Despite having grown up in the country and attended my fair share of the agricultural shows and carnivals that pass for entertainment, I have never known the joy of chasing a pig—greased, or otherwise.
I no longer feel a desperate need to address this glaring gap in my childhood experience, because the last couple of years have given me a pretty good sense of what it must feel like.
Happiness is a greased pig chase. I don’t mean this in the ‘happiness is a warm blanket’ Peanuts sense, although clearly rural folks get a big kick out of it. I mean it in the sense that the concept itself is extremely difficult to pin down.
One of the reasons it took so long to finish my book is that I wanted to start with an empirical foundation for what is worth pursuing in life. As I found out in the course of my research, the science of happiness is anything but straightforward. As soon as you think you have a firm hold on the concept, it squeals and squirms out of your grasp, and you run around chasing it until you’re exhausted and filthy and still no closer than you were to begin with. I’m sure this is well-known amongst actual experts, but as a layperson who had only previously encountered breezily confident stories about happiness research, I was taken by surprise.
Most of what I found out didn’t end up making it into the book, but it was fascinating in a sort of perverse way; kind of like encouraging children to wrestle small animals for entertainment.
Having already slid and fumbled and coated myself in muck, I thought I might as well put my notes together in one place. Be warned that this is a wonkier post than usual, but it might be interesting for anyone who wants to apply a more critical filter to the usual headlines.
So: what are the problems with happiness research?
- Definitional problems (what is happiness, anyway)
- Data collection problems (are we even asking about the same thing)
- Comparability problems (how do we compare my happiness to your happiness to some guy in Bangladesh’s happiness)
- Methodological problems (uh oh, all our findings crumble to dust if our underlying assumptions are wrong)
- Relativity problems (how are we meant to measure something when the goalposts are always shifting)
- Prescriptive problems (what happens if we set policy on the basis of all the above)
What the Heck is ‘Happiness’, Anyway?
What is this happiness thing? Everyone has experienced it at some point, so this should be an easy question. It feels as silly as asking what is ‘the colour red’. But putting these kinds of internal states into words turns out to be surprisingly hard. Let’s try a few definitions:
Definition 1: The balance of pleasure over suffering
This is the classic ‘hedonic calculus’ which laid the foundation for the science of happiness. But it’s a deeply flawed definition: as discussed previously, the logical endpoint of perfect pleasure and no suffering is a bunch of indeterminate blobs floating in vats with electrodes wired to their brains.
When you look at how people actually behave, the good life often involves deliberate deprivation and striving, rather than just loading up on fun stuff and avoiding anything that might be hard or unpleasant. So hedonic pleasure is great, but it’s only one narrow facet of the overall recipe for happiness.
Definition 2: Experiential wellbeing
We could try a definition that flattens out the extremes of hedonistic pleasure-seeking into something like ‘just generally feeling good in the moment’. I like this better than the first definition, except…that’s not what people want, either!
Happiness researchers were so frustrated by the finding that people refuse to be boxed into experiential definitions of wellbeing that some of them quit the field. Here’s how the great Daniel Kahneman explained his own departure: “They actually want to maximise their satisfaction with themselves and with their lives. And that leads in completely different directions than the maximisation of happiness.”
Definition 3: Life satisfaction
OK, so people want to feel satisfied with their lives, and are willing to do hard or unpleasant things in order to get there. This must be an important piece of the puzzle. But it can’t be the definition of ‘happiness’, because it actually makes people less happy in the moment.
Definition 4: Sense of purpose
The meaning we ascribe to our actions and the stories we tell ourselves has the power to convert even the most abject misery and suffering into something noble. But again, connection to a higher power or purpose is not synonymous with happiness: you can lead a rich and meaningful life despite living in conditions that are objectively miserable.
So now we have to acknowledge that, as per usual, words are just a series of arbitrary grunts strung together by monkeys wearing pants. There is no Platonic entity floating around out there for us to discover. There’s just a cluster of loosely-related mental states and experiences, none of which have any firm boundaries, that we have agreed to gesture at vaguely with the word ‘happiness’.
Different people will put more or less weight on these various definitions, with the mixture changing with their mood, environment, stage of life, and so on.
First we have to separate all these variables. Then, when we’re asking people questions about how ‘happy’ they are, we have to try to figure out which definition they’re using in response—which depends on which self we’re actually talking to.
Remembering vs Experiencing Self
One of the curious divides in our society of mind is between what psychologists call the Experiencing Self and the Remembering Self. Basically, there’s one part of you that actually lives your life, and another part of you that makes notes for later. These two often end up so completely at odds that it’s hard to believe they’re both swirling around in the same gray matter.
