The takeaway
While gross domestic product measures all that is produced in a country, net domestic product does not count the production that replaces depreciated goods.
NDP versus GDP
Gross domestic product (GDP) measures all the economic activity in a country. It does not make a distinction about whether that activity is beneficial or not, only that it happened. It measures production at market value, except for government expenses, which are counted at cost.
But say that a country is hit by a major hurricane that destroys quite a bit of infrastructure. Right after, there's a lot of construction. That drives up investment figures, which show up in GDP. So the country may look like it's firing on all cylinders, but in fact the additional activity is just replacing broken capital.
Net domestic product (NDP) measures instead only the economic activity that is not replacing broken capital. The difference between GDP and NDP, gross versus net, is the consumption of fixed capital. In other words, the investment that replaces depreciated capital. Our FRED graph above shows that these two series track each other quite closely, because the US economy is so large and diversified that most of depreciation is regular and catastrophes tend not to generate any visible change.
But if you look into the details, you can see some differences between GDP and NDP. In our second FRED graph, above, we take the first graph and look at growth rates from quarter to quarter, which dramatizes fluctuations. (We stop before the COVID-19 pandemic to prevent the graph from being dominated by that episode.)
We can see that sometimes GDP grows more than NDP and sometimes the opposite occurs. The differences are not huge, up to a half percentage point, but not negligible.
How these graphs were created: Search FRED for and select "Real Net Domestic Product." Click on "Edit Graph," open the "Add Line" tab, and search for real GDP. You have the first graph. For the second graph, click on "Edit Graph," change units to "Percent Change," click on "Apply to All," and end the sample period on 2020-01-01.
Suggested by Christian Zimmermann.