The most important contribution of prospect theory to the concerns of our political era.
The pioneering prospect theory work of Daniel Kahneman and Amos Tversky was appropriately lauded in a McArthur award to Amos in 1984 and in the 2002 Nobel Prize awarded to Danny shortly after Amos’ death. A decade later Kahneman’s (2011) account of that work in Thinking Fast and Slow attracted attention far beyond academia, and the tale of the remarkable collaboration and life-long friendship between Amos and Danny was brilliantly told by Michael Lewis a few years later in The Undoing Project. I doubt that any body of work in psychology since Skinner and Freud has been as widely known within the academic community. I am sure that none has generated as much in the way of theory and research within what is now the thriving area of behavioral economics (a label which I and many of my colleagues in psychology consider a “name change for business reasons” that fails to acknowledge the primacy of the pioneering work done within our discipline of psychology).
Kahneman and Tversky (1979, 1984; see also Kahneman, 2011) combined deep insight about human motivation and decision-making with a talent that we recognize as the particular forte of experimental social psychologists, especially those who dominated our field in its golden age. That talent involved the design and execution of clever experiments—experiments that captured and “bottled” phenomena in a way that illustrated more general theoretical insights through non-obvious demonstrations. The most famous of these demonstrations (McNeil et al, 1982) showed effects of framing on judgment and decision making that violated normative principles, most notably a between-condition difference in preferred medical treatments for lung cancer (surgery vs radiation) made by physicians when the risks and benefits of the two procedures were framed in terms of lives lost rather than lives saved.1
Later investigators showed similar framing effects for the desirability of mammograms framed in termed of advantages for undoing that test versus the disadvantages of not undergoing it (Banks et al, 1995), the attractiveness of retiring at 65 versus 68 framed in terms of the reward for “delaying” retirement versus the penalty for “early” retirement (Fetherstonhaugh & Ross, 1999) and a variety of other decisions. The decisions explored in these studies were hypothetical, but they dealt with important tradeoffs involving costs, benefits, and risks, and made the potential real-world relevance of decisional framing clear. At this point in our history, the framing of advantages and disadvantages of opening the economy in the face of the continuing COVID 19 threat—and who does the framing—are literally matters of life or death. Looking only a bit further into the future, framing issues relating to the economic and ecological tradeoffs involved in combating climate change, and other issues including immigration and international trade, may be even more important determinants of the future facing our country.
Nevertheless, I believe that the feature of Kahneman and Tversky’s work on Prospect Theory that ultimately is most important is not the effects of framing. Rather, it is the explication of loss aversion and the closely related endowment effect. The theoretical point they were making was best illustrated in the case of financial gambles. Both research findings and exercises in thinking experiments confirm that while decision-makers are “risk aversive” in the domain of potential financial gains they are “loss averse”, or “risk tolerant” when it comes to prospective losses. That is, they prefer a certain gain to a gamble that has an equal or even somewhat higher expected return when one multiplies the size of that gain by the probability of realizing it; but they prefer to gamble rather than accept a certain loss given equivalent magnitudes and probabilities. If any illustrative figure in the field of psychology or behavioral economics is to be featured in textbooks in the decades to come, it is surely the pair of convex curves illustrating these risk-taking preferences regarding prospective gains and losses.
Expanding the domain of prospect theory
Prospect theory in general, and loss aversion in particular, were explored by Kahneman and Tversky & Co. and those following in their footsteps using decisions where prospective gains and losses were readily quantifiable and explicitly specified —in particular, dollars or gained or lost and deaths accepted or prevented. Furthermore, to compare choices made under different framing conditions and identify departures from relevant normative prescriptions, the exact probabilities associated with the prospective outcomes were presented explicitly. One can hardly question the wisdom of that research strategy given the impact the work and the recognition it has received. However, it is important to recognize that the everyday choices and decisions that people make individually and collectively—including many of the most important personal and political ones—involve prospective losses and associated likelihoods that do not involve money or time or other things we value and that cannot be readily quantified. Prospective and experienced losses of health, friendship, and meaning in one's life are obvious cases in point. Furthermore, those everyday choices and decisions often pit gains in one domain against losses in another domain. At this point in the history of our country, three social-political issues involving prospective risks, losses, and gains losses are of major concern.
