Is There a Predominance of Suffering?
An Empirical Study on the Human Condition
B.Contestabile First version 2022 Last version 2026
Table of Contents
1. Introduction
2. The Asymmetry Thesis
3. Population Ethical Intuitions
4. Application: The Global Balance of Life Evaluations
5. Objections Concerning the Measurement of Suffering
6. Objections Concerning the Measurement of Asymmetries
7. Implications
8. Limitations and Future Directions
9. Conclusions
Appendix: Caviola et al.’s Pilot Study
Does suffering predominate on a global level? The intuition that it does has deep philosophical roots, from the First Noble Truth of Buddhism to Schopenhauer’s pessimism and modern negative utilitarianism. Yet contemporary surveys, such as the World Happiness Report, suggest the opposite: there are more happy than suffering people, and progress studies claim steady improvements in global well-being. This paper argues that the apparent predominance of happiness does not arise from the data itself but from a methodological assumption in how well-being is interpreted and aggregated. Standard approaches assume that happiness and suffering are equally weighted—a commitment that lacks empirical justification. Recent empirical research on population-ethical intuitions indicates that most people assign greater weight to suffering than to happiness. Applying this asymmetry to global life-evaluation data reverses the aggregate balance, revealing a predominance of suffering.
Because research on population-ethical intuitions is still at an early stage, and debates about measuring well-being remain ongoing, this result has a provisional character. Nevertheless, it challenges the widespread assumption that happiness predominates and calls for the use of weighted averages in well-being reports.
A short version of this paper is available on
1. Introduction: Ancient Intuition versus Contemporary Surveys
While theories of well-being differ considerably on what makes life go well, there is wide agreement that suffering makes life bad (Phillips, 2006). This near-universal consensus makes the question of suffering's predominance particularly significant for how we understand progress and how ethical priorities are derived.
The question has ancient roots. Early Buddhist philosophy maintained that suffering (dukkha) fundamentally characterizes human existence and cannot be compensated by happiness (Beckwith, 2015, 26–32). In Western tradition, a similar intuition appears in Schopenhauer's metaphysics and resurfaces in twentieth-century ethics through negative utilitarianism (Popper, 1945; Smart, 1958; Contestabile, 2014).
Yet the alleged predominance of suffering faces an empirical challenge. The World Happiness Report consistently finds that there are more happy than suffering people. Recent work in the progress-studies literature goes further, arguing that human well-being has been steadily improving (Rosling et al., 2018; Pinker, 2011, 2018). Does the ancient intuition of suffering's predominance collapse under empirical scrutiny?
This paper argues that it does not. The apparent predominance of happiness does not arise from the data itself but from how well-being is interpreted and aggregated. Standard well-being reports typically assign equal weight to happiness and suffering when evaluating populations—an assumption that is rarely defended on empirical grounds. Once this assumption is questioned, the picture changes. While average living conditions may have improved since the time of Buddha and Schopenhauer, they have not improved enough to reverse the global balance. The currently reported predominance of happiness rests on a methodological assumption rather than on a substantive philosophical justification.
The argument unfolds in three stages. Section 2 examines the philosophical grounds for an asymmetry between happiness and suffering. Section 3 analyzes how this asymmetry appears in recent research on population-ethical intuitions. Section 4 applies these findings to global life-evaluation data.
2. The Asymmetry Thesis: Philosophical and Empirical Grounds
2.1 The Philosophical Tradition
The claim that suffering and happiness are asymmetric—that suffering carries greater weight in ethical evaluation—has ancient roots. The First Noble Truth of Buddhism asserts that suffering pervades existence and cannot be compensated by pleasure or contentment (Bodhi, 2005; Beckwith, 2015, 26–32). In Western ethics, Popper argued that pain cannot be outweighed by pleasure (Popper, 1945, 284). Negative utilitarianism—which is usually associated with Karl Popper’s notes on ethics—encompasses forms of utilitarianism that model an asymmetry between happiness and suffering (Fricke, 2002, 14). Contemporary philosophers including Parfit (1991, 101) and Arrhenius (2000, 138) have recognized asymmetries between the positive and negative dimensions of well-being.
2.2 Psychological and Behavioral Evidence
Contemporary psychology and behavioral economics provide empirical support for the asymmetry thesis. The principle of negativity bias holds that negative experiences and emotions impact us more profoundly than positive ones (Baumeister et al., 2001; Rozin & Royzman, 2001). Negative events tend to be more memorable, more influential on judgment, and more resistant to habituation than positive events of equivalent intensity.
Prospect theory demonstrates that losses loom larger than equivalent gains—a phenomenon called loss aversion (Kahneman & Tversky, 1979). This asymmetry appears not merely in monetary contexts but across a wide range of domains in decision-making. Psychometric evidence further suggests that the scales of suffering and happiness are not experientially symmetric (Diener & Emmons, 1984).
There may be evolutionary grounds for this asymmetry: failing to avoid harmful actions could lead to death, whereas failing to pursue beneficial actions does not have similarly bad consequences (Caviola et al., 2022, 20). Recent results from behavioral science confirm that pleasure and pain in animals are asymmetrical experiences, with pain exerting a stronger and more enduring influence on behavior than pleasure (Neville, 2020; Fraser, 1998)
2.3 The Methodological Problem
Standard approaches to aggregating well-being, however, ignore this asymmetry. The World Happiness Report and similar surveys employ linear point scales that assign the same weight to all score levels. Thus, on a 0–10 point scale, a score of 3 (three steps below the midpoint) is weighted the same as a score of 7 (three steps above). In principle, one could monitor the different well-being levels separately, but the weighting problem emerges as soon as averages are calculated across levels (Contestabile, 2016, 54–57).
Well-being aggregation based on linear point scales is not a neutral methodological choice. It embodies a commitment to the equal weighting of happiness and suffering—a commitment that is typically undefended and often unacknowledged in well-being research. But how much weight should suffering receive? To answer this question, we must turn to empirical investigations of population-ethical intuitions.
