I’ve been getting a lot of questions about this: "Solar panels a ‘loser,’ professor says."

Severin Borenstein is an economics professor at UC Berkeley. He did an analysis of California’s solar program and found that if you compare the current cost of distributed generation solar PV, which delivers retail power, with the wholesale power cost of a gas peaker running on pre-Katrina natural gas prices — and leave global warming and environmental benefits out of the equation — then solar “isn’t cost effective.”

Quick, someone call the Nobel Committee.

We can argue about faulty assumptions and apples-to-oranges comparisons, but that would continue to miss the forest for the trees. The point of the California Solar Initiative, and other solar programs, is not to deliver the cheapest kWh amongst all possible kWhs in year one. It’s about market transformation. The entire premise of the program is to take an expensive but useful technology and make it cheap.

A great historical analogy is the commercial development of the integrated circuit, as described by Denis Hayes in this 2001 report (PDF) for the Energy Foundation:

Perhaps the best analogy to the role the government could play for solar cells is the role that it played in commercializing integrated circuits. In 1961, Texas Instruments began producing integrated circuits for small, specialized applications. The earliest versions were very expensive: They cost $100 but replaced just a couple dollars’ worth of larger electronics, perhaps two transistors and three resistors. There was essentially no market for such devices in the private sector. Other electronics companies sneered at them. But the American military recognized the potential importance of small, lightweight, low-power integrated circuits. The Department of Defense began to purchase integrated circuits in large quantities. Following a predictable “learning curve” (see the accompanying article by Robert Williams), the price fell dramatically. As the price fell, numerous private market niches opened up.

Sales of Integrated Circuits
Year Unit Price Military Percentage
1962 $50.00 100%
1963 $31.60 94%
1964 $18.50 85%
1965 $ 8.33 72%
1966 $ 5.05 53%
1967 $ 3.32 43%
1968 $ 2.33 37%

And in fact, that’s the path we’re on.

At last year’s American Solar Energy Society conference, First Solar gave a presentation (PDF) that made a credible case that if things continue — market growth and cost reductions — it will be selling its product at $1.25/W by the 2010-2012 timeframe. That’s grid parity in a lot of places. Sunpower makes a similar case (PDF).

Do yourself a favor and page through the slides. The cost reductions are not coming from places that can be invented and directly commercialized straight from a lab bench. It’s process efficiency all through the value chain, combined with economies of scale in materials.

And let’s not forget — materials are only half of the cost equation. Workforce development is the other half, and Lord of the Rings notwithstanding, that doesn’t come from a lab either. It is created through a market.

Borenstein thinks that if we sit on our hands long enough, a government lab somewhere will invent a magic solar bullet. Well, we could give that shot. But given the stakes, and the low level of funding our government directs at the problem, that’s a rather risky proposition.

A better approach, I believe, is to recognize that NREL and friends have done a fantastic job — the technology works. While government R&D efforts should be redoubled, creating markets:

  • leverages orders of magnitude more private equity to the problem. In 2007, publicly-listed solar companies raised about $7 billion globally. Over $600 million in VC money went to California based solar start-ups alone.
  • addresses issues — process engineering, workforce development, material economies of scale — that don’t find their solutions in a lab.

In the late ’80s and early ’90s, a 354 MW solar thermal power plant was installed in 50 MW chunks in the Mojave Desert. The first 50 MW went in at 28 cents/kWh. The last 50 MW went in at 14 cents. Same technology, same material — just learning by doing. Shortly thereafter, policy support was pulled, and Luz went bankrupt.

Imagine where we would be today if that generation’s Borensteins hadn’t gotten their way.

Update [2008-3-28 10:4:22 by Adam Browning]: My post was focussed on how, technical issues notwithstanding, Dr. Borenstein’s paper misses the point. For those further interested, here’s a paper (pdf) that discusses some of those technical issues in dispute.