PERSPECTIVE PERSPECTIVE Social dimensions of fertility behavior and consumption patterns in the Anthropocene Scott Barretta,b, Aisha Dasguptac, Partha Dasguptad,1, W. Neil Adgere, John Anderiesf, Jeroen van den Berghg,h,i,j, Caroline Bledsoek, John Bongaartsl, Stephen Carpenterm, F. Stuart Chapin IIIn, Anne-Sophie Crépino, Gretchen Dailyp, Paul Ehrlichp, Carl Folkeo,q, Nils Kautskyr, Eric F. Lambins,t,u, Simon A. Levinv, Karl-Göran Mälero, Rosamond Naylorw, Karine Nyborgx, Stephen Polaskyy, Marten Schefferz, Jason Shogrenaa, Peter Søgaard Jørgensenq,bb, Brian Walkercc, and James Wilendd Edited by B. L. Turner II, Arizona State University, Tempe, AZ, and approved February 12, 2020 (received for review June 14, 2019) We consider two aspects of the human enterprise that profoundly affect the global environment: population and consumption. We show that fertility and consumption behavior harbor a class of externalities that have not been much noted in the literature. Both are driven in part by attitudes and preferences that are not egoistic but socially embedded; that is, each household’s decisions are influenced by the decisions made by others. In a famous paper, Garrett Hardin [G. Hardin, Science 162, 1243–1248 (1968)] drew attention to overpopulation Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. and concluded that the solution lay in people “abandoning the freedom to breed.” That human attitudes and practices are socially embedded suggests that it is possible for people to reduce their fertility rates and consumption demands without experiencing a loss in wellbeing. We focus on fertility in sub-Saharan Africa and consumption in the rich world and argue that bottom-up social mechanisms rather than top-down gov- ernment interventions are better placed to bring about those ecologically desirable changes. | | fertility consumption socially embedded preferences The biosphere is the part of Earth that is occupied by regenerative system. Also, like its constituent parts, the living organisms. A useful way to partition it is in terms biosphere is diminished as a stock, in quality or quantity of interconnected constituents we call ecosystems. (or both), when our demand for its goods and services Ecosystems combine the abiotic environment with bi- exceeds its ability to supply them on a sustainable basis. ological communities (plants, animals, fungi, microor- ganisms) that form self-regenerative functional units. The Commons as Externalities They differ in their spatial reach (a hedgehog’s gut is Being in large measure a common property resource, an ecosystem as is a tropical rain forest) and rhythmic many of the constituent parts of the biosphere are time (minutes for bacterial colonies; decades for a boreal freely available. Moreover, people everywhere are forest), but even as an aggregate, the biosphere is a self- capable of recognizing that a commodity is free. Thus, a School of International and Public Affairs, Columbia University, New York, NY 10027; bEarth Institute, Columbia University, New York, NY 10027; c Population Division, Department of Economic and Social Affairs, United Nations, New York, NY 10017; dFaculty of Economics, Cambridge University, CB3 9DD Cambridge, United Kingdom; eGeography, College of Life and Environmental Science, University of Exeter, EX4 4RJ Exeter, United Kingdom; fSchool of Sustainability, Arizona State University, Tempe, AZ 85287; gInstitute of Environmental Science and Technology, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; hInstitució Catalana de Recerca i Estudis Avançats, 08010 Barcelona, Spain; iSchool of Business and Economics, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands; jInstitute for Environmental Studies, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands; kAnthropology Department, Northwestern University, Evanston, IL 60208; lPopulation Council, New York, NY 10017; mCenter for Limnology, University of Wisconsin, Madison, WI 53706; nInstitute of Arctic Biology, University of Alaska Fairbanks, Fairbanks, AK 99775; oBeijer Institute of Ecological Economics, Royal Swedish Academy of Sciences, SE-105 05 Stockholm, Sweden; pDepartment of Biological Sciences, Stanford University, Stanford, CA 94305; qStockholm Resilience Centre, Stockholm University, SE-106 91 Stockholm, Sweden; r Department of Systems Ecology, Stockholm University, SE-106 91 Stockholm, Sweden; sSchool of Earth, Energy & Environmental Sciences, Stanford University, Stanford, CA 94305; tWoods Institute for the Environment, Stanford University, Stanford, CA 94305; uGeorge Lemaître Earth and Climate Research Centre, Earth and Life Institute, Université Catholique de Louvain, 1348 Louvain-la-Neuve, Belgium; vDepartment of Ecology and Evolutionary Biology, Princeton University, Princeton, NY 08544; wEarth System Science, Stanford University, Stanford, CA 94304; xDepartment of Economics, University of Oslo, 0317 Oslo, Norway; yDepartment of Applied Economics, University of Minnesota, St. Paul, MN 55108; zDepartment of Environmental Sciences, Wageningen University, NL-6700 AA Wageningen, The Netherlands; aaDepartment of Economics, University of Wyoming, Laramie, WY 82071; bb Global Economic Dynamics and the Biosphere, Royal Swedish Academy of Sciences, 104 05 Stockholm, Sweden; ccLand and Water, Commonwealth Scientific and Industrial Research Organisation (CSIRO), Canberra, ACT 2601, Australia; and ddDepartment of Agricultural and Resource Economics, University of California, Davis, CA 95616 Author contributions: S.B., A.D., P.D., W.N.A., J.A., J.v.d.B., C.B., J.B., S.C., F.S.C., A.-S.C., G.D., P.E., C.F., N.K., E.F.L., S.A.L., K.-G.M., R.N., K.N., S.P., M.S., J.S., P.S.J., B.W., and J.W. wrote the paper. The authors declare no competing interest. This article is a PNAS Direct Submission. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND). 1 To whom correspondence may be addressed. Email: pd10000@cam.ac.uk. First published March 12, 2020. 6300–6307 | PNAS | March 24, 2020 | vol. 117 | no. 12 www.pnas.org/cgi/doi/10.1073/pnas.1909857117 if population size relative to the cost of harvesting the biosphere is overshoot in demand that humanity has been exercising over the large, then other things equal, people draw excessively on it (1). biosphere’s goods and services in recent decades. The “tragedy of the commons” is a widely used metaphor for the We do not claim that behavior is necessarily based on con- phenomenon (2). Those excessive demands reflect the desires of scious choices. A reviewer has suggested that, in preliterate so- individual households to convert the biosphere into their own pri- ciety, behavior was purely norm based and that basic female vate sources at rates that are higher than the rates they would education has helped to bring fertility into the calculus of con- choose were households to be guided by their collective interest. scious choice, with an attendant reduction in fertility desire. One expression of that excess is the difference between the dis- However, conscious choice even in modern societies does not count rates