This month, Germany's Federal Environment Agency has published an English translation of Environmentally Harmful Subsidies in Germany, a 2008 report that had previously only been available in German.

Divided in three parts, it begins by introducing readers to some basic concepts underlying environmentally harmful subsidies: how subsidies can be defined, how they affect the environment and why it is important that they be reformed. It then discusses of how significant such subsidies are in Germany, reviewing and adding up the subsidies with environmentally harmful effects in four sectors of the economy: housing and construction, energy, transport and agriculture. It ends with practical advice on how governments can monitor the effects of their subsidies and change or abolish them if necessary.

Interestingly, transport comes out as the sector with the highest amount of harmful subsidy spending in Germany (€19.6 billion), largely thanks to exemption of kerosene from energy tax, an energy tax reduction for diesel fuel and a VAT exemption for international flights. But at least two big-ticket items, nuclear power subsidies and European Union agricultural subsidies are listed as ‘not quantifiable' - evidence in themselves of the challenges facing governments who want to take environmentally harmful subsidies seriously. The total spending on environmentally harmful subsidies in 2006 is estimated at €42 billion.

The paper also offers some useful insights into the complex policy challenges facing governments, noting, for example, that some subsidies are not possible to change in the short term. Bilateral air transport agreements prevent Germany from implementing kerosene taxes, and would need to be renegotiated for the system to change.

The report draws a number of interesting lessons from its analysis, many of which are well-summarized at the end, in a 10-point list of "principles behind an effective, efficient and environmentally sound subsidy policy".

It can be downloaded from the Federal Environment Agency's website.