If I gave you $5,000 right now and asked you to spend it in whatever way would help people the most, what would you do?
That amount of money can buy 1,000 malaria-preventing bed nets that will, on average, end up saving one kid’s life. Or it can boost the income of five poor families by giving each $1,000. Or it can treat 30 people for depression by paying for a few months of group therapy that evidence suggests can have lasting effects.
Which is best?
Nobody knows for sure, and that’s a problem facing everyone, from regular people wanting to spend a few bucks on charity to governments with millions to invest in public policy. It’s hard to figure out which does the most good — saving one life, reducing poverty for a few, or treating mental illness for dozens — and even asking the question can feel morally icky. Yet we can’t escape it if we want our money to help people as much as possible.
Economists love things they can measure objectively, like the number of deaths in a village or the number of dollars in an account. So over the past century, they’ve focused on measuring health and wealth. The best policy programs for society are deemed to be the ones that save the most lives, say, or increase gross domestic product (GDP) by the widest margin.
And there’s a good rationale for using a metric like GDP as a shorthand for well-being: There is a very high correlation between a nation’s GDP per capita and its self-reported life satisfaction. Just look at this chart, showing the low levels of well-being among the world’s poorest nations.
In other words, GDP is a very strong predictor of life satisfaction.
But a strong predictor is not a perfect predictor. As we’ve gathered more data on the happiness of different populations, it’s become clear that increasing wealth and health do not always go hand in hand with increasing happiness. By the economists’ objective measures, people in rich countries like the US should be doing great — and yet Americans are only becoming more miserable. And people in some higher-GDP European countries like Portugal and Italy report lower life satisfaction than people in lower-GDP Latin American countries.
What’s going on here? How do we explain the gaps in life satisfaction that objective metrics like GDP don’t explain?
Nowadays, a growing chorus of experts argues that helping people is ultimately about making them happier — not just wealthier or healthier — and the best way to find out how happy people are is to just ask them directly. This camp says we should focus a lot more on subjective well-being: how happy people are, or how satisfied they are with their lives, based on what they say matters most to them — not just based on objective metrics like GDP. Subjective well-being can tell us things that objective metrics can’t.
It’s a revolution in thinking that’s gathering force in policy and charity circles alike, and it’s starting to upend conventional wisdom about the best ways to do good.
Here’s one eye-popping example from the charity world. The Happier Lives Institute, a research center aiming to identify evidence-based ways to improve happiness, wanted to find out which is better at boosting well-being: giving people cash or giving them therapy. So HLI ran a direct comparison — involving evidence from some 80 studies and 140,000 participants — between GiveDirectly, an organization that gives cash transfers, and StrongMinds, an organization that provides group therapy for people with depression. Both serve very poor people in sub-Saharan Africa.
Surprisingly, HLI found that the therapy was nine times more cost-effective than the cash at improving self-reported well-being. Some evaluators disagree with this finding, taking issue with, for example, how HLI’s analysis assumes that improving someone’s depression score by some amount is equivalent to improving their overall life satisfaction by that same amount; these evaluators argue StrongMinds is less cost-effective than HLI’s findings suggest. Nevertheless, the findings have served as a challenge to how empirically driven philanthropy allocates funding.
Changes are afoot in the policy world, too. At the international level, UN Secretary-General António Guterres made a push in May for nations to track progress beyond GDP. Governments from Scotland and Wales to Finland and Iceland have already begun to move well-being to the center of their policymaking.
In 2019, New Zealand became the first country to make well-being the organizing principle of its national budget. The budget required all new spending to go toward specific well-being goals, such as bolstering its population’s mental health. That prompted immediate results, like an influx of over $200 million into services for survivors of domestic and sexual violence, the nation’s largest-ever investment in that issue.
Around the world, a new focus on improving subjective well-being is causing a paradigm shift in how we try to do good. But the approach is beset with controversy. Some say it goes too far. And others say it’s not going far enough.