3. Population-Ethical Intuitions: The Weight of Suffering
Recent research by Caviola et al. (2022) provides the first systematic empirical investigation of how people weigh happiness against suffering when evaluating entire populations. The theoretical framework of this research is provided by hedonic welfare theories, that is, theories concerned with ranking populations according to their happiness and suffering (Caviola et al. 2022, 2). The first study (1a) examined the trade-off between happiness and suffering. Participants were asked to imagine a world that contains 1000 people, composed solely of extremely happy and extremely unhappy people. The people who are happy consistently have extremely positive experiences. The people who are unhappy consistently have extremely negative experiences. Participants were then asked the following question:
“Given this information, what percentage of happy people would there have to be for you to think that this world is overall positive rather than negative (so that it would be better for the world to exist rather than not exist)?”
The average response was approximately 76% (SEM ± 1.11%).1 In other words, roughly three times as many extremely happy people are required to balance a given number of extremely suffering people (Caviola et al., 2022, 5). This represents a substantial departure from the symmetric weighting assumed by classical utilitarianism.
Study 1b and 1c investigated whether this asymmetry is sensitive to intensity levels. Caviola’s team found that for mild suffering and happiness, the asymmetry factor was approximately 1.5, about half the value observed for extreme intensities (Caviola et al., 2022, 6–8).
Finally, the results of a pilot study suggests that under conditions of heightened personal responsibility for outcomes, the asymmetry factors may become substantially higher (see Appendix).
3.2 Philosophical Interpretation
What do these findings mean philosophically? Given equal intensity of happiness and suffering, does the asymmetry reflect experiential differences (people judge suffering to be more urgent) or normative judgments (people judge suffering to deserve greater moral weight)? A second group within studies 1a-c—focusing on intra-personal compensation—helps to answer this question. In this group, the percentage of happiness required to compensate for suffering closely matched that required for inter-personal compensation (Caviola et al., 2022, 4–5). This result strongly supports the thesis that the same risk-based perspective was applied to intra- and inter-personal compensation. The inter-personal asymmetry is therefore best understood as an experiential measurement rather than a normative judgment.
The result of studies 1a-c suggests that participants’ intuitions fall between two extremes:
- Classical utilitarianism: treating happiness and suffering as equally weighted.
- Strict negative utilitarianism: considering only suffering in evaluation.
Most people, it appears, endorse a moderate negative utilitarianism that assigns weight to both happiness and suffering but weighs suffering more heavily. This intermediate position is intuitively attractive: it avoids both the radicality of ignoring happiness entirely and the implausibility of easily compensating one person’s suffering with another person’s happiness.
4. Application: The Global Balance of Life Evaluations
4.1 The World Happiness Report Data
The World Happiness Report, published by the UN Sustainable Development Solutions Network, provides the most comprehensive global data on well-being. Its best-known reports draw on the Gallup World Poll, covering more than 150 countries and representing over 98% of the world's population (Gallup, 2009). The Gallup World Poll employs the so-called Cantril ladder (Cantril, 1965), with the following wording (Gallup, 2009):
- Please imagine a ladder with steps numbered from zero at the bottom to 10 at the top.
- The top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you.
- On which step of the ladder would you say you personally feel you stand at this time? (ladder-present)
- On which step do you think you will stand about five years from now? (ladder-future)
The ladder-present question alone is widely regarded as a valid and reliable measure of subjective life evaluation (Cheung et.al., 2014). Most of Gallup’s own reports—and most external research—disregard the ladder-future question. We follow this convention in the analysis that follows.
Fig.1 compiles responses from all available Gallup World Polls conducted between 2005 and mid-2011, weighted by each country’s population aged 15 and older. The figure contains 11 columns, each representing a possible answer to the ladder-present question. The height of each bar corresponds to the frequency of responses for that score (Helliwell et al., 2012, 25, Figure 2.1).
Fig.1 World Distribution of Cantril Scores
The distribution in Fig.1 yields an average score of 5.3 on the 0–10 scale.2 Subsequent reports show a slight decline to 5.1 by 2019 (Helliwell et al., 2020, 25, Figure 2.2), followed by stability through 2022–2023 (Helliwell et al., 2023, 3), and a recovery to 5.4 in the most recent period 2023–2025 (Helliwell et al., 2026, 21–23, Figure 2.1).
4.2 The Symmetric Interpretation
How should we interpret this 5.3 average? Bipolar rating scales, such as the Cantril scale, are designed with symmetry in mind: the upper part of the scale is intended to represent positive categories, the lower part negative ones. The midpoint (around 5.0 on a 0–10 scale) can be seen as a rough benchmark between positive and negative life evaluations. A global average of 5.3 therefore suggests that, on balance, people rate their lives slightly above the midpoint—indicating modestly positive overall life evaluation.
In the following, happiness denotes positive life evaluation, or evaluative happiness—a term coined by the World Happiness Report (Helliwell et al., 2012, pp. 6, 11). Suffering denotes negative life evaluation and is used in line with Gallup’s interpretation of low Cantril scores (Gallup, 2009). Both concepts are treated as continua ranging from minimal to maximal intensity, consistent with contemporary psychological theory (Cowden, 2021).
4.3 The Predominance of Suffering
A straightforward inspection of Fig. 1 confirms a surplus of happiness. This conclusion, however, holds only under the assumption that happiness and suffering are given equal weight. Once the asymmetry revealed by population-ethical intuitions is applied, the picture changes substantially.
According to Caviola et al.’s population-ethical studies 1a-c (2022, 5–8), suffering should be weighted more heavily than happiness, by a factor ranging from 1.5 (for mild intensity) to 3.0 (for extreme intensity). For a rough estimation, it is sufficient to assume an average factor of 2.0. Doubling the weight of level 0–4 scores in Fig.1 leads to a negative overall balance. Table 1 shows a more detailed estimation.3
This estimation is likely conservative, given higher asymmetry factors in the pilot study, the positivity bias from underrepresented and excluded groups, and the empirical midpoint above 5.0.