of time that are inferred from our market transactions preclude the influence of social norms (17). More importantly for and the rates people would choose to discount the future were we fertility surveys, women selling goods in the marketplace in the able to negotiate among ourselves over the demands that we make poorest countries may be illiterate but are never innumerate. We of the biosphere (3). Thus, estimates of discount rates based ex- will make use of the distinction subsequently. clusively on market data (4, 5) mislead: they are biased upward. Behavior directed by social preferences has externalities built Another expression of our excessive demands on the biosphere is into it. Here, we study the biospheric implications of those ex- procreation, for freedom in the commons provides people with a ternalities. In contrast to competitive social preferences and to the stimulus to reproduce their numbers, other things equal (3). The problem of the commons, conformist social preferences give rise stimulus, which may of course be negligible for some communities to social environments that have the structure of coordination depending on their social mores, provides a direct link between the games, possibly possessing multiple stable equilibria, all parties commons and aggregate population size. No doubt the demands being better off in some than in others (14). Although conformist that each of us makes of the biosphere are small, but when summed preferences over reproduction can amplify the tragedy of the across all people and over time, they are substantial. The commons commons, they appear in the literature infrequently, perhaps create a social environment that resembles the classic prisoners’ because of concerns about state intervention in what are regarded Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. dilemma in game theory. as personal matters. Social preferences over consumption goods The excessive demands that we make of the biosphere (which in contrast are mentioned in the media on occasion, but they have we will also refer to as nature) are a manifestation of externalities, not yet entered policy discourse, perhaps because they too are which are the unaccounted consequences for others, including thought to be in the private domain. However, because social future people, of actions taken by one or more persons. As ex- preferences give rise to externalities, behavior responding to ternalities reflect institutional failure (e.g., the failure of nations to them should not be regarded as purely personal matters. reach agreement on the use of the global commons; the failure of Moreover, behavior based on such preferences can be shifted by a community to protect its local natural resource base through social mechanisms that do not involve state directives but instead, checks and balances; the failure of a society to award equal power bottom-up social mechanisms. The latter are likely to be better for women and men), the commons can only be rescued by rea- placed to bring about ecologically desirable changes. soned collective action. It has been customary to study ways to In this paper, we present a framework for identifying the factors avoid the tragedy of the commons by considering taxes and that have contributed to humanity’s overshoot in the demand for regulations at the national level (1), communitarian practices at the the biosphere’s goods and services. We then apply the notion of local level (6–8), and treaties at the international level (9). social preferences to the framework to argue that the social costs The underlying assumption about human behavior in many of of eliminating the overshoot may prove to be far less than is those studies has been that people are egoists. Here, we study currently feared. ways that are based on social norms that arise when people look to the decisions made by others when making their own decisions The Impact Equation and Unsustainable Development (10–16). For the purposes of managing the commons, there is thus Population size, living standards, and the technologies and insti- a role not only for governments but also, for communities and tutions in play shape humanity’s demand for the biosphere’s civil society. goods and services (18). The composition of the biosphere mat- We use the term socially embedded preferences or “social ters greatly, but here, we simplify by conceptualizing it as a self- preferences” for short to identify someone’s behavior and prac- organizing regenerative resource measured in, say, tons of bio- tices that are influenced by the behavior of others. Two broad mass. The biosphere’s primary productivity is the rate at which it classes of social preferences have been found to be empirically regenerates itself. As we are measuring the biosphere in units of significant: competitive and conformist. The former as in the biomass, primary productivity is the amount of carbon that is fixed phenomenon of conspicuous consumption [first studied by the (net of respiration) by autotrophs per unit of time, which for our sociologist Thorsten Veblen (10) and more recently given wide purposes here, amounts to the net rate at which carbon is fixed empirical support in refs. 11 and 12] displays a desire for a high from photosynthesis. Primary productivity is the source of energy social status relative to that of others. The latter embodies a desire for all living matter and is a flow. In contrast, the biosphere is a to be like others, meaning that a person prefers to be close to stock and is the locus of the products and services that we extract others in their choices rather than be distant from others (13). The from it. former gives rise to a pressure for higher consumption; the latter Humanity’s impact on the biosphere per unit of time is called leans toward conformity. Our behaviors in some spheres of life the “global ecological footprint.” To construct a measure of that have been found to be competitive (11, 15); in others, they have impact, let N be human population and y an index of human ac- been found to be conformist (13, 16), and in still others, they seem tivity per person per unit of time. Estimating average human ac- to resemble those of the egoist, familiar in economics textbooks. We tivity per person poses huge measurement problems; therefore, study the implications of competitive behavior over consumption in for tractability, we suppose that it corresponds to the standard the rich world and conformist behavior over reproduction in the of living as measured by gross domestic product (GDP) per capita. world’s poorest countries as templates for understanding the The move almost surely yields an underestimate of what we are Barrett et al. PNAS | March 24, 2020 | vol. 117 | no. 12 | 6301 after because there are many human activities that are not cap- In recent decades, Ny/α has exceeded G, thus drawing down tured in the market value of all final consumption goods and S. Studies of biogeochemical signatures of the past 11,000 y have services in a year. Global GDP is thus a mere proxy for what we are tracked