It is a matter of debate whether governments should play an active role in stimulating industrial upgrading. But it strikes me as highly unlikely that an activist role for government has much benefit for products low on the value chain. A new policy note from ODI on four product markets in five developing countries seems to bear this out. The market for sugar is a particular object of abuse:
...the state is heavily involved in the sugar industry in some countries, including Bangladesh, Kenya and Viet Nam. In all three countries, however, the state-led sugar industries exhibit low productivity and poor performance, and the use of obsolete technology and inefficient farming methods mean poor cane yields and sugar outputs. All three are struggling to compete and survive in the face of competition from sugar that is either privately produced or imported. They need substantial levels of costly government subsidisation, which is unlikely to be sustainable in the long run, thus jeopardising many livelihoods.
In stark contrast, Zambia, which has a private sector-led sugar industry, produces the highest amounts of sugar per hectare of the five countries, (three times higher than Viet Nam which is the next most efficient country).
The five data points don't add up to an econometric study, but the case-study approach employed by ODI provides a much more granular understanding of the political economy issues involved. And it's very hard to ignore the huge difference in sugar prices between Zambia and the other countries. Much more on competition policy is available in ODI's comprehensive 107-page report, plus related case studies for each of the countries covered in the report.