1990), the tipper has higher income and education (Lynn and Thomas-Haysbert, 2003), and the tipper is Asian or White as opposed to Black or Hispanic (Lynn and Thomas-Haysbert, 2003). It is possible that these variables moderate the relationship between dollar tip amount and bill size because they reflect differences in awareness of the restaurant tipping norm.
Supporting this possibility, one study found that Blacks are half as likely as Whites to know that the customary restaurant tip is 15 to 20 percent of the bill and additional, unreported analyses of that study’s data indicated that awareness of the norm increases with income and education (Lynn, 2004b). While dollar tips increase with bill size, percentage tips decrease with bill size (Green, Myerson and Schneider, 2003). This effect – known as the “magnitude effect in tipping” -- is due to a positive intercept in the relationship between dollar tips and bill sizes rather than to a marginal decrease in the positive relationship between these two variables (Lynn and Sturman, 2003).The positive intercept has been attributed to: (1) a tendency to leave a minimum tip when bill size is very small (Lynn and Bond, 1992), (2) a tendency to add a constant amount for the mere presence of the server to the standard percentage tip (Green, et al, 2003), (3)) a tendency for some people to be “flat dollar tippers” while others are “percentage tippers” (Lynn and Sturman, 2003), and (4) a tendency to round-up tip amounts (Azar, 2004a). Of these explanations, however, only the “flat dollar tipper” explanation has received any empirical support. National surveys indicate that about 20 percent of restaurant tippers leave a flat dollar amount rather than a percentage of the bill (Paul, 2001; Speer, 1997) and a computer simulation by Lynn and Sturman (2003) demonstrated that this fact is sufficient to produce the magnitude effect in tipping.
Payment Method Restaurant patrons paying with credit cards generally leave larger bill-adjusted or percentage tips than do those paying with cash (Feinberg, 1986; Garrity and Degelman, 1990; Lynn and Latane, 1984, Lynn and Mynier, 1993). These credit card effects on tipping could be due to: (1) the reduced psychological cost of delayed payments, (2) pre-existing differences between cash and credit-card customers, and/or (3) conditioned responses to credit-card stimuli (Feinberg, 1986). Consistent with the latter of these explanations, McCall and Belmont (1996) found that people tipped more when the bill was presented on tip trays embossed with credit card insignia than when it was presented on plain tip trays and that this effect occurred even when people paid the bill with cash.Dining Party Size Large dining parties leave smaller percentage tips than do small dining parties (Freeman, Walker, Borden and Latane, 1975; Lynn and Latane, 1984; May, 1980). This effect has been attributed to: (1) a diffusion of the shared responsibility that each group member has for the server (Freeman, et al, 1975), (2) an equitable adjustment for the smaller per-person effort involved in waiting on larger tables (Snyder, 1976), (3) a cost-reducing adjustment for the larger bill sizes acquired by larger tables (Elman, 1976), and (4) a statistical artifact produced by a positive intercept in the relationship between dollar tips and bill sizes (Lynn and Bond, 1992). Of these explanations, only the statistical artifact explanation has been empirically supported (see Lynn and Bond, 1992).
Service Quality Dining parties that rate the service highly leave larger tips than those who rate the service less highly (Lynn and McCall, 2000a). Furthermore, this relationship remains statistically significant even after controlling for customers’ food ratings, customer patronage frequency, and many other variables (Conlin, et al, 2003). The robustness of the effect after controlling for many potential confounds suggests that it is causal – i.e., that receiving better service causes people to leave larger tips. Despite its reliability and robustness, however, the service-tipping relationship is weak (see Bodvarsson and Gibson, 1999; Bodvarsson, Luksetich and McDermott, 2003; Lynn, 2000c, 2004c).
Customer service ratings account for only 1 to 5 percent of the within-restaurant variability between dining parties in tip percentages (Lynn and McCall, 2000a). Similarly weak relationships between service and tipping have been observed at the server and restaurant levels of analysis (Lynn, 2003b). Several studies have examined potential moderators of the service-tipping relationship. A quantitative review of those studies testing the service by patronage frequency interaction found that the effects of service on tipping do not vary with the tipper’s frequency of restaurant patronage (see Lynn and McCall, 2000a).However, studies testing other interactions have found that the effect of service on tipping is moderated by customer ethnicity (Lynn and Thomas-Haysbert, 2003) and day of the week (Conlin, et.
al., 2003). Changes in service ratings are associated with larger changes in tip percentages among Asians and Hispanics than among Blacks and Whites. Changes in service ratings also have a bigger effect on weekday tip percentages than on weekend tip percentages. This latter effect may be attributable to the greater control over service delivery that servers have on weekdays (which are comparatively slow) than on weekends.