How subjective well-being went from hot to not — and back again
If you asked me, “What’s the ultimate purpose of life?” and I answered “wealth” or “health,” you might find that kind of weird. We value those things not in themselves, but because they tend to improve our well-being — how happy or satisfied we are with life. Well-being is the thing we really care about.
This perspective, common among philosophers for centuries, reached peak popularity in the 18th century. Early utilitarians like English philosopher Jeremy Bentham argued that happiness should be the single currency by which we compare the value of different goods, including wealth and health. Adam Smith pored over economic methods with the goal of fostering happiness for the people. And Thomas Jefferson wrote that “the life and happiness of the people is the first and only object of good government.”
But the 1920s saw a huge reversal, as behaviorism became fashionable among psychologists like Ivan Pavlov (yes, the dog guy) and economists like Lionel Robbins. The behaviorist school of thought argued that we can’t know what people are thinking or feeling inside, so trying to understand people’s subjective experiences is pointless.
“Basically, economists wanted to be more scientific,” explained Michael Plant, who leads the Happier Lives Institute. “They thought something only counts as science if it’s objectively measurable. Feelings aren’t objectively measurable, therefore they are not science.”
So economists turned away from squishy concepts like happiness and toward objective proxies for well-being, like GDP. In the postwar period, GDP became the go-to way for measuring well-being, even though the concept’s inventor, Simon Kuznets, warned that “the welfare of a nation can scarcely be inferred from a measurement of national income.”
Economists also embraced metrics used to assess the value of different health policies. One of these is the quality-adjusted life year (QALY). One QALY is equal to one year of life in fantastic health. So the more QALYs a policy secures for a population, the better it’s deemed to be.
Starting in the 1970s, though, the intellectual pendulum began to swing back toward subjective well-being, thanks to a few key developments.
First, economists started to rethink their assumptions about human behavior. They had long been assuming that people behave “rationally” — simply making the choices that will get them what they want. But in the 1970s, the behavioral economics researchers Daniel Kahneman and Amos Tversky argued that each of us is actually a pretty irrational mess, shot through with cognitive biases, often behaving in ways that don’t maximize our happiness because we’re not great at predicting what will make us happy. That means you can’t just look at people’s behavior to determine what makes them happy — you have to survey them about how happy they are and try to tease out the causes.
Second, economist Richard Easterlin showed in 1974 that richer doesn’t necessarily mean happier. Case in point: since World War II, Easterlin noted, the US had enjoyed a lot of economic growth, yet Americans’ level of well-being had barely budged. This became known as the Easterlin Paradox, and although it remains contested, it was pivotal in getting people to realize that it’s important to measure subjective well-being and not just GDP.
Partly as a result, in the 1970s, countries began collecting survey data on their populations. A typical survey question asked: “Overall, how satisfied are you with your life nowadays?” For 50 years, the data trickled in.
Subjective well-being stages its comeback
With the exception of Bhutan, whose king asserted in the 1970s that “Gross National Happiness is more important than Gross Domestic Product,” nobody made much noise about the well-being data until around 2009. Then the global financial crisis hit.
Amid all the suffering, economists realized they knew very little about how people adapt to crises and which policies would best mitigate negative impacts on well-being. So national happiness became the subject of policy conferences and college courses. France commissioned a study on it from economists Amartya Sen, Joseph Stiglitz, and Jean-Paul Fitoussi, who concluded that an overemphasis on GDP had blinded policymakers to other aspects of societal well-being.
In 2011, the OECD released its first well-being report on its member countries, and in 2012, the UN began releasing its annual world happiness report. And individual countries began to collect data on well-being in more sophisticated ways.
First of all, they designed surveys carefully, to avoid priming respondents in ways that might bias their happiness scores. For example, asking someone to think about politics right before you ask them about their life satisfaction is a big no-no. Surveys were also better able to control for culture-specific reporting differences, like whether it’s socially acceptable to complain about life.
