Happiness & Public Policy

I’ve got a new piece at The American on how money saves us from unhappiness in old age. A slice:

Easterlin, a pioneer of the study of happiness in the field of economics, set out to chart the trajectory of happiness over an ordinary person’s life-span. He discovered that, on average, happiness rises slowly from our early twenties, peaks at about forty-five, and then declines as slowly as it rose. But the smooth arc of happiness over the life-cycle obscures dramatic action in average satisfaction within the main domains of life—family, work, health, and finances—that together compose the overall trend.

Easterlin, drawing on the massive General Social Survey, reports that health satisfaction heads steadily south from eighteen on, while family satisfaction peaks at about fifty then tails off determinedly. Job satisfaction hits a crescendo at about sixty and slopes off with retirement. Only financial satisfaction, like Matlock reruns, gets better with old age. Financial satisfaction, Easterlin finds, dips until the mid-thirties, levels off, then heads skyward, soaring ever higher each remaining year of life. If not for sharply rising financial satisfaction, the mild downward slide from midlife would be a sharp drop into a well of gray-haired despair. Money does make us happy in at least this one way: as a firewall against an otherwise soul-sapping senescence.

But Easterlin—a vocal critic of the money-happiness link—does not interpret his findings quite this way. Why not?

Why not find out?

Here’s a good overview of Richard Lucas’s recent work on adaptation. Lucas has found that adaptation is generally not as complete as previously thought.

The thrill of marriage generally lasts about two years.

The agony of divorce, on the other hand, can last a lifetime.

That’s according to Richard Lucas, a psychology professor at Michigan State University.

Lucas studies happiness - “the thing we call subjective well-being to make it sound more scientific,” he said - and how our levels of happiness can change.

For years, the accepted wisdom among psychologists has been that those levels don’t change. Much.

The article offers a Lucas quote that emphasizes the importance of invidividual variation and the hazards of cross-sectional studies:

Lucas said there is an optimistic message to be gleaned from his results and that message has to do with the variety of individual reactions to any life event.

“For example, when we look at marriage, on average, people come back to where they were in about two years,” he said.

“But there’s a lot of variability in the amount of change that occurs. There are actually a lot of people who get a large boost in happiness following marriage, but they’re balanced out by those people who don’t get any change or even have a decline in happiness.”

The important thing to know is what kind of person you are, and the kinds of conditions under which your kind of person will find lasting satisfaction, not what happens to people on average.

The famous 1978 paper by Brickman, Coates & Janoff-Bulmann on paraplegics and lottery winners created a sort of conventional wisdom around a very strong view of adaptation or reversion back to an ex ante hedonic “set-point” after an big external positive or negative hedonic shock. The new conventional wisdom is that some but generally not all big hedonic gains and losses dissipate due to adaptation. Here’s a good summary of recent work from Diener and Oishi:

In 1978 Brickman, Coates, and Janoff-Bulman presented evidence from paraplegics and lottery winners to offer empirical support for the idea that adaptation brings us all back down to hedonic neutrality, irrespective of how good or bad the event was originally experienced. It should be noted, however, that in a closer inspection, the evidence of Brickman et al. for adaptation was mixed (i.e., paraplegics were not as happy as others). Our recent studies offer stronger support for adaptation, as well as the modifications that must be made to the original theory. First, it appears that people adapt over time, but not always completely back to the point where they started. For instance, we found that both widowhood (Lucas, Clark, Georgellis, & Diener, 2003) and unemployment (Lucas, Clark, Georgellis, & Diener, 2004) led to lower levels of life satisfaction even many years after the event. Although people showed adaptation over time after the event occurred, they had not adapted completely back to their former levels of life satisfaction even after five years. Despite the fact that people may not adapt to all conditions, we have found that they adapt to the smaller rewards and setbacks of everyday life (Suh, Diener, & Fujita, 1996).

A couple recent papers by Andrew Oswald and friends speaks directly to the old conventional wisdom. In this paper, Oswald and Nattavudh Powdthavee “provide longitudinal evidence that individuals who become disabled go on to exhibit recovery in mental wellbeing. Adaptation to severe disability, however, is shown to be incomplete.” And in this paper, on medium-sized lottery winners, Oswald and Jonathan Gardner

offer new evidence by using longitudinal data on a random sample of Britons who receive medium-sized lottery wins of between £1000 and £120,000 (that is, up to approximately U.S. $200,000). When compared to two control groups — one with no wins and the other with small wins — these individuals go on eventually to exhibit significantly better psychological health. Two years after a lottery win, the average measured improvement in mental wellbeing is 1.4 GHQ points.

