2. Gifts, deadweight loss, and static inefficiency: Waldfogel and his critics

Nearly all individuals have received a gift that they found little value in or did not want at all. For instance, your aunt may have purchased you a ceramic statue for several hundred dollars as a wedding gift. You and your spouse considered this art piece aesthetically atrocious and would have preferred an envelope with cash. Waldfogel's (Reference Waldfogel1993, Reference Waldfogel2002, Reference Waldfogel2009) reasoning about the inefficiency of gifting quantifies this lost value as deadweight loss. Waldfogel argued there is a significant spread between resources expended by a giver and the value placed on the gift by the recipient, a result of asymmetric information between the two parties about what each person values. He noted:

The implications of this analysis are straightforward: the more information the gift giver has about the recipient's preferences, particularly immediate needs or wants, the more efficient the outcome will be (Waldfogel, Reference Waldfogel1993: 1330). In contrast, the more socially distant a giver is from the recipient (and hence the less information they have), the greater the discrepancy between the value the recipient places on the gift and what the giver paid for it. Old married couples are expected to have higher-quality information about preferences than occasional acquaintances or distant cousins. The informational costs of gift-giving are one main source of inefficiency. Waldfogel supported his argument by surveying college students and showing that recipients place roughly 16% lower value on gifts received than what the giver paid for them, with gifts from more socially distant individuals being approximately 33% less valuable (Reference Waldfogel1993: 1334). From this he concluded that cash is more efficient than in-kind presents. In a static analysis, cash allows the recipient to maximize their want satisfaction, holding the giver's expenditure constant.

Not unexpectedly, critics quickly jumped on this finding. Solnick and Hemenway (Reference Solnick and Hemenway1996) noted the constrained sampling frame of Waldfogel's experiment (i.e. rich, dependent, and relatively homogeneous Yale undergraduates). They further claimed this early experiment neglected measurements of the sentimental value of receiving a gift. They pointed to research by Belk (Reference Belk1976) that ‘many items give recipients higher utility if they are purchased by someone else’ (cited in Solnick and Hemenway Reference Solnick and Hemenway1996: 1299). Sentimental value increases the total value created by the gift, producing a higher value from which any inefficiency due to poor information is subtracted. Expanding the research design to include college students, university staff, and the general public with a greater age range, they found that ‘rather than a deadweight loss … most gifts created positive value. The mean yield was, at minimum, greater than 2, and the median gift increased value by 11 percent’ (Solnick and Hemenway, Reference Solnick and Hemenway1996: 1301). Another study, by List and Shogren (Reference List and Shogren1998), using mixed methods seemed to support both Waldfogel (Reference Waldfogel1993) and Solnick and Hemenway (Reference Solnick and Hemenway1996). List and Shogren's point estimates replicating Waldfogel's study with different undergraduates generated a very mild deadweight loss of 1–2% of the gift's cash value, whereas an auction experiment demonstrated that gifts ‘produced a plausible welfare gain, as real selling prices were at least 27 percent higher than the hypothetical price’ (Reference Waldfogel1998: 1354). It should be noted that all these studies examined gifting as a single-period event without recognizing that it is an iterated experience involving norms of reciprocity.

Waldfogel responded with a follow-up study in 2002 that expanded the sample sizeFootnote ⁶ and set of explanatory and control variables to include race, religion, and frequency of contact between givers and receivers. He found deadweight loss to average roughly 7% of the gift's value (Reference Waldfogel2002: 417). An updated study with alternative measures by Principe and Eisenhauer (Reference Principe and Eisenhauer2008) also found a 7% deadweight loss associated with in-kind gifts and, surprisingly, a 14% loss connected to gift cards (see also Offenberg, Reference Offenberg2007).Footnote ⁷ Gifts from more distant relatives were shown to have a greater mismatch in value as would be expected in a theory emphasizing asymmetric information. The study revealed that the categories of individuals who were more likely to give gifts with a larger spread (i.e. higher deadweight loss) were also those more likely to give cash, although cash transfers still accounted for a small fraction of giving (approximately 13.5% of all gifts). Grandparents were the most likely to give cash (59.2%), with aunts and uncles being the next most likely (25.7%). Friends and ‘significant others’ only gave cash about 2% of the time (Waldfogel, Reference Waldfogel2002: 418). The likelihood of giving in-kind gifts increased with regularized contact: those who were in daily contact with a recipient, and presumably knew the recipient's preferences better, only gave cash about 6% of the time, whereas those who had just yearly contact gave cash roughly one out of three times (Waldfogel, Reference Waldfogel2002: 418). Again, these results correspond with common experience and a theory based upon asymmetric-information problems.

