Monday, January 04, 2010

Are Taxes the Root of Unhappiness?

According to Allysia Finley, of the Wall Street Journal, the states with the highest tax rates also rank as the “unhappiest.” Finely quotes a study that ranks states according to how happy they are. She notes that New York, Connecticut, Michigan, Indiana, New Jersey, California, and Illinois are all among the unhappiest states, who also have the highest tax rates.

Eight of the ten happiest states lean right while eight of the ten unhappiest tilt left. While the study by no means proves that being liberal makes people unhappy, it does reflect some of the unfortunate implications of living in a blue state.

But first a note on the study. Using data from the 2005-2008 Behavioral Risk Factor Surveillance System and a 2003 economics paper examining quality-of-life indicators, economists regressed the subjective measure of well-being (how people rate their satisfaction) against the objective measure (states' quality-of-life rankings based on compensating differentials). A compensating differential in labor economics refers to the additional amount of income an employer must pay a worker to compensate for the undesirability of a job or the location's lack of amenities (e.g. local and state tax levels, climate, environmental conditions, quality of schools, and crime rates).

For example, employers in New York would have to pay higher wages to compensate for New York's high taxes, traffic congestion, cold weather, and poor schools. Due to these "disamenities," New York ranked lowest on the quality-of-life index.

What's noteworthy about the study is that states' quality-of-life rankings (measured by their compensating differentials) correlated exceedingly well with residents' satisfaction ratings. The correlation between quality of life and satisfaction is statistically significant (P=0.0001; r=0.6; r2=0.36). The coefficient of determination r2 shows how well the regression line fits the data points. While an r2 of 0.36 may not seem large---and in some studies may not be statistically significant---it is unusually high by the standards of behavioral science. To give an idea of the magnitude of this correlation, the r2 of people's satisfaction ratings taken two weeks apart is also 0.36.

Continue reading at

Blog Archive