The GHQ is a measure of general psychological well-being that ranges from 0 (best) to 36 (worst). Healthy individuals score from 10-13. To put the result from the lottery study in perspective, about the worst possible thing that can happen to you in terms of psychological well-being is the death of a spouse. The average loss in GHQ points from such a devastating loss is 5 GHQ points. So, the size of the positive effect of winning a medium-sized lottery is about 1/3 of the negative effect of losing a spouse. I’d call that significant.

They conclude:

The paper’s main result — that a windfall is followed eventually by a significant improvement in mental health — contrasts with standard interpretations of the work of Brickman et al (1978). An advantage of the present study is that we follow the same individuals through time and do not have to rely on cross-section comparisons.

Better research methods are consistently showing that adaptation is not complete, and that persisting gains in a number of kinds of psychological well-being flow from gains in material well-being.

There’s an interesting, but rather strangely house-size obsessed article (the author has written a book on building your own house) on happiness in last week’s Washington Post to which Robin (and my Cato boss, David Boaz) alerted me. The author interviews economist Luis Rayo, who has written a fascinating theoretical paper [pdf] with Gary Becker formally modeling, among other things, the way an idealized process of natural selection would fit organisms with a strong desire for good feelings while also ensuring that the good feelings don’t last very long. In an analogical nutshell: satiation just can’t last long; we’ve got to get hungry all over again to be motivated to get off the couch and look for the next meal. The way I interpret the paper, they nicely show that the process of psychological “adaptation” or “habituation” — the alleged basis of the so-called “hedonic treadmill” — is more a precondition for running at all (like friction) than a way of running in place. Anyway, in the Post article, Rayo points out that not all satisfactions are subject to adaptation.    

More important, [Rayo] went on to say, the psychology literature and surveys clearly show that not all happiness is ephemeral and geared to endlessly moving targets. With nonmaterial things, the target does not move.

“Exercise will absolutely make you feel better. Your social network, family and friends can bring permanent happiness. Longtime relationships can bring long-term satisfaction.”

The claim here is… what? Satisfaction from money is hit hard by adaptation, but satisfaction from health and social embeddedness isn’t?

It’s truly hard to know what to make of the claim. Because I’m certain Rayo knows what he’s talking about, I’m sure he didn’t say “nonmaterial things” are not subject to adaptation. I assume the author intends something like “non-pecuniary,” since exercise is a material thing, as are our friends and family members (to be pedantic about it). And, of course, money can buy both gym memberships and the leisure to nurture our emotionally sustaining relationships. “Material things” and “things money can buy” are not well-defined categories about which one can make useful psychological generalizations.

Suppose tomorrow a Swiss bank account was opened in my name and one billion dollars was deposited in it — but I didn’t know it. It would be pretty surprising if this had any effect on my feelings or my satisfaction with life. Having money and knowing it does affect “happiness” (as construed by survey research) by providing a sense of security and control; we gain something simply in knowing we could convert our cash to consumption. But money affects happiness mostly through actual consumption. No doubt some patterns of consumption are more subject to adaptation than others. Sadly, happiness research has fixated almost entirely on income levels, and almost not at all on consumption levels, much less on differently composed patterns of consumption at different levels. It is safe to say that we know almost nothing about this.

A number of happiness researchers are souring on the strong adaptation thesis popularized by Brickman and Coates’ famous paper on lottery winners and amputees. Richard Easterlin claims to show that the fairly stable level of average self-reported happiness over the life cycle (rises slowly and slightly from about 18 to 45 and then declines slowly and slightly) is a function of offsetting changes in life-domain satisfaction, and not so much adaptation. So, for example, average satisfaction with health declines sharply from middle age. But satisfaction with finances rises sharply. Strangely, Easterlin takes declining health satisfaction as evidence that we do not adapt fully to changes in health, but he does not take sharply rising financial satisfaction as evidence that we do not adapt fully to gains in financial resources. Here is what he says about the latter:

While people’s incomes typically rise during their prime working years, and then level off and decline, satisfaction with their financial situation is, on average, fairly constant until almost age 40, after which it begins to rise, with the largest increase in late life. What must be happening is that conceptions of material needs are being readjusted as actual life circumstances change. Relative to income these needs are lowest in late life, and financial satisfaction correspondingly greatest.