The estimation above understates the predominance of suffering because the Gallup World Poll is subject to systematic sampling biases. Individuals facing structural, social, or health-related obstacles are less likely to participate (Collins et al., 2024). Some groups are not captured at all, most notably people experiencing advanced illness or terminal conditions. Because these groups are precisely those most likely to report low life evaluations, their omission artificially inflates the global average.
The four largest groups in this context are the following:
- Extreme poverty: Approximately 800–830 million people live below the international poverty line, recently updated to about $3/day (World Bank, 2025).
- Homelessness: An estimated 300–318 million people experience strict homelessness (used here for the aggregate estimate), while broader forms of inadequate housing or slum conditions affect roughly 1.1–2.8 billion (UN-Habitat, 2025).
- Severe chronic illness or cognitive impairment: Disability affects approximately 1.3 billion people globally, many of whom face substantial barriers to survey participation (Institute for Health Metrics and Evaluation, 2025).
- Conflict-affected populations: Around 117–123 million people are forcibly displaced due to conflict or violence (UNHCR, 2025).
Taken together, these categories yield a gross total on the order of 2.5–2.6 billion people, corresponding to approximately 31% of the global population (about 8.2 billion in 2025).
The degree of underrepresentation or exclusion, and the well-being levels of those concerned, are largely unknown. Moreover, these groups substantially overlap: individuals living in extreme poverty, for example, are disproportionately likely to also experience chronic illness, homelessness, or displacement due to conflict. For these reasons, the positivity bias is not used to claim a predominance of suffering in this paper. Instead, it serves to interpret the estimation in Section 4.3 as conservative.
The Cantril scale specifies only its two endpoints—the best and worst possible life—without defining the midpoint. This raises the question of how respondents interpret the scale. Empirical evidence suggests that the score 5 is often viewed as negative rather than neutral (Tay & Diener, 2011, 356–359; Kahneman & Deaton, 2010). A pilot study by the Happier Lives Institute (n = 128) asked participants to identify the point on a 0–10 scale that corresponds to being neither satisfied nor dissatisfied (Samuelsson, 2023). The mean response was 5.3, indicating that the empirical midpoint may lie slightly above the scale’s theoretical midpoint. Any upward shift reinforces the case for a predominance of suffering. Considerations about the empirical midpoint lead naturally to questions about measurement, which the next section addresses.
5. Objections Concerning the Measurement of Suffering
5.1 The Commensurability Objection
Objection: The entire argument depends on treating Cantril scale responses as interpersonally comparable. But suffering is inherently subjective. How can we meaningfully compare one person's "score 4" with another's "score 4," especially across cultures?
Reply: This is a general problem for any empirical research on well-being, not a difficulty unique to the present argument. The Gallup World Poll addresses it through several strategies: linking subjective responses to objective life conditions (health, income, social support), analyzing large samples that mitigate individual variation, and testing for consistency across time and contexts (Helliwell et al., 2012, 16–22). While perfect interpersonal comparison remains impossible, consistent patterns across large populations provide meaningful information.
Moreover, the measurement challenge cuts both ways. If we cannot meaningfully aggregate Cantril scores, then we also cannot support claims about global increases in well-being or the predominance of happiness. The progress-studies literature (Pinker, Rosling) depends on the same data. Skepticism about measurement undermines both types of conclusions equally.
5.2 The “True” Suffering Objection
Objection: The definition of suffering as “the lower half of a symmetric scale” has been contested by MacAskill, who argues that scores below the midpoint do not “truly” correspond to lives in which suffering dominates (MacAskill, 2022, 195–196).
Reply: The assumption that well‑being levels below the midpoint should be interpreted as states of happiness is not supported by empirical research. Individuals reporting such scores typically experience significant dissatisfaction, stress, or negative affect (Gallup, 2009; Kahneman & Deaton, 2010; Tay & Diener 2011, 356–359). The more plausible concern, therefore, is whether mildly negative states should be excluded from the definition of suffering.
Yet, by methodological parity, such exclusion would also require that mildly positive states be excluded from the definition of happiness. The imprecision inherent in classifying well‑being states into categories of “happy” and “suffering” cannot be used to selectively discount only one side of the scale. When this symmetry principle is respected—no matter how narrowly the term suffering is defined—the sign of the overall balance remains unchanged. Table 1 shows that, for each symmetric level of the well-being scale, the amount of suffering exceeds the amount of happiness.3
5.3 The Alternative Metrics Objection
Objection: The Cantril scale has been criticized—due to its ladder imagery—for evoking associations with wealth and social status, potentially overlooking other important dimensions of well-being (Nilsson et al., 2024). Would the predominance of suffering disappear if alternative measures of subjective well-being were used?
Reply: The problem in answering this question is finding surveys with the same preconditions (Helliwell et al., 2012, 13). When comparing the Cantril scale with life satisfaction measures for the same respondents, and in the same survey, the latter yields a slightly higher mean score—about 0.5 points on an 11-point scale (Helliwell et al., 2012, 14). This difference may be due to framing effects. The question about life satisfaction asks: “All things considered, how satisfied are you with your life as a whole, nowadays?” Studies show that positively valued words systematically push responses upward (Schuman & Presser, 1996). The World Happiness Report considers the Cantril scale to be the more reliable tool in terms of stability, cross-cultural neutrality, and evaluative depth (Helliwell et al., 2017, 10–13).
Gallup’s life evaluation index (Gallup, 2024) is not a different interpretation of the Cantril scale, but a different metric. It represents a complex three-category system (thriving, struggling, suffering), which correlates present and expected life evaluations with living conditions and emotional states. Peer-reviewed literature characterizes the three categories as points on a continuum from positive to negative life evaluation (Riley et al. 2021). Empirical findings suggest that the struggling category correlates with a negative—rather than neutral—life evaluation (Tay & Diener 2011, 356–359; Kahneman & Deaton, 2010). However, Gallup’s life evaluation index only shows the proportion of people in each category, and not the category’s level of well-being. It is therefore unsuitable to calculate overall well-being.
