Commentary: When People Behave Badly

Preston, Samuel H.

doi: 10.1097/EDE.0000000000000049
Author Information

From the Population Studies Center, University of Pennsylvania, Philadelphia, PA.

The authors report no conflicts of interest.

Correspondence: Samuel H. Preston, 289 McNeil Building, 3718 Locust Walk, University of Pennsylvania, Philadelphia, PA 19104. E-mail:

Article Outline

Martikainen et al1 show convincingly that alcohol consumption and smoking are shortening national life expectancy in Finland. Together, alcohol and cigarettes have reduced Finnish life expectancy at age 25 by 4.3 years for men and 1.2 years for women. For men in the lowest income quintile, the sacrifice is much greater at 8.6 years, while poor women lose 2.9 years. Losses are much lower in the highest income quintile, at 1.8 and 0.6 years. Alcohol and smoking account for 60% of the income differential in mortality among men and 36% among women. And they account for the large majority of the increases in income differentials for both sexes from 1988–1992 to 2003–2007.

These estimates are based on Finnish vital statistics linked to a sample from the population registry of Finland. The alcohol-attributed deaths are estimated using causes of death assigned on death certificates, including contributing and underlying causes. The estimates of smoking-attributable deaths are based on deaths assigned to lung cancer that are then translated into all-cause mortality, using statistical associations between lung cancer and all-cause mortality that have been observed over time and space.

Comparable estimates can be made in other countries using national aggregates. What is especially valuable about the analysis by Martikainen and colleagues1 is that distinctions are drawn by income class, which can be accomplished only through data linkages. And the authors allow us to see how these differentials have evolved. Relative to other developed countries, Finland does not appear unusual in its proportion of deaths attributable to smoking or in its patterns of smoking by social class.2 However, the authors suggest that deaths attributable to alcohol are more prominent in Finland than in most European countries, which may contribute to what appear to be unusually sharp social class differences in mortality in Finland.3

The main measure of differentials used in the article is a comparison of mortality in the bottom and top quintiles. I applaud the use of quintiles in this analysis, especially for purposes of viewing changes over time because it averts many issues associated with changing patterns of selection. A widely publicized claim that poorly educated Americans have declining life expectancy4 was shown to be incorrect when the bottom quintile of the educational distribution was used rather than years of schooling per se.5 However, the use of only the top and bottom quintiles leaves out three-fifths of the population. Using the entire set of quintiles would enable the reader to see whether the lowest group had fallen into a ditch or the highest group climbed a skyscraper. More comprehensive measures such as the Gini coefficient or the index of dissimilarity would have been a useful supplement.

This article adds to a growing literature, suggesting that behavioral factors are playing a large and increasing role in variations in life expectancy within and across populations of rich countries. An earlier study concluded that smoking accounted for more than half of the social class disparity in male mortality in the four countries investigated.6 A recent monograph of the US National Research Council concluded that high levels of smoking and obesity were the main reason why the United States had fallen behind peer countries in life expectancy.7 Fenelon8 recently demonstrated that the lower mortality of Hispanic Americans—the so-called Hispanic paradox—is attributable largely to the fact that Hispanics smoke much less heavily than non-Hispanics.

Access to health insurance, rather than behavioral issues, has dominated health policy discussions in the United States in recent years. A careful analysis concluded that the absence of health insurance is producing 45,000 excess deaths in the United States each year.9 That is a sobering number, but a relatively small number when compared with the Centers for Disease Control’s estimate10 that cigarette smoking is producing 443,000 excess deaths per year—about 10 times as many.

Back to Top | Article Outline


What is the proper social response when people behave in ways that threaten to harm themselves? One response is to try to identify and address the “causes of the causes”—the constellation of factors that produce the behavior itself. More direct routes are also available.

The most draconian option is outright prohibition of a particular behavior. Bans against the sale and use of heroin, other opioids, and marijuana (with exceptions for medical applications) are in place throughout nearly all developed countries. In 1919, the 18th amendment to the US Constitution established a ban on the production, transport, and sale of alcoholic beverages. This ban was repealed in 1933 by the 21st amendment. A conventional reason for the repeal is that prohibition did not work to reduce consumption, while greatly expanding the scope of organized crime. However, it appears that a stronger motive was the desire to tax the sale of alcoholic beverages when government revenues were flagging during the depression.11

Prohibitions extend to other forms of unsafe behavior. In the United States, only New Hampshire lacks a law requiring adults in the front seat to wear a seat belt. One might expect the same logic to apply to the use of motorcycle helmets. However, because of organized lobbying effort by Hells Angels and other groups of motorcyclists, only 21 states require adult motorcyclists to wear a helmet—and that number is falling.

A second major way to discourage bad behavior is through financial disincentives, including the so-called sin taxes. One of the best-studied examples of the effects of tax increases on alcohol consumption and mortality is Mikhail Gorbachev’s effort to reduce alcohol abuse in the Russia.12 In 1985, Gorbachev embarked on a campaign to raise the price and reduce the availability of alcohol. Prices in Russia ultimately increased by 75%, and sales were reduced by two-thirds. The effect on national mortality levels was massive and immediate.