Many parents identify raising children as one of the happiest and most satisfying experiences of their lives. But if you ask them about it in the moment—exhausted from sleep-deprivation, social life obliterated, forced to endure torturous violin recitals—they are less happy than their childless peers. Having kids seems to make life worse for the Experiencing Self, but the rose-tinted Remembering Self converts it into a majestic and life-affirming experience.
So which self is filling in the happiness surveys?
Asking people how happy they are in general is an appeal to the Remembering Self’s whimsy. Without a convenient score of life satisfaction out of 10 conveniently to hand, we tend to grasp for some easily-accessible accomplishment instead: if we have kids, of course we must be happy, because that was our expectation beforehand, and it’s a deeply-ingrained social narrative.
This availability bias makes it hard to tease apart the connection between happiness and other variables we might be interested in, like money or career paths. You ask a young lawyer if they’re happy, and they look around them, and say, ‘um, yeah, I guess I earn a lot and have a prestigious job‘, and then get back to preparing briefs for their 14 hour day. You’re asking them about Definition 1 or 2, and they’re replying about something completely different.
There’s also social pressure in answering these questions: at least some of what is measured as ‘happiness’ comes from people who have been conditioned against complaining, especially if there is a strong social expectation that someone in their situation is meant to be happy.
Let’s assume we can straighten out the definitional and collection problems so that both researcher and subject are actually talking about the same thing, and the data collected is genuine and unbiased.
The next enormous problem is in assuming that these responses can actually be compared in any meaningful fashion.
My ‘Ecstatic’ is Your ‘Meh’
The classic happiness rating system looks something like this. Go ahead and rate yourself:
Oh, you’re a 6? Neat!
But…like…what does that mean, exactly?
Maybe you have a very low baseline happiness, and your 6 is actually the same as my 3. Or maybe my 8 would barely even register on your scale. Who knows? Certainly not the happiness researchers, because there is no reliable way to compare these internal mental states.
There’s not even consistency at the level of an individual. Maybe what I called a 5 yesterday is more like a 6 today. When I’m old and creaky and have lower expectations from life, maybe my 8 will be closer to what I currently consider to be a 4.
And there’s sure as hell no consistency between groups. Is a Bangladeshi 5 comparable to a Canadian 5? How could we possibly tell? What would it even mean to compare these numbers, given that they’re relative to the conditions in these countries, and the relevant set of social norms? I come from a culture in which it would be considered gauche after, say, discovering the cure for cancer, to describe your feelings in language stronger than “yeah not too shabby, full credit to the team, always room to improve”. How does that compare against superlative American culture, in which everything is awesome, the best, yuuuge, etc?
Happiness scales like the ladder above usually collect ordinal rankings (first, second, third…). This is more honest than using cardinal numbers (1, 2, 3…), because there is no pretence of interchangeability or equal intervals—maybe the first rung is the ground, the second is a two-storey house, and the third is the penthouse floor of a skyscraper. But if we actually want to compare these highly idiosyncratic rankings from one person to another, we still have to somehow convert them back into cardinalised numbers.
And this is the coup de grâce. Even if happiness really can be meaningfully measured, the way we cardinalise the numbers—and by extension, the assumptions we make about how happiness is distributed—can turn everything we know upside down.
To Assume Makes an Ass Out of U and Me (and Happiness Researchers)
Does happiness cluster together in a nice bell curve, with a logarithmic drop-off in the tails to the left and right? Maybe. After all, that’s how a bunch of other variables are distributed in the population: height, intelligence, and so on.
But we’ve wrongly assumed a normal distribution before, with disastrous consequences. Maybe there are a few freakishly happy people distorting the average. Maybe there are lots of incredibly depressed people way out on the left tails, and fewer comparatively ecstatic people. If we cardinalise our numbers in such a way that it skews the distribution to the left or the right, we get wildly different outcomes.
This is what economists Timothy Bond and Kevin Lang set out to do in a couple of recent papers. It’s amazing that their work hasn’t had more publicity, because it potentially upsets every ‘known’ result in the field. From the researchers:
If happiness is normally distributed – Mexico, Trinidad and Tobago, Great Britain, Ghana, and Colombia are the happiest countries, suggesting the relation between development and happiness is weak or negative. A right-skewed log-normal transformation (h’ = e2h) strengthens this result, replacing Britain and Colombia with Guatemala and South Africa. However, if happiness is left-skewed (h’ = –e(-2h)), the happiest five countries become New Zealand, Sweden, Canada, Norway, and Great Britain, all members of the OECD.
It’s not only that countries drop in and out of the top or bottom 10 depending on the distribution. The paper goes through all the famous effects we’ve heard about—the U-shaped curve of happiness through life, the observation that increasing wealth doesn’t increase happiness, the paradox of declining happiness among women, and transforms them under different assumptions.