the human-induced evolution of soil nitrogen and phos- seeking. On occasion, national income statisticians have offered phorus in inventories of sediments and ice (21). They have estimates of the magnitude of economic transactions that are reported that the hockey stick shape that has been found in time missing in GDP, but they are too scanty to be of use here. series of mean global temperature also characterizes many global Degradation of the biosphere can hasten the depreciation of biogeochemical signatures over similar timescales: they display a other assets (rising sea level submerges coastal infrastructure, flat trend over millennia until some 250 y ago followed by a slow hotter weather lowers labor productivity, and so on). Here, we increase until the middle of the 20th century and then, a sharp rise focus instead on the demand that we make of the biosphere’s that continues at a fast rate to this day. We interpret that feature of goods and services. That demand takes two broad forms. We the time series as a continual and growing rise in Ny/α. Biological harvest nature’s goods and use nature’s services for consumption extinctions, increases in greenhouse gas concentrations, con- and production. (Fish, timber, and fresh water constitute goods; tamination of marine species with persistent pollutants, nutrient pollination, water purification, flood protection, and sequestration overload in soil and water, and oceanic dead zones are reflections and storage constitute services.) We also degrade natural capital of the biospheric overload, meaning that S has declined, with the stocks through overharvest and through overuse as pollution attendant reduction in G, which in turn, has increased the gap sinks. Let X denote harvest demand and Z denote the demand between Ny/α and G. Because the decline in S would appear to that we make of the biosphere as a pollution sink. Both are have been dramatic since the middle of the last century, mid-20th functions of Ny, and therefore, we may write X = X(Ny) and Z = century has been proposed as the time that we entered the Z(Ny). This way of partitioning the global ecological footprint says Anthropocene (21). Crude estimates (22) suggest that the ratio of that pollution is the reverse of conservation (19). Ny/α to G(S) is today 1.7. Let αX be a numerical measure of the efficiency with which the Over the years, a large body of work has found that bio- Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. biosphere’s goods and services are converted into GDP, and let diversity is a major determinant of the biosphere’s productivity αZ be a numerical measure of the extent to which we are able to and its ability to withstand stresses and shocks (22, 23). Alarm- treat our waste products before discharging them. Therefore, we ingly, current extinction rates of species in various orders appear have X = Ny/αX and Z = Ny/αZ. Humanity’s demand from the to be 100 to 1,000 times higher than their average rate (about biosphere in a given year is then (Ny/αX + Ny/αZ). Define α = (αX + 1 per million species per year) over the past several million years αZ)/αXαZ. Then, (Ny/αX + Ny/αZ) ≡ Ny/α is our proxy measure of the (24, 25). In the last 40 y, there has been a 60% decline on average global ecological footprint. If the footprint exceeds the bio- in the populations of mammals, birds, fish, reptiles, and amphib- sphere’s regenerative rate, the stock diminishes. Similarly, if the ians, mostly in the tropics (26). Biodiversity loss and climate footprint is less than the biosphere’s regenerative rate, the stock change are now widely regarded to be the two most significant increases. (Bifurcations leading to regime shifts and irreversibil- global environmental problems that we face today. Moreover, ities can be incorporated into this analysis.) However, population, each influences the other (27). To the extent that they reinforce GDP per capita, or both could increase without making additional one another, policies that curb one can be expected to temper demands on the biosphere provided either αX or αZ and thus, α the other. were to increase correspondingly. Improvements in technology These findings are consistent with macroeconomic data (3). (e.g., substituting degradable waste for persistent pollutants; From 1950 to 2019, global GDP per capita increased nearly decarbonizing the energy sector) as well as institutions and prac- tices (e.g., establishing protected areas; reducing food waste) and fivefold from about 3,500 international dollars (i.e., US $3,500 appropriate redistributions of wealth are among the means by purchasing power parity) to about 17,000 international dollars in which α can be raised. 2011 prices, even as population trebled in size (from ∼2.5 billion Let G denote the regenerative rate of the biosphere. G de- in 1950 to around 7.7 billion in 2019). A nearly 15-fold increase in pends on the biosphere’s stock, which we write as S. Thus, G = global economic activity in 70 y, taking place from a base year G(S). The G function can be affected by policy. The application of global GDP per capita that was higher than ever before, may give biotechnology in agriculture is one avenue. Another is “ecosystem one hope of a future of perpetual material advancement, but engineering.” A recent experiment in American Samoa, for exam- when the economic data are paired with environmental indicators, ple, found that transplanted heat-tolerant corals were more likely to it would appear that the rise in economic activity has been built on survive a bleaching event than less tolerant local corals, enabling unsustainable foundations. Produced capital and human capital quicker recovery of the ecosystem after such an event (20). In the have no doubt increased by many fold, but as a stock of natural range of stocks we are concerned with here (stocks below the level capital, the biosphere has shrunk considerably. capable of sustaining a maximum yield), dG/dS > 0. In the long run, humanity’s ecological footprint must equal the Humanity’s ecological footprint does not have to equal the biosphere’s regenerative rate. That equality represents a sus- biosphere’s regenerative rate because the difference would be tainable state of affairs. The widely discussed United Nations’ accommodated by a change in the biosphere’s stock S. A world (UN) Sustainable Development Goals (SDGs) set in 2015 were rich in healthy ecosystems could, on utilitarian grounds, choose to formulated on the assumption that they can be attained, but draw down the biosphere and use the goods and services that it the background documents did not probe the question whether supplies so as to accumulate produced capital and human capital. the goals are sustainable in a world that simultaneously enjoys That is what economic development has come to mean among growth in global GDP. many thinkers, but the scenario comes in tandem with an over- We, therefore, work backward by first identifying a condition shoot in our demands from the biosphere. The overshoot cannot the global economy’s treatment of the biosphere