Reciprocity & Risk

RECIPROCITY AND RISK

Intro

This lecture is titled "Reciprocity & Risk," but an alternative title might be "The ecology of sharing"

Topical outline:

1) What are "reciprocity" and "risk"? (Definition of the key concepts)

2) What ecological conditions favor reciprocal sharing?

3) What social conditions are required to make sharing persist?

The Key Concepts

What is Reciprocity?

Reciprocity can be defined broadly as the voluntary exchange of goods, services, or information between 2 or more individuals or groups, such that A gives to B and B gives to A (though not necessarily the same thing, or in equal amounts)

This so broad as to encompass most forms of social interaction, so let's narrow our focus down to phenomenon of delayed reciprocity: voluntary exchanges of the sort just defined that have a significant time lag

Delayed reciprocity is captured in simple expression "You scratch my back, I'll scratch yours," but of course ranges over much more complex interactions, such as military alliances, ceremonial feast exchanges, various aspects of marriage, etc.

Delayed reciprocity is not unique to humans, though it is certainly most highly elaborated in our species: non-human species lack elaborate division of labor, complex rules of reciprocity, and extensive extra-familial sharing (social insect colonies, which might seem like an exception to this, are not -- they are essentially family groups, just much larger than vertebrate families ever get)

Indeed, humans are unique in the degree to which exchanges of goods, services, and information dominate social life, and degree to which these often involve delays in reciprocation, some quite lengthy

There's no doubt that human capacities for culture (social transmission of values and beliefs) and language (symbolic communication) are critical for understanding human elaboration of reciprocity

But frequency & intensity of reciprocity varies within and between cultures/societies, so in order to explain this variation we have to go beyond ascribing practices of reciprocity to cultural values, and ask "Why do those values (and not some others) exist?"

Hence, it is worth examining the ways in which ecological theory can illuminate general human propensity for delayed reciprocity, as well as the variable expression of this propensity in particular socioecological contexts

What is Risk?

To do this, we need to look at our second key term, "risk"

This has two distinct colloquial meanings:

1) Dangerous ("If you climb that ladder, you're taking a risk")
2) Chancy ("It's a risky investment")

Risk also has a technical meaning in economics and decision theory that formalizes this second colloquial meaning, and refers to variation in yield or payoff of some decision or action

Notice that risk in this sense can be either valuable (e.g., the "risky investment" could turn out to do really well) or costly (the investment does poorly)

Thus, economic risk is a probabilistic concept, measured in statistical units such as variance or standard deviations; and it applies to the consequences of some action, such as the variation in yield ("income") from a given choice or action

Risk in this sense can vary independently from average values: thus, investing in a single company's stock or a single agricultural crop will produce an average yield that is lower or higher (or the same) than can be obtained from a diversified portfolio of stocks or crops, but it will almost certainly have a higher risk (greater variation in yield over the course of a given period of time)

It is this meaning of risk which ecologists have employed in an attempt to understand the adaptive significance of sharing and reciprocity (as well as other topics)

To illustrate why risk might have ecological relevance, we can revisit the optimal prey choice model: it assumes that a forager will always prefer the suite of prey types yielding the highest average rate of return, regardless of the risk (variation) involved

But if this suite ("diet breadth") exposes you to a high probability of failure (which can be true even if on average it's high-yielding), most foragers will prefer an alternative yielding a lower average rate but greater reliability of daily bread (i.e., lower risk) -- which is probably why few if any subsistence systems specialize in hunting whales or growing a single crop

But is risk always a bad thing, to be avoided to the extent possible? What about risk-takers?

To answer this, we have to look a little deeper into economic logic of risk, and distinguish value from yield

Consider a simple graph of food income in relation to some broad measure of value [see Boone 1992 reading, p. 321, Fig. 10.5]

Now ask the question "What is the value (to the consumer) of a given amount of food?"

("Value" here can be measured as "reward" or utility -- the economist's or psychologist's approach, or as fitness value -- the evolutionary currency)

If you only care about the average yield, then the graph of value as a function of yield will be linear (x amt. of yield = y amt. of value; 2x yield = 2y value; etc.) [Boone 1992, Fig. 10.5(a)]

But now ask yourself: Do you get as much satisfaction from the 3rd chocolate sundae (at one sitting) as from the second, let alone the first? Unless you're a sundae fanatic, the answer probably no; and this becomes increasingly likely as you extend this out for a while (e.g., 10th sundae of the day may make you barf)

This basic logic is represented graphically by a diminishing returns curve: each additional unit is worth less and less [Boone Fig. 10.5(b)]

