What is the future of solar power?

Joshua Gans nails it:

If we believe Moore’s Law in solar, then the safe bet in terms of behavioural reactions is not to react. Within a decade or two, energy will be socially as cheap as it is privately as cheap now. That means that changing habits for environmental austerity is not the way to go.

I would make a simpler but less optimistic point.  If a solar breakthrough is now likely, in which market prices do we see it reflected?  It is true that fossil fuel prices took a steep tumble in the last few months, but I’ve never heard anyone suggest that price plunge had to do with a forthcoming solar revolution.  It seems cyclical in nature, or perhaps related to the spreading news of further fossil fuel discoveries, including natural gas.  For better or worse, those shale oil and natural gas discoveries — which by the way will create lots of jobs — will further raise the bar against solar power, and it’s not just the Republicans who will promote them.

Alternatively, is there a bubble in the stock prices of solar power specialists?  What’s the total market cap of companies selling solar panels?  Or is there a bubble in the share prices of companies which supply cheap and reliable power storage?  The evidence on these points seems weak to say the least.  Keep in mind that other countries can make the switch even if you think political conspiracy will prevent it here.  And solar panels can be cheap in the sense that my bicycle is cheap, the real question is whether we see industry-wide price changes as would befit a systematic solar energy revolution.

Is there any reason, based in industry-wide market prices, to be optimistic about the near-term or even medium-term future of solar power?  I don’t see it.


How many years is the near-term and how many years is the medium-term? I see solar becoming competitive with fossil fuels in the second half of this century.

If it will ever be competitive, it will be competitive in the next 10 years.


"Ever" is a long time.

Solar cells are currently (roughly) half the cost of solar power (assuming no storage, it's worse with storage), so even if solar cells were 'free' we would only see the price drop by at most a factor of two. So the cost ratio of solar cells to the price of solar power is declining and therefore any additional drops in price aren't liable to have much effect. Most of the price decline has already happened and the cost curve is flattening out, though admittedly the current production glut on the market has dropped it faster than expected.
Either way, if it's not cheap enough within the next 10 years, it's unlikely to ever get there. For the record, I expect it to reach limited cost parity within that time frame.

And of course, I'm assuming no drastic increases in the price of other forms of energy. But a drastic increase in the price of nuclear or wind seems unlikely.

I don't see the decline in price per Watt of installed of solar power flattening. It has been dropping at least 10-15% per year, with bursts to 20-25%. The US price of panels alone in September was $2.60/ Watt peak.

Slide 5 year.

Do you expect the price of power inverters, metal framing, wiring and labor to drop? If not, then the price of installed solar can not keep dropping for much longer. Almost, all of those historical charts only show the price of the solar cells, because the rest of the kit is fairly fixed in price.

As I said, the price of solar cells is roughly half the cost. Even if they were free, that still doesn't make the rest of the equipment cheaper. So, there is a lower boundary of about one half of the current cost.

Yes I do expect further drops there, as I expect the _aggregate_ cost of the non-PV parts of the installation to continue to drop since, as PV _efficiency_ increases, less non-PV (frame, wiring, land, etc) is required for a given power output. Note too that the conversion efficiency of the semiconductor portion is not the only place for efficiency gains, as concentrated luminescent solar and similar technology would allow collection of sunlight via very inexpensive materials, leaving the actual conversion of sunlight to electricity for a relatively small area.


JWatts, you say "Do you expect the price of power inverters, metal framing, wiring, and labor to drop?" The answer is yes. Because one of the drivers to lower cost per watt has come with higher efficiency per cell: the average house needs fewer panels and that means less metal and less labor. Also, installation hardware is benefiting from efficiencies of scale, so even the price per bracket is coming down.

The real problem with solar is not the cost of solar cells, it is the sun, which does not shine 4380 hours per year, because of night. That does not include twilight, clouds, and dirt. Therefor, every watt of solar must be
supplemented by an equal or greater amount of generating capacity or of storage. So you have to pay for two systems instead of one. That is a real deal killer.

I would guess something closer to 2050 or 2075. By then we will probably have exhausted a lot of the fossil fuels that can be extracted for less than $200/bbl oil or equivalent.

Of course, that assumes cheap fusion isn't around, and then there's the thousands of years' worth of fissionables which we've barely touched...

you're missing a point: you still have to burn a lot of diesel fuel to mine the materials, & you still have to burn a lot of coal to make glass & the metal equipment & the transmission lines...if you want solar, better do it now when fossils are still available...

You have to burn a lot of diesel just to mine the coal, too. If you count it on the one hand, you have to count it on the other hand as well.

When the prices are high enough, the market will allocate resources accordingly. They aren't going to run out overnight, there are huge amounts of them that just aren't profitable to exploit at current prices.

Solar power will never be largely used, because the maximum potencial of solar power harvesting is 1400W/mˆ2. If we want to consume 15 terawatts of solar power (2008 consumption according to Wikipedia), we would have to build 10000 square kilometers of solar cells, assuming we capture all the little watts (what is impossible).

I doubt we consumed 15 TW, the units are wrong. It should be in TW-h or GW-h.

Roger's units are right, yours are wrong. TW is a measure of power consumption -- while your measurement of TW*h is a measure of energy consumption and needs a timeframe to be meaningful in terms of power (maybe TW*h per hour :-).


So an array 100 km x 100 km would meet the entire power requirements of the USA?

Let's give it 10% efficiency ie 1/10th, or about half of what solar cells achieve now.

1000 km x 100

Honestly if you remember the Minuteman X 'racetrack' proposals (oval highways in the Utah desert around which MX missiles would move on portable launchers, and thus be immune to Soviet first strike) that was going to take up more of the USA.

The problem is storability *not* taking up 10,000 or 100,000 km2 of some sunny part of the USA.

In fact, since many/ most solar panels will be on roofs, on car parks or shading roads, the space already exists.

"So an array 100 km x 100 km would meet the entire power requirements of the USA?"

15TW would be enough to power the world....

I figure if you put an 1 array 50 KM x 50KM in China's Gobi Desert, Australia's Desert, The United States Southwest, Saharan African, Saudi Arabia, and Somewhere in South America and have the very best in transmission lines throughout the world, and the capacity to store lots of electricity at one time...

Maybe we will eventually have house batteries like we have phone batteries, computer batteries, car batteries...

The city of Phoenix would be half the size of the array....

In fact I advocate putting arrays on top of existing buildings first and foremost.

Financial Times today (November 10 2011)

Solar is here now.

Solar power at retail now $4/watt and in some commercial US installations below $3/ watt.

At the retail end, the point of grid parity is getting close.

Since we buy things for *benefits* not costs, then:

- if solar cells keep falling at 7% pa (but that is say roughly half the cost of solar installation) then those with discount rates below 7% can afford to wait

- however those facing unusually high electricity prices (California, Florida, many European countries) have the incentive to move *sooner*

Sooner because the future positive cash flows lost are outweighed by the time value you capture by moving sooner, unless your discount rate happens to be below 7%.

Note the UK has just halved solar feed in tariffs from next year, aiming at a 5% rate of return for solar installations (inflation linked). FIT from 42p/ kwhr (8 for large installations) down to 21 projected. Retail electricity price is about 13p kwhr.

Solar is *here*. Now. And it's economic in many situations.

Does it need to be a breakthrough? Can it not be a gradual, incremental change?

The argument against gradual, incremental progress in solar tech being disruptive in energy markets is that other forms of energy, particularly fossil fuels, have not experienced the sort of price movements that one might expect were this the case. Presumably, markets can price this sort of incremental progress better than a breakthrough, which is likely difficult to predict. Thus, the argument goes that, unless there is a breakthrough, solar technology will not appreciably change the energy markets in the foreseeable future.

Other forms of energy - fossil fuels, etc. - get better subsidies than solar so market prices are skewed. And if you count the cost of the externalities and transportation and distribution costs solar comes out way ahead.


Sorry Loren, I call BS on the subsidies. Solar is and remains massively subsidized compared almost all other energy technologies. You'll need some pretty solid evidence to take that allegation further.

Maybe you are not up on the latest info about the costs of externalities. According to an August 2011 paper from the AER (link below) oil and coal fired power plants total costs are greater than their value added when you account for externalities.

Uncompensated environmental damage is the equivalent of a subsidy.



According to International Energy Agency (report released yesterday) fossil fuel subsidies, globally, are roughly 5 times renewable energy subsidies (over $500bn for the former).

Solar power needn't displace other power sources. It's just one more power source, so cheaper solar power is just a shift outward of the energy supply curve. There's also a kind of Jevons effect.