Another challenge with surveys, pointed out by Sen, was that a person’s self-reported well-being could simply adapt to whatever social environment someone has grown accustomed to, so it may not reflect real deprivations in that environment — especially ones that have been around long enough to be normalized.
“Just because somebody poor in Guatemala says they’re happy at a daily level, that doesn’t mean we should say they’re all fine, let’s not worry about them,” said Carol Graham, a public policy professor at the University of Maryland and a senior scientist at Gallup, who has written prolifically about how to measure subjective well-being.
This had been a serious problem with early surveys, in part because they only used a single, general measure of well-being. They didn’t make distinctions between a long-term measure like life satisfaction and a short-term, affective measure like contentment or stress. So researchers weren’t able to tease out the difference between enduring life satisfaction and a brief bout of cheerfulness.
Experts developed ways to deal with this. Take the UK, a global leader in measuring subjective well-being. About a decade ago, the government’s Office for National Statistics began including these four questions (dubbed the ONS4), each with a 0-10 score, in their annual population survey:
- Overall, how satisfied are you with your life nowadays?
- Overall, to what extent do you feel that the things you do in your life are worthwhile?
- Overall, how happy did you feel yesterday?
- On a scale where 0 is “not at all anxious” and 10 is “completely anxious,” overall, how anxious did you feel yesterday?
The first question is designed to get a broad sense of people’s overall life satisfaction, which tends to be pretty stable over time. The second is about meaning or purpose in life. The third and fourth questions are more like a snapshot in time, intended to give a short-term measure of affect — happiness or contentment on one hand, and stress or anxiety on the other.
Now, the UK government gathers two types of well-being data: the subjective data based on the ONS4 questions above, and more objective data on things that are known to feed into societal well-being, from health and wealth, to the unemployment rate and crime rate, to time spent in nature and volunteering. The government uses both the subjective and objective measures — 52 in total — to guide public spending.
Other nations are also making policy based on a mix of subjective and objective measures of well-being. After New Zealand passed its groundbreaking well-being budget, for instance, its government began measuring progress using 61 well-being measures that track things like loneliness and trust in government institutions. It did not do away with GDP, but it went beyond it. As then-Prime Minister Jacinda Ardern noted, “GDP alone does not guarantee improvement to our living standards.”
Lingering skepticism about subjective well-being data — and how experts are tackling it
Lots of people are still skeptical of efforts to measure subjective well-being. That’s understandable — the word “subjective” is right there in the name.
Take Elie Hassenfeld, the co-founder and CEO of the charity evaluator GiveWell. He’s used to metrics like QALYs, and he’s definitely got doubts about measuring subjective well-being with questions like “Overall, how satisfied are you with your life nowadays?”
“What are you measuring exactly when you ask this question? I don’t really know what it means for someone to say ‘I’m a 6 out of 10’ in the way that I know what it means for someone to not have a broken arm,” he told me.
“And to the extent that I could know what it means for someone, I wouldn’t know what it means for someone else,” he continued. In other words, does one person’s 6 correspond to another person’s 6? This is an open question that needs further research.
Hassenfeld also questions whether a measure of subjective well-being gets at the things we really care about, things that make life worthwhile. “My grandfather survived the Holocaust and came over and worked seven days a week, 18 hours a day, owning a diner for so many years,” he said. “And I don’t think that in any of those years, he would’ve given a high answer to a subjective well-being question. But he saved up enough money to send my mom and aunt to college, and there’s meaning there.”
What Hassenfeld is getting at is this: We don’t have to feel happy in the moment to feel that our lives are worth living. Part of our happiness rests not in the present, but in the meaning we derive from working toward our goals in the future — and in the hope we have that we’ll be able to actually achieve those goals.
