Is the VAT a money machine?

This is a very useful paper, full of facts and figures on VATs around the world.  Here is one bit:

As shown in the first column, all OECD members other than the U.S. have adopted the VAT over the last 30 years or so, beginning with France continuing through adoption by Australia in 2000. The (unweighted) average standard rate of VAT is about 17 percent, but with considerable variation. Within the EU, it varies between 15 percent (the minimum permissible under the union’s rules) in Luxembourg, and 25 percent (the maximum) in Denmark and Hungary. And several non–EU countries apply far lower standard rates than this, the most striking being the fi ve–percent rate in Japan. Most also apply a reduced rate to some commodities, with domestic zero–rating being quite widespread. The fourth column shows that revenue from the VAT is also typically substantial– averaging a little over seven percent of GDP–but again with considerable variation, from a high of over 12 percent of GDP in Iceland to a low of around 2.5 percent in Japan.

The authors — Michael Keen and Ben Lockwood — conclude that a VAT is a "weak" money machine in the sense that increases in a VAT are partially offset by declines in other tax rates.  They also note:

In a purely statistical sense, there is, thus, no strong evidence that the VAT has in itself caused the growth of government.

I saw this on Twitter somewhere, though now I forget whom to thank; sorry! [Update: It is probably "the wisdom of Garett Jones"]