Even though the tendency to give cash gifts increases as the giver knows less about the recipient, cash remains rare. Waldfogel attributed this to the cultural stigma of giving cash. The nature and enforcement of this stigma was not theorized but inferred from the measured gap between those who should be giving cash because of their infrequent contact with the recipient and the fact that they give in-kind gifts (Waldfogel, Reference Waldfogel2002: 421–426). Potential motivating reasons for giving in-kind gifts beyond cultural pressure to avoid the stigma of giving cash include ‘altruism, exchange, and paternalism’:

The paternalistic motivation presumably requires monitoring of the in-kind gift's use, implying it is either for immediate consumption in the presence of the gift giver (e.g. eating a fruitcake) or for regularized consumption where the giver can observe its use (e.g. wearing a sweater to family gatherings). Paternalism produces an asymmetric effect. The giver is likely to further increase their value from gifting (Brennan and Walsh, Reference Brennan and Walsh1977, Reference Brennan and Walsh1980), while the value of information the giver can obtain about the recipient's preferences erodes (if they do not outright disregard the recipient's preferences), lowering the value to the recipient. Immediate consumption would not seem to guarantee the change in long-term behavioral patterns that paternalism usually suggests; a paternalist does not want to modify behavior just once, but on an ongoing basis. Paternalist gifting thus suggests some degree of monitoring and more likely comes from those in routine contact with the recipient. Distant acquaintances giving paternalistic gifts do not know whether their efforts were successful given their lack of contact. Moreover, those with little knowledge of the recipient do not know whether the recipient needs to be ‘nudged’ into ‘better’ behaviors. It is plausible, though, that the giver could imagine (or hope for) regularized use if not able to monitor behavior directly.

While admitting it is empirically difficult to disentangle altruistic from exchange-based motivations (given data limitations), Waldfogel (Reference Waldfogel2002) posits that ‘[o]n its face the low frequency of cash gifts is evidence of paternalistic motivations for giving … if the givers were primarily concerned with recipient utility, they would give cash, allowing the recipients to make utility-maximizing choices’ (416). However, he finds that since ‘cash gifts are more likely from the sorts of givers who would be expected to give unwanted noncash gifts, the decision to give cash thus appears to reflect givers’ concern for recipient utility. In other words, the giving in this sample does not appear to be paternalistic, despite the low frequency of cash gifts’ (426). If this inference is true, that leaves altruism and exchange as the more dominant motivations.

It is conspicuous that the gifting-as-exchange thesis is not explored further by Waldfogel since exchange is a cornerstone of economic analysis. Smith (Reference Smith1976 [1776], book I, chap. 2) observed early on that we are a species that loves to truck, barter, and exchange. But Waldfogel's survey research and his critics' follow-up studies never probed whether the recipients of gifts intended to reciprocate or whether the gifts they received were given in response to presents they had given others; all these studies analyzed ‘one-shot’ events in which the subjects were only recipients. But exchange by its very nature is reciprocal. While we acknowledge that gifting certainly generates deadweight loss in the short-term, the ‘burnt offering’ component is a sacrifice that creates an opportunity for relationship. When it is reciprocated in an iterated exchange, becomes publicly visible, and is ritualized in ceremonial behavior, it can enhance welfare for society in the long-term through its promotion of generalized trust. It is the repeated, ritualistic, and public nature of gift giving that we turn to next, a feature that helps to explain the institutional endurance of this supposedly ‘wasteful’ cultural practice.

