Giffen Goods

A Giffen good is a product for which demand rises when its price rises and falls when its price falls, the opposite of the usual link between price and quantity. The classic candidate is a cheap staple food eaten by very poor households: when the staple becomes dearer, the family can no longer afford better foods and buys even more of the staple to avoid going hungry. Giffen goods are rare and mainly of interest in economic models and in the study of poverty.

The idea is named after the Scottish statistician Sir Robert Giffen; Alfred Marshall attributed it to him in the 1895 edition of his Principles of Economics. Giffen goods are often confused with Veblen goods, which are also bought more when prices rise but for status reasons, and with positional goods. The mechanisms are completely different: Giffen behavior comes from poverty and necessity, Veblen behavior from display.

Imagine a family that lives on rice with a little meat when money allows. Rice is cheap and fills stomachs; meat is a treat. One month the price of rice jumps. The family is now poorer overall, so the first thing to go is the meat. To get through the month without going hungry, they replace the missing meat with more rice, the cheapest filling food left.

So although rice became more expensive, the family ended up buying more of it. That odd result is what economists call Giffen behavior.

A price change has two effects on demand. The substitution effect always pushes people away from a good that has become relatively dearer. The income effect reflects the fact that a higher price makes the buyer poorer in real terms. For most goods, buyers cut back when they are poorer, as marginal utility theory predicts, so both effects work together and the demand curve slopes down, as in standard supply and demand analysis.

An inferior good is one people buy less of as their income rises. A Giffen good is an extreme inferior good with very inelastic demand: it takes up such a large share of a poor household's budget that the income effect of a price rise outweighs the substitution effect. The decomposition was formalized by Eugen Slutsky in 1915:

• $\partial x / \partial p$ is the total change in quantity demanded $x$ when the price $p$ changes.
• $\partial h / \partial p$ is the substitution effect, which is always zero or negative.
• $x \, \partial x / \partial m$ is the income effect, where $m$ is income.

For an inferior good $\partial x / \partial m$ is negative, so the income term is positive. If it is large enough, the total effect turns positive. A simple example: a household has $100 a week and buys 60 kg of rice at $1 and 10 portions of meat at $4. Rice rises to $1.25. To keep up its calories the household buys 64 kg of rice, costing $80, and can afford only 5 portions of meat with the remaining $20. Rice became 25% dearer, yet rice purchases rose from 60 to 64 kg.

For decades economists debated whether real Giffen goods exist. Potatoes during the Irish famine of the 1840s are often cited, but the famine cut potato supply so sharply that total consumption fell, and Sherwin Rosen argued in 1999 that the data do not support the claim. The strongest modern evidence comes from Robert Jensen and Nolan Miller, who studied households living not far above the international poverty line and published a randomized field experiment in the American Economic Review in 2008. They gave randomly chosen households in two Chinese provinces vouchers that cut the price of their staple. In Hunan, cheaper rice led households to buy less rice, which implies that a higher price would raise demand: Giffen behavior. The evidence for wheat in Gansu was weaker. The authors found the effect only among households that were poor but not the very poorest, since the poorest could not substitute away from the staple at all.

Giffen behavior matters for food policy in developing countries. A rise in staple prices, for example from a drought or inflation, may not reduce staple consumption, but it can damage diets by squeezing out protein and vegetables, so overall consumption quality falls even when calories hold up. Policies such as a food subsidy have nutritional effects that cannot be read from price changes in the staple alone.

The concept also has limits. It applies to individual households in narrow circumstances; market demand for a staple usually still slopes downward because richer households behave normally. Giffen goods cannot exist across all price ranges, because a household that grew poor enough would be unable to buy more of anything. In practice, apparent cases of rising demand alongside rising prices often reflect other forces, such as higher incomes, expectations of further price rises or speculation, rather than true Giffen behavior. The concept is a reminder that the standard rule that price elasticity of demand is negative, and other rules in microeconomics, are strong tendencies rather than laws without exception.

• American Economic Association: Jensen and Miller, Giffen Behavior and Subsistence Consumption (2008)