Plus, solar power butts up against fundamental physical limits a *lot* sooner than computing speed: for example, you can't get more energy from a solar panel than the sun delivers to it. (Naturally, the limit is much worse: there's the light-to-electricity conversion efficiency, etc)

Yes, but how close are we to that theoretical solar flux limit? I'd guess we've a long way to go.

But it's a logarithmic return on investment.

The best panels are currently around 20%, which I believe is up from around 5%. They might get to 50% someday, but obviously they can never double that, and further progress will cost more for less and less gain.

I think the best usable cells are around 30 percent, the best seen in labs are around 40 percent, and there's a theoretical limit around 85 percent and various practical limits at lower percentages.


I suspect any material progress is not going to come from increasing cell efficiency, but rather from lowering manufacturing, installation, and maintenance costs.

Good points. But remember, the cost side is logarithmic too (as it approaches zero mfg cost, further cost improvements are harder to get and less and less useful because of other costs, such as land, maintenance, etc).

> But remember, the cost side is logarithmic too

I'm in complete agreement.

I think the only way to viable solar power at a large scale is space-based solar power. But I don't think that's an especially near-term thing. I'm optimistic about space launch costs, but not that optimistic.

The real world efficiency of CPV is pushing 40% at 1000 sun concentration. Keep in mind that this is achieved with a very small cell (1 cm^2) and a large but inexpensive Fresnel lens, so we're less tied to the cost of the semiconductor material.

The limit is low -- solar energy flux on the earth is on the order of 1kW/m^2. Add in the ideal conversion loss and it's like 300 W/m^2. Nothing to write home about.

You have left off the corollary - you need to have the area to place a collector, and enough space to generate enough power - if you want 10 kW, you need at least 10 m^2.

Solar is as unlikely to be individual any more than our water system is individual, and for many of the same reasons. Whether solar can be harnessed to displace the often messy and wastefully inefficient systems we currently use to generate power is an entirely different question.

Obviously, some owners of decent sized swathes of cloudless and sunny waste lands are going to make incredible amounts of money by selling electricity from solar panels - as motivated by profit as any other business. Much the same way Iceland makes large amounts of money through electrical based smelting, without needing to burn anything. Possibly large enough amounts of money to create an undersea cable to sell the power to other parts of Europe, for an even larger profit.

That's more like 200W/M^2 with today's mass market silicon PV panels and that's only for the equivalent of five hours per day, or 40W/M^2 over the 24 hour day. Replacing the entire annual US electric generation (~4.2 peta Watt hours) would require a total land area of 100 miles by 100 miles, or 10,000 square miles.

On the other hand, that "Meter squared" can come very, very cheaply for solar photo-voltaic: it does not need access to vast amounts of water, does not need major road or rail conduits to transport tons of fossil fuel, does not require any arable land.

Comparable large land area usage in the US:
o One military base, White Sands Missile Range in NM. - 3200 sq mi.
o All US building rooftop area residential and commercial - 6600 sq mi
o US total land area used to produce corn for corn ethanol - 60,000 sq mi.

If this is the amount generated, we do not need to displace all of it. The transmission loss is something like 66%. Install the solar at the point of use and you cut the generation requirement to 1/3 of what is needed in the central distribution system.

Agreed. Fossil fuel based electric generation is ~60% of capacity, nuclear, hydro, and wind supply the rest. Transmission losses are no more than ~7%.

There are some hard limits you run up against from basic physics which we're getting pretty close to. Here's a great treatment of it:


I think Silas has a good point.

Think of it this way.

First, ask the question: Is any part of the energy price in the market determined by a cartel? Think OPEC.

Second, how much will a cartel cut back on its own production to maintain prices from supplies from multiple other sources of energy and non-Opec members?

Third, will such cutbacks be sustainable in countries with growing populations and will there be accomodations between, say, Saudi Arabia and Iran, a toppled Venezuala, a growing Indonesia, an unexplored Russia, Africa or Brazil.

Everyone presumes the starting price is a competitive price.

Is it.

Two points:
1) Demand for electricity is incredibly inelastic. Maybe that will change with the installation of a smart grid, but its not here yet. Besides, even if supply were elastic, generators can build solar in lieu of other planned generating capacity projects that are scheduled to replace capacity that is retiring, with no net increase in supply.

2) You also can't get more energy from a coal plant than the Btu content of the coal you shovel into it. So what? You can always shovel in more coal and you can always get more sunlight. Physical limits apply to all kinds of generating capacity. That's why most coal plants have efficiency rates well under 50% (aka heat rates over 6800 btu/kwh). The issue is costs per kwh, not physical limits.

demand for electricity is short run price inelastic.

Long run it is not. Californians use c. 40% less electricity per capita than other Americans. Partly that is conservation measures like negawatts but also party it is high prices.

It is however income elastic. That said, other rich countries use less electricity per capita than Americans, even stripping out the air conditioning effect.

What an odd post. Krugman isn't arguing for some sort of miracle breakthrough, he's arguing that prices have been declining steadily and are approaching competitiveness. This doesn't show up in market caps because supplying solar panels is a very competitive commodity business with massive excess factory capacity. Stock market prices tell you very little about how competitive the price of solar electricity is. If you actually look at how much solar panels are selling for, you'll see that prices have plummeted in the range of 40% this year. We also saw a major milestone this year in that solar companies now charge consumers less per kilowatt hour than the typical utility rate in some markets (Hawaii, admittedly low hanging fruit). I don't mind your skepticism, but at least use the right metrics when trying to assess the state of play.

There are so many problems with the chart we are arguing about. All the data is made up in the second half. It's extrapolation. It's tough to be sure which point is the last real one (look at Naam's original for clarification.)

Just before the extrapolation, the price hits a plateau, with one data point for 2009 suggesting it's back on "trend". Naam acknowledges this plateau, but makes a bet that it's not a plateau. Recent points even suggest the cost goes back up with the cost of natural gas and oil! The last data point for 2009 could be China fudging the true costs. Wouldn't you like to be more certain about the most recent data point before using an extrapolation?

Krugman is arguing for what Krugman is always arguing for as far as energy as concerned, that the price of fossil fuels are way too low because external costs are not included in the price. But the external costs of solar aren't included in the price either (like the cost of back up systems when the sun doesn't shine) so his conclusion should be that solar still isn't approaching competitiveness with fossil fuels.

Beware 'Krugman derangement syndrome'

If Paul Krugman says it, it must be wrong, and it must be leftist socialist and incorrect. Oh and he's too partisan. Remember that when 'serious' commentators speak about Paul Krugman. The guy who attacked Obama during the last campaign is 'too partisan'. The guy who annoyed the Clinton people to bits all through the 90s is 'too partisan'.

'Solar derangement syndrome' is a related disease.

Real Men (TM) do NOT believe in global warming. Nor do they believe in carbon free technologies.

The ONLY low carbon technology Real Men (TM) believe in is nuclear. And of course the hundreds of billions of subsidies the world civilian nuclear power technology has received does not count. Because it's nuclear.

So it boils down to 'I don't believe in Climate Change, carbon taxes etc. But even if I did, nuclear is the only solution.'

Real men build nuclear reactors. Leftist girly men build solar panels.

As they said in 2004. Anyone can go to Baghdad, real men go to Tehran.

"is there a bubble in the stock prices of solar power specialists?"

The solar stock bubble has already happened and popped. Evergreen went from $5/share in 2003 to $100/share in late 2007 to zero (bankruptcy) today. SunPower went from $25 to $130 to $9 today. FSLR went from $25/share in 2005 to $300/share in 2008 to $50/share today.
The broader global solar ETF (TAN) started out at $30 in 2008, fell to $5 during the recession, recovered to $10 in 2010, and is now down another 70% over the past few months.

What about Solyndra? I heard they got a nice government loan guarantee.

What is the future of solar power?


Partly cloudy.

I don't know if there is a reason to be optimistic about solar power, but I would be hesitant to state that fossil fuel prices took a steep tumble in the last few months. Natural gas--yes, which seems to be in response to fraking. But oil prices -- still the driver of the economy -- is back up to $94-95. I too hope that oil prices would come down, it has risen $15 is less than 1-2 months.

Perhaps it depends on what solar power is competing with. If it is competing with natural gas that is one thing, but if it is competing with oil that is another. These are two very different scenarios.

Solar doesn't compete with oil, since (except for a handful of places, mostly islands like Hawaii), oil is not used much at all in power generation.

Insolation (and solar output) typically peaks when the sun is highest, in the middle of the day, which coincides with peak power demand. In other words, solar power tends to displace power produced by natural-gas-fired peaker plants. These plants are typically the most expensive marginal-cost plants in the conventional generation mix (and so are up and to the right on the supply curve)...but still have a substantially lower marginal cost per kWh than solar. Wind, by contrast, tends to generate power in the morning and evening when the gas-fired plants aren't running, and so wind tends to displace power generated by (much dirtier) coal-fired plants.

