In Defense of Tipping, Part I: Principal-Agent Problems

The cultural norm of tipping is once again prompting dissatisfaction among consumers, as indicated by a recent topline story in the Wall Street Journal. This is nothing new as the practice has always drawn the ire of the public. Samuel Gompers considered it a form of extortion, an attitude shared by many others. Steve Buscemi’s Reservoir Dogs character monologued for three minutes on why he hated gratuities, reflecting the views of the screenplay’s writers. Some authors have even claimed the norm is a legacy of America’s racist past (which AIER’s Phil Magness demonstrated was incorrect). 

In the mid-2010s, a movement led most famously by restaurateur Danny Meyer tried to eliminate gratuities at restaurants in favor of paying servers a stable “living wage” to the applause of many social critics. Although this trend failed to take off, with many prominent restaurants backtracking due to staff quitting, South Park creators Trey Parker and Matt Stone recently announced that their Casa Bonita restaurant will ban tipping.

Current anger, however, is being channeled in new directions. First, there is the spread of tipping to facets of commercial life where it hasn’t been before, including self-serve yogurt shops. Second, people are stressing over the social pressure created by digital payment methods, prompting one to choose a tip while the wait staff stares at you. 

While much of this frustration is understandable and shared by me (as revealed below), I will defend the cultural norm of tipping as an ingenious institution that helps free markets work better. In a series of three separate articles, I will argue that tipping solves important principal-agent problems, leverages price discrimination to the advantage of customers, employers and employees, and instills critical social values that are essential to market economies. I shall also point out some of the problematic features of gratuities including how changes in public policy and the norm itself are undermining an economically useful social practice.

Of Principals and Agents

The principal-agent (P-A) problem is a central concept in economics. It arises when one individual (principal) employs another person (agent) to perform a task. The principal wants the agent to do the job efficiently and in accordance with the principal’s desires. Unfortunately, the agent may have an incentive to shirk. If the agent is paid by the hour, there is an incentive to work slowly so as to be paid more hours. On the other hand, paying the agent by the task incentivizes the employee to rush the job or cut corners on quality.

The inability to monitor an agent’s performance directly is the main source of the problem. If a manager was to constantly observe an employee it would be a waste of resources since the manager could just do the job herself, eliminating the agent. However, the efficiency gained from the division of labor requires we segment out tasks to different people. As such, the P-A problem is ubiquitous in modern economies. 

Solving a P-A problem requires some creativity. Different compensation methods (e.g., wages, salaries) are used based upon the nature of the task. Technology assists as well. Closed-circuit cameras in workplaces help identify workers who are slacking (or gives them the sense that somebody is always watching). 

Gratuities as a Solution to the P-A Problem

Tipping turns out to be an ingenious mechanism for solving principal-agent problems in a number of service-oriented industries. To understand this, let’s take the most common environment where tips are given – full-service restaurants.

Restaurant owners want their guests to have a pleasant experience. Happy customers generate repeat business and spread good reviews via word-of-mouth recommendations. To satisfy their clientele, restaurateurs want to provide good food (relative to price), a nice ambience, and amiable service that is attentive to customer needs. The first two aspects of the dining experience (food and ambience) are relatively stable and easy for the manager to evaluate. Incorrect food orders are typically called out by the customer, and the ambience is simply what it is.

Table service is the one factor that is highly variable and difficult for management to evaluate continuously. Contrary to the notion that restaurant work is a low-skill job, waiting tables requires a nuanced ability to read subtle signals from diners. Servers need to determine if customers are in a rush to get to a show or want to be left alone on a romantic date. Attention must be paid to refilling water glasses, asking if they would like another drink, and providing recommendations that they believe fit the profile of the customers. Moreover, the wait staff needs to put on an air of helpful friendliness even when diners are irritable or pose other problems (e.g., families with unruly children). Anyone who has ever worked in a restaurant knows this well.

Tipping incentivizes a server to provide friendly, helpful, and customized service knowing that meeting a customer’s implicit needs often yields a larger gratuity. Even if a waiter is hungover or must deal with screaming kids, putting on a pleasant face typically nets a decent tip. Alternatively, a waiter will deliver a minimally tolerable level of service to avoid getting “stiffed” (i.e., no tip). Thus, when the manager cannot constantly monitor the amiability of every server, the possibility of the good tip nudges the wait staff into better behavior that benefits the restaurant’s clientele and owner. An establishment that provides quality service will also receive more business, ceteris paribus, ensuring future employment for the staff. This is a win-win-win situation for diners, owners, and the staff. 

Gratuities also incentivize better food quality. This is because another P-A problem exists between the front-of-house (wait staff) and back-of-house (kitchen). A diner who receives an incorrect order or sub-par food may not tip. As such, waiters have an incentive to monitor what goes on in the kitchen. If an order is wrong, servers want to ensure the kitchen staff will rush the correct entrée. A good waitress often will share tips with the back-of-house to ensure a good working relationship, even if the establishment doesn’t practice “tip pooling” (which is oddly illegal in some states). 

