Iser.essex.ac.uk

Dean Lillard, Cornell University and DIW

You can't always get what you want: observations on self-reported satisfaction,
consumption, and underlying utility

Although social science surveys have long asked respondents to report levels of life
satisfaction or happiness, economists have only begun to use those data in the last ten
years or so. Other social scientists often joke that, rather than simply ask what makes
people happy, economists would rather use data on observed consumption to infer the
shape and arguments in a utility function. The hesitation is founded on an inherent
skepticism on the part of economists about outcomes that cannot be verified through
objective behavior and on the very powerful notion that preferences are revealed
through observed behavior.
In their survey of the growing economics literature that studies happiness, Frey and
Stutzer (2002) suggest that economists can learn much if they would directly study
data on life satisfaction. Among other things, they suggest that measures of life
satisfaction or subjective well-being can proxy for utility. If one accepts these
measures as a proxy for utility then one can then estimate those goods and activities
that directly yield utility. Frey and Stutzer point to numerous possible determinants of
life satisfaction and review the empirical associations that have been identified to date
in the literature. These include consumption, both absolute and relative income and
wealth, social status, unemployment (both individual and aggregate), inflation, and
the democratic institutions and processes in the society in which a person lives. Frey
and Stutzer suggest that studying the determinants of life-satisfaction will yield
implications for economic theory and public policies. It is possible to use measures of
life-satisfaction to shape public policy if one is willing to assume that life-satisfaction
measures can be used to construct a social welfare function. From this leap one can
then examine how social welfare increases or diminishes with policies such as taxes,
income redistribution, levels of inflation, and unemployment.
This logic can be extended to analyze how individual well-being is affected by being
unemployed or by the consumption of various goods. Data from both Britain (Clark
and Oswald 1994) and Germany (Knut and Gesine 1996) suggest that observationally
equivalent individuals suffer lower subjective well-being when they are unemployed
than when they are employed. Others have argued that irrational consumers suffer
lower levels of life-satisfaction when they become addicted to goods such as
cigarettes. The basis for such inferences is an assumption that consumers of addictive
goods fail to make decisions consistently over time and that they suffer self-control
problems. These assumptions form the basis of the model of cigarette addiction of
Gruber and Köszegi (2001).
Gruber and Mullainathan (2002) adopt these assumptions to argue that cigarette excise taxes may make smokers better off. They use life-satisfaction measures from the General Social Surveys in the United States and in Canada to show that higher cigarette taxes are associated with higher levels of self-reported happiness among people who are predicted to be smokers. They argue that predicted smokers are happier when they face higher taxes because it helps them abstain from smoking. Gruber and Köszegi (2004) adopt the assumption that these results are causal to investigate what level of cigarette excise taxes would maximize social welfare.
This paper presents evidence that cautions against the rush to such exercises. The
evidence challenges the assumption on which the above analyses are shakily founded
- that cigarette use causes people to be less happy. Using panel data from Great
Britain (British Household Panel Study) and Germany (German Socio-Economic
Panel), I show that unhappy people are more likely to start smoking and that people
are more likely to quit when they become happier. These results are consistent with a
model in which individuals are not addicted to cigarettes but instead use cigarettes to
medicate symptoms of a preexisting mental condition. That model implies that
happiness determines cigarette consumption not the reverse. It also implies that
reduction in smoking brought about by an increase in taxes will reduce consumer
welfare.
These findings are consistent with the idea that smokers rationally choose to smoke
and rationally choose to quit smoking. The idea that smokers use cigarettes to
medicate mental states is consistent with clinical findings that buproprion - a known
antidepressant - is effective in getting smokers to quit. Until recently, buproprion,
was the only non-nicotine replacement product that was approved by the FDA for use
as a smoking cessation treatment (marketed under the name Zyban). The findings are
also consistent with observed patterns of quitting among smokers - that is smokers
make several attempts to quit.
This pattern has long been interpreted as evidence that individuals are addicted to
nicotine - and indeed it may be - but the important policy implication is that consumer
well-being will not be improved if we could magically eliminate cigarettes from our
economy. Quite the contrary, consumer well-being would drop drastically because
the nicotine mitigates the underlying mental condition and the removal of a ready
source of nicotine would imply that millions of smokers would suffer greater mental
anguish.
These observations should not be interpreted to mean that there are not costs
associated with smoking. Incontrovertible evidence links smoking to reductions in
health and increased risk from a host of illnesses. I also don't claim that everyone
who smokes does so to medicate himself. Indeed, elsewhere I show that advertising is
very effective in getting youth to smoke (Lillard and Sfekas 2004). What is
interesting from that research is that only a small fraction of youth who try cigarettes
remain as smokers. The question remains as to who and why some youth continue to
smoke while others do not.
More generally this paper shows, using longitudinal data that allows us to document
levels of and changes in life satisfaction before an individual starts or before he or she
stops smoking. This evidence can then be used to assign the direction of causality
between life satisfaction and one type of consumption smoking.
References

Clark, A.E., and Oswald, A.J. (1994), Unhappiness and unemployment, Economic
Frey, B.S. and Stutzer, A. (2002), What can economists learn from happiness research, Journal of Economic Literature 40, 402-435 Gruber, J. and Koszegi, B. (2001) “Is Addiction Rational? Theory and Evidence” Quarterly Journal of Economics, 116(4), 1261-1303 Gruber, J. and Koszegi, B. (2004) “Tax Incidence When Individuals are Time Inconsistent: The Case of Cigarette Excise Taxes” Journal of Public Economics, 88(9-10), 1959-1988 Gruber, J. and Mullainathan, S. (2002), Do Cigarette Taxes Make Smokers Happier?, Gerlach, Knut & Stephan, Gesine, 1996. "A paper on unhappiness and unemployment in Germany," Economics Letters, Elsevier, vol. 52(3), pages 325-330, 9. Lillard, Dean and Andew Sfekas. 2004. “The Effect of Advertising on Smoking Experimentation” Manuscript. Department of Policy Analysis and Management, Cornell University.

Source: https://www.iser.essex.ac.uk/files/conferences/bhps/2005/docs/pdf/abstracts/lillard.pdf