It strikes me that there is an obvious hypothesis Easterlin neglects. Financial satisfaction shoots up right about the time in the life-cycle when income plummets — retirement from the work force. You can see the perversity here in looking primarily at income as a proxy for material well-being. On average old people have relatively small incomes, because they are retired, and relatively huge stores of wealth, because of compound interest. And wealth increases non-linearly due to compound interest, with the biggest gains coming later in life. Forty-something is about when expenses on children fall sharply and the compounding effects of interest starts to get good. And retirement provides the leisure time to enjoy the consumption of accumulated wealth. I conjecture that experience helps us figure out what we really like, and so old people are more likely to consume in patterns that are truly enjoyable.

Click here for an image of the charts from Easterlin’s paper. Financial satisfaction is the only thing keeping us from being miserable in old age! The lasting satisfactions from friends and family all plummet! Now, it seems to me pretty arbitrary to interpret the permanent negative effect of declining health satisfaction as disconfirming the adaptation-setpoint hypothesis but to interpret the permanent positive effect of increasing financial satisfaction in terms of some kind of complicated story about shifting conceptions of material needs. Both explanations should have the same form: Declining health makes us less happy, and we don’t get used to it. Increasing wealth, and the increased leisure and consumption it enables, make us more happy, and we don’t get used to it either.

[Cross-posted from Overcoming Bias.]

In Vegas at the APEE meeting, Larry White gave me his paper “Is There an Economics of Interpersonal Comparisons?,” Advances in Austrian Economics 2a (1995): 135-51, which is really outstanding. Larry does a great job cutting through all the confusion about the various sense of utility, etc. So, since actuality implies possibility, I know it’s possible for economists not to be at all confused about what “economics” does an doesn’t say. But the last three papers I just read (by significant figures in economics) simply refuse to not be confused.

For example, Easterlin writes: “In contrast, economics places particular stress on the importance of life circumstances to well-being, particularly one’s income and employment situation.” I hadn’t realized that economics includes a theory of well-being! (Economics says income and employment have a lot to do with utility; it is not a part of the theory that utility is what makes a life go well.) Easterlin here is working into a contrast between setpoint-adaptation theory with his aspiration readjustment theory. Which is fine. But there is nothing about economic theory that is inconsistent with setpoint-adaptation theory. See for example Becker and Rayo’s paper “Evolutionary Efficiency and Mean Reversion in Happiness” which elegantly and ingeniously models adaptation using orthodox analytical tools without fuss. So why bring it up? Easterlin is really saying that adaptation is not part of economic folk theory. Which is correct, but maybe not relevant? Anyway, I think I’m just about to give up. Pointing out the difference between economic theory and regularities in opinion among economists isn’t hairsplitting. (Imagine political philosophers confusing regularities of opinion among political philosophers for a theory. Oh no, wait!) But it’s tiresome, and after a while, I start to feel like I am hairsplitting.

The new Journal of Economic Perspectives [walled] contains two interesting looking happiness papers, one on “Developments in the Measurement of Subjective Well-Being” by Kahnemann and Krueger, and “Some Uses of Happiness Data in Economics” by Di Tella and MacCulloch.

I’m halfway through the Kahneman-Krueger paper. Worthwhile data on self-reports versus experience sampling method versus day reconstruction method. While Kahneman is spot on about what’s wrong with global self-reports, I think he’s a bit too blithe in assuming that ESM and DRM get at the objective subjective situation (”experiened utility” in Kahneman’s terms). And the argument against shifting aspiration levels (i.e. Easterlin) and in favor of adaptation strikes me as too quick. The argument, in a nutshell, is that if shifting aspiration levels explained flat self-report trends, then net positive affect, as measured by ESM/DRM, ought to be rising with income, but it’s not. Therefore, adaptation. They may be right. But I’m not convinced that the way people score their feeling in moment to moment, or day to day, self-reports don’t also suffer from readjusting expectations. I am convinced that people are going to be pretty accurate about self-reporting the valence of affect, i.e., whether it is good or bad. So I may not have much problem with Krueger and Kahneman’s “U-Index,” which just measures the proportion of time spent having negative experiences. But I’m not convinced that improvements in the quality of experience aren’t being lost in even fine-grained self-reports. Again, I think we need good work on hormonal and neuronal correlates of good feelings to be sure. (They cite an interesting-sounding study that took cortisol samples, which is the right approach.)