Some would argue that you have to compare the inside-the-fence cost of solar versus the fully-delivered cost of peak power (including transportation and distribution, or "T&D" charges, which on average are an 8-10% add-on to the wholesale price of power). ***However***, some utilities pass on T&D charges as a fixed cost per meter, so a solar installation may not reduce the T&D charge to the end user on a per-kWh basis.

15-ish% conversion efficiency for production solar cells is currently state-of-the-art (recent First Solar average for CdTe cells is 12.4%, Suniva is selling crystalline silicon cells with 19% stated efficiency). Sharp recently announced a complex triple-junction cell that would have an efficiency of 36%, but that's for commercialization in 2014 and for use in very expensive applications (e.g. satellites).

All good points especially re wind v. solar and displacement.

Just on oil fired electricity generation, in a lot of the world oil *is* the key generation fuel: diesel powered generators are ubiquitous and in many countries there is not the infrastructure to import coal or a natural gas production and distribution infrastructure (only rich countries have LNG terminals).

The problem is of course one of storage: power demand peaks after sunset in Africa, not before.

Solar hot water is a big winner, because the alternatives are either charcoal which is increasingly expensive and ugly environmentally, or electric immersion heating. Certainly you see lots of SHW rigs in Botswana (if not in poorer places).

The "Moore's Law" reference is a red herring. Transistors are getting cheaper, not because silicon wafers are getting cheaper, but because the transistors are getting smaller. But with solar power, it is only the area that counts. Economic solar-electric power (in the temperate zone) needs two things: 1) solar panel costs to drop by a factor of three or so 2) cheap electricity storage. It is the latter that is the "unobtanium". The "retail" storage method of choice today is still the lead-acid battery; the "wholesale" storage method is pumping water uphill. Both low-tech, expensive, wasteful. But nothing better (cheaper) is in clear sight.

I think you're off base in calling pumped-hydro expensive. There are over 70 GW of large scale pumped hydro projects installed or in the planning stages. Most of the American dams were built decades ago, when there were no renewables on the grid at all.

It's expensive because there are 2 further conversion losses (at at least 10% each way), and that's assuming completely sunk capital costs which are perfectly co-located with your solar PV. Which of course they aren't in most cases. If you have to send the power to a dam, your conversion costs run closer to 40% each way, I think. This does not help cost-effectiveness.

It gets worse. Whilst there is enough hydro capacity to store the spare energy when renewables are a small part of the mix, once that % rises above 2-3% then you simply don't have the dams, pumps, and turbines to "store" the energy required.


Moore's Law has absolutely nothing to do with Solar power. The cost of solar is related to the efficiency of the cells and we are in the 20-30% range now. The most we could possibly see is a roughly 4 fold improvement in the process and that's almost certainly not possible.

There are lots of things you can do with good, cheap solar panels that don't need batteries: run air conditioners, for instance. Then you would have solar panels displacing the "peaking" plants of today, which tend to be on the dirty end of the spectrum. So that would be a dandy thing to have.

Solar doesn't have to become a universal energy solution; it just has to win one of many seats at the table of modern energy sources. And I suspect it's close to doing that: a decade or two.

Please, Craig, do a basic calculation for me -- tell me how many square meters of solar cell it takes to run a central a/c unit in the middle of the USA. Hint: once you figure out the number, you'll realize you are being silly.

I'll even let you ignore starting current.

Well starting current would have to be shared across the grid to be affordable. However, Solar does have a very useful niche in being a sunny day 'peak load' shaver. Hot day, peak power is expensive. It's quite possible we could see widespread, economical use of 'peak' solar power. Since, our current peak power matches with solar's peak power it could act to save a lot of natural gas for better use.

Wow: informed AND gracious! What a charmer you must be at cocktail parties.

It's not my field, so don't be shy about correcting the numbers on the back of my envelope here. It seems to me that my 2 1/2 ton air conditioner pulls about 3500 Watts. I also read that I'm supposed to average 700 Watts the square meter of solar radiation hitting my roof. So a magical 100% efficient solar array would need to be 5 m^2 to run the unit during sunlight. You tell me what you think is reasonable for 10 or 20 years from now. Forty percent? Twenty? That gives you...hang on...12.5 m^2 (135 sq. ft.) and 25 m^2 (270 sq. ft.), respectively. That's against the 1300 sq. ft. of my charming 1927 bungalow, so we'd be talking about this fitting comfortably on the southern slope of the roof.

Or I can make an even quicker and dirtier guess by noting that it's no big deal to see a 1 m^2 panel today that's supposed to give me 150 Watts at peak sunlight. So two dozen of them would go a certain distance, and again, we're only haggling over the price, ten or twenty years from now.

Ooh, someone got served.

I onced lived in a place where the utilities offered to put my ac unit on its own meter. On peak days the utility company would cut off the power for the AC for an hour or two. I would be refunded for the inconvenience. I think 20 bucks but I forget. So I could see utilities subsidizing rooftop solar cells to offset AC. One day. Maybe.

Craig: A bit of Googling suggests your sunlight/m^2 figure is a very low estimate for peak hours. A closer figure would be in the 1000 Watts+ range, if you're only looking at peak hours. (here's one source: http://acrim.com/TSI%20Monitoring.htm)

Of course, this means you can't run the AC at night (or if you could store energy during the day, the 1000 Watts, or even 700 Watts figure, would be too high, since some of that energy is being stored).

Careful about mixing up watts with kilowatts.

I really doubt anywhere on the Earth's surface gets 1000 watts per square meter of power at any time ever. There's only about 1366 hitting the upper atmosphere.

Dan--I'm not an expert, and I'm prepared to be educated. But, all due respect: "Over the course of a year the average solar radiation arriving at the top of the Earth's atmosphere at any point in time is roughly 1,366 watts per square meter... The Sun's rays are attenuated as they pass through the atmosphere, thus reducing the insolation at the Earth's surface to approximately 1,000 watts per square meter for a surface perpendicular to the Sun's rays at sea level on a clear day." (http://en.wikipedia.org/wiki/Insolation) I mean, that is a 27% reduction, even at local noon.

I also don't see where I got my prefixes mixed up, but if I missed something, please let me know.

Or again, to keep things practical: you can buy one meter square panels today that throw off 150W in full sun, no problem. This isn't Flash Gordon stuff or theoretical physics. The only question is what happens to the price in coming years.

Here's a question worth trying to answer: If we had solar cells that were 100% efficient (at turning the sunlight into electricity) and could be manufactured (exclusive of mechanical packaging) at zero cost, what portion of our energy would they end up supplying? Only when we know the answer to this can we meaningfully talk about "breakthroughs".

That's really the salient question, and the answer is probably "surprisingly little." Solar has the same problem as wind: unreliable power is not very useful, because people place a high premium on power being available 100% of the time.

If you need to keep a reliable backup system available 24/7, that usually more than offsets the cost savings from eliminating the fuel. I doubt the wind/solar equation will change until the annual average oil and nat. gas prices have more than doubled from current levels.

You don't need power available 100% of the time. Obviously, solar coincides with daytime; in places like California, Arizona and Nevada, this is largely predictable, even in the case of partly cloudy skies. Peak solar can coincide with peak cooling hours, which is a benefit. (By contrast, wind has the unfortunate propensity to blow more at night, when demand is low). But it is fair to say that almost all renewables (except perhaps hydro) are subject to inherent variability, whether wind speed, sunshine, tides or currents; and this requires either loading balancing from either additional power generation (generally gas-fired plants) or demand reduction (ie, the "smart grid").

In terms of power prices, per kWh, (levelized cost unless otherwise noted):

- onshore wind: 10-12 cents
- offshore wind: 20 cents
- solar thermal: 20 cents
- solar PV: 20-27 cents, recent numbers quoted at 17 cents.
- combined cycle gas: 6-7 cents

- ERCOT (Texas) spot power market, Nov. 8th: 3.4 cents / kWh

Peak solar can coincide with peak cooling hours,f

That's great until you want power on a hot, cloudy day.

drop in insolation is less than 50%. Even when it's raining it's only c. 50-60% I believe. Does depend on which type of cells you use.

50% peak-time reliability is better than zero, but it still means you need twice as much as a reliable supply would need, unless you're willing to accept not having enough power sometimes.