Established measures of affective happiness are closely related to life evaluations (Helliwell et al., 2013, 10), but they are not directly comparable. The Gallup experience index and the Gallup net affect measure the frequency of ups and downs, rather than levels of well-being. A possible predominance of suffering cannot be assessed without measuring the latter.
More sophisticated measures of well-being include multi-item scales (Ryff, 1989), the Day Reconstruction Method (Kahneman et al., 2004), the Experience Sampling Method (Csikszentmihalyi & Larson, 1987), and the Diary Method (Bolger et al., 2003). However, it remains unclear whether these methods would significantly alter the distribution shown in Fig.1 (Hudson et al., 2020; Shiba et al., 2022). Moreover, any improvement in quality must be weighed against the global practicality of the method. Currently, no other measures of life satisfaction match the Gallup World Poll in terms of global coverage, methodological consistency, and empirical rigor.
5.4 The Data Collection Objection
Objection: How do we know that people really evaluate their condition in ways that match the actual quality of their mental state?
Reply: Self-reporting of subjective well-being has been studied extensively (World Happiness Report 2012, 16-22). Overall, research suggests that this method – even if not perfect – provides meaningful information, showing consistency over time, and agreement with non-self-report methods (Sandvik 2009).
6. Objections Concerning the Measurement of Asymmetries
6.1 The Hedonic Framework Objection
Objection: Caviola et al.’s studies 1a–c operate within the framework of hedonic welfare theories, whereas the Cantril scale incorporates non-hedonic components. Evaluations of the best possible life, for instance, may reflect factors such as achievements or relationships, and not only pleasures.
Reply: Three lines of evidence support applying Caviola et al.'s findings to life satisfaction measures like the Cantril scale.
First, life satisfaction bridges hedonic and non-hedonic dimensions. Even when satisfaction stems from non-hedonic sources—achievements, relationships, autonomy—it manifests experientially as a form of happiness. Empirical research confirms this connection: measures of affective happiness correlate closely with life evaluations (Helliwell et al., 2013, 10; King et al., 2006; Disabato et al., 2016). The World Happiness Report's framing reinforces this broader understanding by treating Cantril scale scores as measures of happiness. Gallup's research extends this framework by describing low Cantril scores as suffering, based on correlations with adverse life conditions and negative emotional states (Gallup, 2009). The distinction between hedonic and non-hedonic well-being may be less sharp in experience than in philosophical theory.
Second, research on the philosophical Asymmetry—the view that creating unhappy lives is morally significant while creating happy lives is neutral—is debated independently of strict hedonism (McMahan, 2009; Frick, 2014).
Finally, loss aversion—a key psychological mechanism underlying the asymmetry—extends beyond hedonic value (Kahneman & Tversky, 1979). The inclusion of non-hedonic values in the Cantril scale may amplify rather than diminish the asymmetry. The loss of achievements, relationships, autonomy, or dignity may weigh more heavily than the loss of pleasure.
In sum, the asymmetry disclosed by Caviola et al. likely extends to life satisfaction measures. The driving force appears to be a fundamental judgment that negatives (suffering, dissatisfaction, loss) carry greater weight than positives (happiness, satisfaction, gain).
6.2 The Scale Structure Objection
Objection: Caviola et al.’s scale is a linear point scale (Caviola et al., 2022, 6). By contrast, the Cantril scale is ordinal rather than cardinal: it reflects rank order, but the intervals between levels may not be equal. “What if the Cantril scale is nonlinear? (Wodak, 2019).”
Reply: A nonlinear structure of the Cantril scale would challenge the present analysis only if it disproportionately increased the weight of positive scores. Currently, there is no empirical evidence supporting such an assumption (Philippi, 2024).
6.3 The Neutral Level Objection
Objection: MacAskill (2022, 196) argues that the neutral level for existence—the point where life becomes worth living—lies well below the midpoint of the Cantril scale, perhaps at levels 1–2, contradicting Caviola et al.’s assumption.
Reply: This objection conflates two distinct questions:
(1) Hedonic neutrality: where happiness and suffering balance.
(2) Existential neutrality: where life becomes preferable to non-existence.
Caviola et al.'s research concerns (1). Their study frames the evaluation from an impartial perspective, distinguishing it from participants’ personal preferences about existence. If the value of existence is exclusively derived from happiness, then lives with positive scores are worth living and lives with negative scores are not. On this basis, the neutral level must lie at the scale's midpoint (Caviola et al., 2022, 8, 22).
MacAskill's research addresses question (2). People may prefer existence even when suffering outweighs happiness (Metzinger, 2017, 243). In this case, participants do not compare happiness with suffering; they compare the two evils of suffering and non-existence. The neutral level for existence therefore lies below the scale’s midpoint.
The comparison between (1) and (2) is not about two different measured values of the same neutral level, but about two distinct kinds of neutral level. Caviola et al.’s level is neutral in terms of happiness and suffering, while MacAskill’s level is neutral in terms of tolerable and intolerable suffering.
For illustration, suppose existential neutrality was found at level 1 of the Cantril scale. In such a case, one cannot conclude that level 2–4 scores are states of happiness. That would contradict the correlations found in empirical research (Gallup, 2009; Kahneman & Deaton, 2010; Tay & Diener 2011, 356–359). It simply means that most people prefer moderate suffering to non-existence. On this basis, it follows that most people would likely accept a global predominance of suffering when participating in a corresponding survey.
6.4 The Early-Stage Research Objection
Objection: Research on population-ethical intuitions is still in its early stages. Caviola et al.'s studies 1a-c represent a first attempt, involving a total of about 1,000 participants—a figure comparable to the average national sample size in the Gallup World Poll (Helliwell et al., 2023, 32). Is it not premature to base philosophical conclusions on such limited data?