It’s a critique that Graham takes seriously. “People in some poor countries tend to score higher than you’d think they’d score, and part of that is because they have hope for progress — they think things can get better,” she said. “Whereas in rich countries, where you have big sectors of the population that are declining relative to their peers, then they don’t have hope for the future.”
Graham’s most recent book, The Power of Hope: How the Science of Well-Being Can Save Us From Despair, argues that hope needs to be part of well-being metrics. She has already convinced one pretty important reader: the UK government, which recently expanded its battery of survey questions. “I’m really thrilled that hope is now a measure,” Graham told me.
Plant, of the Happier Lives Institute, agrees that how much meaning or hope you have in your life may affect how happy you are. But, he said, that’s probably already reflected in people’s answers to happiness questions. “Part of the virtue of the subjective approach is that people can bring whatever matters to them into their assessments. So, how much meaning you have in your life could be an input into that.”
And when he looks at a survey like the World Happiness Report, he feels confident that it’s capturing something pretty close to reality. The data makes sense. Many of the top 10 happiest countries, according to that survey’s 2020–2022 rankings, are the ones with strong social safety nets combined with high GDP per capita, especially the Nordic countries. By contrast, the bottom 10 countries are nations that have been wracked by poverty, violence, or both:
Looking at this type of data, Plant said, “I don’t have much time for feelings-measurement skepticism. I think the counterarguments there just aren’t very strong.”
Even skeptics like Hassenfeld do believe that subjective well-being data is getting at something real. In a recent podcast interview, he said, “I think the pro of subjective well-being measures is that it’s one more angle to use to look at the effectiveness of a program. It seems to me it’s an important one, and I would like us to take it into consideration.” This warming toward subjective well-being is a big deal, given that his organization, GiveWell, influences how over $500 million is spent on charity each year.
Should subjective well-being be the only measure?
According to Plant, well-being data is now robust enough that we should feel emboldened to go further than nations like New Zealand have gone.
“To be a full-blown well-being budget, you would actually compare things in terms of units of well-being, which the New Zealand government isn’t doing,” Plant told me. He believes each policy should be evaluated according to how many well-being units it produces, and policymakers should invest in the intervention that comes out on top.
Right now, New Zealand and other nations are using dozens of measures to assess policy outcomes, some that capture subjective well-being directly and others that capture objective facts that relate to well-being. But how is anyone supposed to compare two policies with precision, Plant reasons, when one outcome is in units of well-being and the other is in dollars? That’s apples to oranges.
Instead, Plant argues we should compare how much good different things do in a single “currency” — specifically, how many well-being-adjusted life years, or Wellbys, they produce. Producing one Wellby means increasing life satisfaction by one point (on the 0-10 life satisfaction scale) for one year. It’s a metric that some economists, including those behind the World Happiness Report, are coming to embrace. If we were to evaluate every policy in terms of how many Wellbys it produces, that would allow for direct apples-to-apples comparisons.
“I’m pretty bullish about just using well-being as the [single] measure,” Plant told me.
So far, however, no government has fully thrown away more conventional metrics in favor of using only Wellbys. Graham thinks doing that would be a bad idea. “The Wellby is a reasonable measure. Many of my colleagues are in favor of it,” she said. But she’s adamant that it should not be the only measure. “I’m not interested in replacing GDP or all the other objective measures,” she explained, because what’s most useful is “when you find gaps between objective and subjective measures, and then you can ask what’s driving that.”
“Take the paradox of unhappy growth in rapidly growing countries like China and India, where life satisfaction falls and suicides rise at periods of rapid growth,” Graham said. “GDP captures the growth rate, and subjective well-being measures capture how people are experiencing and feeling about life. Rapid growth brings uncertainty, cultural change, and rises in inequality, all things people do not like. These things usually even out over time but periods of change, especially rapid change, tend to be unhappy ones.”
