Abstract
We investigated the link between tipping, an altruistic act, and bribery, an immoral act. We found a positive relationship between these two seemingly unrelated behaviors, using archival cross-national data for 32 countries, and controlling for per capita gross domestic product, income inequality, and other factors. Countries that had higher rates of tipping behavior tended to have higher rates of corruption. We suggest that this surprising association may be accounted for by temporal focus—people may tip and bribe others in order to receive special services in the future. Indeed, in a pair of follow-up survey studies, we find evidence that the link between tipping and bribery can be partly accounted for by prospective orientation.
Which country is more likely to be plagued by corruption: one where people generously tip their waiters, or one where they walk away without leaving any gratuity at all? Many people might assume the latter, but we advance the counterintuitive idea that the former is more likely to be true. It may seem strange to associate corrupt behaviors, such as bribery, with altruistic behaviors, such as tipping. But, in fact, tips and bribes can possess striking similarities that may lead to their positive association. In a sense, both are gifts intended to strengthen social bonds and each is offered in conjunction with advantageous service. One could even argue that the main difference between the two acts is merely the timing of the gift: Tips follow the rendering of a service, whereas bribes precede it.
Although the line between tips and bribes is clear in modern-day Western societies, an examination of the historical origins of tipping practices paints a blurrier picture (see Lynn, Zinkhan, & Harris, 1993). According to one perspective, tipping customs trace back to the Middle Ages, when feudal lords traveling beyond their territory would toss coins to beggars whom they passed along the way—the lords hoped that such an act of kindness would help ensure safe passage (Schein, Jablonski, & Wohlfahrt, 1984). According to another perspective, tipping originated from the custom of “vails,” which was popular in Tudor England; guests who stayed overnight were expected to leave payment for their hosts’ servants after their stay, to compensate them for the extra work created by their visit (Shamir, 1984). While the latter perspective may appear more in line with popular connotations of tipping (the gratuity follows the rendering of a service), the former may appear more in line with the notion of a bribe (the gratuity precedes the rendering of a service).
Although we expect to find a connection between tipping and bribery, we also expect that this connection may not seem intuitive to the layperson. As evidence of this, we asked 51 participants from a national online pool to consider several variables that might be related to the incidence of bribery in a particular country and to describe the nature of this relationship using four options: (1) probably positively related, (2) probably negatively related, (3) probably not related, and (4) I don’t know. The list of variables included gross domestic product (GDP), income inequality, rate of violent crime, tipping frequency, and a host of other factors. For the relationship between tipping and bribery, just 5.9% of the respondents (n = 3) thought they were “probably positively related.” Another 11.8% (n = 6) thought they were “probably negatively related,” and 78.4% (n = 40) thought they were “probably not related.” The remaining 3.9% (n = 2) indicated they did not know.
In the current research, we examine whether the link between tipping and bribery differs from these intuitive predictions. Specifically, we expect to find a positive association between the two practices. We also propose a possible psychological explanation for this surprising link—an individual’s temporal focus. Psychologists have studied temporal focus in several ways: investigating the relative influence of distant-future versus near-future actions (Trope & Liberman, 2003), demonstrating systematic biases in representations of future events (Liberman & Trope, 1998), and identifying the distortion of expectations based on past experience (Johnson & Sherman, 1990). Here we investigate whether temporal focus in the form of retrospective versus prospective differences can account for a connection between bribery and tipping. In short, to the extent that people provide tips to encourage better service in the future (as opposed to, say, rewarding good service in the past), we expect to find a positive connection between tipping and bribery.
Study 1
To test our hypothesis about the connection between tipping and bribery, we use archival data about both practices drawn from independent sources. Our data on national differences in tipping practices come from a previous study (Lynn et al., 1993) that estimated the prevalence of tipping in 32 countries. Using data gathered from The International Guide on Tipping (Star, 1988), Lynn, Zinkhan, and Harris coded 33 different service professions (e.g., waiters, cabdrivers, hairdressers, etc.) in each country as likely to be “tipped” or “not tipped” (1 = tipped; 0 = not tipped). We include every country for which data are available and construct our index using the same method used by Lynn et al. (1993): The tipping score is equivalent to the number of professions coded as likely to be “tipped.” A higher score indicates a greater prevalence of tipping in the focal country (the scores range from 3 to 31).
