This post was co-authored with Vijay V. Vaitheeswaran, who is an award winning correspondent for The Economist and currently the magazine’s China Business Editor.

It seems like a really difficult time to be an entrepreneurial innovator in America. The venture capital industry has grown risk averse, and angels are only half-heartedly stepping into the breach. Crowdfunding may help, but this has its own risks. Meanwhile, entrepreneurs in Europe are getting lots of support from industrial policy—and that’s not to mention China’s lavish government-led push for “indigenous innovation”.

If only politicians in Washington could get their act together and come up with a serious industrial policy, America could get its innovation mojo back, right? Wrong! The arguments were captured in a feisty discussion on this topic between us held at Singularity University as part of its provocative Which Way Next series of events.

America is not losing the innovation race to China or to anybody else, and its state capitalism does not trump our entrepreneurial variety. That is because innovation is at heart a bottom-up process—a Schumpeterian dance of risk, failure, resilience and reward that is foreign to all-knowing bureaucrats but second nature to entrepreneurs.

Greed is not only good but can also do great good— if, that is, there are clear incentives to tackle the big global challenges. But first, governments must get out of the way. History shows that official attempts to pick technology winners, no matter how initially promising, usually end in tears. Three decades ago, the French government developed Minitel, a national communications network that allowed users to send messages, book train reservations and so on. This was a popular invention, and for its time a clever one. The snag was that bureaucrats insisted on keeping this a closed system, even after it became clear to everyone else that the future belonged to open networks like the internet. Minitel inevitably proved a dead end, and France remains a laggard in the global internet race. This top-down approach was always misguided, but it is absurdly out of place given the speed at which innovation happens today.

Thanks to globalization and the rise of the information economy, new ideas move to market faster than ever before. This is in large part due to the shift from top-down innovation (of the sort made famous by AT&T’s Bell Laboratories and other secretive corporate silos) to more open, networked and user-driven models of innovation. Think, for example, of the iPod: the research behind it was done by firms the world over, but Apple has reaped huge rewards from its skills in design, marketing and systems integration.

Besides, innovation is truly a global enterprise today. Much of the value created by firms is in the form of intangibles such as knowledge networks and open business models. And yet, most governments cling stubbornly to national industrial policies that offer perverse incentives for local firms to squander resources on parochial technologies and outmoded business models. India’s flourishing software and business-process outsourcing industries took off not because of help from the government, but precisely because its License Raj did not understand these nascent fields well enough to choke them off.

That is good reason to think that innovation works best when government does the least. However, that is not to say that governments should do nothing at all. On the contrary, there is an essential—but carefully circumscribed—role for the state in fostering innovation. For a start, only governments can ensure the framework for innovation is sound. America’s success at maintaining the rule of law, encouraging risk capital and applying pragmatic bankruptcy codes all played a role in the spectacular rise of Silicon Valley, for example. Governments should also encourage investment in knowledge-supporting infrastructure, which ranges from smart electricity grids and broadband internet networks to basic research and university education. But investing in the innovation framework must not mean picking technology winners.

Indeed, the biggest boost governments can give to national competitiveness is to stop doing some things. The first thing is to end perverse policies that discourage collaboration outside one’s own firm or country. America’s tax credit offered for corporate research is appropriately generous for work done inside a firm’s labs (and Congress should act on Barack Obama’s recommendation to make this stop-go policy permanent), but stingy if that same work is done with, say, university researchers. In contrast, Canada’s tax law does not punish collaboration in this way.

The second thing is to stop creeping protectionism. International coordination of technical protocols can be a good thing: Europe’s embrace of the GSM standard helped its firms grab an early lead in mobile telephony, for example. However, there are worrying signs that some proposals for such standards (for example, rules on smart-grid protocols or electronic health records) are really covert attempts to produce rules favoring local technology firms. History shows that such efforts can lock local firms into dead-end technologies and leave them unable to compete globally. The most important thing that America must do is to stop closing our door to skilled immigration. Flexible labor markets are essential for a vibrant economy.

Given the democratization and globalization of innovation seen in recent years, companies must be allowed to tap freely into the brainpower of billions of innovators-in-waiting worldwide if they are to remain competitive. Studies of innovation clusters in Israel, Taiwan and southern India have shown that the catalyst sparking the rise of those aspiring Silicon Valleys was the constant flow of talented researchers, entrepreneurs and venture capitalists to and from the actual Silicon Valley (in which 52% of the startups were founded by immigrants). In sum, there are some useful things that government can do to boost innovation. Most center around the framework conditions that allow market forces to function properly, so that the seeds sown by investors and entrepreneurs fall on fertile soil. Given the dismal failure of past efforts, though, governments should approach even these policies with humility. In the end, the best industrial policy is probably no industrial policy.

A video of our entire discussion from Singularity University’s Which Way Next series is posted below: