Thursday, July 12, 2012

Non-Sedentary Economics

This week a study appeared on sedentary behavior and life expectancy.  The conclusions are that if sitting could be limited to less than 3 hours per day and watching television could be limited to less than 2 hours per day the life expectancy would increase.  The results were based on a meta-analysis, i.e, a study combining results from other studies, of five studies where people were asked about their behaviors like sitting and watching TV and then followed for some amount of time with mortality as an outcome.

A few thoughts on this issue.  First, the studies were not all US studies.  Is there reason to suspect that the effects of a sedentary lifestyle may not be the same everywhere?  Perhaps.  It may depend on the entire range of activities an individual undertakes.

Second, it is not clear whether the studies tested for any interaction terms.  The results have been interpreted as meaning that even people who exercise regularly are at risk if they also sit a lot or watch TV a lot.  The effects seem to be independent.  But what about a person who exercises regularly (perhaps even pretty hard) and also sits a lot.  Take myself.  Does running 25-45 miles per week (there is a lof of variation in my schedule depending on the temperatures, travel, vacation, work intensity, etc.) while sitting easily 8 hours a day at work, make a difference.  Perhaps the running has an independent effect (I sure hope it does). But do the two interact in more complex ways that are difficult to capture?  And how does sitting and working on a computer compare with watching TV? Or is watching different because one's body is more completely relaxed?  And what about those who multi-task rather than just "vegging out" in front of the TV?

Regardless of the study's limitations, (and the authors are up front about a number of others and no study is perfect), what if we take the results at face value?  First, is there a government role?  Is there a market failure for activity?  Is there a market failure with respect to TV watching?  Or are these just choices that people make that they should be allowed to make?  And what if people with jobs that are largely desk jobs want to change their behaviors?  what options do they have?  Is there any way to facilitate these individuals being productive while not sitting at work?  Is there suddenly going to be more of a market for the standing work desks?  And, if so, whose responsibility will it be to buy them?  The employer?  The employee?  Is there a place for government intervention here?

In the end, for me personally, there are a lot of tradeoffs and a lot of possible behaviors.  No easy answers. I'll just take my chances that my combination of activities, job responsibilities, food choices, and sleep quantities is right for me.  I suppose that is all most economists suggest--that people understand their choices, understand the implications of their choices, and are left as free as possible to make them.   

Thursday, July 5, 2012

Income Inequality and Health

Some researchers hypothesize that there is a relationship between income inequality and health. The hypothesis suggests that in countries with less income inequality health outcomes will be better. This can provide an argument in favor of income redistribution from upper income individuals to lower income individuals. One interpretation would be higher taxes on higher income individuals with some type of extra support/benefits for lower income individuals.

Many of the studies that have supported this hypothesis have focused on cross-sectional data from multiple countries. One concern that other researchers raise is that when using this type of data, the countries with a great degree of income inequality also have more generous social support programs. In this case, it may not be the lower income inequality that actually results in better health outcomes. Instead, the argument may be in favor of more generous social programs--however we may find the resources for such programs Understanding what is only a correlation and contrasting that with causation is key.

One way to re-assess this question is to use panel data. A recent study by Dr. Mauricio Avendano at the London School of Economics tests this hypothesis. He used data over a number of years from the Organization on Economic Cooperation and Development (OECD). The data are advantageous for addressing the research question of interest to Dr. Avendano as he can look at how changes in income inequality over time within a country leads to any changes. He focused on infant mortality data. He did not find a strong relationship between income inequality and infant mortality.

So, does this completely rule out the possibility that programs leading to less income inequality will benefit health? No. But it does suggest that we should look harder for alternative hypotheses to explain what has been observed in cross sectional data and use this information to motivate policy. Policy that is based on correlations rather than causation is not likely to be efficient policy.

As with medicine, our focus in developing policies with respect to the economy and public health should be evidence based. This is simply one more study that suggests that the evidence to support specific programs that explicitly are aimed at reducing income inequality is not there. To clarify--it may be that the only way to find resources for more generous support programs would be to impose higher taxes on higher income individuals. That could result in less inequality in income that can be consumed. The key to the interpretation of Dr. Avendaon's finding is that while we could argue that reducing the income of high income individuals by a relatively small amount should not hurt them much while giving more resources to those with lower incomes should benefit them a lot may be true, there is nothing specifically about reducing income inequality itself that leads to better outcomes.