The COVID 19 pandemic. The first, and most immediately salient issue, relates to the losses of life and the various social and financial opportunities imposed by the covid 19 virus. The willingness of so many Americans to restrict their freedom of movement and endure financial losses no doubt reflected the fact that the pandemic, in its initial stages was new—the “reference point” was zero cases of the virus and no loss of life. By contrast, many of those same Americans have long been unwilling to pay the much smaller price in terms of dollars and inconvenience that would be required to reduce the toll of all the other medical risks that our society tolerates. Notably, there is not overwhelming support for any of the various plans for universal coverage of medical costs that would do much to reduce that toll. Indeed, if a decade ago we had freed poorer Americans from having to choose between costly medical insurance and the risk of bankrupting illnesses, and between such health care needs and the need to meet everyday costs of housing, food, and other necessities, the collective gains would have outweighed the collective costs. Certainly, the number of Americans highly vulnerable to the COVID virus because of hypertension, diabetes, and heart disease would have been much lower.
For his part. President Trump initially gambled that delay in taking measures to prepare for, and potentially reduce, the spread of the virus would have a lower cost to his political standing—albeit not to the health of our nation—than would immediate acknowledgment of the seriousness of the threat. That acknowledgment would have immediately depressed stock market indices and the economy as a whole. But those effects, especially the ones most affecting working-class Americans would have been less pronounced, less persistent, and in the long-run less costly than the ones our nation has endured.
Today, as the 2020 Presidential Election approaches, President Trump is urging state governors and mayors, to lift restrictions on commerce and social contact. He is similarly demanding that public schools and universities reopen, and even that sporting events that put thousands of attendees at risk, “go back to normal” Moreover, he is also encouraging his supporters and even citizens who do not support him, to rebel against and defy enacted restrictions. All trusted experts acknowledge the inevitable, enormous, costs of such re-openings in terms of human lives and illness, and they are not even denied by the Trump administration. Yet much of the public seems willing to bear those costs and even the risk that they will prove even greater than our already dire best guess models predict. A quick look at the convex loss curve helps explain why the American public seems numbed to such costs and risks. For most of us an increase from 0 deaths to 1,000 prompts demands for action; likewise an increase from 1,000 to 100,000. But an increase from 100,000 to 200,000 seems to be producing little more than weary nods of recognition and seems a price many Americans are willing to accept to restore a semblance of normalcy.
Would President Trump, back in January or February have dared to take equivalent risks to achieve a large boost in the economy, and improve his re-election prospects from pretty favorable to very favorable? More importantly, would the body political have been willing to accept such a risk? It seems clear that President Trump and his supporters have endorsed a level of risk to avoid and recoup losses that they would not have endorsed to achieve equivalent gains—a gamble that virtually all medical authorities not tied to the administration deemed short-sighted and foolhardy, but one that much of the loss-averse public now seems willing to take. Tellingly, the Swedish government, in contrast to that in all of the other Scandinavian countries, made a similar gamble in the face of the pandemic and paid a steep price in terms of lives lost to avoid the relevant social and economic costs. Time will tell whether Swedes will be happy about the consequences of that gamble and the government that took it. Time will also tell whether citizens in other countries experiencing renewed waves of the virus as a consequence of re-openings will look harshly at the leaders who threw caution in the wind and opted for short-term gains that would well-received by business interests and many segments of their societies.
