Somewhere in this paper there is a number for Apple Cinnamon Cheerios.

An economist sat down, took the 1992 introduction of Apple Cinnamon Cheerios to the American breakfast, and calculated what its arrival was worth to consumers in welfare terms. The answer is 0.002 percent. That is the size of the gift. Two thousandths of one percent of your wellbeing, delivered in a yellow box.

I mention this because Hans-Joachim Voth and Jonathan Hersh put that figure in a table next to coffee, sugar, and tea, and the comparison is the entire argument. Their number for the three colonial goods, arriving in England between roughly 1600 and 1850, is 17 percent.

Not 17 percent more coffee. Seventeen percent more life, measured the way economists measure such things, which is to ask how much income you would have to hand someone to make them indifferent to losing the good entirely.

Economic history has a standard story about the centuries before industrialization. Nothing happened. Real wages in England peaked in the 1450s, when the Black Death had killed enough people to make the survivors valuable, and then fell. By 1600 they were somewhere between two thirds and two fifths of that plague era peak. The recovery afterward was, in the authors' word, anaemic. As late as 1760 the English worker was earning less in real terms than his medieval counterpart.

Voth and Hersh do not dispute the index. They dispute what the index can see.

A real wage series prices a fixed basket of goods over time. It is very good at telling you whether bread got cheaper. It is structurally incapable of telling you what happened when a good that did not exist in your basket showed up in it. The new good enters the index at whatever price it settles to, and all of the enormous consumer surplus from the years when it went from unavailable to merely expensive is simply invisible. The measurement does not undercount the gain. It never counts it at all.

So the question becomes empirical. How big were the new goods?

Before roughly 1700 the European diet was bread and beer, and that is not a figure of speech. The average European ate 182 kilograms of bread and drank 182 litres of beer a year. Half of all household spending and three quarters of all calories came from those two items alone. This is a diet with no second act.

Then the prices of three imported goods collapsed.

By the end of the period, sugar, coffee, and tea together were 7.2 percent of an average English household budget and roughly 10 percent of food spending. For scale, and this is the comparison that does the work, American spending on personal computers in 2004 was 0.6 percent of consumption expenditure.

Thirty three pounds of sugar per person per year. That is a nation that has discovered dessert and has no intention of going back.

Running the equivalent variation calculation over the whole period, the authors get sugar at 7.58 to 8.03 percent, tea at 7.28 to 7.85 percent, and coffee at 1.45 to 1.54 percent. Roughly 17 percent combined. A different and cruder method, following Hausman, gives 13.5 percent for sugar and tea alone, which is reassuring in the way that two bad clocks agreeing is reassuring.

Then they line it up against the modern era.

Coffee, sugar, and tea beat the internet, personal computers, satellite television, and mobile phones put together, with room left over.

You are permitted to find this suspicious. I do, a little. But the mechanism is not mysterious. The internet arrived into a life already stuffed with goods, and had to compete for the marginal hour against television and the telephone and the newspaper. Sugar arrived into bread and beer. The marginal value of the first genuinely new thing in a life of two things is not the marginal value of the ten thousandth thing.

The title is "Coffee, Consumer Choice, and the Consequences of Columbus," and it is worth being precise about which consequences are being counted.

The sugar price fell because Caribbean production scaled in the mid seventeenth century. The mechanism by which it scaled was chattel slavery. The column notes that sugar became affordable once Caribbean centres developed, and then proceeds to the welfare arithmetic for the English consumer, which is the arithmetic it set out to do.

That is a legitimate scope. It is also a one sided ledger, and the title invites the other side in. An equivalent variation calculation asks what you would have to pay someone to give the good up. Nobody ran that calculation for the people producing it, because the framework has no way to represent them. They are not consumers in the model. They are the supply curve.

I do not think this invalidates the measurement point, which is correct and generalizes well. I think it means the paper answers a narrower question than its title advertises, and that a reader should hold both facts at once. English living standards rose more than the wage series admits. The accounting for how stops at the harbour.