3. A Mauss tale: institutional endurance and ritualized practice

If gifting is a major source of social inefficiency, economists face two puzzles: first, why has the practice endured for so long? Second, why has it been common across widely disparate cultures? Gifting is a near-universal institution, an observation made by Marcel Mauss (Reference Mauss2011 [1954]) nearly a century ago. Mauss's interest was to explore exchange in ‘primitive’ societies, ranging from Melanesia to ancient India, and medieval Germany where no formal monetary exchange occurred. But this cultural institution exists even in the most-advanced market economies, which lends credence to Mauss's claim of its universality. Beyond Christmas, there are numerous holidays where gifts are given – for example, Hanukkah, Kwanza, Eid al-Adha, Raksha Bandhan, and Chinese New Year. Birthdays, bridal showers, and anniversaries are common cultural gifting moments as well. Add to this everyday gift-giving such as bringing a bottle of wine to a party, a welcome basket to new neighbors, or small presents that begin business negotiations (common in Japan). If gifting was inefficient as Waldfogel claims, why would this cultural institution exist so ubiquitously and resist replacement? In a static sense, Waldfogel is right. In the dynamic sense, however, we would expect many aspects of gifting to be selected out of the set of institutional options via some evolutionary mechanism (Alchian, Reference Alchian1950).Footnote ⁸

Mauss's take on gifting is relevant here, as he viewed such customs as a nonmarket form of exchange. Gift relationships existed prior to the formalization of exchange in ‘primitive’ societies. Mauss did not define market exchange explicitly, although his understanding of gifting indicates that market transactions involve voluntary exchange that benefit both parties (i.e. produce gains from trade). Gifting, in contrast, involves ‘prestationsFootnote ⁹ which are in theory voluntary, disinterested and spontaneous, but are in fact obligatory’ (Reference Mauss2011 [1954]: 1). The obligatory nature is important. Something is obligatory when it forms normative expectations beyond a simple quid pro quo transaction. An individual feels obligated to give (and reciprocate) based upon cultural norms and a desire to maintain a solid reputation within a community. Those who fail to participate in the office white elephant gift exchange or bring wine to dinner parties will soon find themselves ostracized from the group. Following such rituals signals an individual wants to be a member of some network, be it social or commercial.

Mauss's most noteworthy illustration – the Melanesian kula – reveals the linkage to commercial transactions and points to several other key features common to many institutional gifting rituals. The kula was a ‘whole system of inter-tribal and intra-tribal commerce’ (Mauss Reference Mauss2011 [1954]: 19). It ‘is distinguished from the straightforward exchange of useful goods’ and ‘like the American potlatch, consists in giving and receiving, the donors on one occasion being the recipients on the next’ (20). The main kula ritual involved gifting only, but during ‘the lesser kula … the sea voyage also serves an opportunity for exchange of cargoes; the nobles themselves do business; numerous objects are solicited, demanded and exchanged, and many relationships are established in addition to kula ones; but the [main] kula remains the most important reason for the expeditions and the relationships set up’ (20–21). It is important that this gifting is associated with trade not only among the elite, but all members of society as ‘all kula transactions are an opportunity for ordinary exchange … which does not necessarily take place between established partners’ (25, latter emphasis added). Here we see the importance of building connections between quasi-anonymous individuals not directly tied to explicit trade relationships but could potentially be partners later.

The kula signified a desire for formal relationships between tribes on neighboring islands and involved intricate procedures of presenting gifts via prescribed methods (i.e. rituals) that were obligatory. The specific rites associated with the kula were ornate and often involved ‘non-utilitarian’ gifts (e.g. shell jewelry). Failure to abide by the ritualistic norms signaled that the giver or recipient was not acting in a trustworthy manner and may have been harboring ulterior motives (Mauss, Reference Mauss2011 [1954]: 18–27). Given the possibility that wealth-destroying raids and war could break out easily between rival tribes, gifting rituals served to signal information about one's future intentions and a willingness to sacrifice wealth for the sake of peace and prosperity (see also Leeson, Reference Leeson2014a). Sacrifice was important in reducing the uncertainty that was ubiquitous in those societies. A willingness to ‘burn resources’ in the present was akin to posting a bond to deliver on one's promises in the future. The ritualistic nature of gifting also emphasized the importance of iteration; repeat play sets expectations and reduced uncertainty. Breaks in the pattern raised suspicions and hinted at future conflict.