On the other hand, one can imagine why the bottom of the curve might show accelerating returns (each additional unit worth more than last); if you've gone long without food, the first bite whets your appetite and leaves you hungrier than before! [Boone Fig. 10.5(c)]

Keeping our example of yield as food income but shifting our value currency from satisfaction or utility to fitness, consider why the same general logic might apply: an adult human of average size limited to 100 kcal a day will soon perish, but at 500 kcal/day may be able to hang on a few weeks; by 1500 kcal/day will be hungry but surviving; by 2000 kcal/day will have enough to be active (and resist diseases and other insults), at 2500 kcal/day has enough to reproduce, and beyond that may gain little fitness benefit (though even 4000 kcal/day may offer some marginal fitness gain in a subsistence context with variable income, since one can store some for rainy day, exchange some of it for other benefits, support more dependent offspring, etc.)

Now add some risk (variation in food income) to this picture [Fig. 10.2 in Boone 1992, p. 307]

A little mathematical analysis shows that if one's income variation is primarily in zone where marginal value is diminishing, risk reduces utility or fitness: value of lows (a) plus value of highs (ß) averages less than value of their average (i.e., [V(a) + V(ß)]/2 < V[(a+ß)/2]) [study the graph carefully to understand this point -- it's not obvious!]

But the situation is reversed when marginal value is accelerating (i.e., the curve is bending upwards): then variable outcomes are more valuable than stable ones w/ same mean, thus favoring risk-seeking choices or behavior

(This is counter-intuitive, but easier to grasp if you think of what a starving person would prefer: a sure bet of 500 kcal a day for a week [= likely death] or a gamble with equal chances of getting zero or 1,000 kcal/day; clearly, the best chance of surviving is to take the gamble!)

Ecological Conditions Favoring Delayed Reciprocity

So what does any of this have to do with reciprocity?

According to ecological theory, quite a lot: given diminishing fitness returns to food consumption (i.e., a downwards-bending curve), risk (variation in outcome) can make reciprocity very adaptive

To see why, consider a simple little scenario, involving a population of two individuals (I said "simple"!) who survive through foraging or farming

Each person's food income is likely to fluctuate from day to day or year to year, due to chance events (in weather, animal behavior, injuries that might prevent you from working on some days, etc.), and we might suppose this could be unpleasant (you might go hungry some days) or even disastrous (you might starve, or succumb to disease in your weakened state)

But suppose this variation is not synchronized (e.g., each individual forages or farms in different but nearby areas, each has an independent chance of getting injured, etc.)

Then they have a way of reducing their risk: they can buffer variation in food income by pooling the harvest and each consuming half

This way they can get the benefit of averaging with less of the ups & downs associated with risky outcomes -- V{(a+ß)/2} rather than {V(a) + V(ß)}/2

Of course this will reduce risk only to the degree that the two parties have failure rates that are "unsynchronized" or out of phase (perfect synchrony means Sally gets exactly what Jane gets; perfect asynchrony means whenever Sally gets a, Jane gets ß, and vice versa; in the real world, we can expect synchrony to be somewhere in between these extremes)

But as we increase the size of the sharing network, even partial asynchrony will be very effective at reducing variation in individual income (following the "law of large numbers"), a result that is fundamental to insurance plans

In sum, the ecological conditions favoring resource-sharing to reduce risk are clear:

1) harvest is relatively unsynchronized between individuals or households (so that for any given harvest some get a lot and others little);

2) food comes in large packages (resulting in diminishing marginal value to individual consumer)

Evidence from several hunter-gatherer societies agrees with this argument: food most often shared when harvest is unsynchronized, and when it comes in large packages (e.g., big game); much less emphasis on sharing plant foods, small game, etc.

A particularly detailed test of risk-reduction model conducted by Kaplan & Hill with Aché Indians (hunter-gatherers in Paraguay forest) [Aché sharing graph]

Aché practice considerable variation in amount of sharing: share game resources very widely, nearly as much for honey, and less so with plant foods

K&H found that variation btwn resources in amount of sharing correlated quite well with the 2 features the risk-reduction model predicted: package size (the bigger, the more diminishing returns), and daily variation in harvest synchrony between families (greater variation = less synchrony)

Analogous findings apply to subsistence agriculturalists, but here risk-reduction more often involves storage (self-insurance) and field-scattering (spreading your agricultural eggs among several baskets)

Stabilizing Reciprocity: Make it Conditional

It would seem that the risk-reduction model provides a neat and tidy explanation of delayed reciprocity, sufficient to the task

But it's not that simple: the risk-reduction model considers only the benefits of reciprocity, and we must also attend to its costs

In the present context, the important costs have to do with the delay inherent in any form of delayed reciprocity, which requires that two (or more) parties cooperate sequentially to provide a collective benefit (such as risk reduction)