And in the US, Graham noted, “GDP levels and unemployment rates suggest prosperity and stability.” But they mask extremely high levels of inequality — one of the reasons why Americans are becoming more miserable — and high levels of despair, including rising mortality due to deaths of despair. “GDP alone cannot tell that story.”
On the flip side, subjective well-being alone also may not tell the full story. People in poor countries sometimes score higher on happiness than their income levels would predict, but their happiness could stem from the hope that things will improve. We shouldn’t let it occlude the fact that they’ve been forced into situations of deprivation because of global injustice. “You still need to look at their life expectancy, their disease rate, and so on,” Graham said. “That’s why I think you need both kinds of measures. They capture what we wouldn’t otherwise notice.”
Likewise, when I asked Hassenfeld if he could imagine GiveWell ever completely doing away with objective measures like QALYs and only using subjective well-being to compare cost-effectiveness across charities, he demurred. “I mean, never say never,” he said. “But I think it’s unlikely that we would ultimately go all-in on this metric.”
Should we prioritize saving lives or improving them? And who gets to decide?
When you’re deciding what to fund, it’s definitely helpful if you can directly compare how much good different things do in a single currency. But whether the currency you use is GDP or QALYs or Wellbys or something else, the good life isn’t reducible to a universal mathematical equation. It’ll always depend on your philosophical assumptions.
If you have to choose between extending and improving lives, one philosophical question you’ll have to face is this: Is it worse for a little kid to die, or an adult? In Western philosophy, there are three main views one could take on this.
According to deprivationism, it’s best to save the youngest kids, because they stand to be deprived of the most good years if they die now.
According to the time-relative interest school of thought, it’s best to save someone older because they’ve got the psychological capacity to think about their future self, so they have more interest in getting to become that older self.
And according to Epicureanism, death isn’t actually bad for the person who dies (how can something be bad for them if they don’t exist?). So it makes sense to give more weight to living well, not living long.
Which of these philosophical views you subscribe to can dramatically alter the cost-effectiveness of the charities or policies you’re evaluating.
Plant’s team showed this by comparing how much good three charities do — StrongMinds, GiveDirectly, and the Against Malaria Foundation (AMF) — in terms of how many Wellbys each produces. It turned out saving lives by preventing malaria was a bit more cost-effective than StrongMinds, but only if you held the philosophical view most favorable to saving lives. If you embraced a different assumption, like deprivationism, AMF’s cost-effectiveness nosedived to 12 times less cost-effective than StrongMinds, or around the same as GiveDirectly.
The crucial point here is that because no one philosophical view is objectively right — this comes down to your own personal values and opinions — the issue can’t just be mathematized away.
And there’s a growing consensus that those on the receiving end of charity or policy decisions should be the ones whose values and opinions carry the day; anything else would be paternalistic. Governments like those of Wales and New Zealand have directly sought the input of their populations to ask what matters most to them. And the charity world is catching on, too.
A few years ago, GiveWell asked extremely poor people in Ghana and Kenya about their “moral weights.” How do they weigh saving lives versus improving lives (say, by reducing poverty)? The respondents came out very strongly in favor of saving lives, especially young lives.
It was a good impulse to survey recipients about their preferences, but there were two unfortunate things about how it was done. First, the questions were formulated in a complicated way that could have confused respondents, since they required a solid understanding of probabilities. And second, instead of just basing its spending decisions on the stated preferences of the respondents, GiveWell also polled its own staff and donors for their personal views and folded those into its decision-making.
“They decided to weight [the views] on the basis of the answer they thought was most sensible in the first place. That’s circular,” Plant said. “I’m inclined to go the other way — to say, this is really complicated, it’s unpleasant to think about, and people will make different choices.”
That probably feels like an annoying place to end up. But it’s true. We’ve now got more sophisticated methods for measuring subjective well-being, but it’s still complicated. It feels morally uncomfortable to think about the questions this type of measurement raises. And, most importantly, these questions are not answerable through math alone — some of this will always be about values, and those are, well, subjective.