Thread by @FBareille

It's finally out ! With Alex Gohin (@SmartLereco), we have a new paper in @AnnalsOf in which we measure the market and environmental impacts (noinpoint source pollutions, imported deforestation, carbon emissions) of French #pesticide policies:
jstor.org/stable/10.1560…

👇

What did we do on this paper?

First, we conducted a structural econometric estimation of the impact of output and input prices on the farmers' production decisions, including land-use decisions and pesticide applications.

What's new there?

Well, we used a detailed database from @InseeFr where all farming types (including vineyards and fruit production, which use many pesticides, but also breeding farms) are included for the whole France from 1990 to 2015.

Basically, we have all french farmers.

The price to pay to use this database is that we have to deal with limited number of observation. We thus ran a generalized maximum entropy estimations on each of the 21 metropolitain French regions (22 - Corsica).

What did we find?

1 - The own-price elasticity of pesticide application is estimated to be -0.82. This is higher than what it is found in the litterature. It is explained because we included forage, fruit, vegetables and vineyards (which present elasticities higher than 1).

2 - The own-price elasticity for cereals is -0.34, i.e. more or less what is found in the literature (which has mainly focused on crop farms ✔️).

Does it suggest that a pesticide tax would be more effective than what has been previously estimated?

3 - The price elasticities of land-use are lower than for pesticides: it is more difficult for French farmers to modify their land use than their practices.

It means is that breeding farmers will reduce their pesticide applications on forages rather than start growing crops.

Second, once this econometric step is done, we developed an original CGE framework based on GTAP-Agrmodel.

This model does not isolate pesticides from other chemicals, we thus specified the role of pesticides and fertilizers in France and introduced the estimated elasticities.

We then simulated an ad valorem pesticide tax of 50%, as it is done in Danemark for instance. We found interesting results that highlight the complex trade-offs when targetting pesticide reduction.

1 - As expected, the tax has a negative economic impact on the French farm and food processing sectors (- 900 million euros/year but only - 100 million euros for France welfare) but leads to a reduction of pesticide use by 37%.

On this dimension, the policy is a success.

2 - However, the farmers tend to react to this tax by changing their land use. They now prefer crops that use marginally less pesticides but marginally more #fertilizers.

The incidence is that the nitrogen nonpoint pollution increases by 5% in France (#GreenAlgae).

3 - The reduction of French production (by about 6%) is compensated by the increased production in the ROW, notably in Brazil. How do they increase their production? Mostly replacing forests by arable lands.

We have here an illustration of a potential "imported deforestation".

What is interesting is that, if you add deforestation + the increased use of fertilizers in France, the agricultural sector has increased his carbon emissions by about 10% of the French agricultural sector.

So, for -37% of pesticides, you have +5% of #fertilizers, +10% of #carbon and - 5% of farm #revenues.

This is a partical exercise, but our paper illustrates the complex trade-offs faced by the regulators.

A complete valuation of the pesticide value is needed to go futher.