To me, happiness research is looking rather like the social capital research of the 1990s - intellectually interesting, but yielding very little in the way of worthwhile policy recommendations. But the Australian population, long inclined to looking towards the state, does not agree. According to a new poll reported in today’s Sydney Morning Herald:

77 per cent agreed with the proposition that a government’s prime objective should be promoting the greatest happiness of the people rather than the greatest wealth.

I’m briefly quoted in another story in the today’s happiness coverage disagreeing with the idea that happiness should be an objective of government. This isn’t because there is nothing at all happiness research can tell us about policies conducive to well-being, but because I think there are a range of reasons why, given Australia’s current situation, this objective is not likely to lead to long-term changes in average national happiness:

1) Few significant and lasting benefits for most people. On a 0-10 scale, about three-quarters of the Australian population will rate themselves as 7 or above in happiness or life satisfaction surveys. They are happy already, and while various positive events or changes in their lives could give them a boost, they are likely to adapt back to their genetically-influenced set range. Many of those below 7 will be suffering temporary setbacks, from which they will recover in any case. The main scope for improvement is among the people enduring persistently low levels of well-being.

2) Not easily accessible by policy. People without partners tend to have low well-being (pdf, p.11), particularly in the middle years of life. But what can government sensibly do about this? For a start, there is some reverse causation - not surprisingly, people who were unhappy to begin with find it harder to form relationships. So the government’s job will be to match Australia’s least attractive personalities with mates, and be more succesful at it than all the traditional methods plus the new online search tools. Even the left’s quixotic faith in the state’s capacity to do good isn’t likely to extend to thinking government can achieve anything here. Short of re-introducing arranged marriages, I can’t see that there is much scope for even massaging the statistics. And if divorce laws were tightened, it probably would not increase happiness - people would be unhappy in marriage rather than unhappy in divorce.

3) Redundant additional arguments. There are some causes of ill-being that can be affected by government policy, such as removing causes of unemployment or providing medical assistance to those suffering from illnesses that affect their well-being. But governments are already trying to alleviate these problems. Another argument for trying to remedy them is redundant; the only interesting issues surround how we should go about dealing with these issues.

4) Dubious arguments. Many of the various odds-and-ends policy ideas for improving happiness collected in documents like the Wellbeing Manifesto are not clearly supported by research. For example, it proposes a maximum 35 hour working week, when the research (pdf, p.5) does not show even those working much longer hours than that with low average well-being and, as reported in yesterday’s Australian, most people settle into hours that suit them over time. Nor do I know of any research showing that restricting advertising would have any discernible effects on well-being, or that stopping ‘turning universities into businesses selling degrees and make them the critic and conscience of society’ would help - indeed, it would probably reduce the number of people capable of carrying out the ’secure, rewarding jobs’ the manifesto, in all its banality, thinks are a good thing. Being a ‘critic and conscience’ of society no doubt appeals to the kind of people who write well-being manifestos, but demand for such services is limited.

Governments should focus on what they can realistically achieve. Given the nature of the instruments they have available to them, this means that their activities will be biased toward the material - redistributing money, creating incentives etc. Changing how people feel is much tougher, and attempts to do so are likely to join the long lists of failed government policies.

According to the latest Sensis consumer report 32% of us are ‘more stressed’ than we were a year ago, and 22% of us are ‘less stressed’, creating an overall trend toward greater stress. On the other hand, 28% of respondents in the Sensis survey report being happier than they were last year, while only half that proportion, 14%, felt less happy than twelve months ago.

But as Daniel Gilbert argues in his book Stumbling on Happiness, we are very bad at remembering our past emotional states. When researchers keep contemporaneous records of how people feel, and then ask them how they felt at earlier points in time, the respondents often misreport their own previous state of mind. This, he argues, is because rather than using direct recall of our emotions we use theories of how we would feel in particular circumstances, which are often wrong or neglect factors that would have affected us at the time.

The Sensis survey also asked its respondents why they felt less or more stressed or less or more happy, and we can see here mostly plausible theories as to why they may feel differently to a year ago. Leading the stress factors is a ‘heavier workload’, followed by ‘financial concerns’, a rather catch-all ‘everyday life’, and ‘more study’. It might be the case that these things have actually caused more stress, but also that because we are working harder, or facing high interest and petrol bills, we assume that we must be more stressed than we were before.

For the question on happiness, it seems to be that while more money is a minor factor explaining why people think they are happier than a year ago (9% ‘better financially’ perhaps plus 15% ‘new job/promotion’, though there other reasons beside money for a new job improving happiness), it is a big factor explaining why people feel less happy than twelve months ago (25% ‘cost of living’, 12% ‘financial worries’, 11% ‘petrol prices’, or nearly half - 48% - of all nominated reasons). This could be because of loss aversion, our tendency to feel losses more intensely than gains. But it could also be that in identifiying what has changed over the last year some respondents note their added financial pressures and assume that these have made them less happy than in the past.

Without keeping records of particular respondents it is hard to sort out exactly what’s going on here, but the large discrepancy between those who say they are happier and those who say they are less happy is a little suspicious, given what else we know about subjective well-being measures.

Subjective well-being tends to be quite stable over time at the population level - that is, the averages will be very similar from survey to survey. For example, in the Australian Unity Well-being Index (pdf) it varied through 14 surveys only by 3.1 points on a 0-100 range, and usually no more than 1 point survey-to-survey. In the HILDA survey, which does actually track the same respondents, their annual report (pdf) indicates that through three surveys the life satisfaction average varied by .1 on a 0-10 scale. So while about a quarter of respondents did show a significant shift in life satisfaction over a 12 month period, the ups and downs tend to cancel each other out to get the same average. The Sensis result would seem to suggest that ups significantly outnumber downs, pushing average happiness up. It could be that this is in fact the case, that the last 12 months have been a good year. But memory playing tricks is also a likely explanation.