Determinants of obesity in Turkey : a quantile regression analysis from a developing country

This study investigates the factors that may influence the obesity in Turkey which is a developing country by implementing Quantile Regression (QR) methodology. The control factors that we consider are education, labor market outcomes, household income, age, gender, region and marital status. The analysis is conducted by using the 2008, 2010 and 2012 waves of the Turkish Health Survey (THS) prepared by the Turkish Statistical Institute (TURKSTAT). The obesity indicator in our study is the individual's Body Mass Index (BMI). QR regression results provide robust evidence that additional years of schooling has negative effect on individual's BMI and this effect significantly raises across different quantiles of BMI. QR results also indicate that males tend to have higher BMI at lower quantiles of BMI, whereas females have higher BMI at the top quantiles. This implies that females have higher tendency to be obese in Turkey. Our findings also imply that the positive effect of age on individual's BMI levels raises across the quantiles at a decreasing rate. In addition, the effect of living in urban or rural areas do not significantly differ at the highest quantile distributions of BMI. Our results also reveal that the negative effect of being single on BMI increases gradually in absolute value across the quantiles of BMI implying that single individuals have less tendency to be obese or overweight compared to the married or widowed/divorced individuals. Moreover, the negative effect of being in labor force on individual's BMI increases across the quantiles of BMI implying that an individual is more likely to be obese if he/she is out of labor force. Finally, the impact of household income on BMI is positive and significant at all quantiles.