But as Joe Romm (http://thinkprogress.org/romm/2011/11/07/362705/krugman-solar-power/) points out, it's the user price that matters. My marginal rate last month (California Coastal rate from Pacific Gas and Electric) is 34 cents per kWh for the first 17 kWh, then 30 cents for the next 173 kWh. Both of these make plenty of economic sense now. Many people are paying high marginal costs from their utility, so rooftop solar doesn't need to compete with the ERCOT spot market or the cost of coal at the power station.

Just as TallDave continues to assert that a power source that isn't 100%-on-demand is worthless, I continue to assert that variable power sources have value.

Its value depends on what other things are supplying your grid and what things are drawing from your grid. If your power grid supplies are heavily solar already, and the demand is an industrial factory that cannot be interrupted, then the marginal solar unit is not that valuable. If your grid supplies aren't a lot of solar but have a lot of hydro, and your demand includes things that can be interrupted like battery charging and AC units, then the marginal solar unit is pretty valuable.

Economics will answer most of these questions. My power company gives me a discount in exchange for being able to disable my compressor remotely.

Again, that's great if you don't mind being hot on a cloudy, windless day. Most people very adamantly do not want that.

FUD notwithstanding, that's not even close to being required.

Hey, it was your idea.

Drat, I forgot to stipulate that the storage problem had been solved (actually we know an international DC grid solves this problem)...

We could get all of our energy from solar/wind if we had unlimited, perfectly efficient batteries. Energy storage is the main thing holding back these renewables from replacing base load sources, and we're pretty far off from being able to provide energy storage on the scales needed (see: http://physics.ucsd.edu/do-the-math/2011/08/nation-sized-battery/).

A good question. And one factor that always gets ignored is that the Cost of solar cells is roughly half of the installed cost of solar power.

Ergo, even if solar cells were 'free' the cost of solar power would not be. You still need power inverters, a scaffolding structure, wiring and labor to put it up. So solar will not lead to 'free' energy, though it will probably establish an energy cost upper limit for day time peaking power.

How many square miles of land are you willing to lose? The energy density is pitifully low. See upthread.

The 48 states get hit by about 40,000 TWh of sunlight every day. The US uses 80 TWh of energy (all forms) each day.

Oh, and the reason there's no bubble in companies that provide "cheap and reliable storage"? Those firms aren't public or don't exist yet. Not to mention the fact that the grid storage market won't be worth anything for several years yet. You need a much higher percentage of intermittent resources on the grid than what we have currently before much storage is needed.

A lot of electricity markets see rapidly fluctuating prices. Just look at Texas, or even PJM.

Storage can be highly profitably even in the absence of renewables. Imagine you're a nuclear or coal plant operator who sells for low prices at night and high prices during the day. Now image that instead you can charge a battery at night to sell that electricity during the day when prices are higher. An efficient battery would make you a ton of money.

If an efficient battery existed, I'm sure power companies would be making a ton of money. Since they aren't.........you figure it out. Our local nuclear power plant does store energy at night. It pumps water uphill into a lake.

If these advances are more than 20 years out at market discount rates we wouldn't expect them to have a big effect on valuations of risky assets today.

Also, it could be that the players who will profit from this cheap energy are difficult to identify. If solar panel production has low barriers to entry over the long run then even the panel producers need not benefit.

Comparing IC production and solar cost efficiency strikes me as a bizarre comparison. Moore's law for ICs states that the number of transistors on a lowest-cost IC will double every 18 months. This is due to mostly to lithography minimum feature size decreases, and also to lots of magic that semiconductor wizards, err, engineers, perform. To the first order, this would predict doubling the performance of such a device, while cost should remain the same. Both of these aren't quite the case today but usually it's close enough.

Solar, as I understand it, has constrained efficiency. That means for a single piece of silicon solar panel, the efficiency is roughly proportional to area, and that this efficiency hasn't changed much in the last few years. The efficiency has nothing to do with feature size; increases in *cost* efficiency are motivated by using cheaper materials/manufacturing methods and more clever setups (like using mirrors to focus solar radiation). I have a feeling that while the progression might follow an exponential curve, in reality it has nothing to do with increases in IC integration and is subject to a set of completely non-related variables.

Cliff's: Just because a series looks exponential, doesn't mean it follows, or has anything to do, with Moore's law.

Agreed. I'll add that Moore's law is also about engineers managing the complexity of integrated circuits and their fabrication. Photovoltaic price declines are driven by different factors, especially efficiency of raw material usage and the like. Material usage matters almost not at all in the price of integrated circuits.

This is why I think Keystone XL is great. Shale oil is not cheap oil, just through for land and energy reasons. What is is less risky oil. Rather than making bets on buried deposits, you have stuff very near the surface which you know is there in large swaths of land.

As oil embraces risk-free but expensive oil, risky but potentially cheap oil (in say, Africa and elsewhere) becomes less desirable.

This embrace keeps oil prices well above $80 a barrel for the long term, which makes it easier for Solar Power to gain a foothold.

What does oil have to do with solar?

>What does oil have to do with solar?

They both fill liberals with such intense emotion that when they speak of either, it is very high comedy.

But other than that, nothing whatsoever. Now, back to the show!

"Is there any reason, based in industry-wide market prices, to be optimistic about the near-term or even medium-term future of solar power? I don’t see it."

I would believe this insight more if you could point to where current market prices showed the coming ubiquity of the IC. (Not just desktop computers.)

Power is relatively cheap right now (I do not even think about electricity costs ... data centers do, but it counts there), so it seems it would be a slow transition. The difference between getting solar or using fossil fuels in the long run will be a minor choice for your home compared to e.g. location.

If they could become cheap enough for the average Joe to demonstrate status, you might get a Prius effect. Cost differences are a minor component compared to lifetime costs of owning a Prius, but the status you demonstrate is key. Remember, the Prius started with Hollywood celebrity stories ... such a thing is hard to price in in advance.

I think Krugman should examine Krugman's 7% yearly price drop claim. This seems like a false trend or at least an unsustainable trend. When huge swaths of an industry start to go bankrupt it is very common to see dramatic price drops for their products. That has nothing to do with technology or efficiency improvements. Its also not going to last. Invoking a Moore's law equivalent seems stupid here (in light of the other insightful comments here as well).

By the way, the future for solar is in space. If we can roll out a huge solar panel away from the earth, one, it can receive sunlight 24/7, two, it can keep its incidence to the sun at a constant perpendicular angle (not possible here on earth), and, three, it would receive sunlight up to 8X stronger than that received on Earth's surface. The energy would just have to be beamed down to earth. We just have to do a Manhattan project "Summon Technology" spell to come up with a good energy beaming tech (along with some of the solar-space tech).

Energy beaming tech is not the problem, launch cost is the problem. And the minimum necessary scale is a problem - the smallest practical sizes are quite large for antenna focusing reasons.

Not to mention, the who do you trust to control large directed energy beam's floating above your head issue. ;)

I know that the designs all call for a diffuse output, but that's a tuning issue and can be modified.

Actually it can't. The low power density of a "standard" solar power satellite is an intrinsic consequence of the diffraction limit as applied to the size of the antenna and the wavelength of the microwaves, hard-coded into the laws of physics. Microwaves are inherently fuzzy. Even if the on-site engineering staff were Mad Scientists in the nominal employ of the Pentagon but secretly working for Al Qaeda, there would be no way to "modify" or "retune" it into a Death Ray. An Unseasonabe Warmth Ray, perhaps, but that doesn't quite have the same cachet.

Finch is correct, though, that the scailing issues are a problem. If you need electricity in ten-gigawatt chunks, lots of them, and you have a trillion or so dollars to spend on infrastructure development, solar power satellites may be the right answer. Nobody has found a credible way to make it pay on any significantly smaller scale.

If you're talking about the recent price trend, the drop is more like 40% in a year. That's not sustainable. However, the 7% trend goes back pretty far in time, especially if you look at the supplier's costs rather than the market price (to control for the significant variation in margins).

One indicator that might show it would be the flexibility of new thermal plant investments. Flexible thermal plants (for example gas peaking plants) are less efficient than baseload generation. The expected price of flexibility, which increases in the expected supply of unreliable renewable energy, must be offset against the opportunity cost of greater efficiency.

These investments are forward looking so would be a good indicator of market expectations.

Unfortunately it would probably not be very easy for an outside observer to "solve for expected price of flexibility".

Some have attributed the recent tumble in fossil fuel prices to Qadafi's ouster, and the resumption of oil flows from Libya.

Is there any market price which would foreshadow for us what the market believes will be viability of solar in the near term? Or would there have to be an Intrade market to tell us? And what would the contract say? Something about solar becoming a part of the baseload of electricity providers? Something about the price of electricity falling? Something about battery cars becoming competitive with the internal combustion engine car?

