It’s difficult to track self-destructive behavior

If you want to understand economics you have to know something about human behavior, with all its imprecision, uncertainties, and wonderful eccentricities. Needless to say, it is a lifetime work.

A recent working paper by Susan Stewart and David Cutler of the National Bureau of Economic Research looks at the costs and benefits of our behavior as it affects our health and quality of life. They examined specifically the effects of six behavior patterns surrounding: smoking; food; alcohol; motor vehicles; firearms; and drugs. These patterns, called “modifiable behavioral risk factors,” accounted for 43 percent of the deaths in our country in 2000.

The authors of the working paper looked into whether our behavior had changed in any significant ways. What they found was interesting, important, exasperating, predictable, sad, hopeful, and funny. It is, in other words, a portrait of us as mirrored by our behavior.

The short, almost equation-like form of the findings reveals that the things which improve the quality of our life expectancy are almost exactly offset by the things we do to ruin our health or otherwise end our lives prematurely. Our better behavior adds 1.82 years to our quality-adjusted life expectancy but our bad habits subtract 1.77 years.

The working paper was never intended to be a best-seller to attract movie offers. It was prepared for the National Institutes of Health, which is compiling a list of potentially fruitful areas of research into better health. A good deal of the authors’ work, then, involved clearing a path for further research.

In the process of path-clearing, though, they make some sharply focused observations themselves, which adds a lot of value and readability to the research paper. They attempted, for example, to sort out the medical and non-medical factors in extending or contracting life expectancies.

In one way this was made easier by wealth of data available on medical improvements. On the other side, though, the data on non-medical factors was so jumbled together so that they were unable to tell whether an improvement in health and longevity was due to behavioral improvements, public health measures, or some other causes.

If you enjoy reading mystery novels or watching “police procedurals” on television, you understand the researchers’ problem. Not knowing the time of death, for example, is a significant obstacle to a police investigation, but as we all know, the story doesn’t end there.

The story doesn’t end there for the research into our behavior’s effects, either. In the case of motor vehicle data, for example, it turned out to be impossible at this point to isolate the impact on quality of life and the researchers had to limit their focus to the mortality tables. Much the same situation prevailed in firearms incidents, and they again limited the scope of their analysis to include only homicides and suicides.

These examples typify the kinds of limitations and choices that researchers, and detectives, are often faced with. The result, then, is an amalgam of solid procedure and hard data seasoned with a compound of imagination, intuition, and experience.

In research papers like these there are also what we might call “embedded probabilities.” These aren’t the probabilities that statisticians normally work with, or can derive from the data. Instead, they make up a kind of tally sheet of credibility in our minds as we evaluate each assumption, each “work around” that the researchers use to overcome obstacles the data limitations toss in front of them. It is the same process we’ve seen the police lieutenant on a TV show use to evaluate the progress of an investigation.

The authors of this NBER working paper feel confident that “… substantial improvements in health” through public health programs and behavioral changes are possible. Using their model they estimate, for example, that improvements in obesity and firearms limitations would add 1.09 years to our quality-adjusted life expectancy.

They have done an excellent job of presenting and interpreting the data realistically. Before accepting their conclusions, however, we should also consider some factors that do not show up in the data at all. The first is that the substitution of drugs for alcohol may be an individual and cultural choice that public health programs will find easier to reverse than eradicate. The second, related factor is that we humans can be remarkably imaginative when it comes to self-destructive behavior. It’s a gift and certainly one that keeps on giving.

There are limits to the effectiveness of public health programs and, importantly, times when they become intrusive meddling that invites resistance and push back. The value of accurate, reliable information and education, while a slower process than intervention, should not be ignored as an effective means of encouraging healthy behavior.

James McCusker is a Bothell economist, educator and consultant. He also writes a column for the monthly Herald Business Journal.