We link these data on tipping practices to each country’s score on the Corruption Perception Index (CPI), a ranking produced by Transparency International (2005) that has been used in past research on national differences in corruption (e.g., Fisman & Miguel, 2005). The CPI is calculated using data from 13 sources originated from 11 independent institutions (e.g., Global Insight and Merchant International Group). Country experts (both resident and nonresident) are asked to assess the frequency and magnitude of bribery in the public and political sectors. Reliability is generally strong as demonstrated by the high average correlation between sources (.78). We use a reverse coding of the CPI index as our measure of corruption. The range of this variable is 1 to 10, with 1 indicating almost no corruption and 10 indicating extremely high levels of corruption.
Table 1 presents correlations and summary statistics for the variables we include in our model. As shown in the table, the correlation between tipping and corruption is positive and statistically significant (r = .60, p < .05). Figure 1 depicts the link between tipping prevalence and corruption levels in more detail. Further, we present the results of linear regression models in Table 2. The table shows both unstandardized coefficients and standardized (β) coefficients for the estimated effect of tipping on corruption (we report only the unstandardized coefficients reported for other covariates to preserve space). For all coefficients, we report White standard errors, which are robust to heteroscedasticity.
Country level association between tipping and corruption.
Correlations and Summary Statistics.
Note. GDP = gross domestic product.
*Denotes significant correlations, p < .05.
Country Level Predictors of Corruption.
Note. GDP = gross domestic product. Unstandardized regression coefficients are reported, with robust (White) standard errors in parentheses. β coefficients and p values reported in boldface for 33-Profession Tipping Index.
*p < .05. **p < .01.
We first estimate a model including only our measures of tipping prevalence and corruption. The estimated coefficient of the 33-Profession Tipping Index is b = .18 (β = .65, p < .01), meaning that one additional profession receiving tips is associated with an increase of .18 on the corruption scale. We recognize there might be alternative explanations for the association between tips and bribes. For example, differences in national cultures could potentially drive variation in the prevalence of both tipping and bribery. Indeed, Mazar and Aggarwal (2011) found that collectivism is positively associated with the prevalence of bribery across nations. We therefore include Hofstede’s (2001) Individualism Index as a control variable. We also include another one of Hofstede’s cultural dimensions, the Power Distance Index, which captures the extent to which power is distributed unequally among a nation’s citizens and the extent to which citizens who are less powerful accept that disparity. Power distance could be a driver of both tipping and bribery practices, given that bribes and tips typically are passed between people who hold different levels of power and influence. Consistent with these ideas, we find that individualism is negatively associated with corruption and that power distance and tipping are positively associated. However, even after controlling for each country’s score on the Power Distance Index and the Individualism Index, the association between corruption and tipping remains positive and significant (b = .08, β = .27, p < .01).
We also recognize that variability in economic necessity could lead to different opportunities and incentives for initiating both tipping and bribery transactions. People who live in wealthier countries might find it easier to part with tips and bribes because they have more resources to spare; alternatively, poor people might have a greater need to solicit both tips and bribes. To rule out this alternative explanation, we use information on both GDP per capita (logged) and income inequality. We obtained information on real GDP per capita from the Penn World Tables 6.3 for the year 2005 (Heston et al., 2009). To capture income inequality, we used the Gini Index (United Nations Development Programme, 2009), a commonly used measure of dispersion. The Gini coefficient is a number between 0 and 1, where 0 indicates perfect income equality (everyone has the same income) and 1 indicates perfect inequality (one person has all the income, and everyone else has zero income). As shown in Model 1, controlling for GDP per capita and income inequality reduces the strength of the relationship between corruption and tipping, but the relationship remains positive and significant (b = .05, β = .18, p < .05).
