I showed the graph to illustrate to the group how demand for organic milk was lower for people that placed a higher relative importance on food safety than it was for people who placed a lower relative importance on food safety. But, almost in passing, I told the audience that they might take the figure with a grain of salt because it shows that even if organic was the same price as conventional (i.e., the organic premium was $0), the demand curves predict market shares for organic of only about 8% and 14% (depending on the importance of food safety), which I thought was implausibly low.
After my presentation, an individual who worked for a European food retailer asked why I thought the figures were implausibly low. I said that I presumed most people would choose organic if it were priced the same as conventional. He said, however, that his retail experience was fully consistent with the graph I showed - even when he substantially lowered the price premium for organic, the market share remained relatively low.
I've had those anecdotal thoughts in my mind for a while and recently was able to test them out in a more controlled, survey setting where we could vary product price in a way that there aren't confounds. One of the "confounds" with the European's observation was likely the fact that the organic attribute was likely to appear on less-well-known brands, so we don't know if it was the lesser-known brand or the organic attribute causing the low market share. Our attempt to remove these confounds is this new paper in the journal Applied Economics Letters co-authored with Seon-Woong Kim and Wade Brorsen.
We conducted studies with apples and with milk. In the studies, people made choices between different apples that varied by color (red or green), condition (bruised or not bruised), price, and production method (organic or conventional). Alternatively, people made choices between milk that differed by fat content (skim, 1%, 2%, or whole), package type (cardboard or plastic), price, and production method (organic or conventional).
We used the choices to infer the demand curves for organic vs. conventional, allowing for the fact that different consumers are likely to have different preferences for organic and other milk/apple attributes. Here's what we found.