Reply: This caution is reasonable, but it does not undermine the argument.
First, the asymmetry findings are consistent across multiple studies with different framings. Studies 2a-b elicit symmetric intuitions, indeed, but these framings are not suitable for evaluating entire populations (Caviola et al., 2022, 10–11).
Second, the estimation in Section 4.3 is likely conservative, given higher asymmetry factors in the pilot study, the positivity bias from underrepresented and excluded groups, and the empirical midpoint above 5.0.
Third, the findings align with extensive evidence for asymmetries in psychology and behavioral economics.
Finally, the findings accord with longstanding philosophical intuitions and arguments.
The more serious problem lies on the opposing side: many contemporary well-being reports assign equal weight to happiness and suffering without empirical justification. This paper calls that default into question. Even if Caviola et al.’s findings remain provisional, they nevertheless shift the burden of proof. Those who wish to maintain equal weighting must explain why these findings—together with extensive psychological and behavioral evidence of asymmetry—should be disregarded.
7.1 Implications for Well-Being Measurement
If the argument holds, it has implications for how human well-being is assessed and reported. Recognizing the predominance of suffering would shift the focus from the maximization of happiness to the reduction of suffering, with corresponding consequences for how progress is conceptualized and how ethical priorities are derived. Current well-being reports tend to obscure both the magnitude and distribution of suffering.
Even if future research were to disprove the assumed predominance of suffering—whether through new well-being measures or alternative interpretations of existing scales—the asymmetry between happiness and suffering would remain a serious challenge to standard aggregation practices. Weighing happiness and suffering equally implies that the suffering of a minority can easily be offset by the happiness of a majority. However, Caviola et al.’s research suggests that such compensation is not self-evident.
Simple averages of life evaluations should therefore be replaced by weighted averages that reflect the greater weight of suffering. The use of signed scales, however, is of secondary importance: an average below 5.0 on a 0–10 point scale conveys the same information as a negative value on a -5/+5 scale.
7.2 Implications for Population Ethics
If the argument holds, then—measured by average folk intuitions—the predominance of suffering is a defensible position and not necessarily pessimistic, overly risk-averse, or even irrational. The conclusion that humanity should rather not exist, however, faces two strong objections: First, MacAskill's findings suggest that most people prefer a life of moderate suffering to non-existence (see Section 6.3). Second, the argument does not rule out a future predominance of happiness.
The assumption that global well-being is positive plays a central role in population ethics. Many population axiologies rely on this assumption and are therefore confronted with the Repugnant Conclusion, according to which a very large population with minimal happiness can be morally better than a small population with maximal happiness (Parfit, 1984, chapter 19). In the negative domain, population ethics faces the analogue Negative Repugnant Conclusion (Broome, 2004, 213–214), which holds that a very large population with minimal suffering can be morally better than a small population with maximal suffering.
If the hedonic neutral level is replaced by MacAskill’s neutral level—which population ethicists often call a critical level—then the Repugnant Conclusion is mitigated, while the Negative Repugnant Conclusion is aggravated. A radical response to this trade-off is to disregard population size and focus on average well-being. Although average utilitarianism is not free from theoretical difficulties (Arrhenius, 2000, 54–57), these shortcomings are largely theoretical rather than practical. Average welfare measures remain dominant in applied welfare economics (Arrhenius, 2000, 53).
In conclusion, population ethics should operate with two distinct notions of neutrality, corresponding to two different evaluations of existence. MacAskill’s neutrality is suffering-tolerant or life-friendly compared to hedonic neutrality.
8. Limitations and Future Directions
8.1 Concerning the Measurement of Suffering
The estimation of global well-being faces methodological limitations.
One concerns the individuals who are systematically underrepresented or excluded in the Gallup World Poll. Current methods omit precisely those individuals most likely to experience severe suffering (see Section 4.4).
Another limitation arises from age coverage. The Gallup World Poll excludes individuals younger than 15 years. There is currently no truly global and representative survey that provides a worldwide average well‑being measure for this age group.
For children under 11, well-being is measured mainly through the frequency of affective states. These measures cannot be credibly transformed into a 0–10 scale estimate—not even as a defensible approximation.
A further limitation concerns the exclusive focus on human well-being. Both Buddhism and negative utilitarianism extend moral concern to all sentient beings, raising the question of whether suffering predominates beyond the human case. With respect to domesticated animals, especially those in industrialized systems, welfare assessments often indicate a negative hedonic balance due to chronic stress, confinement, and restrictions on species-typical behavior (Robbins et al., 2018). Wild animals, by contrast, are subject to severe natural stressors such as predation, disease, and starvation (Dawkins, 1995), and some researchers argue that their aggregate hedonic balance may also be net negative (Tomasik, 2015).
From a population perspective, wild animals—particularly fish and invertebrates—vastly outnumber domesticated animals, potentially amplifying the moral significance of even low-intensity suffering when aggregated across individuals (Tomasik, 2015). At the same time, domesticated animals may experience more intense and longer-lasting hedonic states due to prolonged confinement and sustained exposure to adverse conditions. These considerations raise unresolved questions: to what extent are hedonic states commensurable across species, and can human intuitions about population ethics be extended to animal populations?
8.2 Concerning the Measurement of Asymmetries
Two issues deserve particular attention. The first concerns the definition of asymmetry factors. These factors do not merely compare the numbers of happy and suffering people; they compare amounts of well-being, calculated as the number of people multiplied by the level of well-being. In Caviola et al.’s studies, for example, the level of well-being is operationalized in terms of intensity (mild or extreme). A basic question is how the asymmetry scales with different levels of well-being. However, consolidating the range of asymmetry factors, and estimating the positivity bias (Section 4.4) are more immediate concerns than refining scaling functions. In the meantime, it makes sense to work with simple linear interpolation.