We also include a number of measures that capture variations in governmental involvement and order, which could be driving this connection between tipping and bribery. In particular, we control for the national homicide rate (United Nations Office on Drugs and Crime, 2009), which is the crime statistic that is most directly comparable across countries. We also control for the impact of civil liberty restrictions (Freedom House, 2009) with a measure that captures the extent to which the government is involved with its citizens in a coercive manner. Finally, we include three variables that capture the extent of governmental welfare policies: We obtained the highest marginal tax rate in each country (Nation Master, 2011), the legal minimum wage (Various Sources, 2011), and the extent to which a state provides public funding of health care as a fraction of total health care expenditures (World Health Organization, 2011). The positive link between tipping and corruption holds when we add these variables to the model (b = .05, β = .17, p < .01).
A final consideration is potential uneven variance, or heteroscedasticity, in the data. Because all regression results use White standard errors, which are robust to heteroscedasticity, this does not cause issues with statistical inference. However, it raises the interesting question of whether tipping prevalence interacts with other characteristics of a country. For example, one might wonder whether tipping attitudes are especially important in lower income countries or those that are less effectively governed, presumably because such countries lack alternative means of suppressing bribery. Given the limited number of observations, and multicollinearity among variables, it is difficult to answer these questions using subgroups or moderation analysis, but this is a promising avenue for future research.
Discussion
The results of Study 1 offer evidence that countries with a higher incidence of tipping also have a higher incidence of bribery and vice versa. While provocative, these results offer little explanation for why tips and bribes might be positively associated. In the next pair of studies, we highlight an underlying psychological mechanism more directly. In particular, we suspect that one link between tipping and bribery is temporal focus—a preference for tipping with a prospective orientation (i.e., to encourage good service in the future) is likely to be associated with higher tolerance for bribery practices.
In Study 2, we begin by isolating two countries in the sample from Study 1 that had nearly identical scores on the tipping index but had very different scores on the corruption (i.e., bribery) index: Canada and India (corruption scores of 2.9 and 8.1, respectively). We suspect that despite the similar prevalence of tipping in the two countries, Canadian and Indian citizens’ attitudes toward the practice are different in one critical respect: Indians are more inclined than Canadians to tip prospectively, that is, in the hope that a tip will bring about better service in the future. We suggest that this prospective orientation can account for cross-national differences in their attitudes toward bribery.
Study 2
Method
A total of 95 Canadian citizens (39 men and 56 women) and 157 Indian citizens (116 men and 41 women) participated in this study. Following are brief descriptions of the measures included in the survey, which was completed in English.
Future tipping motivation was measured using 5 items rated on a 7-point scale ranging from not at all to very much. Participants were asked to think about “all the situations in which you chose to tip someone,” and rate the extent to which each statement described their reasons for tipping. Sample items include “I want to motivate this person to give me good service in the future” and “I want to establish a personal relationship with the person, as this may be helpful down the line.” The items were averaged to create an overall score (Future Tipping Motivation,
Participants were also asked to complete a 5-item bribery attitude measure taken from a Gallup (2006) survey. Each of these items was rated on a 7-point scale ranging from morally acceptable to morally wrong. The Gallup bribery measure instructed participants to evaluate the morality of several purportedly corrupt actions, including “bribing a policeman to avoid a traffic ticket” and “bribing a government official to avoid a tax audit.” The items were averaged to create an overall score (Bribery Attitudes,
Results
Men had reported more favorable attitudes toward bribery than did women, F(1, 249) = 14.52, p < .001. This effect remained significant after controlling for country F(1, 249) = 5.97, p < .05. We therefore included gender as a control variable in each of our analyses.
As expected, when controlling for gender, Indians (
Mediation Analysis
Finally, we tested our hypothesis that Future Tipping Motivation accounts for the effect of nationality on bribery attitudes using the bootstrapping method for testing indirect effects (see Preacher & Hayes, 2004; Shrout & Bolger, 2002). In accordance with our mediational account, the coefficient for the indirect effect of Future Tipping Motivation on Bribery Attitudes was .60, and the 95% bias-corrected confidence interval ranged from .0043 to .1628 (1,000 bootstrap re-samples). Given that the confidence interval excluded zero, we can conclude that a prospective orientation toward tipping mediated the link between nationality and bribery attitudes.