In addition to the sacrificial component of the kula and similar practices, Mauss observed how rituals developed to foster the reciprocal nature of gift exchanges. These rituals were publicly observable allowing participants to express a willingness to give and receive graciously. This bolstered broad social trust because even if an individual was not involved in a specific gift exchange, they would witness others making sacrificial offerings. And gifting required not only returning sacrifices in kind, but also a willingness to accept the gift (even if one had no use for it): ‘Total prestation not only carries with it the obligation to repay gifts received, but it implies two others equally important: the obligation to give presents and the obligation to receive them’ (Mauss, Reference Mauss2011 [1954]: 10–11). Refusal to accept an offering threatened to unravel the gifting process and stir doubts about someone's intentions, raising uncertainty, dampening trade, and possibly resulting in conflict.

The kula example reveals that the norm-based process of gifting has value beyond the static utility (or disutility) of any particular gift received. The sacrificial, reciprocal, public, and ritualistic, components of gifting institutions serve to lower one of the most common transaction costs within an economy: uncertainty. Our theory of gifting emphasizes how congruent expectations reduce uncertainty, promote peaceful association, build extended social networks, and create social order irrespective of the deadweight loss generated in the short term. Mauss's analysis shows how these benefits emerge from a social context rather than through top-down legal and political changes. We now turn our attention to situating gifting within the broader institutional framework needed to reduce uncertainty and promote economic exchange.

4. Trust and transactions: from the gift relationship to generalized trust

As Mauss reveals, gifting is a robust social mechanism useful for reducing uncertainty, minimizing conflict, and extending relational networks. In this way, the emergence of trust precedes commercial transactions. But why would this be? Alas, Mauss does not tackle this question in depth. An economic approach provides insight. To understand why gifting may enhance commercial efficiency, recall Smith's (Reference Smith1976 [1776]) basic formula for economic prosperity: the wealth of nations is determined by efficiency gains from the division of labor, which in turn is conditioned upon the extent of the market. Extending the market (i.e. expanding trade) is tricky. As the market expands, trade becomes more quasi-anonymous; the more people we know and interact with, the less we tend to know about them. Dunbar (Reference Dunbar1998) argues that humans can only know the intimate details of no more than roughly 200 other individuals. Beyond that, it becomes increasingly difficult to tell whether any one person will engage in trade honestly or intends to cheat us. Indeed, uncertainty represents one of the most significant transaction costs within a market economy; if I am uncertain whether I can reap the rewards of my effort and will be treated fairly and honestly by a trading partner, I have less incentive to truck, barter, and exchange. If reciprocal trust is required for trade, markets will be extremely limited. Reducing uncertainty in quasi-anonymous markets to a level requires generalized trust, which is necessary for an economy of any size to hum along efficiently (Fukuyama, Reference Fukuyama1995; Munger, Reference Munger2018).