To see why, let's return to food-sharing, a form of reciprocity that anthropologists have found to be extensively practiced in politically egalitarian systems like hunter-gatherer bands (but not necessarily in hunter-gatherer societies with other kinds of sociopolitical organization)

Here, we typically observe a strong ethic of generosity that encourages people to share temporary abundances of food with other households in their camp -- to let anyone go hungry while you have plenty would be "unthinkable"

Actually, evidence indicates people do think of it (e.g., ethnographers being asked to hide someone's secret stash, or to secretly give delicacies or rate items only to them)

More important, plenty of evidence of social sanctions designed to discourage hoarding: gossip, ostracism, withdrawal of cooperation, witchcraft accusations, even threats of violence against anyone who refuses to share

These sanctions are clearly designed to reduce the temptation to behave selfishly (which implies that the sharing ethic is not enough to motivate generosity)

How can we explain this coexistence of cooperation (ethic of generosity & sharing) side by side with coercion (sanctions against selfishness, which imply temptation to behave selfishly)?

The risk-reduction model correctly points out that if people reduce their risk through sharing, they will benefit (assuming diminishing returns)

But if one can get away with strategy of receiving shares when empty-handed but hoarding your own harvest when successful, you'll be even better off

Even if the value individuals obtain from large harvests diminish rapidly or reach zero marginal value, so there is little temptation to hoard, free-riding can arise through shirking (failure to bother working at all -- just stay back at camp and collect your share of others' catch; or pretend to be foraging, but just go sit under a shade tree a mile from camp)

Thus, in any system involving delayed reciprocity (e.g. "You share with me today, I'll share w/ you tomorrow") there is a potential for "free riding" (i.e., sharing in the benefits without paying the costs)

This simple logic suggests (and formal models in economics and evolutionary ecology predict) that indiscriminate altruism (unconditional sharing) can't win against selfish free-riders who accept gifts but don't reciprocate

However, more complex models predict that delayed reciprocity can win out if it takes the form of a conditional strategy: I'll scratch your back (or share with you) today if you'll do same for me tomorrow

To succeed, this must be backed with threat of holding a grudge: if you don't scratch my back today, I won't scratch yours tomorrow

But because we can never be sure about tomorrow, the best we can do is rely on past evidence & reputation of the potential reciprocator: since you scratched my back or shared with me yesterday, I'll do the same for you today (and conversely)

What we end up with, then, is not the golden rule ("Do unto others as you would have them do unto you"), nor the "brass-knuckle rule" ("Do unto others whatever you can get away with"), but a rule that game theorists call tit-for-tat ("Do unto others as they do unto you") [see Boone reading]

This tit-for-tat rule is a form of conditional reciprocity, and is only stable under certain conditions:

1) good individual recognition & memory (otherwise one can't distinguish trustworthy individuals from cheats)

2) high likelihood of future interactions (otherwise, grudges are ineffective)

3) small social groups (otherwise it's too easy to get away with free-riding)

Since the conditions for extensive reciprocity are so prevalent among modern hunter-gatherers, and the practice so frequently observed, it's a good bet that this was true throughout most of human evolution, when all humans lived in similar situation

Some anthropologists & psychologists deduce from this that our cognitive machinery has evolved to be both eager to engage in reciprocal interactions, and very good at detecting cheating within these contexts; and considerable empirical evidence supports this deduction

There are some problems with this solution, however:

1) it breaks down if groups are large (where "large" means >5 or so!);

2) delayed reciprocity is rare in other species, even ones meeting the 3 conditions listed above;

3) there is growing evidence that people will violate the predictions of reciprocity theory, and specifically that they will engage in unconditional generosity (giving more than they have to, even in anonymous settings) and will punish free-riders even when it is not in their immediate self-interest to do so

These 3 critiques of the standard view on the evolution of reciprocity may not be fatal, but they are spurring a serious re-examination of the paradigm in both animal behavior and some areas of the social sciences (economics, game theory, and ecological anthropology in particular--see some of the references below, particularly Henrich et al. 2004 and Gintis et al. 2005, for further details)

References

Alvard, Michael S. and David Nolin (2002) Rousseau's whale hunt? Coordination among big-game hunters. Current Anthropology 43(5):*.

Axelrod, Robert (1984) The Evolution of Cooperation. New York: Basic Books.

Axelrod, Robert and Douglas Dion (1988) The further evolution of cooperation. Science 242:1385-90.

Boone, James L. (1992) Competition, conflict, and the development of hierarchies. In Evolutionary Ecology and Human Behavior, ed. E. A. Smith and B. Winterhalder, pp. 301-337. Hawthorne, NY: Aldine de Gruyter.