Discussion
The results from Study 2 are consistent with our hypothesis that the link between bribery and tipping can be accounted for by cross-cultural differences in temporal focus, particularly as it relates to tipping practices. These results also suggest a potential extension—If cross-cultural differences in attitudes toward bribery are driven by temporal focus, it seems reasonable to assume that differences in temporal focus, at the individual level, will relate to bribery attitudes in the same way, even for citizens of the same country. Thus, adopting a more prospective focus toward tipping practices should lead directly to a greater tolerance toward bribery practices.
Our final study, then, is intended to examine this possibility by examining the link between temporal focus and attitudes toward bribery using participants from a single country (the United States in this case) in a controlled experiment, in which we manipulate temporal focus, as it relates to tipping practices. In this final study, we also control the context in which individuals are asked to consider gratuities, thereby addressing the possibility that the relationship between tipping and bribery observed in Study 2 may have been due to Canadians and Indians considering systematically different situations. Finally, we have proposed that a prospective focus toward tipping can account for differences in bribery attitudes, but perhaps the same could be said for any temporal focus. To eliminate the possibility that these results could be driven by any temporal focus, we manipulate both prospective and retrospective motivations to tip in our experimental study.
Study 3
Method
A total of 40 undergraduates at a large west coast university (21 females) participated in this study. Participants were given $10 for their time.
Participants were told they would participate in two separate studies, and that the first study involved attention span and reading comprehension. To manipulate temporal focus in tipping motivation, participants were asked to read an article about tipping that was ostensibly published on the Emily Post website (Emily Post is a recognized authority on proper etiquette in the United States). We randomly assigned participants to one of two conditions. In the Future Motivation Condition, the article was titled “Basic Tips on Tipping: Encouraging Good Service” and described how tips can encourage future good service. In the Past Motivation Condition, the article was titled “Basic Tips on Tipping: Rewarding Good Service” and described how tips can reward past good service. For example, the article in the Future Motivation Condition noted that residents in a doorman building should tip the doorman in order to “encourage particularly good service,” while the article in the Past Motivation Condition noted that people should tip the doorman in order to “reward superlative service.”
Participants were then told that they would be taking part in a separate study concerning media and attitudes. At that point, they were asked to watch a 1-min news report produced by SmarTrend News (a company that produces online stock market update videos) in 2010. According to the reporter pictured in the video, the U.S. Justice Department and the Securities Exchange Commission were investigating a possible $10.3 million bribe that Hewlett-Packard paid to Russia to win a €35 million contract. Participants reported the extent to which they thought the bribery described in the video was morally wrong (on a 7-point scale anchored by Not At All to Extremely), how harshly the company should be punished if the company was found guilty of bribery (on a 7-point scale anchored by Not At All to Maximum Possible Punishment), and the extent to which they thought the crime described in the video was severe (on a 7-point scale anchored by Not At All to Extremely). Given that the overall reliability for these 3 items was adequate (.78), we averaged them to create an overarching measure of bribery attitudes.
Participants then completed a separate, validated measure of bribery attitudes (Napal, 2005). In this measure, participants read a vignette that described a wealthy businessman who was convicted of a crime that he claimed he did not commit. According to the vignette, the facts seemed to confirm that the businessman was guilty, and that the man paid a judge to be partial to his case. Participants then reported their attitudes toward the businessman’s actions on seven 7-point scales ranging from Fair to Unfair, Just to Unjust, Culturally Acceptable to Culturally Unacceptable, Violates an Unwritten Contract to Does Not Violate an Unwritten Contract (reverse coded), Traditionally Acceptable to Traditionally Unacceptable, Violates an Unspoken Promise to Does Not Violate an Unspoken Promise (reverse coded), and Acceptable to Family to Unacceptable to Family (Napal, 2005). The overall reliability of these items was .86, and the items were averaged together for analysis.
Participants also completed the Balanced Inventory of Social Desirability Responding (BIDR). The BIDR is a 20-item scale that measures the extent to which respondents may be engaging in conscious and unconscious deception (Paulhus, 1991). For example, the scale includes items that require participants to report whether “I always obey laws, even if I’m unlikely to get caught” and “There have been occasions when I have taken advantage of someone” (reverse coded) on 7-point scales anchored by Not True to Very True. The overall reliability of the scale was .76, and the items were averaged together for analysis.
