I have just returned from Rome, where I participated in an FAO meeting of technical experts on agricultural investment, particularly foreign direct investment. The meeting was convened to review evidence on agricultural investments in several countries, identify gaps in information and analysis, and think about possible directions of support by development partners. This reflected continuing concern about issues raised by large-scale agricultural investments, which I discussed in a recent ODI blog.

I brought away four main messages from the meeting:

First, there is a real mixed bag of projects out there, with different business models, products and country contexts that need to be unpacked to a greater degree before really strong recommendations can be made from comparing different projects.

Second, it is still pretty early to judge the impact and success of these projects. It can take a long time to get a business off the ground and to reach a point where it starts to achieve what it set out to do.

In the meantime, governments and other stakeholders need the capacity to ensure that the best projects are picked and that the lemons in the bag don’t get to the stage of being allocated land and the whole raft of investment incentives that often accompany large-scale investment projects. This means meeting investors much more on their own terms, armed with the questions and information to negotiate better deals, understanding companies’ strategies and objectives in investing, and being able to assess whether proposals are sensible.

And finally, don’t forget domestic investment. While foreign direct investment is important in terms of its potential impact – particularly with the dynamics surrounding the current spate of large-scale agricultural investments – the vast majority of agricultural activity is financed by domestic resource flows (the Policy Knowledge Platform is set to publish a report that will provide more evidence regarding this). Agricultural activity is financed from income generated both on and off the farm by rural households with multiple activities.

In addition to these main messages, the meeting also provided some useful food for thought about the current spate of agricultural investments – supported by the conclusions from an interesting World Bank study using data collected on the Commonwealth Development Corporation’s (CDC) investments since 1947.

This study showed that a proportion of CDC’s investments failed, so we need to be prepared to accept the fact that some of the agricultural investments currently underway may also fail. Only about one third of CDC’s investments generated moderately attractive internal rates of return (>12%). Looking at the reasons why these investments failed, just under two thirds were down to flawed concepts and then two thirds of these, it was felt, could have been spotted by the people assessing whether to go ahead with the projects or not. So, while this highlights the need for a much more rigorous evaluation process, it also raises an important question: if a reasonably well-resourced professional financial institution can’t always spot the dummy projects, can we realistically expect governments in developing countries always to get it right when deciding which projects to approve? And what kind of capacity do we need to help build in order to reduce the risk of bad decisions?

Where businesses initially failed, however, they were often picked up at a cheaper price by another investor and taken forward with a reasonable degree of success. So an initial failure may not mean the end of the venture.

Out of the four models used within the projects funded by the CDC, the most viable was that of a nucleus estate with outgrowers, followed by a pure processing unit. The two ‘opposing’ models of pure plantations and wholly independent growers had about the same success rate.

The World Bank study showed that it can take a long time before the developmental impacts of investments really come through and can be assessed adequately. Some of the CDC projects had a positive impact over the long term even when it didn’t look as though they would fulfil their potential initially. This reinforces the idea that it may simply be too early to gauge the impact of many of the agricultural projects currently being launched (although we can clearly learn from previous experience in thinking through the likely impacts).

Finally, the data demonstrated that new ventures are the most risky and returns to CDC’s projects were usually better where investments built on existing ventures in a step-based approach. So, a strategy of approving mega farms in the uncharted waters in which some of these investments are sailing (such as biofuels) may simply be too risky.

So, overall, what can we say? Approach large projects with caution, with realistic expectations and armed with the right set of questions. Learn from the past and look at successful models for guidance. And, finally, don’t let attention on foreign direct investment push domestic investment issues into the shadows.