Funny line:

A duration-weighted measurement of affect will uncover that conditions such as paraplegia or marriage are not full-time states; they are experienced part-time.

And, Bishop Berkeley says, good thing God is experiencing us all full-time, or we’d all disappear!

Louis Rayo and Gary S. Becker, “Evolutionary Efficiency and Happiness,” Working Paper, University of Chicago.

Abstract. We model happiness as a measurement tool used to rank alternative actions. The quality of the measurement is enhanced by a happiness function that adapts to the available opportunities, a property favored by evolution. The optimal function is based on a time-varying reference point —or performance benchmark— that is updated in a statistically optimal way. Habits and peer comparisons arise as special cases of this process. This also results in a volatile level of happiness that continuously reverts to its long-term mean. Throughout, we draw a parallel with a problem of optimal incentives, which allows us to apply statistical insights from agency theory to the study of happiness.

This looks pretty promising. I’ll comment when I’ve read the paper.

Happiness ain’t a thing in itself—it’s only a contrast with something that ain’t pleasant…. And so, as soon as the novelty is over and the force of the contrast dulled, it ain’t happiness any longer, and you have to get something fresh.

– Mark Twain, explaining adaptation, in 1907.

Tyler Cowen of Marginal Revolution kindly links to us and reminds of us his posts on happiness. This important passage is among them:

The flat-lining of the happiness-wealth relationship may in part reflect a framing effect. The literature usually focuses on aspiration or treadmill effects, whereby the wealthy expect more. Their greater wealth therefore translates into less happiness than might have been expected. But this is not the only adjustment occasioned by growing wealth. The wealthy also recalibrate how they should respond to questions about our happiness. If happiness itself is subject to framing effects, surely talk about happiness is subject to framing effects as well. The wealthy develop higher standards for reporting when they are “happy” or “very happy.”

So let us assume that both framing effects – concerning both happiness and talk about happiness – operate at the same time. This will imply that even a constant measured level of reported happiness implies growing real happiness over time. Life improvements do usually make us happier, while both our expectations and our reporting standards adjust upwardly. This is the most likely interpretation of the aggregate data. Most individuals strive to earn higher incomes, even after they have experienced the strength of “aspiration” and “treadmill” effects. [emphasis added]

Tyler’s point about the effect of framing on questionnaire responses is very important and rather underappreciated. Work based on SWB studies too often reflects a naive assumption that self-reports reliably track or indicate the objective nature of internal states. This view is naive because there is almost no reason whatsoever to believe that we have cognitive access to some fixed or invariant scale against which to compare the quality of the subjective experience. Our experience of temperature, pain, time, color, brightness, and on and on, is contextually variable. So, we can’t just introspect and say “I’m experiencing 32.6 hedons of positive affect,” because there is no external public standard, like the standard meter for a hedon. Instead, a person must go through a fairly complex chain of inferences. The self-report problem is exacerbated by the fact that happiness is quite likely a culturally loaded notion, as opposed to, say, simple sensations of physical pain.

So, I must monitor my own affect, experiential quality, or hedonic tone, make an inference about how the quality of my experience influences my behavior, commit to a generalization about how behavior tends to relate to the quality of subjective experience, observe other people’s behavior, and then, on the basis of my generalization, ascribe to them a certain quality of subjective experience. I must access conventional standards of how it is that a “happy” person feels and behaves, and compare my feeling and behavior to that standard. Then, I conclude that I’m “Very Happy,” “Pretty Happy,” “Not so happy,” etc…

Kahneman, lord of framing, naturally recognizes both sorts of framing effects, and pushes for the study of “objective” happiness, which does not rely so heavily on self-reports. And he pushes for a purer Benthamism in policy. In “Objective Happiness” Kahneman writes:

Policies that improve the frequencies of good experiences and reduce the incidence of bad one should be pursued even if people do not describe themselves as happier or more satisfied. The recognition that aspiration levels adjust and that people will never be fully satisfied does not mean that they cannot be made (objectively) happier. The implication of this analysis is that the goal of policy should be to increase measures of objective well-being, not measures of satisfaction or subjective happiness.