The second issue concerns the participant pool in Caviola et al.’s studies. In studies 1a-c, the participants were recruited in the United States via the crowdsourcing platform MTurk (Caviola et al., 2022, 4, 6, 7). The total number of participants was approximately 1,000—a figure comparable to the average national sample size in the Gallup World Poll (Helliwell et al., 2023, 32). While data from a single country can yield insights of broader relevance, the United States does not represent the global average: it consistently ranks among the happiest countries in the world (Helliwell et al., 2012, 30). Moreover, participants recruited via crowdsourcing platforms tend to be younger and more educated than average. Occupying relatively secure and comfortable circumstances, they may be psychologically prone to underestimating the intensity of suffering in distant others (Nordgren et al., 2011). The asymmetry factors reported by Caviola et al. may therefore represent a conservative lower bound. A cross-national study with a more representative participant pool would be a logical next step to investigate this thesis.
More generally, the question arises how the quality of population-ethical intuitions might be improved. In everyday life, empathy is often biased and can lead to unjust decisions (Bloom, 2016). However, when combined with an impartial perspective—as in Caviola et al.’s studies—empathy can improve the assessment of other people’s happiness and suffering. Notably, Caviola et al.’s pilot study found that when the framework elicited stronger empathy, the asymmetry factor increased dramatically—reaching a value of 100 (see Appendix). It is also plausible to assume that individuals with greater life experience form more realistic intuitions about compensation than those who are less experienced.
This raises a methodological question: should population-ethical studies privilege the intuitions of highly empathetic and experienced individuals? Historical debates suggest caution. Philosophers who emphasized the weight of suffering—such as Buddha and Schopenhauer—were plausibly both empathetic and experienced. Yet their views were often dismissed as pessimistic, overly risk-averse, or even irrational. For broader acceptance, population-ethical research must therefore engage with folk intuitions, even if these intuitions yield more conservative asymmetry estimates.
Acknowledgment
I would like to thank Magnus Vinding for his helpful comments and suggestions on an earlier version of
this paper.
1. Standard error of the mean for study 1a:
SEM = SD / √N = 1.10
SD = standard deviation = 16.99
N = sample size = 237
2. Average well-being calculation based on the distribution shown in Fig.1:
Level multiplied by frequency = (0×1.7) + (1×2.5) + (2×4.7) + (3×9.5) + (4×12.5) + (5×26.2) + (6×14.7) + (7×11.9) + (8×9.7) + (9×3.3) + (10×3.3) = 533.2
Average = 533.2 / 100 ≈ 5.3
Frequencies add up to 100%
3. Table 1 Estimation of happiness and suffering:
Cantril level 0 is weighed with factor 3.0, level 4 with factor 1.5, and the factors in between are linearly interpolated. The asymmetry range 1.5–3.0 is empirically motivated (Section 3.1), while the linear interpolation is a simplifying assumption (Section 8.2).
Symm. Level = Distance to the midpoint level 5
Amount = (Symm. Level) × (Frequency %) × (Factor)
Suffering share % = Amount % of levels 0–4 = 58.6%
Happiness share % = Amount % of levels 6–10 = 41.4%
Midpoint (level 5) has zero symmetric distance and therefore contributes no amount.
1. Arrhenius, Gustaf. 2000. Future Generations: A Challenge for Moral Theory. PhD diss., Uppsala University, Department of Philosophy.
2. Baumeister, Roy F., Ellen Bratslavsky, Catrin Finkenauer, and Kathleen D. Vohs. 2001. “Bad Is Stronger Than Good.” Review of General Psychology 5(4): 323–370.
3. Beckwith, Christopher I. 2015. Greek Buddha: Pyrrho’s Encounter with Early Buddhism in Central Asia. Princeton University Press.
4. Benatar, D. (2006). Better Never to Have Been, Oxford University Press
5. Bloom, Paul. 2016. Against Empathy: The Case for Rational Compassion. Ecco/HarperCollins Publishers.
6. Bodhi. 2005. The Connected Discourses of the Buddha: A Translation of the Saṁyutta Nikāya. Wisdom Publications.
7. Bolger, Niall, Angelina Davis, and Eshkol Rafaeli. 2003. “Diary Methods: Capturing Life as It Is Lived.” Annual Review of Psychology 54: 579–616.
8. Broome, John. 2004. Weighing Lives. Oxford University Press.
9. Cantril, Hadley. 1965. The Pattern of Human Concern. Rutgers University Press.
10. Caviola, Lucius, David Althaus, Andreas L. Mogensen, and Geoffrey P. Goodwin. 2022. “Population Ethical Intuitions.” Cognition 218: Article 104941.
11. Cheung, Felix C.K., and Richard E. Lucas. 2014. “Assessing the Validity of Single-Item Life Satisfaction Measures: Results from Three Large Samples.” Quality of Life Research 23(10): 2809–2818.
12. Collins, Elliott, Isabella Contreras, Savanna Henderson, Caroline Lamke, Michael Rosenbaum, and Shana S. Warren. 2024. “Can Remote Survey Methods Yield Nationally Representative Samples in LMICs? A Cross-National Analysis of Pandemic-Era Studies.” Innovations for Poverty Action.
13. Contestabile, Bruno. 2014. “Negative Utilitarianism and Buddhist Intuition.” Contemporary Buddhism 15(2): 298–311.
14. Contestabile, Bruno. 2016. “The Denial of the World from an Impartial View.” Contemporary Buddhism 17(1): 49–61.
15. Cowden, Richard G. 2021. “Suffering, Mental Health, and Psychological Well-Being.” Frontiers in Psychology 12: 1–4.
16. Csikszentmihalyi, Mihaly, and Reed Larson. 1987. “Validity and Reliability of the Experience-Sampling Method.” Journal of Nervous and Mental Disease 175(9): 526–536.
17. Das, Kirti V., Carla Jones-Harrell, Yingling Fan, Anu Ramaswami, Ben Orlove, and Nisha Botchwey. 2020. “Understanding Subjective Well-Being: Perspectives from Psychology and Public Health.” Public Health Reviews 41(1): 25.