Boone, James L. (1998) The evolution of magnanimity: when is it better to give than to receive? Human Nature 9(1):1-21.

Boyd, Robert and Peter J. Richerson (1988) The evolution of reciprocity in sizable groups. Journal of Theoretical Biology 132:337-356.

Boyd, Robert and Peter J. Richerson (1992) Punishment allows the evolution of cooperation (or anything else) in sizable groups. Ethology and Sociobiology 13:171-195.

Connor, Richard C. (1995) The benefits of mutualism: a conceptual framework. Biological Reviews 70:427-457.

Cosmides, Leda and John Tooby (1989) Evolutionary psychology and the generation of culture, part II: case study: a computational theory of social exchange. Ethology and Sociobiology 10:51-9.

Dugatkin, Lee Alan (1997) Cooperation among animals: An evolutionary perspective. NY: Oxford U. Press.

Fehr, Ernst and Simon Gächter (2002) Altruistic punishment in humans. Nature 415, Jan. 10:137-140.

Gintis, Herbert, et al., eds. (2005) Moral sentiments and material interests: On the foundations of cooperation in economic life. Cambridge, MA: MIT Press.

Gurven, Michael, et al. (2000) "It's a wonderful life": signaling generosity among the Ache of Paraguay. Evolution and Human Behavior 21(4):263-282.

Hames, Raymond B. (2000) Reciprocal altruism in Yanomamö food exchange. In Adaptation and human behavior: An anthropological perspective, ed. Lee Cronk, Napoleon Chagnon, and William Irons, pp. 3397-416. Hawthorne, NY: Aldine de Gruyter.

Hawkes, Kristen (1992) Sharing and collective action. In Evolutionary Ecology and Human Behavior, ed. E. A. Smith and B. Winterhalder, pp. 269-300. Hawthorne, NY: Aldine de Gruyter.

Hawkes, Kristen, James F. O'Connell, and Nicholas G. Blurton Jones (2001) Hadza meat sharing. Evolution and Human Behavior 22(2):113-142.

Henrich, Joseph, et al., eds. (2004) Foundations of human sociality: Economic experiments and ethnographic evidence from fifteen small-scale societies. Oxford: Oxford U. Press.

Kaplan, Hillard and Kim Hill (1985) Food sharing among Ache foragers: Tests of explanatory hypotheses. Current Anthropology 26(2):223-46.

Kaplan, Hillard, Kim Hill, and A. Magdelena Hurtado (1990) Fitness, foraging and food sharing among the Ache. In Risk and Uncertainty in Tribal and Peasant Economies, ed. E. Cashdan, pp. 107-143. Boulder, CO: Westview Press.

Nesse, Randolph M., ed. (2001) Evolution and the capacity for commitment. NY: Russell Sage Foundation.

Pusey, Anne E. and Craig Packer (1997) The ecology of relationships. In Behavioural ecology: an evolutionary approach, ed. John R. Krebs and Nicholas B. Davies, pp. 254-283. Oxford: Blackwell.

Richerson, Peter J. and Robert Boyd (2001) The evolution of subjective commitment to groups: a tribal instincts hypothesis. In Evolution and the capacity for commitment, ed. Randolph M. Nesse, pp. 186-220. NY: Russell Sage Foundation.

Sigmund, Karl, Ernst Fehr, and Martin A. Nowak (2002) The economics of fair play. Scientific American 286(1), January:82-87.

Smith, Eric Alden and Rebecca L. Bliege Bird (2000) Turtle hunting and tombstone opening: Public generosity as costly signaling. Evolution and Human Behavior 21(4):245-261.

Smith, Eric Alden and Robert Boyd (1990) Risk and reciprocity: hunter-gatherer socioecology and the problem of collective action. In Risk and Uncertainty in Tribal and Peasant Economies, ed. E. Cashdan, pp. 167-191. Boulder, CO: Westview Press.

Wedekind, Claus (1998) Give and ye shall be recognized. Science 280:2070-2071.

Wedekind, Claus and Manfred Milinski (2000) Cooperation through image scoring in humans. Science 288:850-852.

Wilkinson, Gerald S. (1988) Reciprocal altruism in bats and other mammals. Ethology and Sociobiology 9:85-100.

Winterhalder, Bruce (1997) Gifts given, gifts taken: the behavioral ecology of nonmarket, intragroup exchange. Journal of Archaeological Research 5:121-168.

Winterhalder, Bruce (1990) Open field, common pot: harvest variability and risk avoidance in agricultural and foraging societies. In Risk and Uncertainty in Tribal and Peasant Economies, ed. E. Cashdan, pp. 67-87. Boulder, CO: Westview Press.