Let’s say you have a market for…something. Call it booze. Booze is delicious and nutritious – the cause of (and solution to) all of life’s problems.

Market demand for booze (booze is an undifferentiated commodity; units are "flagons") is defined as Q=100-P and supply is Q=X/10. Some easy math, you get supply of (1000/11) (so about 90.9) and a price of (100/11) (so about 9.09); traditionally calculated consumer surplus and producer surplus are 4132.23 and 409.504, respectively.

But, oh no – supply shock! Let’s say a distillery explosion. Nobody was hurt, but many good flagons were tragically lost, and the supply curve shifts left to Q=(P/10)+1. What happens to the market? Some simple math and:

Q drops to 90 flagons.
P rises to 10 gil.
CS drops to 4050.
PS drops to 405.

And there we have our market adjustment – moving right along!

Or should we? Let’s take a closer look at what our economic theory says really happened here regarding economic surplus. Producer surplus is simple – it is a matter of accounting, the difference between costs and revenue (and likely captured in Ricardian rents). But consumer surplus is on more metaphysical ground. It is the sum of the gap between each market participants maximum willingness to pay and the price paid. But what, really, is "willingness to pay?" And how can we distinguish it from ability to pay?

Let’s look at this example. The loss in consumer surplus is divided into two components – the increase of price for continuing market participants and the shrinking of the market itself. The latter is especially important to this analysis – the market has shrunk by roughly a flagon. So let’s make an assumption here going forward – the market demand curve is the sum of individual demand for booze, but each person is sated with only one flagon. If their income increased, they would spend it on other things. Now, we know this is quite contrary to fact, but let’s roll with it just to see what happens.

Now that we have our market demand curve divvied up into individual curve, we can see that the approximation of this supply shock is that of the original 91 booze-buyers, one is priced out and the other are all out roughly 1 gil. Therefore, the loss in consumer surplus is roughly equal to the loss of 1 gil from everybody still in the market, plus our poor sober sod who was priced out entirely – and, importantly, his total loss in consumer surplus is measured at nearly zero, since he was the marginal market participant. Since his initial consumer surplus from his flagon is next-to-nothing, he is considered to have "lost" very little at all.

Now, let’s imagine two worlds:

In World One, all market participants have identical incomes but different preferences. Some of them crave that week-ending flagon of sweet, sweet booze, but some just don’t care very much at all, and maybe only even participate in Friday Booze-Swilling because it is the socially encouraged thing to do.

In World Two, all market participants have identical preferences but different incomes. All of them "want" the flagon of booze equally, but some of them are quite wealthy and some quite poor.

So when our supply shock hits, in World One the priced-out individual would be the one who had the least innate desire for booze; in World Two he would be the one least able to pay (perhaps he makes but 10 gil a week and needs one for food and board). But, importantly, our diagrams of demand and supply would look identical, and thus would computer economic surplus in our market, despite very different underlying realities. In World One, metaphysical loss to consumers is slim; but in World Two, it is much more severe. Therefore, consumer surplus is not a very good measure of what people want assuming significant income elasticities of demand across market baskets.

This has implications, right?