To understand how real GDP is calculated, imagine an economy in which only two goods, apples and oranges, are produced and in which both goods are sold only to final consumers. The outputs and prices of the two fruits for two consecutive years are shown in Table 7-1.

The first thing we can say about these data is that the value of sales increased from year 1 to year 2. In the first year, the total value of sales was (2,000 billion × $0.25) + (1,000 billion × $0.50) = $1,000 billion; in the second it was (2,200 billion × $0.30) + (1,200 billion × $0.70) = $1,500 billion, which is 50% larger. But it is also clear from the table that this increase in the dollar value of GDP overstates the real growth in the economy. Although the quantities of both apples and oranges increased, the prices of both apples and oranges also rose. So part of the 50% increase in the dollar value of GDP from year 1 to year 2 simply reflects higher prices, not higher production of output.

To estimate the true increase in aggregate output produced, we have to ask the following question: how much would GDP have gone up if prices had not changed? To answer this question, we need to find the value of output in year 2 expressed in year 1 prices. In year 1 the price of apples was $0.25 each and the price of oranges $0.50 each. So year 2 output at year 1 prices is (2,200 billion × $0.25) + (1,200 billion × $0.50) = $1,150 billion. And output in year 1 at year 1 prices was $1,000 billion. So in this example GDP measured in year 1 prices rose 15%—from $1,000 billion to $1,150 billion.

Now we can define real GDP: it is the total value of final goods and services produced in the economy during a year, calculated as if prices had stayed constant at the level of some given base year. A real GDP number always comes with information about what the base year is.

A GDP number that has not been adjusted for changes in prices is calculated using the prices in the year in which the output is produced. Economists call this measure nominal GDP, GDP at current prices. If we had used nominal GDP to measure the true change in output from year 1 to year 2 in our apples and oranges example, we would have overstated the true growth in output: we would have claimed it to be 50%, when in fact it was only 15%. By comparing output in the two years using a common set of prices—the year 1 prices in this example—we are able to focus solely on changes in the quantity of output by eliminating the influence of changes in prices.

Table 7-2 shows a real-life version of our apples and oranges example. The second column shows nominal GDP in 2005, 2009, and 2013. The third column shows real GDP for each year in 2009 dollars. For 2009 the two numbers are the same. But real GDP in 2005 expressed in 2009 dollars was higher than nominal GDP in 2005, reflecting the fact that prices were in general higher in 2009 than in 2005. Real GDP in 2013 expressed in 2009 dollars, however, was less than nominal GDP in 2013 because prices in 2009 were lower than in 2013.

You might have noticed that there is an alternative way to calculate real GDP using the data in Table 7-1. Why not measure it using the prices of year 2 rather than year 1 as the base-year prices? This procedure seems equally valid. According to that calculation, real GDP in year 1 at year 2 prices is (2,000 billion × $0.30) + (1,000 billion × $0.70) = $1,300 billion; real GDP in year 2 at year 2 prices is $1,500 billion, the same as nominal GDP in year 2. So using year 2 prices as the base year, the growth rate of real GDP is equal to ($1,500 billion − $1,300 billion)/$1,300 billion = 0.154, or 15.4%. This is slightly higher than the figure we got from the previous calculation, in which year 1 prices were the base-year prices. In that calculation, we found that real GDP increased by 15%. Neither answer, 15.4% versus 15%, is more “correct” than the other.

In reality, the government economists who put together the U.S. national accounts have adopted a method to measure the change in real GDP known as chain-linking, which uses the average between the GDP growth rate calculated using an early base year and the GDP growth rate calculated using a late base year. As a result, U.S. statistics on real GDP are always expressed in chained dollars.