# What are Giffen Goods?

What exactly are Giffen goods and are they of purely theoretical interest or has there been empirical evidence of their existence?

• Giffen goods actually came about as an economic concept as a response to the existence of some goods that violate the first law of demand. – rosenjcb Nov 19 '14 at 23:32
• This question appears to be off-topic because the answer is readily available on Wikipedia. – Steven Landsburg Nov 29 '14 at 18:38
• @StevenLandsburg: It would probably be better if you raised this objection over on Meta (i.e. whether "off-topic if on Wikipedia" should be a valid reason to close a post). – Steve S Nov 30 '14 at 3:12
• Have you considered accepting an answer? – luchonacho Aug 24 '17 at 10:30

Consider the Slutsky equation, $$\frac{\partial x}{\partial p} = \frac{\partial x^c}{\partial p} - \frac{\partial x}{\partial I} x.$$

A giffen good is the case where the income effect $\frac{\partial x}{\partial I} x$ is negative and large (in magnitude) enough so that $\frac{\partial x}{\partial p} > 0$.

From Wikipedia:

There are three necessary preconditions for this situation to arise: (1) the good in question must be an inferior good, (2) there must be a lack of close substitute goods, and (3) the good must constitute a substantial percentage of the buyer's income, but not such a substantial percentage of the buyer's income that none of the associated normal goods are consumed.

(Sufficient when (1) also adds that the good is so inferior that the income effect is greater than the substitution effect.)

A Giffen good does not generate utility directly through its price. Contrast this to a Veblen good where there the user actually derives utility directly from the price.

• Why is (3) a requirement? Does the income effect's size relative to the substitution effect depend on the good's share of income? For example, if rice were a Giffen good and price rose, would only households who initially spent a large share of their income consume more rice, and others consume less? Or if thatched roofs are sufficiently inferior, and require rebuilding every so often but don't consume a large share of income, could it not be a Giffen good? – Max Ghenis Feb 24 '18 at 15:17
• @MaxGhenis Yes, the income effect's size relative to the substitution effect depends on the good's share of income. Consider two extremes. Suppose rice is a Giffen good for you. At the same time, it might not be a Giffen good for me. If I don't initially consume much rice (say, almost none) and the price of rice goes up, then the price change doesn't effect my budget because I don't actually consume any meaningful amount of rice. So the income effect is near zero. Now consider the other extreme. If my budget is 100% rice, then when the price of rice goes up, I can't possible consume more rice. – jmbejara Feb 27 '18 at 18:32
• But won't you also barely substitute any rice for something else when it's a small initial share of budget, and substitute much more rice when it's a large share of your budget? i.e. aren't both the income and substitution effects proportional to the initial consumption level? – Max Ghenis Feb 27 '18 at 19:35
• Maybe this would be easier to see if you write the equation in terms of elasticities (again, see en.wikipedia.org/wiki/Slutsky_equation): $\epsilon_{p,x,x} = \epsilon^h_{p,x,x} - \epsilon_{I, x} b_x$ where $\epsilon_{p,x,x}$ is the uncompensated own-price elasticity, $\epsilon^h_{p,x,x}$ is the compensated own-price elasticity, $\epsilon_{I, x}$ is the income elasticity, and $b_x$ is the budget share of good $x$. The price elasticities show the pct change wrt the pct change in price. The wealth effect, as you'll see, depends on the pct change in income and the budget share of the good. – jmbejara Feb 27 '18 at 19:40

There is rather low probability for demand of a good to exhibit the Giffen property at market level, where averaging over heterogeneous preferences, different income levels and consequent differentiated behavior, will usually offset Giffen phenomena.

Looking at @jmbejara answer, goods that are likely to satisfy all three necessary conditions are drugs like heroin. Heroin is a "good" alright, since, although it may have negative side effects and eventually grave effects, its consumption provides positive instantaneous utility. Here we observe that:

1) Inferior good (=income elasticity negative). This comes from data: the bulk of heroin consumption is done by lower-income individuals (that were "lower-income" prior to any income-detrimental effects that the habit itself may eventually have).

