Dani Rodrik's teaching forms the basis for how I approach industrial policy - namely that governments should intervene where markets fail, and implement policies that target those failures.  This principle also forms the basis for an issue I have with how he approaches policies to promote economic growth through innovation.

In short, Rodrik's position is that market failures lead to an inefficient allocation of resources, which is particularly acute in developing countries.  He argues that entrepreneurs do not invest in new activities due to externalities associated with the 'self-discovery' process.  Namely, entrepreneurs bear the cost of failure, but shares the benefit with new entrants if an activity is deemed profitable.  This leads to an under provision of innovation.  This externality is the basis for patents, but here we are talking about adapting existing activities in a new environment, not a new invention.  Industrial policies that promote structural changes then, can promote economic growth.

This all makes sense to me, but there's a problem.  Rodrik is taking too narrow a view of innovation with his focus on the entrepreneur instead of the broader ecosystem.  Entrepreneurship is an extremely high-risk endeavor, where decisions are made based on expected returns.  Rodrik assumes entrepreneurs do not innovate because the expected payoff is too low due to self-discovery externalities, but there are potentially other market failures within the economy that affect returns.  In particular, it is crucial to understand the absence of venture capital in most developing countries.  Without it, entrepreneurs must either invest their own capital or raise debt, both of which require them to bear greater risks than raising equity and as a result demand higher returns.  
 
Venture capital financing directly reduces the risk entrepreneurs face by transferring it to the investor.  Venture capital also reduces the chance of failure by improving the human capital of the firm through the VC's role as an advisor, providing access to its business network, and sending a positive signal to the market with its investment.  The absence of the option to raise equity thus radically changes the expected returns for the entrepreneur, and potentially has a much larger impact on decision-making than the information externality central to Rodrik's approach.

Beyond its impact to the entrepreneur's expected payoff, venture capital plays a direct role in the ecosystem for innovation.  Venture capital firms are willing to take on more risk in a single firm since high-risk investments are diversified across their portfolios. Furthermore, since venture capital firms invest within industries or in complementary industries, they directly benefit from the success of a single investment and mitigate inefficiencies due to the self-discovery information externality.
 
Since venture capital is a critical component in the ecosystem for entrepreneurship, understanding its absence in developing countries is essential to formulate industrial policy for innovation.  Is the problem low returns, or do market failures contribute?  There is reason to believe it is the latter.  Venture capital suffers from the same externalities of the discovery process as entrepreneurs, albeit to a lesser degree.  In addition, there may be a coordination failure between entrepreneurs and venture capital firms:  entrepreneurs are absent because of the lack of equity financing options, and venture capital firms must have a certain level of entrepreneurship to be profitable.  There also may be a coordination failure between venture capital firms themselves.  Such a coordination failure could exist if the information asymmetry between entrepreneurs and VCs is reduced in a competitive environment by precluding the groupthink that could emerge with a single firm.  Venture capital may also display herd behavior, wherein the most important criteria for making an investment is whether other venture capital firms want to invest as well.  

These arguments begin to provide a basis to believe that multiple market failures contribute to the dearth of venture capital in developing countries, and governments may have a role to play in promoting innovation through policies targeting the venture capital market.  Given venture capital's significant impact on entrepreneurial decision-making, an understanding of failures in the broader ecosystem for innovation, inclusive of both entrepreneurs and venture capital firms, is necessary for the formulation of industrial policy to foster structural change and economic growth through innovation.