The climate change crisis. Scientists martial increasing evidence of climate change, putting us at heightened risk today and holding the prospect of far more dire consequences in the not-too-distant future. Had we acted sooner to curb the practices that produce greenhouse gases the costs of mitigation would have been lower. The chances of avoiding calamity grow ever slimmer. Yet individually and collectively, we fail to rise to the challenge
Climate change skeptics, and the economic and political powers that encourage them, challenge the certainty of a much less habitable planet for our children to say nothing of generations yet unborn. In suggesting that climate change is not really occurring, or that it is not a product largely human activity, they are essentially proposing that we gamble, indeed that we back a longshot, rather than pay the price today for curbing the activities that are putting us and those who will follow us at tremendous risk. In other words, we as a society are showing exactly the kind of loss aversion postulated by Kahneman and Tversky.
Of course, loss aversion is not the only barrier to the mobilization of effective mitigation efforts. Temporal discounting plays a role insofar as the costs would be borne today and are certain, whereas the benefits will be realized only in the future. The human capacity to rationalize also plays a role. Those who oppose such actions, whether out of genuine doubt or cynical motivation, exaggerate the amount of disagreement in the scientific community. They further encourage us to believe that some technological miracle will address the problem at low cost, suggest that a warmer planet will produce benefits that offset the costs, or insist that we will adapt to it more readily than is prophesized. They also make the issue a matter of tribal politics, a test of loyalty to political factions that denigrate scientific expertise. (The corresponding litmus tests for political progressives who spearhead efforts to combat climate change, it could be argued, is the willingness to enact and adopt measures to reduce energy use that is driving climate change, whatever the costs, and regardless of who will pay the greatest price)
Work on endowment effects and willingness to pay versus demand to receive that flowed from prospect theory is worth recalling in the context of the climate change debate. Most important were the studies by Kahneman Knetch, and Thayler (1990) showing consistent violations of the Coase theorem. In their study, some research participants were endowed with some object of value, such as a mug emblazoned with a University crest or an attractive pen, while other participants were given a sum of money. A market was then created for potential buyers and sellers, The Coase theorem suggests that initial (in this case instant) endowments should not determine value attached to the money vs the object, and thus a rate of trades for money versus that object should be fifty percent. However, the actual percentage of such exchanges was than fifty percent because prices demanded by those endowed with the object tended to demand a higher price than the potential buyers who had received cash.
Rather than digress to discuss that now-large body of research, let me just note that I regularly do a classroom exercise where I ask my students how much they would be willing to pay to make their car brakes 10 percent safer. The modal answer is in the range of $300 to $400. Then I ask them to imagine that a researcher offers them the chance to be paid for making their brakes 10 percent less safe. The amounts in question are inevitably an order of magnitude greater than their “willing to pay” amounts. To complete the exercise I ask them how safe their brakes are now compared to that of the safest brakes on the market, and the response is a lot of sheepish grins and confessions that they have no idea about the relative safety of their braking systems.
I believe we would find the same asymmetry when many Americans consider the potential costs and benefits of effective measures to deal with climate change. The public that is willing to pay only a modest price in terms of dollars and inconvenience to lower greenhouse gas emissions would, I believe, demand a huge price to allow an increase in such emissions, even though few have a firm grasp on the magnitude, and imminence of the threat we are facing. The same asymmetry is apt to apply to the costs of reducing the Covid 19 toll. And again, rationalization, denial, and wishful thinking limit the success of campaigns to educate the public about those threats. 1
The political divide among US voters. The operation of loss aversion and risk tolerance is highly relevant when we consider the political divide in the US body politic. Blue state progressives, in particular the financially and culturally coastal elites who seek to understand the support for Donald Trump in the rural areas and small towns in much of our country, often fail to understand the gamble accepted by those providing that support. That lack of understanding begins with a failure to appreciate the losses, non-financial as well as financial, that globalization, increased automation, immigration, and other changes in demography have imposed on many of a huge number of their fellow citizens.