must satisfy if be maintained indefinitely because our life support system would the SDGs are themselves to be sustainable. That condition points be threatened. to ways in which S can be stabilized. To sustain a value of S 6302 | www.pnas.org/cgi/doi/10.1073/pnas.1909857117 Barrett et al. requires that the global ecological footprint equals the bio- difference in their projections of global population size in 2050, sphere’s regenerative rate: that is, and both project a larger population in 2100 than population today. Therefore, our analysis here is unaffected by differences in Ny=αX + Ny=αZ ≡ Ny=α = GðSÞ. [1] the two publications. While the projections in ref. 28 are built primarily on forecasts of women’s education and other objectives Eq. 1, which we call the Impact Equation, applies to the bio- in the SDGs, including the global standard of living, a multitude of sphere as a whole. Although the notion of ecological footprint (the factors is implicit in the UN projections (30). However, because the left-hand side of Eq. 1) can be applied to any group of individuals— UN projections of alternative future trajectories, cast in probabi- from the individual and the household to nations and the global listic terms, are more widely known, we cite them here. population—trade in commodities and services breaks the link The UN’s median projection of world population in year 2100 between demand (Ny/α) and supply [G(S)] for economic units is 10.9 billion, with a 95% certainty range of between 9.4 and smaller than the world as a whole. The ecological footprint of a 12.7 billion (29). More than three-quarters of the increase from nation will not balance the regenerative rate of its ecosystems if its today’s 7.7 billion is expected to be in sub-Saharan Africa, where trade in the biosphere’s goods and services does not balance in population in 2100 is projected to rise from today’s ∼1.1 billion to units of biomass. Of course, it could be that a country pays for its 3.8 billion (Fig. 1). Comprising around 14% of the world’s pop- imports, perhaps even at their appropriate prices, but that is a ulation, the region represents only a bit more than 3% of the world different matter. Here, we are only formulating a way to break economy. Therefore, sub-Saharan Africa cannot remotely be held down the global imbalance of demand and supply of those goods responsible for the global environmental problems that we face and services into imbalances among groups in the global pop- today. However, attempts to raise incomes there even to the ulation; we are not discussing “fair trade.” current global average income (∼17,000 international dollars) in Decoupling the demands that humanity makes on the bio- the face of a near-3 billion rise in numbers will require an increase sphere Ny/α in the left-hand side of Eq. 1 also serves to remind us in the region’s annual output from 3.5 trillion international dollars that measures to reduce environmental pollution, Z, can raise our Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. to about 68 trillion international dollars at today’s prices. That rise, demand for the biosphere’s products, X = Ny/αX. Solar panels assuming that it is possible, is likely to have adverse conse- require minerals such as aluminum, cadmium, and zinc. However, quences for the region’s ecology, contributing to further societal obtaining those minerals usually requires destroying forests. conflicts there and greater attempted population movements The literature on the economics of climate change offers a both within the region and out of it (31). convenient illustration of the need to move to a world where the The SDGs are reticent about population, and yet, it is difficult Impact Equation is reached. Much of that literature has taken Ny to imagine that they can be met without addressing the subject. It as given under various scenarios or pathways, focusing instead on has been argued that the goal to restrict the increase in mean mitigation: that is, on raising αZ by decarbonizing the economy global temperature to 2 °C from that in the preindustrial revolu- and removing CO2 from the atmosphere. However, the reason tion era is unlikely to be met unless population growth is reduced that CO2 concentrations in the atmosphere have increased is not substantially (31), and yet, even the recent Paris Agreement on only that mitigation and carbon removal measures have been climate change has made no mention of population. Population in slow; it is also that growth in both GDP per capita and population sub-Saharan Africa has risen sharply since the middle of the 20th have been strong. In this paper, we point to ways for reducing Ny century because the substantial reductions in the death rate there or putting the world onto a more favorable pathway for limiting were not matched by reductions in fertility. Today, women in sub- climate change. Saharan Africa have on average around 4.7 births over the course Why has αZ not risen more? The reason is a low rate of inno- of a lifetime in contrast to a world average of 2.5. The total fertility vation and investment in nonfossil fuel energy sources and carbon rate in India has fallen to 2.2, while that in China (at 1.7) is well capture and storage technologies. However, these low rates, in turn, have been caused by a persistent and pervasive institutional failure: the failure to achieve collective action in limiting climate 12 Billions change. Despite nearly 30 y of diplomatic effort, the world has been unable to overturn the tragedy of the climate commons. 10 The externalities relating to fertility and consumption have other causes, and they can and should be addressed by other institu- 8 tions. Doing so will not make up for the failure to address climate Total Populaon change directly, but it will help to keep equilibrium global mean 6 temperature closer to the levels advocated for in international conferences. The variables on the left-hand side of Eq. 1, however, point to 4 proximate causes behind the size of our ecological footprint. As ways to raise αZ have been much discussed in the literature on 2 climate change, we explore factors that influence future N and current and future y. - 1950 2000 2050 2100 Fertility in Low per Capita Income Settings Year There are two prominent global population projections for the Oceania Northern America Lan America and the Caribbean Europe Asia Africa future: one emerging from the International Institute for Applied Fig. 1. Population, 1950 to 2020; projections 2020 to 2100. Systems Analysis (28) and the other from the UN Population Human population is by region; data are available at https://population. Division in ongoing work over the past decades (29). There is little un.org/wpp/. Barrett et al. PNAS | March 24, 2020 | vol. 117 | no. 12 | 6303 below replacement level (29). Sub-Saharan Africa’s slower fertility and intersects the 45° line from 0 to F at three points (B, C, and D), decline has been traced to many reasons, including inheritance each of which is a social equilibrium. B and D are stable, while C is