What industries would become the buggy whip industries if there is a revolution in solar, and how are their stock prices doing?

Shorter Krugman: Solar power has been in development for decades. Within a few years of further public support it will be cost competitive. Fracking is not perfectly safe today. Therefore, it will never be safe and we should ban it.

Shows he is a political hack

Shows you're a political hack

Shows your mom is a political hack!

Where did he say they should ban fracking?

He didn't actually call for banning. He did say "that it produces toxic (and radioactive) wastewater that contaminates drinking water", ignoring the fact that the studies that found drinking water contamination generally found that the drinking water was contaminated in the area 'before' any fracking took place.
He also said that "Yet what the industry and its defenders demand is, of course, precisely that it be let off the hook for the damage it causes. Why? Because we need that energy!". That seems to be a straw man argument.

It's disingenuous and it's hard to believe that a Nobel Prize award winner accidentally made those kind of mistakes and he frequently makes 'mistakes' of that type. That's why I consider him to be a political hack. He's way to smart to say the stuff he says frequently unless it's designed for the consumption of the gullible.

The industry could have nipped a lot of the backlash in the bud a year ago by just disclosing what chemicals they use in their fracking fluid. They set themselves up for (possibly) irrational environmental protests by trying to be as secretive in the Northeast as they are allowed to be in Texas and the west.

Nope! that would have sold their IP down the river.

Note for the non-hack above: if you're going to point at "studies," it is necessary to be more specific, perhaps even deign to name them. And also, if you are going to claim that these studies counter certain claims, you might as well explain why this is the case (i.e. the fact that contaminated water was generally contaminated before fracking is transparently irrelevant to the matter at hand.)

It's funny people love to criticize Paul Krugman and call him a hack and yet most of the time when I read these criticism it's REALLY clear the critic either didn't read the article or has a severe reading comprehension problem. I'm not saying Krugman is always right, but he deserves better critics.

You must be thinking of some other Krugman. This one is definitely a hack -- clever, did some great econ work, but the hackiest hacking hack of hack columnists. His own ombudsman attacked his hackery.

Which isn't to say, I'll add, that he's not an effective advocate. He's the left's Ann Coulter -- neither has ever met an argument that was too tendentious for them.

that would be Ann Coulter with the Nobel Prize in economics?

Who was right about Iraq? The housing crisis?

Who was right about government borrowing crowding out private lending?
Who was right about the non-existence of the "bond vigilantes" and about predicting low inflation rates?

He did not win a Nobel prize for his hack columns.

Optimism: We will see widespread adoption of local solar power in Los Angeles and Phoenix ~2015.

Solar will become cost- competitive with on-the-grid _peak_ prices somewhere in the middle part of this decade in the sunny parts of the American southwest. That is, it will be be cheaper for a homeowner to install solar cells on the roof to run than to rely on the power company to run the air conditioner in the summer.

This transition will require no additional infrastructure, no government subsidies, no new power transmission lines, and NO ENERGY STORAGE because the peak power is already matched up with the peak demand at the point of consumption.

Since approximately 5 million homes will be candidates, this solar panel market will explode at this point, further driving down the cost of installation.

When the prices get close enough it will only take one “rolling brownout” of residential air conditioning to drive adoption of the residential solar power technology.

A major focus in solar energy is now the creation of cheap solar cells made out of common materials The dream of Professor Harry Atwater at Caltech, for example, is all silicon solar cells which can be rolled out onto the desert in New Mexico. There's also a big boom in people attempting to make hydrogen out of solar, under the premise that the hydrogen could be transported. See the 122 million dollar JCAP project (http://solarfuelshub.org/).

So I assume this is in response to Krugman's column today. But from what I read Krugman is not arguing for any sort of special government funding for solar power, he is pointing out a variety of negative externalities generated by fossil fuel extraction that are not internalized by extractors and rightfully saying that this amounts to a subsidy of fossil fuel production.

This gets to the heart of why I have trouble taking "free marketers" like Tyler Cowen too seriously, when it comes to things like massive negative externalities very very little digital ink is spilled about how to address them, instead they are sort of brushed under the rug. Whether or not solar power will be successful is not an argument for continuing de-facto subsidies for fossil fuel producers. Tyler brings up job creation in fossil fuel as a reason to live things as they are, but why can't we argue for subsidies in other areas for the purpose of job creation?

The only numbers in Krugman's piece refer to the bad analogy with Moore's Law.

(Not counting "Economics 101")

So what it's a space-limited NYT column not a peer-reviewed paper. He does bring up a list of externalities and as per usual Tyler ignores that completely.

Because he says solar is cost-competitive if the externalities are quantified. So, quantify them.

Maybe he'll put up a blog post, I'm not here to defend everything Krugman writes but he IS writing a column with the limitations thereof so I think your point is unreasonable.

of course if you ignore the massive externalities that solar creates as well.

At best guess, carbon should be $100/tonne (say $28/ tonnne of CO2) rising steadily to around $300/tonne ($100) by, say, 2030. See Stern Review. Australia's new carbon tax is about AUD18/ tonne CO2, I believe. European carbon market around 15 Euros.

That's your externality. Even in rigged markets where too many permits were issued.

Each kilowatt hour of coal fired electricity generates roughly 1 tonne CO2. Each kwhr gas fired, roughly 0.5 tonnes.

I want to know why do environmental or infrastructure costs get totally ignored? But if it's something Tyler doesn't like then he'll do a blog post bringing up all these issues.

What are the externalities of solar panels?

And how much of the fossil-fuel externalities are offset by the current solar subsidies, that is if you want a straight-up head-to-head competition.

I don't know, what am I the solar panel expert?
If there were though Tyler Cowen would be ALL over them on this blog. I'm saying a major point of Krugman's article was that there are unpriced externalities from fossil fuel extraction and I'm asking why Tyler thinks they shouldn't be internalized or why he doesn't even bother to mention them. I mean I KNOW the real answer why he doesn't mention them but I just want to bring this point up.

No, you are talking about something completely different from what TC is talking about, while Krugman doesn't even talk about what he's talking about.

Here's a pretty good rule of thumb for you. When Krugman describes an orange he says "ORANGE!!!" When Tyler describes an orange he talks about juice, pulp, seeds, rind and Krugman supporters say "You didn't say orange!"

I've been reading Marginal Revolution long enough to know that when Tyler only talks about oranges if he likes them and if not he'll just discuss apples.

With all due respect, Tyler is just as likely to link to a story about problems at a particular orange farm and add a few sentences which imply that perhaps we should abandon the field of agriculture entirely.

Well we do know CBBB isn't much of a numbers guy.

Today, all energy sources (coal, wind, solar) compete against the marginal price of natural gas. In the wholesale market (what Electricity Distributors pay to Generators), the conversation stops there. Until the solar technology is cheaper than gas, we will not see much wholesale (read: big, utility-scale 50+MW) projects *without* subsidies. For this reason, the solar industry sees the future of large-scale projects to be fairly limited (20 years or less)

At the retail level, however, there is a compelling argument to be made in favor of disruption. When you look at your electricity bill, you will see Generation charges and Transmission/Distribution charges. When the electricity company buys power in the wholesale market, it must transport those electrons to your business or home. In the coastal regions of the US, it's not uncommon to see those rates equal one another - around $0.07 per kWh for each category. In this context, solar is not solely competing against gas (generation). On the contrary, it is competing against the Transmission and Distribution costs. As our demand for electricity increase and those systems reach capacity, prices escalate, independent of fuel costs. Transmission projects are notoriously expensive projects that take years to develop, suggesting that solar could be a viable competitor because one can deploy it rapidly.

A consumer of distributed generation (contrast with wholesale generation) who has a solar panel at their location avoids the Generation and Transmission/Distribution rate. In a state with expensive electricity rates (CA, NY, NJ, CT, MA), the payback period for a residential system is around 8 years. There are certain zip codes in the US whose current retail electricity rates are equal to the rate implicit in owning a solar PV system... without subsidies.

As the trend in retail electricity prices tends to increase over time, and because technology costs related to solar PV tend to decrease over time, it is reasonable forecast a point at which a significant portion of retail electricity buyers would prefer solar PV at their location. The industry view is that this point is within a decade.

And I'd love to have solar panels, and I will likely have them before the average hippie. So, the strawmanning is too damn high!

If you're receiving AC, you do not even get the electrons, they only wiggle your own electrons around a bit.

I live in the south, and ran the numbers on a solar system earlier this year. Even with the recent rate hikes, and counting in all the federal, state, and local incentives, the ROI on the system is... not good. Break-even is a good 12-15 years out. Combined with the large up-front costs (some of those incentives are rebates) it's just not worth it, and won't be for quite some time unless there is some kind of packaging revolution like you mention.

