No 168: Fun with Google Ngram

THE tools on "Google Labs" are really cool. For example, Google Ngram allows a user to search the millions of books scanned into Google Books, to find how often different words or names have appeared in the past.

This can be very interesting for us economists.


For instance, this is what happens if you search using "inflation" and "unemployment":

 

Let's be clear what this graph shows: how often books contain the words "inflation" and "unemployment" between 1800 and 2008.

Firstly, it is clear that few writers talked about either inflation or unemployment until about 1870. This is probably because early Economics was almost universally concerned with microeconomics.

Inflation is the first of the 2 terms to start to be written about, but it is overtaken by unemployment in about 1905. Both of the terms, but especially unemployment, show a huge increase during the Great Depression of the 1920s and 1930s.

After World War 2, the incidence of inflation increases and that of unemployment falls, perhaps due to the Phillips' Curve and its emphasis of the correlation between the two. In the 1970s, inflation leaps up to the same level as unemployment, at the same time as the world economic crisis of the time which saw high levels of both.

The peak for both came during the recession of the early 1980s.

Since then, the frequency drops, possibly due to the relatively good economic climate until the Credit Crunch. It will be interesting to see how it has increased during our current economic crisis, in a few years when the data is available.

Here are some more Ngram graphs:

Not quite sure about Karl Marx's popularity around 1700, about 120 years before he was born! I guess this also shows how Adam Smith's has continued to be a very important figure, even when Marx was at his most influential.

Two more - war and peace first:

How about tea and coffee:

Come on, then, why don't you have a go? Just go to Google Ngram and then tell us about the graph you made.




(Technical note: choosing the option to search through "English One Million" gives a more even sample over the different years.)