Illustration by Bloomberg View
Illustration by Bloomberg View

The worst and best of U.S. government policy on solar energy has been on display lately. In a cringe-worthy failure, California solar-panel maker Solyndra filed for bankruptcy this week, just two years after winning $535 million of federal loan guarantees.

Solyndra’s rapid rise and fall should have played out entirely in the private sector. Silicon Valley is thick with venture capitalists willing to finance risky, iconoclastic startups.

Indeed, Solyndra raised a staggering $1.1 billion from private sources since its founding in 2005. The extra federal support ended up having the well-intentioned but unfortunate effect of allowing the company to ramp up manufacturing quickly, even as evidence was emerging that it had badly misread the changing economics of the solar-panel market.

A few years ago, prices for the silicon wafers used in most flat-panel solar systems were soaring. Solyndra proposed building an entirely different panel, using tubes coated with thin films of copper-indium-gallium-selenide that would pick up light from any direction.

Solyndra predicted that its tubes would be far cheaper than the silicon alternative. No such luck. Silicon prices have plunged nearly 90 percent from their peak in 2008, making conventional panels the better bargains.

Solyndra now blames its cost disadvantage on the Chinese government’s willingness to subsidize its solar-panel industry. So have two other U.S. solar companies that went belly up in recent weeks, Evergreen Solar Inc. and SpectraWatt Inc. Such complaints sound petulant, given that all three companies benefited from Washington’s support. The truth, as Anthony Kim of Bloomberg New Energy Finance puts it, is that “the technology did not work as well or as cheaply as hoped.”

A day after Solyndra’s failure, it was heartening to see the Department of Energy award a total of $145 million to 69 solar energy projects taking place inside universities, government research labs and major corporations. Many of those grants are for as little as $750,000 apiece, and none is larger than $13 million. Each targets a current impediment to cheaper solar power -- and then provides initial support for a dozen or more different ideas that might lead to breakthroughs.

Steering small amounts of money to many early-stage researchers is the best way for government to operate. Misfires don’t cost much. Successes can be huge. Private-financing alternatives are patchy at best, because these researchers’ ideas might not pay off for a decade or more. The benefits of this sort of pioneering government support can be seen in the Defense Department’s early work on computer networks, which helped make the Internet possible.

Solyndra’s stumbles, on the other hand, underscore how poorly governments fare when they try too hard to pick industrial winners.

To contact the Bloomberg View editorial board: view@bloomberg.net.