Get the daily email that makes reading the news actually enjoyable. Stay informed and entertained, for free.
Your information is secure and your privacy is protected. By opting in you agree to receive emails from us. Remember that you can opt-out any time, we hate spam too!
HomeEconomyAre Humans Inherently Selfish?

Are Humans Inherently Selfish?

One of the principal assumptions guiding economic analysis is that human beings are self-interested utility maximizers. In short, individuals make choices that yield the greatest benefit net of cost.  This assumption allows economists to model human behavior and make predictions related to consumer purchases, corporate decision-making, and public policy.

Students in my Introduction to Political Economy course generally accept this assumption, although they believe that self-interest is synonymous with selfishness, which they further argue is socially destructive and should be countered with government regulation. 

The discussion of whether humans are self-interested, and if this equates to selfishness, seems academically esoteric. Understanding this topic, however, is important for public policy in practical terms. If people are exclusively self-interested, and possibly selfish, cooperation will be more difficult to achieve. This prompts policymakers to rely upon coercive government action (or “nudging”) to obtain “prosocial results.” But if humans demonstrate a willingness to forgo short-term gain in favor of sacrifice and generosity, many of the socio-economic problems may be resolved without government intervention.

A simple game played with students illustrates that while humans tend towards self-interest, they also act with prosocial motivations. A growing body of research is revealing that pro-sociality is found more often in market-based (versus communitarian) societies. 

The Ultimatum Game

To explore self-interested behavior with students, I use the Ultimatum Game

The rules are simple. There are two players who remain anonymous to one another. Player A is given $100 and told to divide it between himself and Player B in one-dollar increments. This split is sent anonymously to Player B who decides whether to accept or reject the offer. If B accepts A’s offer, then both players walk away with the amount of cash according to A’s offered split. If B rejects the offer, neither player receives any money. 

Anonymity assures the offer facing Player B is a one-time choice, eliminating the possibility of strategic reciprocity.

In what economists call “thin rationality,” Player A’s self-interested, utility-maximizing offer should be to keep $99 and offer B just $1. Player B’s thinly rational response should be to accept the offer because $1 is better than zero dollars. This is “thinly” maximizing because we are only concerned with the monetary value of the exchange. In thin rationality, $1 > $0, ceteris paribus.

The choice before Player B when offered a $99/$1 split is simply choosing between $1 (by accepting the offer) or $0 (by rejecting). Anonymity makes the $99 that A kept completely irrelevant (in a thinly rational sense). Player A is giving B a simple choice: Do you want $1 or $0?

To prove that people would rather have $1 as compared to $0 in a simple choice, I randomly pick students from the large lecture hall a few days before the game and ask them whether they want $1 with no strings attached. Over three decades of doing this, not one person has ever turned that offer down.

Given that a strictly rational person would rather have $1 instead of zero, the thinly rational offer that A makes should be a $99/$1 split knowing that B would never turn down a free dollar. Moreover, this offer is mutually beneficial; everybody is better off!

Note that a $100/$0 split is not optimal for Player A because it creates a scenario where B is indifferent between her choices. She would get $0 for accepting or rejecting the offer. Since this presents a “coin flip,” Player A should anticipate an acceptance 50 percent of the time for an expected utility of $50 not $99. By making the transaction mutually beneficial, A can “rationally guarantee” an optimal $99 payout for himself.

This is where I make a definitional difference between “selfish” and self-interest. A “selfish” person has no regard for the well-being of others. A truly “selfish” person would choose to maximize his benefit even if it injured somebody else. Although there would be no harm to Player B with a $100/$0 offer, Player A certainly is not concerned with B benefiting. I would call this “weakly selfish.” Although $1 might not seem much better, a $99/$1 split does improve the well-being of B, hence it is “other regarding” to a small extent. The Ultimatum Game does not test whether one is strongly selfish and willing to hurt others for his benefit, but if we can show that people may not be as self-interested as we think then characterizing people as “selfish” would be more difficult.

Theoretically, if both players are strictly self-interested, the most common outcome should be a $99/$1 offer with both parties accepting.

Is this what happens? Nope, not even close.

People Are Not Thinly Rational

In fact, a $99/$1 split is rarely offered and when offered, it is almost always rejected.

Consider Figure 1, which presents results from the game when played amongst 124 undergraduates at the University of Washington in 2019. Students were informed of the rules of the game and each student made an offer as Player A. The offers were randomized and returned to students who were then asked to accept or reject the offer they received as Player B. The results were astoundingly consistent with every other iteration of the game I’ve played going back to 1994. 

[See the “Addendum” below for additional methodological notes.]

I divided the offers of students into five categories: Selfish ($100/$0 split); Rational ($99/$1 split); Semi-Rational (anything between $98/$2 and $51/$49); Egalitarian ($50/$50 split); and Altruistic (A offered B more of the cash than they kept). 