Smith (Reference Smith1976 [1776]) innately understood the need for promoting certainty and trust within an extended market when he argued that production and exchange required ‘a well-governed society’ (22). Scholars since Smith's time have noted the importance of well-specified property rights and other institutions that alleviate uncertainty. Institutional political economists have devoted substantial attention to the role of government and other formal structures designed to alleviate uncertainty. From constitutional contracts (Buchanan, Reference Buchanan2000 [1975]; North and Weingast, Reference North and Weingast1989) to laws reducing information asymmetries (Akerlof, Reference Akerlof1970; Grossman and Stiglitz, Reference Grossman and Stiglitz1980), government is seen as providing rules and enforcement that reduce uncertainty and allow humans to interact predictably and peacefully. Recently, though, attention has shifted toward the study of culture and community self-governance (as opposed to government) that provide norms and rituals that harmonize social interaction within civil society (Brennan et al., Reference Brennan, Eriksson, Goodin and Southwood2013; Ellickson, Reference Ellickson1991; Leeson, Reference Leeson2014b; Ostrom, Reference Ostrom1990, Reference Ostrom2000; Skyrms, Reference Skyrms2014 [1996]). Norm-based forms of social governance existed prior to, and often have outlasted, all the formal legal structures (i.e. government) that we are familiar with today (Gill, Reference Gill2021). Moreover, norms and rituals account for the vast majority of the rules determining our daily social interactions (Bicchieri, Reference Bicchieri2006).

Cultural rituals and norms assist in economizing on informational costs by creating simple heuristics, or ‘shared mental models’ (Denzau and North, Reference Denzau and North1994). In behavioral psychology, these models (or heuristics) are the working knowledge of cause and effect that guide individual action (Simon, Reference Simon1955). In some areas the actor will be relatively proficient with details necessary to make informed choices. For other choices – ones that are infrequent or rely upon expertise the chooser does not possess – choices still must be made, including deciding whom to trust in a commercial interaction. The economics of information literature suggests that people are willing to sacrifice value to forgo search as uncertainty increases (e.g. Stigler, Reference Stigler1961). When traveling abroad, we seek out trusted information, perhaps from a concierge, on which restaurants to visit or vendors to patronize. This lowers the cost relative to investigating a wide range of restaurants and vendors individually. If the hotel has a trusted brand name to protect, we can assign greater credibility to the concierge's recommendation. The concierge is a useful institution when it reduces uncertainty and facilitates commercial exchange.

Gifting is similar. By extending individual relationships, information about the quality of heuristics increases. Assuming bounded rationality, we seek to economize on all the costs associated with both the selection and increasing quality of those heuristics. Public gifting rituals are a proxy for this selection process because they extend relationships with those whose heuristics have more credence; if we observe others participating in gift exchanges, we can infer that they are willing to engage in sacrificial behavior just as we do. Not all people who engage in reciprocity act according to high-quality heuristics, but on average individuals engaged in multiple-round games of reciprocity possess heuristics that are higher than average quality compared to those used by individuals making idiosyncratic or random choices. A restaurant that hangs Christmas decorations might simply be engaging in relatively low-cost signaling, but signaling at any positive cost is a better indicator of average quality than no signal at all. A more formal heuristic can be provided by clearinghouses such as travel guides or customer-satisfaction rating systems (e.g. Yelp), but such services are specific to the individual and business; they do not convey the generalized trust that quasi-anonymous trade requires. Moreover, gifting rituals long preceded the information technology such modern clearinghouses rely upon. Indeed, gift giving may be the first institution that served to send signals about the trustworthiness of others.

Another means of alleviating uncertainty involves selecting good cooperators over time via repeated interactions. Developments within game theory illuminate how this works. In a simplified format, human interactions (including trade) can be modeled as a cooperative endeavor and structured as a prisoners' dilemma.Footnote ¹⁰ While it would behoove both parties to exchange honestly and fairly, it is in the interest of either party to cheat while the other party behaves honestly. For instance, two parties may contract to exchange 10 bolts of fine linen for a barrel of 12-year-old scotch. The scotch, however, may have been aged only 5 years and the cloth woven of inferior material. At the time of exchange, it is difficult for either party to detect whether the other merchant has violated the negotiated contract. They might have recourse to judicial institutions to rectify this injustice. However, litigation is time consuming and financially costly, and results uncertain. The judicial system may also be untrustworthy, rife with corruption and rent seeking. Such concerns – the potential for cheating and costly litigation – diminish the likelihood of trade; the extent of the market shrinks, and social welfare stalls or declines.