18. Dawkins, Richard. 1995. River Out of Eden. Basic Books.
19. Diener, Ed, and Robert A. Emmons. 1984. “The Independence of Positive and Negative Affect.” Journal of Personality and Social Psychology 47(5): 1105–1117.
20. Disabato, David J., Fallon R. Goodman, Todd B. Kashdan, Jerome L. Short, and Aaron Jarden. 2016. “Different Types of Well-Being? A Cross-Cultural Examination of Hedonic and Eudaimonic Well-Being.” Psychological Assessment 28 (5): 471–482.
21. Fraser, David, and Ian J. H. Duncan. 1998. “‘Pleasures’, ‘Pains’ and Animal Welfare: Toward a Natural History of Affect.” Animal Welfare 7, no. 4 (1998): 383–396.
22. Frick, Johann David. 2014. "'Making People Happy, Not Making Happy People': A Defense of the Asymmetry Intuition in Population Ethics." PhD diss., Harvard University.
23. Fricke, Fabian. 2002. “Verschiedene Versionen des negativen Utilitarismus.” Kriterion 15: 13–27.
24. Gallup. 2009. “Understanding How Gallup Uses the Cantril Scale.” Gallup News.
25. Gallup. 2024. “Indicator: Life Evaluation Index.” Gallup.
26. Helliwell, John, Richard Layard, Jeffrey Sachs, and Jan-Emmanuel De Neve, eds. 2012–2023. World Happiness Report. Sustainable Development Solutions Network.
27. Hudson, Nathan W., Ivana Anušić, Richard E. Lucas, and M. Brent Donnellan. 2020. “Comparing the Reliability and Validity of Global Self-Report Measures of Subjective Well-Being to Experiential Day Reconstruction Measures.” Assessment 27(1): 102–116.
28. Institute for Health Metrics and Evaluation (IHME). 2025. Global Burden of Disease 2023: Findings from the GBD 2023 Study. IHME.
29. Kahneman, Daniel, and Amos Tversky. 1979. “Prospect Theory: An Analysis of Decision under Risk.” Econometrica 47(2).
30. Kahneman, Daniel, and Alan B. Krueger. 2006. “Developments in the Measurement of Subjective Well-Being.” Journal of Economic Perspectives 20(1): 3–24.
31. Kahneman, Daniel, and Angus Deaton. 2010. "High Income Improves Evaluation of Life but Not Emotional Well-Being." Proceedings of the National Academy of Sciences of the United States of America 107 (38): 16489–16493.
32. King, Laura A., Joshua A. Hicks, Jennifer L. Krull, and Amber K. Del Gaiso. 2006. "Positive Affect and the Experience of Meaning in Life." Journal of Personality and Social Psychology 91 (1): 179–196.
33. MacAskill, William. 2022. What We Owe the Future. Basic Books.
34. McMahan, Jeff. 2009. "Asymmetries in the Morality of Causing People to Exist." In Harming Future Persons, edited by Melinda A. Roberts and David T. Wasserman, 49–68. Dordrecht: Springer.
35. Metzinger, Thomas. 2017. “Suffering, the Cognitive Scotoma.” In The Return of Consciousness, edited by K. Almqvist and A. Haag, 237–262.
36. Neville, Vicky, Jessica King, Ian D. Gilchrist, Peter Dayan, Elisabeth S. Paul, and Michael Mendl. 2020. “Reward and Punisher Experience Alter Rodent Decision-Making in a Judgement Bias Task.” Scientific Reports, 10: 11839.
37. Nilsson, August Håkan, Johannes C. Eichstaedt, Tim Lomas, Andrew Schwartz, and Oscar Kjell. 2024. “The Cantril Ladder Elicits Thoughts about Power and Wealth.” Scientific Reports 14: Article 2642.
38. Nordgren, Loran F., Kasia Banas, and Geoff MacDonald. 2011. "Empathy Gaps for Social Pain: Why People Underestimate the Pain of Social Suffering." Journal of Personality and Social Psychology 100 (1): 120–128.
39. Parfit, Derek. 1984. Reasons and Persons, Clarendon Press.
40. Parfit, Derek. 1991. “Equality or Priority?” The Lindley Lecture, University of Kansas. In The Ideal of Equality, edited by M. Clayton and A. Williams, 81–125. Macmillan.
41. Philippi, Christian Larroulet. 2024. “Well-Being Measurements and the Linearity Assumption: A Response to Wodak.” The Australasian Journal of Philosophy 102(2): 512–518.
42. Phillips, David. 2006. Quality of Life: Concept, Policy and Practice. Routledge.
43. Pinker, Steven. 2011. The Better Angels of Our Nature. Viking Books.
44. Pinker, Steven. 2018. Enlightenment Now. Viking Books.
45. Popper, Karl. 1945/1966. The Open Society and Its Enemies. Vol. 1, 5th ed. Routledge.
46. Riley, Carley, et al. “Trends and Geographical Variation in Population Thriving, Struggling and Suffering Across the USA, 2008–2017: A Retrospective Repeated Cross-Sectional Study.” BMJ Open 11, no. 7 (2021): e043375.
47. Robbins, Jesse, Becca Franks, and Marina A. G. von Keyserlingk. 2018. “‘More Than a Feeling’: An Empirical Investigation of Hedonistic Accounts of Animal Welfare.” PLoS ONE 13 (3): e0193864.
48. Rosling, Hans, Ola Rosling, and Anna Rosling Rönnlund. 2018. Factfulness. Flatiron Books.
49. Rozin, Paul, and Edward B. Royzman. 2001. “Negativity Bias, Negativity Dominance, and Contagion.” Personality and Social Psychology Review 5(4): 296–320.
50. Ryff, Carol D. 1989. “Happiness Is Everything, or Is It? Explorations on the Meaning of Psychological Well-Being.” Journal of Personality and Social Psychology 57(6): 1069–1081.