I am rather curious as to why solar thermal is being so ignored in these conversations, though. It's a much better candidate for wholesale power generation that photo-voltaics are.

And the US has some of the lowest retail electricity prices in the developed world (taking the US in the round, California is different).

OK Canada and Australia are lower. But, guys, there's a whole world out there where electricity costs more to the retail customer.

PV is not the only solar:


Solar open-loop A/C using desiccants

Solar closed-loop absorption and adsorption cooling

Raw materials costs are decreasing rapidly. Chinese polysilicon manufacturers are now on the verge of cutting production because of low local prices, an indication that transportation costs are beginning to count for relatively more of the costs of a panel.

Different players with like design technologies (e.g. thin film, polysilicon, etc.) are in the process of consolidation as prices are driven down.

Solyndra used monocrystalline silicon with technology that did not figure into anyone's consolidation plans. On the other hand, players with somewhat alike, commoditized, mass-produced designs have been merging and consolidating like crazy.

To echo the comments of Solar Industry Worker above, home-based generation is already viable in high-electricity rate areas. In the greater NYC area, under half of a typical electricity bill is the actual cost of generation. Even excluding fixed costs, such as the NYS system benefits charge, transmission costs nearly double the size of your typical bill compared to receiving power "over the fence." In addition, up-front costs are increasingly being somewhat mitigated by solar panel leasing models, which one company recently expanded into New Jersey with some success.

The pricing point of electricity at the power plant is, as the commenter above notes, highly dependent on natural gas, which in some corners is not seen to bode well for solar power. On the other hand, with the conversion of LNG import terminals to export, barring government intervention, we will almost certainly see a large increase in the price of natural gas from current, historically low levels. Add on top of that some operators of public vehicles considering investment in natural gas as a transportation fuel, and the pricing point for solar doesn't look half bad.

-----> On the other hand, with the conversion of LNG import terminals to export, barring government intervention, we will almost certainly see a large increase in the price of natural gas from current, historically low levels. Add on top of that some operators of public vehicles considering investment in natural gas as a transportation fuel, and the pricing point for solar doesn’t look half bad.

Your conclusion is precisely the opposite of the correct one. The reason people are considering export terminals for LNG and conversion of vehicles to LNG/CNG is the expectation of sustained cheap natural gas.

Solar companies would pretty much disappear tomorrow if the huge goverment subsidies were taken away. In Hawaii about 65% of the cost is covered by rebates and tax credits

And the Solyndra thing indicates that the technology is changing so rapidly that it's too early to pick winners, the whole point of the statement what Tyler links to.

My understanding is that the problem is that the sun produces very little solar energy per square metre of the earth's surface. Therefore, there can be no equivalent of Moore's law ... just an asymptote of 100% of the sun's energy captured and converted. And even 100% isn't very much energy ... you'd need to cover over a large part of the United States to make a dent in our energy consumption.

Perhaps someone who knows the details could fill us in.

The company I work for mainly builds roof top systems for commercial properies. Kohl's, Staples, Wholefoods, that sort of thing. Big empty roofs doing absolutely nothing, plenty of square meteres to offset the peak load of the facility and more. At some sites we track the net usage and can compare what they are pulilng from the grid to what the panels are putting out. In California you can offset the full load of the building pretty routinely, doesn't work quite so well in New Jersey. Also 10 - 30 MW sized plants are getting build all over the south west. Whatever the debate about the physics they are getting built and cranking away.

I think this Moore's law business for solar is a bit of a stretch. Panel prices are crashing because the Chinese are 'picking winners' as is the popular turn of phrase in this thread. We might as well take advantage of it before they crush all the competition and get back to monopoly pricing. Even with that there is a lot of scope to drop costs besides panels. Micro inverters, better and lighter racking equipment, building the racking into the roof at the time of construction. Low hanging fruit abounds.

that 'large portion of the USA' would be 2-3% of surface area or around what is covered by parking lots, roads and buildings now.

In other words, a south facing wall or parking lot becomes a place to put a solar panel.

I think the right question to ask is this: What should the government subsidize in solar power so as to cause the earliest and steepest replacement of fossil fuels? So far the answer has been to subsidize the installation of technology that's far from market-competitive. But this has only been producing larger volumes of such inadequate technology.

Instead, if the government aggressively funded research, the eggheads would more quickly produce solar technology which would make economic sense. And it's only when that happens that we will experience an inflection point. Present solar panels should be seen as primitive predecessors of what will eventually be worth building at scale. We want to get to that eventually ASAP, but for that, we have to put our foot on the gas with research funding - instead of pretending that what we have now is already worth building for normal applications.

The challenge with solar energy now is purely engineering and manufacturing, more expensive government funded research is definitely not required. Moore's law is really a specific example of the experience curve effect, where simply put, the more you manufacture an item, the more productive you get at manufacturing the particular item. This is a very well known effect, and continues even when the product is very mature (we never see plateauing with the experience cost curve). As lower prices spur increased demand, which lowers costs and so on we get a virtuous circle resulting in rapid reductions in costs. We have seen this with flat screen LCD screens, where very quickly a very mature technology (CRT) has been replaced very quickly. Note that reductions with experience cost curve are a function of the number of items manufactured, not the period of time, so if demand doubles quickly then costs quickly come down. Solar energy should be an almost perfect example of this since current installed base is very low, and the market is very large, so we can have many doubling's in demand in a short period before the market becomes saturated. So I am quite confident that solar energy will continue to lower costs substantially for many years to come. Since PV costs are already near competitive this means we are very likely to achieve parity with NG power in many areas soon.

How to profit from this insight? PV companies are not necessarily a good investment even if demand is rapidly growing, since the market is very competitive. As others have noted, A/C is almost perfectly suited to solar power. The exodus to the sunbelt will continue and accelerate with lower AC costs; the issue of solar power not being a good match for temperate regions will be solved by people moving to sunbelts - which is happening already. Desalination is another technology that is also very suited to solar energy (maybe a mix of concentrator and PV), so water supplies for these areas will no longer be a constraint. Maybe therefore land near to big sunbelt cities might be a good way to profit?

Why don't we see markets taking account of this process - i.e. PV prices falling in anticipation of the future price falls? We may already be seeing this, where people are deferring purchasing of solar systems, even when they would be profitable now, to take advantage of expected future price discounts. This may be driving the price lower for solar energy systems in the short term. But there are still enough buyers to drive the virtuous demand cycle where waiting would not make sense on a discounted cash flow basis.

On subsidies - perhaps the massive subsidies given by governments have kicked started the cost curve process. I myself am skeptical, at most it could have accelerated by just a few years since there was already a commercial PV market for remote locations, I suspect the money spent could have improved human welfare a lot more in other areas. I would eliminate them therefore.

For whatever it's worth: many solar companies have large outstanding short interests relative to there market capitalization. They are a pretty popular short with "smart money" types.

In Australia we have high insolation and we have electricity prices that are considerably higher than the US average. Solar PV will continue to expand here because we are approaching the point where a company with flat roof space can save money by installing solar panels. If the company pays spot prices for its electricity that point should reached sooner, as the periods of highest electricity prices are highly correlated with intense sunshine here.

I don't get it. Why aren't solar companies flocking to the hot, dry Australian outback and installing cheap Chinese solar panels so that you all can bask in your glorious green energy consumption?

Because batteries suck balls and PV panels are too expensive.

Name Redacted, here in Australia the vast majority of PV installations don't use batteries. This is because there is generally a diesel generator already on site for off grid applications and locations that are on the grid are on the grid and so don't need energy storage. As for solar panels being too expensive, they are much cheaper than running a diesel generator and have now declined in price so much they could potentially save money for some grid connected organisations with low installation costs and low discount rates. (And I'll mention that in addition to saving money on grid provided electricity, rooftop solar panels can also result in significant reductions in air conditioning costs.)

Anti-Gnostic, it is now normal for farms and isolated communities in Australia to have solar panels.

Or, you could make electricity more cheaply using other technologies.

Ronald Brak

What's the average retail electricity price in Oz? (I can convert out of AUD ;-))

I had thought, looking at your electricity consumption per capita, big houses, lots of AC, that it would be quite low? I realize your heavy industrial base distorts that, but conversely Oz is a modern, suburbanized, service economy-- tends towards high domestic load.

Huh. I guess 'green' energy ain't so green on a couple of levels.

What is the future of electricity, or energy itself? What about the future of food production? What would happen if food, worldwide, became "free"? Be careful what you wish for, or you may have Utopia on your hands.

Tyler, as the author of the Scientific American post that Krugman cites, and as a reader of your blog, I thought I'd add a few thoughts.