rules, the prevalence of polygamy, lack of access to modern unstable. In the figure, which is purely illustrative, every household methods of contraception, low education among women, and desires d children if all other households have d children each kinship obligations (32–34). Here, we examine a pathway that has and b if each among all others have b. Imagine now that every been less discussed but that offers an explanation for why the household prefers the outcome in which all households have b desired number of children should have remained substantially children each to the one in which all have d children each. As higher in sub-Saharan Africa than elsewhere. The account of the having either b or d children is a stable equilibrium, a fertility rate pathway combines history with social mechanisms as seen through of d would be just as tenacious as a fertility rate of b. the lens of economics. The theoretical observation that reproductive behavior is Fertility practices are not only influenced by private desires and guided by conformist preferences [more particularly, that the wants; they are also shaped by societal mores. Anthropologists drivers of such behavior can give rise to multiple reproductive have reported that women in sub-Saharan Africa acknowledge that equilibria (14)] was given support in a study of contraceptive use in they are able to acquire social status through reproductive success rural Kenya, which found that, in communities having dense social (33). This attitude toward reproduction has been called “Children as networks and a poorly developed market economy, a woman wealth” (35). would be unlikely to use contraceptive methods if contraception Using this as a template, we may say that reproductive be- use in her network was low, whereas she would be likely to use havior is “conformist” when the family size that a household de- such methods if contraception use in her network was high (36). sires is positively related to the average family size in the community The theory was also given support in an analysis of a natural ex- or more broadly, in the world that households come into contact periment that found that state-level fertility rates declined in step with (15). Conformism gives rise to externalities harboring multiple following staggered introductions of cable television across In- equilibria. As long as all others aim at large families, no household dian states in the 1980s (37). Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. will wish to deviate from the practice; if, however, all other house- Further support has been provided in a recent analysis of holds were to restrict their fertility, every household would wish to contraceptive uptake in Bangladesh (38). The study concerned restrict its fertility. A society can thus get embedded in a self- women living in the same community but belonging to different sustaining mode of behavior characterized by high fertility and religious groups. After controlling for individual differences in stagnant living standards, even when there is another potentially education, age, wealth, and the like, the study found that a self-sustaining mode of behavior that is characterized by low fertility woman’s choice to use contraception depended strongly on the and rising living standards, which is preferred by all (14, 15). The predominant choice made by other women in her religious group structure of social interactions arising from conformist preferences is and was unaffected by the predominant choice made by women known as coordination games. belonging to the other group. Conformist preferences are depicted in a stylized form in Fig. Persistent reproductive practices that go counter to present- 2. The curve ABCDE is a household’s desired number of children day interests may of course have had a rationale in the past. plotted against the average number of children per household However, even when circumstances have changed, a society can (the horizontal axis) in the community. The curve is upward sloping remain stuck in a mode of behavior characterized by high fertility (d) even when there is an alternative that is characterized by low fertility (b), which would be preferred by all. This suboptimal sit- uation disappears when tradition breakers (e.g., more educated women) lead the way to move toward smaller families. Newspa- pers, radio, television, and the internet communicate information about other lifestyles. The media can be a vehicle by which con- formism becomes based on the behavior of a wider population than the local community, which disrupts existing practices. Community discussions on the benefits of smaller family size can coordinate households to act on lowered fertility targets. Eco- nomic demographers have interpreted the persistence of high fertility in sub-Saharan Africa to a strong desire for children (39). Conformist preferences point to a different interpretation. Be- havior based on such preferences can be expected to display a strong positive correlation between fertility desires and fertility outcomes (as is shown in ref. 39), but causality should not be at- tributed to the relationship. Conformist preferences tell us that it would be as true to say that fertility rates in a country are high because people there have a strong desire for children as it would be to say that people there have a strong desire to have children because fertility rates are high (15). Fig. 2. Conformist preferences for children. The vertical axis shows a Fortunately, the transition from d to b children per household household’s desired number of children; the horizontal axis shows does not require coercion, taxation, or even education. A con- the average fertility rate in the society to which this household formist behavior can be shifted by changing expectations about belongs. The curve from A to E shows the household’s preferences. The curve from 0 to F shows the 45° line at which the desired number others’ fertility choices. Family planning programs can be so of children equals the average fertility rate. Points B, C, and D designed as to encourage members of communities to share in- represent equilibria. Points B and D are stable; point C is unstable. formation about modern methods of contraception and discuss 6304 | www.pnas.org/cgi/doi/10.1073/pnas.1909857117 Barrett et al. the advantages of smaller families. As routes to fertility transition, billion would be 336 trillion international dollars or nearly three investment in community-based family planning programs should times the current level. The SDGs speak of environmental pro- be regarded as essential (40). tection but would not seem to have been influenced by humanity’s By providing access to subsidized contraceptive commodi- prospective demands of the biosphere’s goods and services and of ties and services, family planning programs were successful in its assimilative capacity. If that demand is to be brought down to accelerating fertility declines in Asia and Latin America in the sustainable levels, either global output (Ny) will have to decline 