NOTE: Doug's comment relates to a French study projecting a 30% reduction in pesticide use within industrial agricultural systems.
http://www.gmwatch.org/latest-listing/1-news-items/13737

COMMENT from Doug-Gurian Sherman: This is a useful study, but it is conservative in its findings. Work in the US, such as at the Rodale Institute, often shows higher yields for organic systems, although more systematic analyses often show somewhat lower yields (typically about 10 percent). A practical issue in countries like the US is that organic is typically more labor intensive, so overall costs can be higher. Where labor is abundant, labor-intensiveness may not be an issue, and generally even a good thing, but should be made easier for workers through new innovations.

There are also other sustainable ag systems that minimize the use of pesticides and fertilizers, but do not ban them, that consistently outperform industrial ag, and can also be compatible with larger farms. For example, work at Iowa State University shows that pesticide and fertilizer use can be reduced by 70 percent or more with no loss in yields (http://arsweeds.cropsci.illinois.edu/Agron_J_2008.pdf ). And as that work progresses, I think there will be continuing dramatic reductions in chemical use.

Because it takes time to rebuild degraded soil, and working with nature means adapting to local conditions, which in turn means more complex farming systems, the biggest farms are usually not interested. They want to maximize yield and acreage and reduce time so they can farm more acres with less labor. These are the same motivations that drive other industries. But it also leads to the owners of capital appropriating more of the profits, and most workers getting less, while mid-sized farms (in the US) are eventually eliminated. So even where these systems work, without political and social change, they are not nearly as widely adopted as possible. Industrial systems are subsidized in the US and Europe, in part directly, and importantly through an entrenched research establishment–both private and public sector–that focuses disproportionately on improving industrial monoculture ag rather than agroecological farming. Not coincidentally, this is in sync with what the big ag companies want, because it allows them to sell more of their products. 

Relatively speaking, agroecology is moving forward with one hand tied behind its back. It is a testament to these farming systems, the farmers using them and their innovations, and the small number of scientists doing research on these systems, that they are as successful as they are. 

Research has an important and under-appreciated role in improving agroecological farming. Scientists working together with farmers can do a lot. Breeding can develop crop varieties that are better adapted to using organic sources of nutrients that are released more slowly than synthetic fertilizers, for example. Pasture crops can be improved for yield, growth in mixtures of species, and for improved nutrition for livestock. Cover crops can be improved for establishment, geographic range, faster growth, and so on. The resilience and yield of crop varieties used in organic and similar systems can be improved. In the US, surveys of organic farmers show that one of the things they most want is improved varieties adapted to organic and sustainable farming systems. And research on how to make all of this fit together, as at Iowa State, is important and can pay dividends.