Solar PV will primarily compete with coal and natural gas, and not with oil. Oil is not extremely fungible with coal and natural gas at this point, so we shouldn't see a particularly large impact on oil prices.

The aggregate impact of solar on electricity markets will lag the point at which solar drops below the price of coal by many years. The cost of electricity from solar is dominated by up-front capital investments which are amortized over at least 20 years (and often 30). That means the solar build-out will not be instantaneous.

Indeed, a simple model (to be found in my forthcoming book next year) shows that if we spent 0.5% of global GDP on building out solar capacity each year, and if solar PV prices keep dropping by 7% per year, it will still take until roughly 2050 until solar could replace all coal and natural gas for electricity generation. (Obviously the future won't be that simple. There will be a role for a great many different energy sources, including wind, hydro, nuclear, and probably some continued use of fossil fuels for many decades to come.)

My point here is that the impact on energy prices will be more closely linked to the amount of solar deployed than it will be to the price of new solar installations. The former will lag the latter by quite some time.

We can see that today. Solar PV provides 0.1 - 0.2% of the world's energy today. Its total contribution is tiny. Yet it's also growing at an impressive pace of around 40% per year, from just over 1 gigawatt of installed capacity worldwide in 2000, to around 40 gigawatts of installed capacity in 2010.

That trend rate, interestingly enough, mirrors the 30-35% annual growth rate of mobile phones worldwide over the last 20 years. And in the developing country in particular, mobile phones succeeded in part because deploying them was far less costly than deploying land lines.

So I would turn the question back to you: In 1990, were there price signals in traditional telephone service that would have led us to believe that mobile phones would relatively soon surpass land lines?

Ramez Naam

Why would we assume solar PV costs will continue dropping 7% year?

Even if they did, why would we spend large amounts of money installing solar before it was cost competitive?

The fact that solar has been growing is entirely a function of government's directing money to expensive outcomes the market would not produce. It is evidence of nothing but the willingness of government to spend money unwisely. With the same $billions spent on solar PV they could have induced huge growth in power generated by hamsters running on treadmills. And no doubt huge cost reductuions in the hamster-on-treadmill technology.

There's no guarantee that solar PV costs will keep dropping at this pace, of course. Indeed, in 2010 and 2011 the pace changed - for the better. If I were to redraw that graph today, the trend since 1980 would be 8% annualized improvement.

The reductions in solar cost, at this point, come down to two things:

1) Technical and process innovations in the design and manufacturing of solar cells. Here, the headroom is quite high. Roadmaps to another halving of prices exist, and reaching that price point doesn't appear to require any dramatic new innovation.

2) Expectation of competitive pressure. Every solar maker out there knows these numbers, and so in their own product goals, they use them to create their own targets for their next generations of products.

Its clear that government price support for solar has increased volume and competition, both of which have likely led to faster innovation in bringing prices down. At this point, the industry has matured to the point that the need for additional price supports is rapidly dropping.

Why would we spend money on solar now? There are several reasons:

1) Off grid applications.

2) A hedge against future coal and natural gas costs.

3) A hedge against future carbon taxes.

4) Regions that are especially high in sunlight (making solar effectively cheaper).

Those are all reasons for private entities to install solar. From a societal perspective, there's a large additional motivator:

5) Reducing the cumulative externalities from coal and natural gas burning.

Ramez Naam

Let's be honest about a couple things:

1) Solar PV is a significantly uncompetitive technology today on a cost per delivered kWh basis
2) There is no reason to assume it will ever be more competitive than a) fossil fuels or b) other renewables (with which it it also significantly uncompetitive today.)

You may hold hope that some day this will change, but today hope is only hope.

If your argument rests on hope that someday something may happen in the future, you should not act as it it has already happened today. In some conditional future it may make sense to spend heavily on the production of solar panels. It does not make sense to spend heavily on the production of solar panels today. Even in your hope-based argument, you at best have a case to spend more money on researching solar PV with a goal of making it economic, not to spend billions subsidizing the deployment of current uneconomic solar PV today.

Second, I am willing to recognize that the "externalities from natural gas burning" is a legitimate argument, but what do you imagine they are? I am asking for a number here. A specific $/Mcf. It is not sufficicent to say "there are externalities to natural gas burning, therefore we should subsidize uneconomic solar". It must be demonstrated that the externalities are of a relevant magnitude before such conclusions can be drawn. And even then, we are left to question "Why solar PV?". It is not even close to competitive with wind.

An economist might accept that if externalities exist with natural gas, a Pigovian tax ought to be applied to natural gas to address those externalities associated with its use. If set correctly, that tax would lead to a more efficient fuel mix, conservation, etc. I do not believe it would lead to any more use of Solar PV. No agnostic view of energy does today.

If you can demonstrate otherwise by reasonable calculations, please do.

Otherwise, we are arguing over matters of faith.

There are no guarantees on future rates of innovation, yet the trend line and future prospects for solar PV look quite good. The trend line goes back to 1954, when solar PV cost $1000 / watt. End of 2011 costs look to be around $1.10 or $1.20 / watt. Should that 57-year-long trend continue for another 5 years, solar PV, without subsidies, will be cheaper than coal and natural gas the American South and West. Should the historical trend continue for another 5 years beyond that, then, without subsidies, solar PV will be cheaper than coal and natural as throughout the bulk of the nation.

As for the externalities of coal and natural gas, Epstein et al estimate the externalities of coal electric in the US at around 18 cents per kwh (with a range of 9 cents - 27 cents per kwh), an amount which, if factored into prices, would make coal electric more expensive than solar PV electric today.

Reference: http://solar.gwu.edu/index_files/Resources_files/epstein_full%20cost%20of%20coal.pdf

I have never seen a comprehensive accounting of the negative externalities of natural gas. They're almost surely lower than those of coal, but it would not surprise me to see them around 10 cents per kwh, which would also make natural gas electric more expensive than solar PV electric today.

Ramez Naam


'expensive outcomes the market would not produce'

Quite. Like fossil fuels. Fossil fuel production and consumption, globally, attracts c. 5 times the subsidies that renewable energies do.

Fortunately this was all ignored for the transistor. The space programme needed microelectronics and so NASA specified them, and from that US companies learned how to make better and cheaper transistors.

And so the world of today is what we now see. Computers. Mobile phones. the Internet. blogging even.

If solar is so great, why isn't solar thermal used to heat water more often? I moved to Florida last year and expected to see more solar panels than I have. I hardly see any (in the Sunshine State!), and I don't even see solar thermal installations. Everyone just has the normal water heater, and it's usually electric (not too many homes here have gas hookups - the infrastructure has not kept pace with the suburban sprawl).

Maybe solar thermal is a no brainer but no one uses it out out habit. But I'd like to see the numbers.

Because you can't just buy it and plug it in. Seriously.

Well in Southern China it is c. 1/3rd of all households. other countries like Greece as well.

US has:

- cheap and widely available natural gas domestically produced. Domestic price of US gas is 1/2 to 1/3rd price in other countries in Western Europe etc.

- cheap domestic electricity where gas is not used for water heating (as above). Propane and fuel oil prices are similarly lowered by that competition (but are fundamentally driven by oil prices).

- (US also has lots of wood heat)

UK estimate for hot water for 2 people is 3500 kwhr pa (3466 kwhr = 1 BTU). Add 1500 kwhr pa for each extra person in home.

That's only a 5th of the average UK gas consumption for heating per household, and our winter temperature seldom drops below 32F/ 0 C. Average electricity consumption is 4500 kwhr pa and US number is c. 12000 (at least half of that is air conditioning).

So at our gas price, retail, of say 4.5 pence (6 cents)/ kwhr, that would be c USD200 saving pa if a solar hot water heater did *all* our hot water. And it does not-- 100% for maybe 5 months of the year on a 2 m2 panel, say 50% over the whole year.

So USD100 pa saving (actual estimates are around £180 pa).

Cost of installation around £3000. So 15 year payback, 4.5% accounting rate of return.

Since estimates of discount rates consumers apply to energy saving technologies range 15-30% pa, you can see why solar water heating, even in a place with twice the gas rates you have, just does not fly.

I should add.

Most households are capital constrained.

Spending $6k say putting in a solar hot water heater, borrowed at 5-8% on a home equity loan, when you are going to sell the house within 10 years is not going to work.

The really imaginative schemes (the Green Deal announced in the UK but not yet implemented) will allow homeowners to borrow against future savings, shared with the utility company.

However humans are herd animals. When 10%-30% of homeowners are doing it, there will be a 'tipping point' and suddenly 80% (those with south facing, unshaded roofs) will be doing it.