And it is worth pointing out that the unreliability of measures of subjective well-being does not really imply that objective well-being should be the standard for evaluating policy, instead. It may be that neither subjective nor objective well-being is the correct standard for evaluating policy, even if some form of well-being is the correct standard for determining the rationality or desirability for individual action.

In any case, once we take framing effects into account, we ought to conclude, as Tyler does, that the SWB data obscure a probable ongoing positive real relationship between wealth and happiness.

In what I believe is the first post to link to this here Happy and Public Policy blog, my Cato colleague Radley Balko gives his take on happiness:

There’s a bias against contentment in our wiring. In hunter-gatherer society — where much of our psychology was developed — contentment meant complacency, which likely meant starvation. Ambition, angst, and worry compelled us to seek more — more status, more food, more mates, more “stuff.” And so those of our ancestors who lacked contentment probably did better at getting those things, and therefore at living longer and reproducing, than those who didn’t. Meaning they passed ambition, angst, worry and other traits at odds with happiness through our ancestry to us.

Which means, I think, that in virtually any setting, we’re (a) fighting against a natural predisposition toward unhappiness, and (b) we’re likely to measure our own success by those who live around us, our peers, and not on what took place before we were born.

I think there’s a lot to be said for this view. Radley worries that self-reports are likely to be fairly useless due to the tendency of happiness to recur to a fairly stable baseline, such that objective improvements in the conditions of one’s life are unlikely to show up in self-assessments. Again, I largely agree. Yet I don’t think that the fact that the wealthiest, most free societies score highest in average self-reported happiness is an accident either. The surveys don’t tell us nothing. There are institutional settings in which the baseline seems to shift up and stay up. That’s worth pointing out.

I also agree with Radley that much of the policy-relevant happiness work seems to be motivated by an anti-market mindset. However, the striking thing about the data, as unreliable as they may be, is they reveals that liberal market societies are good for whatever it is “happiness” surveys are actually tracking. I think it’s important to understand what this data does and doesn’t tell us about human well-being. And It tells us less than a lot of people think, partly for reasons Radley suggests. The truth about someone’s well-being often cannot be revealed simply by asking them. And it’s important to think carefully and systematically about how political institutions and public policy effect human well-being, and to do a better job of understanding the causal mechanisms by which they do so, instead of just piling up mounatains of semi-useful correlations. Therefore, we blog!

I finally received David G. Myers The American Paradox in the mail. Naturally, it begins, "We Americans embody a paradox." After quoting Dickens, "It was the best of times, it was the worst of times," that is. But what’s the "worst of times" about our times? Myers says, "There are those who wring their hands and just as rightly worry that our civilization could collapse on its decaying moral infrastructure." Just as rightly as whom? Those who claim, "We’ve never had it so good." And then we get data showing that, yes, things are as good as ever. Swell! And the evidence that "civilization could collapse on its decaying moral infrastructure?" Well, nothing. Pretty much.

Out of wedlock births are up (down recently, but up over decades). Crime. Violence in the media. "Rampant individualism." Materialism. Etc. But none of this begins to add up to the collapse of civilization. Indeed, we’ve just seen that we’re as rich and happy as ever. Hurray! No? The paradox, if there is any, is that none of this bad stuff has made us discernibly less happy. The inference ought to be that there is no threat to civilization. The moral infrastructure is sound. "Radical individualism" and "materialism" apparently leave us as happy as our grandparents in the imagined communitarian golden age. Our main political and economic institutions are remarkably robust, even as social institutitons evolve. And we’re just about as happy as people get. "Americans Watch TV More, Get Out Less, Are Exactly as Happy as Ever." Why don’t we see stories like that?

This is the nth book in this genre that I’ve read, and I simply no longer understand them. The more intimate I am with the data they present, the more inscrutable I find the overall arc of the arguments. The lesson they each show us is that we are better off in a multitude of different ways, worse off in a few others, and as happy as we’ve ever been. The troubled and disappointed tone has come to stupefy me. It simply doesn’t make sense, relative to common sense, or to the science, to think of individual happiness as an open-ended increasing sum, rather than as homeostatic, a kind of equilibrium state. So it’s just not a mystery why our wealth or anything isn’t making us a lot happier, because we’ve already arrived.

