Sugar vs. HFCS is a natural experiment. Americans are the experimental group, and the rest of the world is the control group. In non-OECD countries, HFCS has le

Sugar vs. HFCS is a natural experiment. Americans are the experimental group, and the rest of the world is the control group. In non-OECD countries, HFCS has less than 1% market share. Is there anything useful that can be derived from that experiment, though? The control group varies in significant ways, both amongst itself and w.r.t. Americans, and includes countries with both high and low levels of obesity. Has anyone attempted to back out any kind of useful estimate by treating this as a natural experiment and adjusting for confounds? Examples: most sugar in both Germany and Mexico is from sugar rather than HFCS, but Germany is less obese than the US, while Mexico is more obese than the US. Mexico is poorer than either the US or Germany, so that could be the confound; but Germans walk and bike more than either the US or Mexico, so that could be a different confound. It's kind of a messy experiment.