1960s to 1980s. The rationale for vigorously expanding the content substantially or technological advances and institutional reforms will and reach of such programs today lies in the 53 million women in have to reduce the environmental impact of high consumption sub-Saharan Africa who report that they want to stop or delay levels (i.e., α will have to be increased). Above, we have explored childbearing but are not using any modern method of contracep- ways in which growth in N could be reduced. However, as the tion (41). Some use traditional methods; nevertheless, it should not remedy of reducing y is also typically neglected, we focus on it here. be surprising that 43% of the 206 million pregnancies in developing Consumption behavior is influenced both by our urge to compete regions in 2017 were unintended: that is, they occurred too soon or with others (Veblen’s “conspicuous consumption”) and by our innate were not wanted at all (42). Many unintended pregnancies end in desire to conform (fads and fashions are only the most striking ex- abortion, a significant proportion of which are performed under amples). Building on a wealth of sociological data, Bourdieu (11) has unsafe conditions. argued that our taste for consumption goods and services is sub- In addition to reducing unintended pregnancies, contracep- stantially influenced by the tastes of others in our peer group and tive use among women enhances their own health and that of aspirational groups. Here, we follow that line of inquiry and regard their children by spacing births and by providing greater oppor- consumption choices as being based on social preferences over a tunity for education. Yet, family planning is a neglected feature of category of goods if the (positive) benefit to an individual of an ad- contemporary public policy. Currently, only about 0.6% of over- ditional unit of consumption of a commodity in that category in- seas development assistance is awarded to it (43). Moreover, creases with the number of people who consume that commodity. Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. developing countries relegate family planning expenditures to To illustrate how an environmentally pernicious category of minor government departments. goods can get established, suppose that there are two categories Despite evidence that family planning reduces poverty, the of consumption goods. Category 1 goods are intensive in the use World Bank has given low priority to family planning. Although the of the biosphere, whereas the production of category 2 goods is SDGs include universal access to sexual and reproductive health intensive in the use of human capital. If the biosphere was freely and emphasize reproductive rights, the family planning indicator accessible, category 1 goods would be cheaper than category 2 that is advanced (SDG indicator 3.7.1) is focused on the satisfac- goods, other things equal. It is then easy to see why people would tion of the individual woman’s expressed desires; it does not value have been drawn to category 1 goods in the past and how that increases in the demand for family planning by community en- would have triggered further demands for category 1 goods, thus gagement (15). The introduction of family planning-related media fueling an ever-expanding demand for them. Nor would we expect messages is credited for part of the rise in contraceptive demand technological advances to be directed at lessening the production and use in countries where such communication is active. Be- costs of category 2 goods. Quite the contrary, competing entre- tween 2000 and 2015, the demand for contraception among preneurs would invest in technologies that economize on factor women of reproductive age rose from 32 to 41% in Ethiopia, from inputs other than natural capital in the production of category 48 to 60% in Malawi, and from 30 to 44% in Rwanda (41). In 1 goods. Also, that would further fuel the global demand for cat- contrast, little or no change in demand was observed in countries egory 1 goods. Moreover, the greater the difference is between the such as Nigeria and the Democratic Republic of Congo, where (market) costs of production of goods in the two categories, the family planning programs are weak and continue to be under- bigger would be the social effort required to move consumers away mined by fragile states and conflict. Moreover, the questions that from category 1 goods to category 2 goods. Difference between are directed at women in questionnaires to elicit their fertility the two costs structures would be accentuated by the existence of desires continue to be ill-formed: They do not ask women what fixed capital in the sector producing category 1 goods. Retiring their desired fertility would be were the fertility choices of their machines and equipment before their intended dates can be costly. peers to change. The acknowledgment that reproductive prefer- We are experiencing that phenomenon even as nations try to move ences are socially embedded should be reflected in the ques- away from industries based on fossil fuels. tionnaires used to identify women’s fertility intentions (15). The current structure of market prices works against our com- mon future; the biosphere is precious but priced cheaply. To shift Consumption by the Rich Minority consumption patterns in the rich world and the aspiring con- The World Bank (44) has reported that the ∼1.2 billion people on sumption patterns of the poor world and those of emerging its list of high-income countries enjoy a per capita GDP of 46,700 economies away from resource-intensive goods and services will international dollars, implying a GDP of 56 trillion international require massive, coordinated actions. In a different context, econ- dollars. World output today is about 120 trillion international omists use the term “big push” to indicate the investment expen- dollars. Assuming that our ecological footprint is proportional to ditures and collective determination that are required to move an the scale of economic activity, close to 50% of humanity’s impact economy from one equilibrium to another superior one. Fortu- ($56 trillion/$120 trillion) on the biosphere can be attributed to nately, when social preferences dominate consumption choice, the some 16% of world population. That is of course the current state psychological cost to a person of a collective reduction in con- of affairs. Output growth in the rest of the world can be expected sumption is likely to be a lot less than what it would be if she was to to make yet further demands on the biosphere. If, as a reasonable reduce consumption unilaterally (15). For example, it has been ar- aspiration, per capita income in 2100 were 30,000 international gued that, as between two societies (one in which everyone owned dollars at today’s prices (a figure close to the average in today’s a large house and had little time for exercising or socializing and middle income countries) global output at a population of 10.9 another in which everyone owned a modest-sized home but had Barrett