A World without Gratuities

Another way to think about the benefits of tipping is to consider what would happen if servers didn’t receive tips and were instead paid a consistent “living wage.” The sentiment of a “living wage” is nice, but does it consistently incentivize quality service? 

Remember, not all diners are alike. Some are pleasant, while others can be downright irritating. People with persnickety preferences or parents with crying children do not make the life of a server easy. Some customers are in a rush, whereas others like to linger undisturbed. If a waiter gets paid the same whether or not they interact amicably, there will be fewer visits to the table with unpleasant customers. Also, there will be no incentive to hurry along the order of the couple rushing to make the theater. 

Moreover, service will tend to devolve to the lowest common denominator; if all staff get paid (roughly) the same hourly wage, why would anybody work harder than the lowest performing crew member? It takes effort to smile when you have a headache, but if you get paid the same no matter whether you smile or not, why bother?

More surprisingly, by eliminating tips you will tend to get less qualified staff working at the most critical times, something that an experienced restaurateur pointed out to me. Every restaurant manager knows that Thursdays, Fridays, and Saturdays are the busiest dining days. A waitress who can work tables in a crowded restaurant efficiently, and do so pleasantly, can rake in big dollars on just those days alone. 

However, it is hard work to jostle dozens of tables. If one is paid by the hour and didn’t receive tips from each customer served, the incentive would be to work the least busy times Monday through Wednesday. Given that senior staff often get first dibs at choosing shifts, they would take the easier days leaving newbies to work during the most hectic periods. Those inexperienced servers will likely make more mistakes during busy times, angering customers. Earning tips incentivizes the best staff to work the busiest times when experienced servers are critical.

There is another surprising benefit for using tips. Servers who do not have good people skills will tend to self-select out of the restaurant industry. Untalented waiters and waitresses will observe their take-home pay is lower than those doing a superb job. This provides them with the nudge they need to either improve their performance (benefiting both diners and managers) or quit. Given that it is difficult to fire people, it is better to have them leave of their own accord. Without gratuities there is a tendency for lesser-skilled servers to remain in the industry, and that benefits no one.

Where Should We Tip and Where Not?

As noted, tipping is most common in full-service restaurants. But other places exist where we do (and should) tip. Gratuities are most useful in jobs where there are variable demands for customized service that are difficult to monitor – i.e., where principal-agent problems are most intense. Additionally, the more complex a task an agent must perform, the more one should tip. Consumers apparently understand this according to a Morning Consult poll cited in the WSJ.

Consider this. We tip at full-service restaurants where a waitress needs to be attentive to our needs, but not at fast-food outlets where the commercial transaction is rather simple. At McDonald’s, I order a hamburger and can expect a very standardized meal delivered within minutes. We tip a barista if we have a five-adjective, specialized latte, but not if we order a basic cup of coffee, which only requires someone to open a tap on a pot. Danny Meyer was correct when he noted that we should not tip for fast food or coffee.

We tip for pizza delivery, but not for garbage pickup or postal service. In the former case, one wants to ensure the food arrives quickly and hot. We do not have a similar concern for our trash or junk mail, so long as it gets picked up or delivered at some point in the day. 

We offer gratuities to our hair stylists, particularly if we have special treatments, but not cashiers at Wal-Mart. Ringing up your shampoo purchases is not a customized service task, even though some folks might like a little bit of chit-chat in line (often to the chagrin of others). 

Considering all of this, it is not surprising that people are becoming increasingly frustrated with “tip creep” – i.e., the extension of gratuities into new realms. It makes absolutely no sense to tip the cashier at a self-serve yogurt shop. The digital checkout screen prompting you to add an additional 10 – 25% only adds to that frustration and a sense that tipping is nothing more than extortion. If there is no P-A problem, don’t tip. Period.

This reveals another problem with gratuities. The new point-of-sale payment devices becoming increasingly common in restaurants adds social pressure to tipping. It is emotionally harder to look someone face-to-face and deny them a tip. It is better to leave it at the table (preferably in cash for reasons I won’t reveal here) where the waiter will pick it up after the customer has left. This allows for poor service to be penalized with a poor tip. That is how the institution works as a proper signal and incentive. If everybody always left a 20% tip regardless of service quality, which unfortunately some people do, the institution would lose its effectiveness. It is also why providing a tip in advance (e.g., for online pizza delivery) also attenuates the social norm; the level of gratuity should be determined only after the service has been provided.

A Final Principal-Agent Mystery

There is one additional benefit of tipping that relates to the principal-agent problem. If you are a repeat customer of an establishment, it serves you well to be a generous tipper (assuming quality service). Servers remember who treats them well and returns the favor in-kind. In essence, the customer (principal) is signaling to the agent (server) that they want to be treated “differently” in the future, something which relates to price discrimination and my defense of tipping in the next installment.

However, there is an empirical problem. If gratuities guarantee better service for the customer in the future, why would anybody ever leave a tip at a restaurant that they know they would never return to? A possible answer to that puzzle will be revealed in Part III. 

Stay tuned.

Related Articles – Art and Culture, Economic Education, Economic Trends