For the “winners” in the new global economy, the results of the changes are mostly positive—high paying jobs and perhaps even a chance to get rich, more diverse and interesting restaurants, colleagues, and next-door neighbors, and lots of new high tech devices that open new possibilities for information sharing and investment. For the “losers” in much of the American heartland, however, the consequences of the new economy are almost entirely negative. The costs involve not just the evident loss of good-paying jobs, but the loss of status and of dignity, the pride that comes from providing security and some modest luxuries for one’s family, and the sense that they are masters of their destiny and central to the life of their community. Small wonder that, as prospect theory writ large would predict, they were willing to undertake the risk of backing, and continue to support a candidate lacking in political experience, with character defect they recognize, who promises to “make America great again”. The most important word in that slogan is not “great”, it is “again”—and implicit promise to restore what they have lost, again, not just good jobs, but restored social and cultural ascendency.
However, I think that the rise and robustness of Trumpism and popularism in the face of his lies and misdeeds reflects loss aversion in a way that most pundits have missed. For people who no longer work side-by-side in factories and belonged to unions, and seen other sources of community, including membership in voluntary local organizations, fade away as their towns decline, political solidary and fellowship fills that void. Participation in Trump rallies, wearing MAGA caps, consuming right-wing talk radio and television, exchanging messages via social media, and proclaiming their support for Trump with lawn signs and banners, provided a source of collective identity. Indeed, it allowed individuals who had felt frustrated, ignored, and disrespected to feel part of a social movement, one that gave them both voice and status. They also were allowed to feel themselves privy to secrets that their fellow citizens lacked, It is hardly surprising that they ignored, resisted, and rationalized their leader’s lies, misdeeds, and personal flaws rather than suffer the loss of that positive identity and community, and special status. Nor can we expect them to do so in the aftermath of Trump’s electoral defeat.
The difficulty of ending protracted conflicts via exchanges of concessions. Appreciation of the asymmetry of the gain versus loss curves explicated by Kahneman and Tversky and the endowment effect helps us to understand one of the puzzling and tragic phenomena we witness on the world stage. In too many parts of the globe, we see protracted and costly conflicts, including bloody ones, persist when rationality suggests that an exchange of concessions and an end to hostilities would better serve the parties. In retrospect, no sane person could contend that either World War 1 adversary benefited from continuing the horrific struggle for four years than reaching a negotiated peace agreement. The same could be said of the war between Iran and Iraq that lasted from 1980 to 1988 and exacted a terrible price in both nations in terms of blood and treasury and many other wars that adversaries have fought over the centuries.
In the seemingly intractable Middle East conflict where the outlines of the kind of agreement that could bring peace and greater prosperity both for Israel and its adversaries have been clear for more than two decades following the signing of the Oslo accords. Yet the tit-for-tat attacks and reprisals continue, Israel pays dearly for continuing the occupation of the West Bank and Gaza, and the Palestinians suffer disproportionately. Many sources of folly could be cited to explain these stalemates. Unrealistic expectations of a quick victory, political miscalculation, rationalization, propaganda, the stoking of hatred, distrust, and calls for vengeance by leaders or political factions whose objectives and self-interests diverge from those of the citizenry bearing the cost of the struggle all may play a role.
An additional barrier to agreement is the requirement that the requisite exchange of concessions involves accepting of losses in return for achieving gains (or at least mitigating future losses). The greater steepness of the loss function than the gain function described by Kahneman and Tversky means that both adversaries will feel that the concessions made by the other side are less valuable than the concessions called for by their own side. As a result, they believe that the required exchange is unfair and unsatisfactory, and that continuing the struggle in the hope of winning more favorable terms to be justified.
These prospect theory factors are exacerbated, of course, by the two sides’ differing views about the source of the conflicts and the legitimacy of the rival claims and aspirations. Each side feels not only that it is making the greater concessions, but that their concessions are not only painful but unjust, whereas the other side is conceding little of anything to which it ever was entitled. Again, such sentiments are stoked by leaders who insist that if they maintain, or better still, redouble, their efforts there will be “light at the end of the tunnel.” The rhetoric includes the claim that God and/or history is on “our” side, that agreement would dishonor those “shed their blood for the cause,” that the other side cannot be trusted to abide by the proposed terms of agreement.