et al. PNAS | March 24, 2020 | vol. 117 | no. 12 | 6305 more time for these other activities), subjective wellbeing would be it is as well to ask whether reductions in inequality could be higher in the second society (45). The point made by the author is expected to raise or lower emissions, all else being equal. How that, to the extent that house size is a positional good, people care would income redistribution affect humanity’s ecological foot- more about how much they have relative to others than with the print? To study that, we decompose the aggregate demand that absolute size of the house that they live in. By limiting house size, we make of the biosphere’s goods and services into the demands however, resources can be devoted to activities that are known to of various income groups. Let i, j denote households. Households increase wellbeing. The argument in ref. 45 extends to shifting differ according to their incomes yi, yj, and so on, but they differ expenditure away from positional goods toward public goods (45), also as regards the efficiency with which they convert the bio- including a healthier biosphere. sphere’s goods and services into income. It is conventional to view Although competitive impulses create a tragedy of the com- inequality in terms of the distribution of household incomes, but mons as regards positional goods, conformist preferences can here, we are interested in the distribution of ecological footprints create a positive feedback leading to a mutually preferred col- as well. The latter is reflected in the distribution of the yi/αi. Also, yi lective outcome. A study in California shows that an additional and yi/αi are not the same. installation of solar photovoltaic panels increases the probability Without loss of generality, let us label households in terms of of adoption within the same zip code by 0.8%, with the effect increasing income. Therefore, yi < yi+1 for i = 1, . . ., N. There are being particularly strong at the localized street level (46). Infor- strong reasons for thinking that ecological footprint is an in- mation campaigns also shape behavior. One US study found that creasing function of income, but the matter has been explored electricity customers receiving information about their consump- mainly for the case of carbon emissions (49, 50). A question of tion relative to that of their neighbors reduced energy consumption interest is whether the functional relation is convex or concave. by 2% (47). Yet another randomized field experiment found that Consider an income interval where the function is convex. An social comparison lowered water consumption by nearly 5% (48). egalitarian redistribution of incomes among households in that That a move to environmentally friendly behavior would be a interval would lead to a smaller global ecological footprint, im- Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. gesture toward our descendants is, of course, an added bonus. plying that there is no conflict between income equality and Social and Technological Transformations to Resolve biospheric integrity. However, in a concave interval, the reverse Population and Consumption Dilemmas holds: egalitarian redistributions of incomes would lead to larger In pushing the biosphere deeper into unsustainable territory, the global ecological footprints, and society would face a cruel choice large and growing scale of the human enterprise as reflected in between income equality and the biosphere’s integrity (15). both per capita consumption and population poses a threat to There are thus three reasons why we should limit both fertility wellbeing worldwide. Improvements in technology and institu- and per capita GDP. First, institutions and technology have failed tions are critical. Conceivably, they can even offset or more than to meet the problem of climate change, which implies that limit- offset the harmful effects of scale. Indeed, changes in technology ing our global ecological footprint will prove to be even more and institutions have successfully transformed some sectors. By difficult unless we are willing to entertain the possibility of limiting banning chlorofluorocarbons, the Montreal Protocol is expec- Ny. Second, per capita consumption and population are shaped ted to restore stratospheric ozone to its pre-1980 level by mid- by different institutions than the ones that have kept α low. century. Lead has been phased out of petrol worldwide. Much less Addressing these elements of the Impact Equation (Eq. 1) requires oil enters our seas than was the case a few decades ago thanks to social change and not (certainly not exclusively) policy change. changes in tanker design mandated by international agreements. Moreover, the appropriate levels for bringing about such change In other areas, including biodiversity loss, degradation of crit- are local and national rather than regional and global, and our ical ecosystems, and climate change, our collective efforts have institutions are stronger at the local and national levels than at the not only been less successful; they have failed. For example, regional and global levels. Recognition that behavior of people calculations by the Secretariat of the UN Framework Convention everywhere is in substantial measure socially embedded in ways on Climate Change show that, even if countries fulfill their Paris that are not always apparent to external analysts should give us pledges, CO2 emissions will continue to increase through 2030, hope that correcting the social externalities associated with con- virtually guaranteeing that the collective goal of limiting global sumption and fertility will improve wellbeing directly and not only mean temperature change to well below 2 °C relative to the pre- by serving to reduce emissions. industrial level will be breached. Costs of wind turbines and solar cells have fallen, but there have also been improvements in fossil Data Availability. All data associated with the manuscript are fuel extraction technology (fracking). Decarbonization of the global provided within the manuscript. energy system is proceeding slowly because institutions like Paris have failed to avert the tragedy of the climate commons. Use of the Acknowledgments atmosphere as a repository for CO2 continues to be priced low, Kenneth J. Arrow participated fully in the meeting in which we discussed the even though the social costs of carbon are high. topic of this paper and in follow-up emails, shared with us “a few observations, A solution being considered now is to address climate change which, I think, have to be taken account of in our paper,” all of which, we believe, are incorporated here. He died on February 21, 2017. Our collaboration was jointly with income inequality in the form of a domestic “green supported by the Beijer Institute of Ecological Economics of the Royal Swedish new deal” (House Resolution 109, 116th US Congress, 2019). The Academy of Sciences. The views expressed herein are entirely those of author motivation would seem to be of a political economy nature, but A.D. and do not necessarily reflect the views of the UN. 1 P. Dasgupta, T. Mitra, G. Sorger, Harvesting the commons. Environ. Resour. Econ. 72, 613–636 (2019). 