Assuming that it is possible to compare happiness across people (and I don’t really see why not), then there is a happiest person alive. That person is probably a genetic deviant, like the tallest person, the smartest person, or the fastest person. And, the thing is, they probably aren’t that much happier than many of us. I think we have to accept the possibility that many people who are alive today are about as happy as people get. We may be banging against the upper limits of our (non-reengineered) hedonic capacity. And that’s precisely why people are looking for something else or more, or whatnot.

Because happiness is just one of the good things that makes a life go well, not the thing that makes a life go well. Being happy is like having a good pair of shoes. They’ll take you lots of places. But you still need somewhere to go. And you still need pants.

More happiness stuff:

Rummel has posted this chart from Inglehart and Klingerman’s important paper showing the relationship between political freedom (as measured by Freedom House) and happiness. It’s very clear that freedom (as well having high GNP growth, and a non-communist past–highly interrelated attributes) is good for happiness. Rummel promises to sort out the colinearit problems.

Pete Boettke and Chris Coyne have posted a draft of a paper discussing a Austrian/Public Choice approach to the happiness lit. Chris points to his LibertyGuide review of Easterbrook’s book, and says something after my own heart:

There is one final point to be made regarding the underlying paradox which Easterbrook sets out to solve - perhaps there is no paradox at all. Most people would agree money and material things are not the equivalent of happiness. Given this, why would we expect to see a correlation between an increase in progress and an increase in happiness? It is not clear the claim has ever been that prosperity will lead to the removal of all uneasiness.

Right on.

So, you know all the paradox flogging books: The Progress Paradox, The American Paradox, The Paradox of Choice. Layard begins Happiness with “There is a paradox at the heart of our lives.”

Now, I’m increasingly baffled by the idea that there is anything paradoxical afoot. Many of these books refer to “Easterlin’s Paradox,” in honor of Richard Easterlin’s famous ‘74 paper showing that average self-reported happiness has not gone up as average income has gone up. Now, like I argued in my last post, this isn’t a paradox relative to orthodox economics, because happiness isn’t a concept in the theory of orthodox economics. There is a widespread folk theory, popular among economists, that says that desire satisfaction brings happiness, and that higher incomes brings more desire satisfaction, and so higher incomes ought to bring greater happiness. But the ideas of adaptation and social comparison most often used to explain the stability of the happiness trends are also part of popular folks theories, and, I think, much more plausible prior to rigorous investigation than the economist’s money–>happiness folk theory. Its pretty obvious from personal experience, not to mention from about the entirety of our literary tradition, that we tend to take what we have for granted, that we tend to measure ourselves against others we imagine to be our peers, the money alone won’t make you happy, that what we really need is each other, etc. So, the data show we’re wealthier and that we say we’re just as happy as we used to be. Where’s the paradox? There is no paradox.

But maybe there are explanations, other than the economist’s misguided folk theory, for the bullheaded insistence in describing as a “paradox” something that is in fact predictable and intuitive relative to intuitively plausible psychological principles.

In ages of yore there was a raging debate over whether capitalism or communism was best at delivering the goods. Capitalism now reigns as undisputed champ. But the conquest of scarcity under communism was also supposed to be psychologically transformative and liberatory. And so, yeah, capitalism delivers the goods. But are we transformed? No! We’re almost exactly the same, and that’s really disappointing. If you were expecting the era of material plenitude to free our minds for higher pursuits, and to enable deeper, more meaningful engagement with our fellow men, then capitalism may seem like a bust. We’re left yearning for something else.

So, we’re wealthier than ever, and have the extra freedom that entails. We’re at least as happy as ever. (Despite what you may read in the papers, the average isn’t flat, it’s just rising very slowly.) Indeed, we’re about as happy as people have ever been, as far as we can tell. Depression, like ADHD, is “on the rise” because simply because it is promiscuously defined and diagnosed. But there must be something wrong. Bowling alone? Ennervated by too many kinds of jam? Something. Because life’s just what it’s always been, only just a little better. And we were hoping for something more, well, dramatic.