Results
Independent samples t tests revealed that participants in the Future Motivation Condition (M = 4.68, SD = 1.16) evaluated the bribery described in the video less negatively than did participants in the Past Motivation Condition (M = 5.30, SD = .55), t(38) = 2.14, p <.05. The effect of gender on evaluations of the video were not significant, t(38) = −0.06, ns, and the effect of condition on evaluations of the video remained significant after controlling for gender F(1, 38) = 5.05, p < .05. To rule out the possibility that the observed results were confounded by differences in social desirability between the two conditions, we controlled for BIDR. The effect of condition on evaluations of the reported bribery remained significant, F(1, 38) = 4.37, p < .05.
Independent samples t tests also revealed that participants in the Future Motivation Condition (M = 4.42, SD = .95) valuated the bribery described in the vignette less negatively than did participants in the Past Motivation Condition (M = 5.26, SD = 1.05), t(38) = 2.66, p < .01. The effect of gender on evaluations of the vignette was not significant, t(38) = 0.33, ns, and the results remained significant after controlling for gender F(1, 38) = 6.76, p < .01. Again, to rule out the possibility that the observed results were confounded by differences in social desirability between the two conditions, we controlled for BIDR. The effect of condition on evaluations of the bribery described in the vignette remains significant, F(1, 38) = 7.62, p < .01.
General Discussion
The present research suggests that the line between tips and bribes is fine and, in many cases, blurred. At first blush, tipping and corrupt practices, such as bribery, may seem completely unrelated. However, we suggest that tips and bribes both emanate from similar norms of exchange—indeed, the timing of the gratuity may be the key distinguishing feature between these two acts. This subtle temporal distinction may help explain why tipping and bribery practices are positively correlated across countries even though many individuals perceive them as diametrically opposed from a moral standpoint. The link between tips and bribes appears primarily for those whose tipping is motivated by a prospective orientation (to obtain advantageous service in the future), not for those who have a retrospective orientation (to reward advantageous service in the past). The studies presented here thus highlight a psychological mechanism that may help explain the surprising association between tipping and bribery both within and across countries.
A voluminous body of research has examined the critical role that temporal focus plays in shaping individuals’ attitudes, decision making, and goal setting (Chandran & Menon, 2004; Johnson & Sherman, 1990; Liberman & Trope, 1998; Trope & Liberman, 2003). In the present research, we examine temporal focus in a novel context—social exchange—and suggest that different forms of exchange may be linked by a common prospective orientation. Building on these results, future research might investigate whether broad individual differences in temporal focus (e.g., as measured by the Zimbardo Time Perspective Inventory, Zimbardo & Boyd, 1999), can help explain the positive association between bribery and tipping. In addition, future work might explore whether other temporal differences, such as short-term versus long-term time horizons, can help explain preferences for exchange. Indirect forms of reciprocity (e.g., generalized exchange) may be better suited to a longer term perspective, whereas direct reciprocity may require a short-term view. Further, differences in temporal horizon might account for specific cultural patterns of exchange, such as the Brazilian Jeitinho (Ferreira, Fischer, Barreiros, Pilati, & Milfont, 2011) and the Chinese Guanxi (Chua, Morris, & Ingram, 2009).
While the current results are provocative, they are not conclusive. Future research should gather additional evidence to examine the robustness of our findings and further clarify the underlying psychological mechanism. For example, using an implicit prime to activate a retrospective temporal focus might make people less tolerant of bribery behavior in the future. Along a similar vein, it might be worthwhile for future studies to attempt an intervention. Perhaps encouraging people to provide tips that reward past service, in general, might simultaneously undermine the motivation to bribe government officials, in particular. In sum, much more remains to be learned about the intriguing link between tips and bribes—what we learn could help us gain new insight into these unique forms of exchange.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors gratefully acknowledge financial support from Columbia University, Harvard University and Stanford University.