Supporting this logic, Seligman, Finegan, Hazelwood and Wilkinson (1985) found that pizza delivery drivers received lager tips for faster deliveries, but only when the tipper believed the driver was personally responsible for the delivery time.Server Friendliness Although service ratings are only weakly related to tip percentages, server friendliness is a moderately strong predictor of tipping. Studies have typically found that servers’ verbal and non-verbal signals of friendliness increase tip percentages by 20 to 40 percent or more (Lynn, 1996, 2003b). For example, servers receive larger percentage tips when they: (1) introduce themselves by name (Garrity and Degelman, 1990), (2) repeat customers’ words when taking food orders (vanBaaren, et al, 2003). (3) touch customers lightly on the arm, hand or shoulder (Crusco and Wetzel, 1984; Hornik, 1992; Lynn, Le and Sherwyn, 1998; Stephen and Zweigenhaft, 1986), (4) give customers big, open mouthed smiles (Tidd and Lockard, 1978), (5) squatt-down next to the table during interactions with customers (Davis, et al, 1998; Lynn and Mynier, 1993), (6) entertain customers with games or jokes (Guegen, 2002;Rind and Strohmetz, 2001b), (7) draw smiley faces or other pictures on the back of checks (Guegen and Legoherel, 2000; Rind and Bordia, 1996), (8) write “Thank You” or other messages on the backs of checks (Rind and Bordia, 1995; Rind and Strohmetz, 1998), and (9) call customer by their names when returning credit card slips to be signed (Rodrigue, 1999).
All of these studies involved random assignment of dining-parties to the different treatments, so they provide fairly strong evidence that tipping is affected by servers’ rapport with customers.Server and Customer Sex Men sometimes leave larger tips than do women (e.g., Crusco and Wetzel, 1984; Lynn and Latane, 1984) and waitresses sometimes receive larger tips than do waiters (e.
g., Davis, et al, 1998), but these sex effects on tipping are not always found (Lynn and Graves, 1996; Lynn and Simons, 2000).It appears that the effect of customer sex on tipping depends on server sex and vice versa. In an unpublished quantitative review of the tipping literature, Lynn and McCall (2000b) found that men tipped more than women in studies where the server was female while women tipped more than men in studies where the server was male.
Furthermore, Conlin, Lynn and O’Donohue (2003) found a significant interaction between server and customer sex such that women tipped more than men when the server was male but not when the server was female. These findings suggest that tipping is affected by the dynamics of sexual attraction.Customer Patronage Frequency The regular patrons of a restaurant base their tips on bill size more than do new or infrequent patrons (Lynn and Grassman, 1990; Lynn and McCall, 2000b), perhaps because they are more familiar with the 15 to 20 percent restaurant tipping norm. They also tend to leave larger average tips than do infrequent patrons (Lynn and McCall, 2000a). This latter effect remains significant even after controlling for customers’ ratings of the food and service (Conlin, et. al.
, 2003; Lynn and Grassman, 1990), so regular customers do not tip more merely because they perceive the food and service more positively than do infrequent customers. Instead, regular patrons may tip more because they are more likely to identify with servers or because they value servers’ approval more than do infrequent patrons.Customer Ethnicity Black restaurant patrons are more likely than White patrons to tip a flat amount rather than a percentage of the bill. Blacks also leave smaller average restaurant tip percentages than do Whites. This latter effect remains sizable and statistically significant after controlling for education, income and perceptions of service quality, so Black-White differences in tipping are not due solely to socio-economic differences or to discrimination in service delivery (Lynn and Thomas-Haysbert, 2003; Lynn, 2004). Instead, they may be due to ethnic differences in familiarity with the restaurant tipping norm.
Consistent with this possibility, Lynn (2003) found that Whites were twice as likely as Blacks (71% vs 37%) to know that the customary restaurant tip in the United States is 15 to 20 percent of the bill amount.Miscellaneous Among the other variables positively related to bill-adjusted tip amounts in at least some studies are: (1) alcohol consumption (Conlin, et al, 2003; Lynn, 1988; Sanchez, 2002), (2) sunny weather or forecasts of sunny weather (Cunningham, 1979; Crusco and Wetzel, 1984; Rind and Strohmetz, 2001a), (3) metropolitan area size (Lynn and Thomas-Haysbert, 2003; McCrohan and Pearl, 1983, 1991), (4) customer income (Lynn and Thomas-Haysbert, 2003; McCrohan and Pearl, 1983) (5) customer youth (Conlin, et al., 2003; Lynn and Thomas-Haysbert, 2003; McCrohan and Pearl, 1983), (6) customer ratings of food quality (Lynn and McCall, 2000a), (7) server personality – i.e., self-monitoring (Lynn and Simons, 2000), (8) server physical attractiveness (Hornik, 1992; Lynn and Simons, 2000; May 1980), and (9) server adornment – i.e.
, wearing flowers in hair (Stillman and Hensley, 1980).Predictors of National Differences in Tipping Norms Tipping varies across nations in terms of who it is customary to tip and how much it is customary to tip them. A handful of studies in the psychology and hospitality management literatures have attempted to measure these national differences in tipping norms and to examine their relationships with other variables. The most commonly studied measure of national tipping norms is the number of different service providers (out of a list of 33) that it is customary to tip in a nation. I shall refer to this measure as the national prevalence of tipping.
Two other measures of national tipping norms are the amounts in percentages of the bill or fare that it is customary to tip restaurant servers and taxicab drivers. I shall refer to these measures as national restaurant and taxicab tiprates respectively. All of these measures of national tipping norms are based on content analyses of international tipping guidebooks. Research on the predictors of these measures has generally focused on national character – i.
e., national values, motives and personality traits. This focus rests on the assumption that tipping norms are primarily determined by consumers.Consumer acceptance of these norms is theorized to vary with the value that consumers place on the consequences or functions of tipping. Thus, researchers have examined the relationships between national tipping norms and national character traits relevant to those consequences and functions. The results of this research are briefly reviewed in the paragraphs below.