2) Lack of close substitutes: heroin becomes a biological need, because it substitutes for the production of opiates that the human body normally produces itself in order to regulate its sensitivity to stimulus (this is why when the habit is not fed, the addict feels actual pain. That heroin may also induce a "get-high" state has to do with the fact that the amount of opiates that is inserted in the body by heroin injection exceeds the level that the body would itself produce). This is a very specific service offered, and substances like methadone for example, don't come close enough. More importantly, they are also not freely available, so the substitution effect, even though it may exist, may not be able to actually materialize.

3) Substantial percentage of income: heroin commands relatively high prices, due to its illegal status (and the associated risk), but also due to the exploitation from the supplier of the necessity and urgency with which the good is demanded. Combined with relatively low income, the condition is likely to be satisfied.

So there is a visible probability that rising heroin prices may lead to an increase in the demand for it, at least for a subset of individuals.

• #1 is definitely a stretch... After all, it's ceteris paribus... If a smack addict finds $10,000 in some back alley, will he decrease his consumption of heroin? No way! That dude's throwing a$10,000 smack party! – Steve S Nov 23 '14 at 6:41
• @Steve S This is the elasticity of a very large arc that you describe. It may very well be positive while the elasticity for an incremental income increase may be negative. Moreover, I don't see why we should impose a single behavior on all heroin users -that tendency to throw away heterogeneity should be led to rest in economics. – Alecos Papadopoulos Nov 23 '14 at 16:38
• @SteveS But most importantly, the situation you describe is a windfall gain, not an income increase, in the more long-run sense that we examine the issue, especially as regards income elasticities. Again, the observed fact that heroin consumption is done mainly by lower-level-income individuals (which comes from the data), is the empirical validation of the claim that heroin is an inferior good. – Alecos Papadopoulos Nov 23 '14 at 16:58
• Copying my question to the top answer: Why is (3) a requirement? Does the income effect's size relative to the substitution effect depend on the good's share of income? For example, if rice were a Giffen good and price rose, would only households who initially spent a large share of their income consume more rice, and others consume less? Or if thatched roofs are sufficiently inferior, and require rebuilding every so often but don't consume a large share of income, could it not be a Giffen good? – Max Ghenis Feb 24 '18 at 15:18

The usual textbook example of a Giffen good (i.e. a good whose demand curve slopes upwards) is the Irish potato famine. The idea is that as potatoes (a staple food) became more expensive, people could no longer afford expensive foods such as meat and so ended up buying more potatoes! However, this example has come in for criticism, not least of all because a potato famine means that the total consumption of potatoes must have been decreasing at the time. You can read a critique of the potato famine hypothesis here.

More recent work has suggested, however, that consumption of rice (also a staple) in some poor parts of China does exhibit Giffen behaviour. This research was published in a paper by Jensen and Miller in the AER in 2008 titled "Giffen Behavior and Subsistence Consumption". They ran a field experiment in which they subsidised rice consumption (i.e. decreased its apparent price) for randomly chosen households and found that the result was a decrease in rice consumption.

Alfred Marshall popularised the notion of a Giffen good in 1895, but it took until the 21st century for fairly robust evidence of Giffen bahviour to emerge. That finding such an example took so long suggests that Giffen goods are, indeed, a very rare phenomenon.

(Correct me if I'm wrong but) a Giffen good is exactly a good whose demand increases as price increases and decreases as price decreases ceteris paribus. In other words its consumption goes down when they become cheaper.

In mathematics: $\frac{\partial x_d}{\partial p_x} > 0$

Do they exists? For me it's obvious there are examples of this in the real world. Just look around, price is also a signal for quality for many (definitely culture dependent), and in casy of very assymetric a little information a higher price can trigger higher demand (in my personal experience). I have however heard many established economists doubt its existance. Here is what one reference book has to say about it. (Mas-Colell et al., 1995, P. 26).

Also make sure to check out a paper [WHICH ANOTHER ANSWER A ALREADY LINKED TO] (Published in the American economic review, 2008) who claims that it "provides the first real-world evidence of Giffen behavior, i.e., upward sloping demand." : http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2964162/