Distributing gains (vs acceptance of losses). Prospect theory dealt with risk-taking under conditions of uncertainty and the separate asymmetric concave functions governing the weight decision-makers give to prospective gains and prospective losses. That theory did not deal directly with a question, or rather an intuition based on real-world observation, that led me and a group of colleagues to explore the way people allocate gains and respond to changes in such allocations. Our study (Diekman et al, 1997) focused on the status quo bias and the willingness of those advantaged by the status quo defend and rationalize the legitimacy and fairness of their advantage. In our research, which alas dealt with hypothetical rather than actual allocations, some participants were asked to propose a division of a sum of money. In one study, the money was performance bonus to be divided between them and a co-worker. In another study it was scholarship monies to be divided between their own and another University (San Jose State and Stanford). Other participants were not asked to propose a division, but told of an allocation, in one condition an equal division, in other conditions an allocation, that gave them or their school either a much larger share of the monies, or a much smaller share. All participants later evaluated the fairness of all three allocations.
The results were clear and can be summarized succinctly. Allocators were virtually unanimous in proposing an equal division of the funds. Participants told they or their school had been allocated an equal or smaller share consistently rated an equal division as fairest, and showed no tendency to think it would fairer for them or their school to receive the lion’s share of the funds than it would be for the other individual or school to receive the lion’s share. The only participants whose ratings of fairness showed an asymmetry were those who initially had been informed that they or their school would receive the larger share. These participants alone rated a division favoring them or their school as fairer than a division favoring their co-worker or the other school, and they offered justifications for the relevant inequality.
In other words, those “endowed” with an advantageous allocation were willing to justify a degree of inequality they would not have proposed or defended if they had not enjoyed that initial endowment. Most notably, they alone indicated that they thought the arguments favoring their school (higher tuition and admission standards versus greater financial need and bleaker prospects for a high-paying job after graduation) were more legitimate than the arguments favoring the other school. They alone treated the prospect of greater equality as a painful loss, and they alone were inclined to rationalize the continuation of inequality. 2
In retrospect, I wish we had explored the allocation of losses as well as gains. Had we done so, I suspect, that we would have seen more self-serving (or in-group serving) allocations and even marked tendencies for the advantaged to justify the dubious argument that they or their group should not bear an equal (or even larger) share of the losses. The prospect theory basis for that prediction is straightforward. If losses loom larger than gains, we can expect that parties will be more willing to share gains than bear equivalent losses. I know of no research findings that test this speculation but real-word observation leads me to be confident it is correct. In economic downturns, we see the advantaged insist that the economy will be better served if their losses are mitigated by government actions (i.e. “bailouts” for large businesses) than if equivalent sums are used to ease the plight of ordinary workers and genuine small businesses. (I remember how, long ago when Stanford was facing budget cuts, I heard a highly paid business school professor argue that equal percentage cuts or reductions in salary increases would be unfair. He justified this contention by noting that equal percentages would mean greater losses for more highly paid faculty.
The steep road ahead.
As I conclude this essay, our society faces two challenges that involve tradeoffs between costs and benefits—the threat of covid 19 and the far greater existential threat of climate change. Both challenges are made more difficult because of temporal discounting. Both tempt Americans to engage in denial, wishful thinking, and rationalization. Both also make them susceptible to reassurances by political leaders and others whose self-interest is served by these psychological defense mechanisms.
In the case of covid 19, the benefits of premature easing of restrictions on work and play will be felt today; the likely costs in terms of new spikes in infections, hospitalizations, and deaths will be borne months from now (and, too many of our fellow citizens rationalize, may not take place at all). Many Americans further convince themselves that they and their loved ones will not be the victims, thst it will be other people whose age or poor health or foolhardy actions place them at greater risk.
The avoidance of burdensome and costly measures to combat climate change spares most Americans from inconvenience and economic costs today; the greatest penalty for inaction will be borne in the decades to come, and if they are anywhere near my age, something for future generations to deal with. Comforting denial and rationalization are hard for most to resist. ( The scientists may be wrong; it is mainly other countries that will be affected; scientists will figure out ways to solve the problem, adaptation to climate change will be easier than doom-spreaders claim; there will be advantages as well as disadvantages of a warmer climate ).