Put bluntly, the entire body of happiness research tells us very little, unless we first make some enormous assumptions:
- Happiness can be cardinalised in some meaningful way (hmm)
- It is normally distributed within populations (hmmmmmm)
- The rate at which experienced happiness correlates with reported happiness isn’t changing (e.g., your older self has the same standards as your younger self)
The Goalposts Are Always Shifting
That third assumption also seems unlikely to be true, because the goalposts for happiness are constantly shifting. This happens on the level of each individual life—you’ve probably found yourself pounding the hedonic treadmill as you adapt to each new thrill, and constantly crave something better. But we also see it at the level of entire countries, or groups within a society (probably women are not getting ‘less happy’—they just have higher expectations than they did in the 1950s.)
The central quandary behind all the problems we’ve mentioned so far is that happiness is relative. It’s not a destination we arrive at, but a constant measuring of distance: the gap between us and our peers, our countrymen, our own past experiences, what we see on TV, our expectations for the future, our biological baseline mood, and so on. It just doesn’t exist as an independent variable in any meaningful way: the question always has to be ‘happy relative to what?’.
Failure to acknowledge this leads us to make some very bad mistakes.
The Easterlin Paradox is the observation that countries don’t get happier as they get richer. There is a big debate as to whether it exists, and Bond and Lang wiped it out of existence with a left-skewed distribution above, but assume for the moment it’s a real phenomenon.
This lends a lot of support to de-growth and anti-capitalism movements, fawning over Bhutan’s propaganda campaign to rebrand itself as the world’s happiness country, and misplaced moralising along the lines of ‘aww look at how happy those poor people are, really makes you think’.
If happiness is purely subjective, then it’s possible that we really aren’t even a smidgen happier than our peasant ancestors were grubbing around in the mud. Does this mean that technology, vaccines, clean water, dental care, and democracy didn’t actually improve our lives? Of course not. It means that happiness is a terrible yardstick, and needs to be burned with fire. Our lives are vastly better by any reasonable definition, even if our dumbass biology refuses to acknowledge it.
The more likely explanation for the Easterlin Paradox is that the definition of happiness is socially constructed, and inflates over time. Even if reported happiness is stable across centuries, or as a developing country gets richer, that actually reflects a real underlying growth in wellbeing, because the bar for ‘being happy’ is continuously getting higher. As Tyler Cowen puts it, it’s better to envy your neighbour’s Mercedes than to envy his horse and buggy.
Moving Past Happiness
There’s nothing wrong with trying to measure things—everyone needs a hobby—but what worries me is when overconfidence in bad models hurts people. I am a big fan of rationality and the Enlightenment, but I think we have crossed into the realm of ‘scientism’ here—the social scientists got jealous of e.g. physicists and tried to copy their precise tools, like a kind of cargo cult reverse-causality, and used them to slap a veneer of credibility onto an inherently fuzzy field.
This fetishisation of measurement has often caused problems: it was a disaster in the finance industry, it was a disaster with the skull-measuring scientific racists, it was a disaster in macroeconomics, and I think it may prove to be a disaster in human wellbeing. The people who are most influential in setting policy and shaping our behaviours are tossing around ‘science’ that is quite possibly doing more harm than good.
To recap: the problem with happiness is that it is a cluster of related-but-different traits, that it changes relative to our peers, our expectations, our environment, and our past experience, that it has no obvious cardinalisation, that it may or may not be normally distributed, and in short, that it resists every attempt to neatly pin it to the page.
No doubt happiness researchers are well aware of all of these problems—after all, they’re often the ones who are pointing them out. But it is shocking that in presenting their findings in papers, news media, TED talks, pop science books, etc, we don’t hear a peep about the limitations of the models.
I don’t want to be one of those assholes who picks holes in things without suggesting alternatives. So: I think there is a better definition of human wellbeing than those stated above, a better framework for setting policy, and a better system for unlocking capabilities and flourishing in our own individual lives. These three things are called ‘eudaimonia’, ‘the capability approach’, and ‘optionality’ respectively, and none of them are subject to the problems discussed above.
I am going to make a minor asshole move in not making any attempt to define or justify these ideas here. This post is long enough, so I’ll either leave it as a tease for my book—in which I explain these concepts in full, and contrast them against the naive conception of happiness—or possibly do a Part 2 at some point down the track.
Incidentally, it looks like greased pig chases are on their way out. I understand that people become deeply-invested in traditions and rituals, which makes it hard to recognise when they have bad outcomes, but this seems like a welcome development. (Also, I’m glad about the thing with the farmyard animals.)