Russia eventually reversed course—in part because of lost tax revenue—and Finland has now lowered taxes on alcohol. After Estonia joined the European Union in 2004, alcohol taxes were reduced in Finland by an average of 33%. Alcohol consumption increased by 12%. As the article by Martikainen et al1 shows, alcohol-related mortality rose, especially among the poor.

It is important to remember that financial disincentives also affect families’ finances. Higher cigarette taxes have a disproportionate impact on household budgets among the poor, both because they are more likely to smoke and because they are less able to absorb higher prices for cigarettes. An episode in New York State shows the distributional consequences of raising taxes on a pack of cigarettes to $4.65, the highest tax in the nation. In 2010–2011, after the tax increase, smokers in New York households earning less than $30,000 spent an average of 24% of their income on cigarettes. This percentage was double that in 2003–2004 before the state tax was raised. And 24% of the low-income group was smokers compared with 10% of those earning over $60,000.13 Perhaps most discouraging, low-income people were least responsive to the rise in cigarette taxes between 2003 and 2010.

The low price responsiveness of smokers near the bottom of the social hierarchy may reflect the fact that a high proportion of them are severely addicted and experience mental illness. A national probability sample found that people with a mental disorder were responsible for 44% of cigarettes smoked.14 The link between depression and smoking appears to have strengthened over time.15 Applying a rhetoric of “choice” to smoking behavior becomes less and less appropriate as the prevalence of smoking declines and leaves behind a hard core of severely addicted and often mentally ill smokers.

There is a reasonable case to be made that an outright prohibition on smoking would be more beneficial to poor people than the present array of rapidly increasing taxes. Certainly, the issue is worth discussing. It is not far-fetched to imagine that, as the proportion of adults who currently smoke declines below 20%, the voting public could be persuaded of the virtues of prohibition of tobacco. But voter preferences are not all that matter when legislative deliberations are dominated by special interests. I suspect that Hells Angels hath no fury like an aroused tobacco lobby.

Back to Top | Article Outline


Samuel Preston is a professor of sociology at the University of Pennsylvania. His major research interest is the health of populations, with particular attention to factors that drive mortality rates. His current research is addressed to the effects of obesity and smoking on past, present, and future mortality in the United States. He is also investigating reasons why life expectancy in New York City has grown so rapidly in the past two decades.

Back to Top | Article Outline


1. Martikainen P, Makela P, Peltonen R, Myrskyla M. Income differences in life expectancy: the changing contribution of harmful consumption of alcohol and smoking to. Epidemiology. 2014;25:182–190
2. Bobak M, Prabhat J, Nyuyen S, Jarvis MJha P, Chaloupka F. Poverty and. smoking. Chapter 3. Tobacco Control in Developing Countries. 2000 New York, NY: Oxford University Press for World Bank and World Health Organization:41–61 In:
3. Mackenbach J Health Inequalities: Europe in Profile. 2006 Rotterdam, the Netherlands: European Union
4. Olshansky SJ, Antonucci T, Berkman L, et al. Differences in life expectancy due to race and educational differences are widening, and many may not catch up. Health Aff (Millwood). 2012;31:1803–1813
5. Begier B, Li W, Maduro G. Life expectancy among non-high school graduates. Health Aff (Millwood). 2013;32:822
6. Jha P, Peto R, Zatononski W, et al. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet. 2006;368:367–370
7. Crimmins E, Preston S, Cohen B. Panel on Divergent Trends in Longevity, National Research Council Explaining Divergent Levels of Longevity in High Income Countries. 2011 Washington, DC: National Academies Press:182 pp
8. Fenelon A. Revisiting the Hispanic mortality advantage in the United States: the role of smoking. Soc Sci Med. 2013;82:1–9
9. Wilper AP, Woolhandler S, Lasser KE, McCormick D, Bor DH, Himmelstein DU. Health insurance and mortality in US adults. Am J Public Health. 2009;99:2289–2295
10. Centers for Disease Control and Prevention. . Smoking-attributable mortality, years of potential life lost, and productivity losses—United States, 2000–2004. MMMR Weekly. 2008;57:1226–1228
11. Blocker JS Jr. Did prohibition really work? Alcohol prohibition as a public health innovation. Am J Public Health. 2006;96:233–243
12. Bhattacharya J, Gathmann C, Miller G. The Gorbachev anti-alcohol campaign and Russia’s mortality crisis. Am Econ J Appl Econ. 2013;5:232–260
13. Farrelly MC, Nonnemaker JM, Watson KA. The consequences of high cigarette excise taxes for low-income smokers. PLoS One. 2012;7:e43838
14. Lasser K, Boyd JW, Woolhandler S, Himmelstein DU, McCormick D, Bor DH. Smoking and mental illness: A population-based prevalence study. JAMA. 2000;284:2606–2610
15. Murphy JM, Horton NJ, Monson RR, Laird NM, Sobol AM, Leighton AH. Cigarette smoking in relation to depression: historical trends from the Stirling County Study. Am J Psychiatry. 2003;160:1663–1669
© 2014 by Lippincott Williams & Wilkins, Inc