"will allow homeowners to borrow against future savings, shared with the utility company."

And we can bundle them into securities and sell them. What could possibly go wrong.

Great site you have here.. Solar, Wind, basically green renewable energies, I'm so glad we have these 2 options and much more to come..http://energyreviewsinfo.com

I don't think the future of solar electric is in base-load power generation, except in a maybe a few areas that are very sunny and hot with large population centers nearby and a mix of power generation methods that allow solar to work its way into the mix. In any event, unless solar electric gets much cheaper, solar thermal will probably be the desired form of large scale solar power generation.

I do think solar has a big future in embedded applications. Think solar powered calculators, only on a larger scale. What's needed is not just a low cell price, but a revolution in solar cell packaging that makes it inexpensive and easy to apply to wide areas. For example, if you could make a cheap composite material that has a tough surface that doubles as a solar panel, you might see it being used to construct car bodies or the trailers of semi trucks, or solar shingles that cost little more than than regular shingles but which can turn a roof into a solar collector. If you can make a thin-film semi-transparent material cheap enough to embed in window glass, you can turn high rises into large solar collectors and use the energy to power the air conditioning during the day.

Even if these cells are relatively inefficient, if the cost per watt-hour is low enough, you could see a material like that make rapid inroads into the market. There is already a big market for off-grid solar power - RV's, camping gear, solar powered attic fans, solar powered landscape lighting, solar powered traffic lights, etc. Make solar an order of magnitude cheaper and eliminate the cost of mechanical support, and solar cells could become ubiquitous, feeding power to our devices and batteries whenever possible.

This would not only lower our consumption of other energy sources, but it would enable entirely new products we haven't even thought of today. The use of such solar panels might eventually reduce our reliance on grid energy by a few percentage points, but they'd also make life more convenient.

But if we're stuck with traditional solar panels and the extensive metal scaffolding and other support hardware they need, it will never be cheap enough, convenient enough, or safe enough.

Safety is actually a big issue. Falls from roofs kill a lot of people each year. Now imagine a country full of solar panels on rooftops - panels that get dirty, collect leaves, and have to be cleaned regularly. They also break down and need to be serviced. Because solar panels have to be up high in an urban or suburban environment so they have a clear shot at the sun, there will be a lot of people working at heights to install and maintain them - including a lot of homeowners with no training and poor equipment. Expect thousands of casualties annually if you put enough solar panels on homes to make up a significant amount of power generation.

You've seen the "deaths per Twh" chart, right? http://nextbigfuture.com/2011/03/deaths-per-twh-by-energy-source.html Rooftop solar is safe, but solar towers are safer still. (And nuclear is the super awesomest stuff ever.)

I wish my Nook were solar-powered. E-ink readers use so little power that you could probably let it charge up on standard indoor lighting. I've seen some third-party products but they can only get the charge up to a certain level.

going by solar cell phone charge rates, it would take a couple of weeks of summer sun in a temperate zone to charge it. Less if they plastered solar panels all over the thing, of course.

Solar panels wont collect leafs if they are placed at the correct angle. Also there is minimal care issues with solar panels and most Solar Panel installers will give a warranty for any problems that you may have with the solar panels.

People need to realise that we need to make the move away from Fossil Fuels and solar is a one of a number of renewable energy sources that can be tapped into quickly and cleanly.

The "correct angle" of the panel depends on your latitude. Where I am it is about 10 degrees tilted South. At the equator, you'd have flat panels. Many installations I see have trackers, where the 10 degree tilt is fixed but the mechanism follows the sun over the course of the day.

Oh, wait, solar panels would never have leaves on them as they are not installed near shade trees. Duh.

what? Sure they are. You're not going to do that at a large solar installation, but my college, for example, sure as hell didn't cut down all the trees around the buildings it stuck panels on.

It’s challenging to find knowledgeable folks on this topic, but you sound like you know what you’re talking about! Thanks

I don't think anybody linked to this page that gives the cost of solar PV per kilowatt based on current market prices:


They give the price per kilowatt-hour of electricity from a commercial 50kw system as 19.72 cents in a sunny climate. As US and Australian dollars are about equal at the moment, this is very close to what many businesses pay for electricity in the daytime here in Australia. The site uses a 5% discount rate while most businesses in Australia have discount rates of at least 7%, but despite this, it seems clear that unless there is a sudden and unexpected decline in retail electricity prices, or a sudden and unexpected halt in the average decline in PV prices, solar PV will continue its rapid expansion here in Australia.

Ronald Brak

Converting AUD to USD at the current exchange rate is, my guess, dangerous.

You really want opportunity cost (ie scale by GDP). Which I suspect would put Australian prices at 10-20% higher (given GDP per capita, PPP, is probably lower by that amount?).

Interestingly, at 13.5p (and entirely ignoring what I just said) UK retail electricity prices are, more or less, 20 cents a USD. But adjusting for relative GDP per capita, that would be more like 28 cents/ USD.

Average electricity use per household (2 people) here is c. 4500 kwhr pa. In the US it's closer to 12,000 I think (but half of that is residential AC, unknown here).

The crime here is that we still have electric (and oil, and propane, and coal) home heating. Replacement with air sourced heat pump would be perfectly feasible.

Certainly for the short term oil and natural gas will continue to be the energy source of choice. We need to remember, however, that they are natural resources and we need only look to other natural resources such as gold to see what happens to the price of resources that are rare. Technology has exposed new oil reserves and will continue to do so, but I doubt that the new sources will keep up with the increasing demand as world modernizes. Even if new sources keep up for the short term, oil is a finite resource and will eventually become scarce. Solar energy is limited only by technology, not by quantity. It will be the primary energy source at some point, it is just a question of when. China seems to realize this, if the US would also we could make more rapid advances in lowering the cost.

One of the advantages of being old is that you have heard all of this before. We had a good round of solar hype during the 70s But the price of oil went down, and we got over it.

The real truth is that that solar is not new and will not develop much more. The photo electric effect was observed in the 1830s, not long after Faraday invented the first electric dynamo and motor. Einstein explained it as a quantum phenomenon in 1905. Bell Labs, which invented the transistor in 1947, invented the semi-conductor photoelectric cell in 1954, which the NYTimes cheered as promising unlimited cheap solar power. By the 1960s, solar cells were being used in satellites. In the late 1970s solar cells were mounted on the White House as a demonstration.

Time has gone on, solar cells have been used in more and more applications, but unlimited cheap power eludes us. The Beulah land seems always to be just around the corner. Why is that?

The real problem with solar power is not the cost of the solar panels, it is the sun, which does not shine 4380 hours per year, because of night. That does not include twilight, clouds, snow, ice, and dirt. Therefor, every watt of solar must be supplemented by an equal or greater amount of generating capacity or of storage. So you have to pay for two systems instead of one. It will never be a good deal, even if the solar cells are free.

And please Moore's Law has nothing to do with solar power. Moore's Law can work because logic gates are scale invariant. A 50 nm logic gate is is so small that it cannot be imaged with visible light, but it works just as well as a gate made of discrete parts that is a million times bigger. With solar cells, size is not irrelevant. What is relevant is the amount of sunlight that the cell can gather. Mirrors and lens are of limited utility, because, they will, if magnification is too high, overheat the cell. Power output, must therefore, be proportional to the size of the cell.

Moore Vs Metcalf

Ok Tyler, here's the skinny, you chose to base your assumption on Moore's law, however I choose to base mine on Metcalf's for the following reasons.

Solar energy comes in two shapes, distributed, panels you or I can install at home or at our businesses and central, large solar plants. In both cases any solar energy gathering device's full potential is to be connected to the grid. And that is where Metcalf's law apply's for solar pretty much as it did for the internet. The more devices are plugged to the grid the greater its worth and the higher its efficiency and resiliency.

I firmly believe that in the foreseeable future, solar energy, precisely photovoltaics will reach mass market, be it thin film or cheaper solar cells. You might have heard of the ongoing trade war between the U.S and China in that regard. And when that happens, solar energy already abundant in its raw form will become cheap. Cheap is actually an understatement if you consider that every hour, the sun radiates more energy onto the earth than the entire human population uses in one whole year, the actual word is free.

I have sometimes ventured in this blog into exploring our next economical model, that I call the free energy ecosystem, where energy will cease to be a luxury for some like here in Africa and become a commodity so mainstream, easy to gather and abundant that it will cease having an actual market value other then for distributors.

So if you ask me about the future of solar power, Its quite bright actually. is there a bubble ? Yes, but there is nothing wrong with a bubble, that is how people make money.

Free energy thanks to the sun, not a matter of if but of when.

Comments for this post are closed