Medical researchers and epidemiologists must take the lead in combating covid and enlightening the public. Scientists who take the measurement and build the models to document the rate and likely consequences of global warm and related threats to the environment have the duty of convincing political leaders to take effective, collective actions. Psychologists and other social scientists who understand the processes that lead people to take unwise risks, and fail to act when action is called for, have a special obligation to design effective interventions to address those shortcomings
References
Banks, SM, Salovey, P., Greener, S., Rothman, AJ, Moyer, A, Beauvais, J & Epel, E (1995). The effects of message framing on mammography utilization. Health Psychology, 14(2), 178-184
Diekmann, K, Samuels, S, Ross, L. & Bazerman, M (1997). Self-interest and fairness of resource allocation: Allocators versus recipients. Journal of Personality and Social Psychology, 72, 1061-1074.
Fetherstonhaugh, D & Ross, L. (1999). Framing effects and income flow preferences in decisions about social security. In H.J. Aaron (Ed.) Behavioral dimensions of retirement economics. Brookings Institution Press.
Jost, JT, & Banaji, MR (1994). The role of stereotyping in system justification and the production of false consciousness. British Journal of Social Psychology, 33, 1–2
Lewis, M. (2016). The Undoing Project; A friendship that changed our minds, Norton
Kahneman, D. (2011). Thinking fast and slow. Farrar, Straus and Giroux
Kahneman, D, Knetsch, JL, & & Thayler, R. (1990). Experimental tests of the endowment effect and the Coase theorem. Journal of Political Economy. 98(6) 1325-1348
Kahneman, D, & Tversky, A. (1979). Prospect theory: An analysis of decisions under risk. Econometrica, 47, 313–327.
Kahneman, D, & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39, 341–350.
McNeil, BJ, Pauker, SG, Sox HC, & Tversky, A. (1982). On the elicitation of preferences for alternative therapies. New England Journal of Medicine 306(21), 1259-1262.
Footnotes
1 In the McNeil et al study, as in many of the most compelling experiments on framing effects, the specific tradeoffs involved probably played a role in the success of the experiment, In particular, the 10% vs 0 % mortality rate attached to the initially risky surgery option versus the initially safe radiation option involved a comparison at the steepest point in the relevant loss curve whereas the 90% survival rate vs 100% survival rate involved a comparison at the point where the loss function was much less steep. This feature of the procedure thus exploited an effect of “reference point”, which is predicted by prospect theory, as well as susceptibility to framing. Rescue schemes that put many at risk to potentially save the few or even a single soldier, or free a handful of miners trapped after a cave-in, reflects the same willingness to gamble, even if the net expected utility of that gamble is negative when there is a chance to avoid all loss of life but not to exhibit the same valuing of saving lives when it comes to measures that would reduce known health risks to mass numbers of people.
2 Appreciation of the concavity of the loss function combined with attention to reference points helps us understand why the public’s willingness to make sacrifices to have the covid death toll be 0 rather than 1000 is greater than its willingness to do so to have the toll be 1000 rather than 2000 and, more relevant to our current situation, to make sacrifices to make the expected death toll at this end of June be 100,000 instead of 101,000. The attractiveness of the first steps to open the economy, by the same token, will be treated as more valuable than steps to improve a later, better, and opener economy. Hence the political strategy of Trump and his close advisors to now endure the risks (to the public) of accelerating the opening of businesses and no longer encouraging social distancing and isolation.
3 Work by Jost and his colleagues (Jost & Banaji, 1994) suggests that there is even a tendency for the economically disadvantaged to justify the system that is allocating so much more to the “haves” or the 1% than the “have-nots” or even moderate wage-earners. Such “system justification”, like many other types of rationalization for maintaining a status quo that serves us badly (including rationalizations for failing to bear the costs of mitigating climate change) is actively encouraged and promoted by interest groups that most benefit from the status quo and are most threatened by changes in the status quo.