51. Samuelsson, Conrad, Samuel Dupret, Michael Plant, and Caspar Kaiser. 2023. "Can We Trust Wellbeing Surveys? A Pilot Study of Comparability, Linearity, and Neutrality." Happier Lives Institute. Report.
52. Sandvik E, Diener E., Seidlitz L. (2009). “Subjective Well-Being: The Convergence and Stability of Self-Report and Non-Self-Report Measures”, Social Indicators Research Series, 39, 119-138
53. Schuman, Howard, and Stanley Presser. 1996. Questions and Answers in Attitude Surveys: Experiments on Question Form, Wording, and Context. Transaction Publishers.
54. Shiba, Koji, Richard G. Cowden, Natasha Gonzalez, Matthew T. Lee, Tim Lomas, Andrew Y. Lai, and Tyler J. VanderWeele. 2022. “Global Trends of Mean and Inequality in Multidimensional Wellbeing: Analysis of 1.2 Million Individuals from 162 Countries, 2009–2019.” Frontiers in Public Health 10.
55. Smart, Richard N. 1958. “Negative Utilitarianism.” Mind 67(268): 542–543.
56. Tay, Louis, and Ed Diener. 2011. “Needs and Subjective Well-Being Around the World.” Journal of Personality and Social Psychology 101 (2): 354–365.
57. Tomasik, Brian. 2015. “The Importance of Wild-Animal Suffering.” Relations: Beyond Anthropocentrism 3 (2): 133–52.
58. UN-Habitat. 2025. Annual Report 2024: Adequate Housing for All. UN-Habitat.
59. UNHCR. 2025. Global Trends: Forced Displacement in 2024. UNHCR.
60. Wodak, Daniel. 2019. “What If Well-Being Measurements Are Non-Linear?” The Australasian Journal of Philosophy 97(1): 29–45.
61. World Bank. 2025. Poverty, Prosperity, and Planet Report 2024: Pathways Out of the Polycrisis. World Bank.
Appendix: Caviola et al.’s Pilot Study
A pilot study conducted by Caviola et al. disclosed that the framing of the question has a major influence on the asymmetry (Caviola 2022, 20). Participants were asked:
“Suppose you could push a button that created a new world with X people who are generally happy and 10 people who generally suffer. How high would X have to be for you to push the button?” (Caviola 2022, Supplementary Materials, 7)
Surprisingly the median response was X=1000, i.e. the asymmetry factor jumped from the range between 1.5 and 3.0 to 100. How can this huge difference be explained? Probably there are several factors involved. Following two theses:
Thesis 1: The difference is caused by deontic intuitions.
- In the first three studies (1a-c) the participants evaluated just outcomes.
- In the pilot study they considered actions that need to be taken to achieve these outcomes.
Participants who imagine taking an active role may follow a deontic rather than an axiological concept (Caviola 2022, 2, 21-22). An axiological concept tells us what has value; a deontic concept tells us what we ought to do. In deontology the rightness of an action is based on rules rather than consequences.
- There are indeed cases where the utilitarian evaluation is declined, for example the Trolley Problem (Thomson 1976), a moral dilemma where a trolley is on course to kill five people down the track, but a driver could intervene and divert the vehicle to kill just one person on a side track. In such a situation people tend to reject doing harm to a single person, even if this action could prevent doing the same harm to several other persons.
- If the harm done to a single person is compared with the happiness (and not with the prevented suffering) of several other persons, the situation is less clear. In a study on the negativity bias of moral judgments participants were asked to imagine that they could press a button that would give one minute of “intense pain” to one person and one minute of “similarly intense pleasure” to ten other people. Every participant refused to press the button. The participants were then asked to state the minimum number of people that would make them press the button. All participants except one stated that no number would be large enough (Rozin 2001, 307) (Caviola 2022, 3). We cannot know the motive of the person, who rejects doing harm. It could be deontic, but it could also be strict negative utilitarian. An example for the latter case is Karl Popper’s controversial statement “…from the moral point of view, pain cannot be outweighed by pleasure and especially not one man’s pain by another man’s pleasure” (Popper 1945, 284), a statement which he wrote in the context of the Second World War and the Holocaust. Popper addressed a well-known discomfort with the classical utilitarian accumulation of suffering and happiness across different people.
Caviola’s pilot study is related to the Rozin study and not to the Trolley Problem. An argument for a moderate negative utilitarian interpretation (and against a deontic interpretation) is the fact, that the rejection is incomplete: 1000 happy people outweigh 10 suffering people, which means that the participants weighed happiness against suffering.
Thesis 2: The difference is caused by the emotional relationship with the assessed populations.
- In the studies 1a-c the participants are in the position of a viewer. Because of this distant role it is unclear to what extent they develop an empathic relationship with the assessed populations.
- In the pilot study, however, the participants take the position of a responsible designer. Responsibility deepens empathy and empathy improves the assessment of other people’s happiness and suffering.
No single person is responsible for human suffering, as in the pilot study, but there is a collective responsibility. The actual and future populations are not the result of an inevitable natural event; they are the result of procreating humans (Benatar 2006).
In the Socrethics papers we discuss several approaches for asymmetric welfare axiologies:
- The Stoic idea which is at the origin of Negative Utilitarianism: the community is considered as a mystic unity. Population-ethical intuitions (section 3) are an imperfect tool to find this unity in a consensus on risk-aversion and compassion.
- The idea of a perfectly empathic observer, as discussed in The Denial of the World from an Impartial View.
- People's intuitions can be distorted by ignorance, bias, or short-term desires. Thus, well-being should be assessed by what an individual would choose under ideal conditions (informed, rational, and free from psychological biases). This approach is pursued in Rawls’ concept of the Original Position and results in the so-called Difference Principle.
- Harsanyi’s risk-based conception of impartiality; further developed by John Broome; see Moderate Negative Utilitarianism.
- The Buddhist tradition, as discussed in Negative Utilitarianism and Buddhist Intuition.
- Prioritarianism, as discussed in Antinatalism and the Minimization of Suffering