Notice that the most common offer was not the (thinly) Rational response. Instead, the modal response was Semi-Rational.  A slim majority of students (~51 percent) gave Player B a Semi-Rational offer of at least $2, though Semi-Rational offers to B were typically between $10 and $40. The next most frequent offer (~30 percent) was Egalitarian. Twelve Altruistic students offered Player B more than they kept. There were two Selfish players who kept all the money for themselves.

In terms of responses, it was not surprising to see both Selfish offers being rejected. Nor was it unusual that all Egalitarian and Altruistic offers would be accepted. An even split seems inherently fair to our sensibilities, and it would seem odd for any rational person to turn down an “altruistic” offer.

The interesting action of this game came with the Rational and Semi-Rational offers and responses. Among eight Rational offers, only one Player B accepted the “free dollar.” Seven students “threw away” money.

As for Semi-Rational offers and responses, I calculated that the average amount of cash needed to entice somebody to accept was $30.90, whereas the average rejection was at $23.16. This implies that it takes roughly $27 (midway between the average accept and reject responses) for player A to feel confident that B will agree to the deal being offered. Again, this was remarkably consistent with previous years where the fulcrum between accepting and rejecting was generally in the $25 – $30 range. Yes, some players actually turned down $20!

What Is Going On? A Thicker Story

All told, this exercise seemingly indicates that people are not inherently self-interested, undermining one of the key assumptions of rationality underlying economic theory. Reality is more nuanced and suggest several practical implications. 

First, this exercise does not repudiate the notion of self-interest in favor of pure altruism. Note that the modal (and majority) offer was “semi-rational.” Indeed, when factoring in the Selfish and Rational offers, a majority of Player A’s wants to receive more money than Player B. Moreover, a solid majority (~70 percent) of those receiving a Semi-Rational offer accepted. Even when adding in responses to Selfish and Rational offers, most Player B’s were content with accepting less than Player A.

Second, that the vast majority (~92 percent) of Player A’s were not willing to make a Selfish or (strictly) Rational offer indicates either they have some degree of altruistic impulse, or they were aware that a Selfish or Rational offer would be viewed as insulting to Player B given social norms related to sharing one’s good fortune (e.g., being picked to play a game). To avoid rejection from Player B (resulting in zero dollars), the safer response is to share more than just $1. Of course, offering an Egalitarian or Altruistic split is essentially a slam dunk, but again it should be observed that the modal offer indicates that Player A wants more, which does tend towards the thin view of economic rationality.

A thicker view of rationality accepts the view that people live in a world with social norms that both condition and constrain our actions. From birth, we are taught lessons about the importance of sharing from our parents, teachers, and popular culture. We enlist these lessons when making decisions about allocating resources, especially when it comes to good fortune (as being chosen for this experiment seems to imply – free money!). More importantly, we understand that others are conditioned with similar social norms and that violating them may harm our reputation. The general culture constrains our thinnest desire to be selfish and grab what we can without sharing. This adds thickness to the preferences we act upon; we want financial gain, but we also want to be loved and be lovely.

Even in an anonymous game, social norms about sharing come through strongly as people are motivated to be generous and/or to worry about their reputation or the guilt they may feel violating “good manners.” The ability of these norms to overcome anonymous interactions is important as so much of our market economy relies upon trading with strangers.

The Ultimatum Game has been played with real cash (and more rigorously) in many different cultural and economic environments. A research team centered around Joseph Henrich discovered, much to their surprise, that market-integrated societies tend to be more generous and cooperative than traditional, communitarian ones. This was reaffirmed in a fascinating paper by Benjamin Enke showing that such market-based societies were much more likely to embed lessons of generosity in their folklore.

None of this suggests that people are not rational. Rather, it reveals the thick rationality of humans who understand that generous behavior towards others (including gifting) benefits society as a whole, which redounds upon them in greater prosperity. Public policy must remember this. Efforts by government to coerce cooperative outcomes may overwrite pre-existing norms and values that accentuate voluntary and peaceful relations among neighbors and strangers. If government policymakers rely on an assumption that individuals are narrowly rational and selfish, they will likely implement policies that undermine the thick rationality that makes free markets work efficiently. We must guard against this and trust our more well-intentioned instincts that people truly are generous, gracious, and cooperative.  

Addendum: Notes on the Ultimatum Game

Students in my political economy course were told to play the game “as if” real money were on the table; they knew there would be no cash payouts. As such, students may have taken the game less seriously, possibly increasing the number of Egalitarian and Altruistic responses. Nonetheless, my results are consistent with experiments using actual cash.

I report 2019 numbers as they are consistent with results dating back to 1995. The 2020 “COVID interruption” made it difficult to replicate the game as previously played. Unfortunately, an error in administering the game in 2021 invalidated the responses. 

In 2022, the game was played again in lecture. Results were oddly anomalous in that students turned out to be more altruistic than previous classes. For the first time in nearly 30 years the modal response was Egalitarian. And roughly 60 percent of Player A offers were Egalitarian (48 percent) and Altruistic (12 percent). Why students would be demonstrably more egalitarian and altruistic than previous cohorts is a mystery. I welcome speculation.