Everyone has heard of Giffen goods in their first Economics class: goods whose demand increases with their price. But do they exist? Marshall, who introduce the concept had the example of bread for poor families, but no such case was found in reality. Samuelson popularize the idea that potatoes in the Great Irish famine were Giffen goods, but again, there is no evidence. So, do Giffen goods exists at all?
Robert Jensen and Nolan Miller claim to finally have found a Giffen good, or rather Giffen behavior, as the good has a regular demand in normal circumstances. After running some field experiments in China by subsidizing some goods, they find that if 1) households are very poor and face subsistence nutrition concerns, 2) households consume a very simple diet of a staple good and a fancy good, 3) the staple good provide cheap calories, comprises a large part of the budget and has no substitute, 4) households are not so poor they only consume the staple good, then one can observe Giffen behavior. In the particular case in Hunan, household eat mostly rice with supplements of meat. When the price of rice increase, households seek to maintain caloric intake by eating less meat (which is even more expensive) and more rice.