Offering solutions to the prisoners' dilemma is an academic cottage industry. One of the most popular solutions involves repeated play. Playing a tit-for-tat strategy in an iterated game can account for cooperation over time so long as each party has a sufficiently extended time horizon (Axelrod and Hamilton, Reference Axelrod and Hamilton1981), an outcome predicted by the folk theorem (Friedman, Reference Friedman1971). Iteration among the same trading partners creates a thick cultural network of information, also known as ‘bonding capital’ (Granovetter, Reference Granovetter1973). Trade historically occurred among individuals within closed tribal networks who knew each other well (Greif, Reference Greif2006). Iteration within thick social settings generates more information about quality, lowers transaction costs associated with uncertainty and information gathering, and provides some ability to cover unarticulated parts of contracts.Footnote ¹¹ If mistakes can be expected to occur in an exchange, it is important to know in advance that your trading partner will graciously fix those errors at their expense. Thick networks contain more information about the graciousness of others. Unfortunately, thick networks tend to limit the extent of the market (and hence prosperity), particularly if individuals are wary of trading with people outside their immediate tribe (Greif, Reference Greif2006). Trade and prosperity require trust in thin networks characterized by infrequent interactions with (quasi-)anonymous others.

Broad cultural gifting traditions that are publicly known build thin-network trust by demonstrating repeated sacrificial exchanges. When we see many others merrily purchasing, wrapping, and exchanging gifts during the holidays (or other gifting events), and doing so on a regular basis, we can assume that most others in society are acculturated into the norm of sacrificing resources for the sake of others and graciously accepting those sacrifices when offered to them. Gifting rituals represent a manifestation of the folk theorem played extensively throughout societies. Although the laboratory experiments on iterated cooperative play often involve the same (sometimes quasi-anonymous) players, publicly visible gift-giving implicitly signals individuals are choosing to cooperate in a game in which they expend resources and potentially create static deadweight loss. Gifting, in other words, is the folk theorem lived daily; we learn to be trustful in commercial interactions by playing out a seemingly wasteful game. Again, it is the iterated process of gifting that is more valuable than the actual gifts themselves.

5. Design principles for dynamically efficiency gifting

Gifting as a cultural institution is uniquely situated to build trust and fosters dynamic efficiency. That there is variation in how this cultural institution emerges across time and space is no surprise. The main feature that makes these practices alike is the way that reciprocal trust is fostered and the extent to which it results in generalized trust at the societal level. In this section we argue that how the gifting institution is ‘designed’ determines the extent to which it can promote generalized trust. We use the concept of ‘design principles’ that was popularized by Ostrom (Reference Ostrom1990) when discussing cultural institutions that solved a variety of common pool resource (CPR) problems. For Ostrom, a design principle is ‘an essential element or condition that helps to account for the success of these institutions in sustaining the CPRs and gaining the compliance of generation after generation of appropriators to the rules in use’ (Reference Ostrom1990: 90). This does not necessarily mean institutions are designed intentionally for specific purposes. Intentional design is possible, but when considering social norms and rituals, we take a more ‘Austrian’ perspective on cultural institutions noting that some actions taken by individuals to resolve certain situations can become embedded in customs that evolve over time (Hayek, Reference Hayek1979: 153–169; Smith, Reference Smith2015). Many of the cultural institutions that Ostrom examined emerged over time, often without explicit conscious intent.

Although Ostrom was concerned with managing the tragedy of the commons, the design principles concept can be applied to other institutions. In our case, we are interested in how four key elements – sacrifice, reciprocity, public visibility, and ritual – assist in promoting generalized trust, the outcome that makes gifting dynamically efficient. The first two features enhance the heuristic function of gifting, whereas the latter two emphasize the iterated process of trust building necessary for thin markets to work. The more that each of these design principles is at work in any specific gifting practice, the more that custom will enhance social trust and lower the transaction costs associated with quasi-anonymous trade. We consider the first one of these principles – sacrifice – to be a necessary condition of dynamically efficient gifting as it is the primary heuristic signal that a person can be trusted. The other three conditions vary across gifting practices and determine the extent and effectiveness of the custom in generating broader levels of trust and economic integration. We now examine each of these design principles in greater length.