2 G. Hardin, The tragedy of the commons. The population problem has no technical solution; it requires a fundamental extension in morality. Science 162, 1243– 1248 (1968). 6306 | www.pnas.org/cgi/doi/10.1073/pnas.1909857117 Barrett et al. 3 P. Dasgupta, Time and the Generations: Population Ethics for a Diminishing Planet (Columbia University Press, New York, NY, 2019). 4 W. D. Nordhaus, Managing the Global Commons: The Economics of Climate Change (MIT Press, Cambridge, MA, 1994). 5 W. D. Nordhaus, A review of the Stern Review of the Economics of Climate Change. J. Econ. Lit. 45, 686–702 (2007). 6 D. Feeny, F. Berkes, B. J. Mccay, J. M. Acheson, The tragedy of the commons: Twenty-two years later. Hum. Ecol. 18, 1–19 (1990). 7 E. Ostrom, Governing the Commons: The Evolution of Institutions for Collective Action (Cambridge University Press, Cambridge, UK, 1990). 8 K. Nyborg et al., Social norms as solutions. Science 354, 42–43 (2016). 9 S. Barrett, Coordination vs. voluntarism and enforcement in sustaining international environmental cooperation. Proc. Natl. Acad. Sci. U.S.A. 113, 14515–14522 (2016). 10 T. Veblen, The Theory of the Leisure Class: An Economic Analysis of Institutions (George Allen & Unwin, London, UK, 1925). 11 P. Bourdieu, Distinction: A Social Critique of the Judgement of Taste (Routledge and Kegan Paul, London, UK, 1984). 12 J. Helliwell, R. Layard, J. Sachs, Eds., World Happiness Report 2013 (Earth Institute, New York, NY, 2013). 13 M. Douglas, B. Isherwood, The World of Goods: Towards an Anthropology of Consumption (Routledge, London, UK, 1979). 14 P. Dasgupta, An Inquiry into Well-Being and Destitution (Clarendon Press, Oxford, UK, 1993). 15 A. Dasgupta, P. Dasgupta, Socially-embedded preferences, environmental externalities, and reproductive rights. Popul. Dev. Rev. 43, 405–441 (2017). 16 O. Johansson-Stenman, F. Carlsson, D. Daruvala, Measuring future grandparents’ preferences for equality and relative standing. Econ. J. (Lond.) 112, 362–383 (2002). 17 C. Bledsoe, Contraception and natural fertility in America. Popul. Dev. Rev. 22 (suppl. 1), S297–S324 (1996). 18 P. R. Ehrlich, J. P. Holdren, Impact of population growth. Science 171, 1212–1217 (1971). 19 P. Dasgupta, The Control of Resources (Harvard University Press, Cambridge, MA, 1982). 20 M. K. Morikawa, S. R. Palumbi, Using naturally occurring climate resilient corals to construct bleaching-resistant nurseries. Proc. Natl. Acad. Sci. U.S.A. 116, 10586– 10591 (2019). 21 C. N. Waters et al., The Anthropocene is functionally and stratigraphically distinct from the Holocene. Science 351, aad2622 (2016). 22 M. Wakernagel, B. Beyers, Ecological Footprint: Managing Our Biocapacity Budget (New Society, Gabriola Island, BC, Canada, 2019). 23 D. U. Hooper et al., Effects of biodiversity on ecosystem functioning: A consensus of current knowledge. Ecol. Monogr. Ecological Soc. Am. 75, 3–35 (2005). 24 D. Tilman, F. Isbell, J. M. Cowles, Biodiversity and ecosystem functioning. Annu. Rev. Ecol. Evol. Syst. 45, 471–493 (2014). 25 S. L. Pimm, P. H. Raven, “The state of the world’s biodiversity” in Biological Extinctions: New Perspectives, P. Dasgupta, P. H. Raven, A. McIvor, Eds. (Cambridge University Press, Cambridge, UK, 2019), pp. 80–112. Downloaded from https://www.pnas.org by Universitat Autonoma de Barcelona (UAB) on March 1, 2022 from IP address 158.109.88.156. 26 M. Grooten, R. E. A. Almond, Eds., Living Planet Report - 2018: Aiming Higher (World Wildlife Fund, Gland, Switzerland, 2018). 27 T. Lovejoy, L. Hannah, Biodiversity and Climate Change: Transforming the Biosphere (Yale University Press, New Haven, CT, 2019). 28 G. J. Abel, B. Barakat, S. Kc, W. Lutz, Meeting the Sustainable Development Goals leads to lower world population growth. Proc. Natl. Acad. Sci. U.S.A. 113, 14294–14299 (2016). 29 United Nations Population Division, World Population Prospects: 2019 Revision (United Nations, New York, NY, 2019). 30 United Nations Population Division, World Population Prospects 2019: Highlights (United Nations, New York, NY, 2019). 31 C. Juma, “Game over? Drivers of biodiversity loss in Africa” in Biological Extinction: New Perspectives, P. Dasgupta, P. H. Raven, A. McIvor, Eds. (Cambridge University Press, Cambridge, UK, 2019). 32 B. C. O’Neill et al., Global demographic trends and future carbon emissions. Proc. Natl. Acad. Sci. U.S.A. 107, 17521–17526 (2010). 33 C. Bledsoe, “‘Children are like young bamboo trees’: Potentiality and reproduction in sub-Saharan Africa” in Population, Economic Development and the Environment, K. Lindahl-Kiessling, H. Landberg, Eds. (Oxford University Press, Oxford, UK, 1994). 34 J. Bongaarts, J. Casterline, Fertility transition: Is sub-Saharan Africa different? Popul. Dev. Rev. 38 (suppl. 1), 153–168 (2013). 35 J. I. Guyer, S. M. Eno Belinga, Wealth in people as wealth in knowledge: Accumulation and composition in Equatorial Africa. J. Afr. Hist. 36, 91–120 (1995). 36 H.-P. Kohler, J. R. Behrman, S. C. Watkins, The density of social networks and fertility decisions: Evidence from South Nyanza district, Kenya. Demography 38, 43–58 (2001). 37 D. T. Jensen, E. Oster, The power of cable TV: Cable television and women’s status in India. Q. J. Econ. 124, 1057–1094 (2009). 38 K. Munshi, J. Myaux, Social norms and the fertility transition. J. Dev. Econ. 80, 1–38 (2016). 39 L. H. Pritchett, Desired fertility and the impact of population policies. Popul. Dev. Rev. 20, 1–55 (1994). 40 J. Bongaarts, Can family planning programs reduce high desired family size in Sub-Saharan Africa? Int. Perspect. Sex. Reprod. Health. 37, 209–216 (2011). 41 United Nations Population Division, Estimates and Projections of Family Planning Indicators 2019 (United Nations, New York, NY, 2019). 42 Guttmacher Institute, Adding It up: The Costs and Benefits of Investing in Sexual and Reproductive Health (Guttmacher Institute, New York, NY, 2017). 43 C. Grollman et al., Donor funding for family planning: Levels and trends between 2003 and 2013. Health Policy Plan. 33, 574–582 (2018). 44 World Bank, World Development Indicators (World Bank, Washington, DC, 2017). 45 R. H. Frank, The frame of reference as a public good. Econ. J. (Lond.) 107, 1832–1847 (1997). 46 B. Bollinger, K. Gillingham, Peer effects in the diffusion of solar photovoltaic panels. Mark. Sci. 31, 900–912 (2012). 47 H. Allcott, Social norms and energy conservation. J. Public Econ. 95, 1082–1095 (2011). 48 P. Ferraro, M. K. Price, Using nonpecuniary strategies to influence behavior: Evidence from a large-scale field experiment. Rev. Econ. Stat. 95, 64–73 (2013). 49 R. Schmalensee, T. M. Stoker, R. A. Judson, World carbon dioxide emissions: 1950-2050. Rev. Econ. Stat. 80, 15–27 (1998). 50 W. T. Harbaugh, A. Levinson, D. Molloy Wilson, Reexamining the empirical evidence for an environmental kuznets curve. Rev. Econ. Stat. 84, 541–551 (2002). Barrett et al. PNAS | March 24, 2020 | vol. 117 | no. 12 | 6307