I was talking to a very knowledgeable friend about whether solar power will become cheaper than coal in the next decade, and he pointed out a problem in how solar advocates price their power. It's well known that you can't just estimate a solar panel's value based on its peak production capability, because much of the time that panel is producing less power or none at all, doing nothing to amortize costs.
Solar advocates admit this, and use Levelized Cost of Energy calculations
to divide energy actually produced by all costs involved, over time and with a discount rate. While solar's over twice as expensive as coal now, the advocates project the cost differential to continue to diminish at the rate it has in the past, and to disappear in a decade.
So my friend's problem is this doesn't distinguish peaking power costs. He's not using the denialist line that baseload power can't include solar at all, but that you still require additional power when solar can't provide it. That additional power is expensive, and this cost externality isn't included in LCOE calculations.
Mulling this over, I think there's an economic and a political angle to take on it. The economic angle is that if we want to consider externalities, then let's by all means consider all externalities - coal isn't going to do so well on that basis.
The political angle isn't whether we should consider externalities, but whether we will
consider the externalities and which ones. Greenhouse gas externalities will not be fully priced in for decades, especially for costs imposed on areas outside of the country where the gases are produced (why should we care about those effects?), but they will start to weigh in, a bit, on costs. The brand new mercury rule
shows some of the other externalities of fossil fuels will start getting price tags as well. Overall, I think the political process will catch up more quickly on fossil fuel externalities, if still very inadequately, than on on LCOE pricing.
One other point my friend made that was a good one - discount rates for future costs mean few companies care about costs more than a decade ahead. I thought new coal plant starts
would be potentially affected by solar price competition, but maybe that price competition is still too far away.
UPDATE: comments point to a good post at Romm's
that discusses the various terms and state of play for solar in the US. Ignoring environmental issues for a second to focus on economics, I think LCOE is fine for any buyer to use to determine whether solar prices out well, but the overall system has to consider other price issues as well. The grid parity at the link works when you're buying electricity at the high retail rates, but it will take a lot longer if you're a utility that can buy wholesale.