Why California’s Push for Rooftop Solar is a Foggy Idea


Would you trust your neighbors with billions of dollars of public money to invest in a clean energy future? If you live in California, this isn’t a hypothetical question.

California Gov. Jerry Brown last month announced his intention to rely on “tens of thousands of little decisions” by Californians to develop a 12 giga-watt renewable energy infrastructure by 2020. In remarks at a UCLA clean energy conference, Brown embraced distributed solar generation in order to avoid the pitfalls that often encumber large-scale renewable energy projects, including the capital costs of transmitting energy from far-flung deserts and hilltops. Furthermore, rooftop solar panels and Cameron-esque windmills also pose little threat to desert tortoises or sacred Native American sites, so they are less apt to be caught up in the kind of litigation that has delayed major renewable projects.

But energy policy that relies on distributed generation has its drawbacks. Perhaps most notably, it forsakes economies of scale. It also places infrastructure investment decisions in the hands of homeowners, who, as this space has suggested, may not make socially optimal—or even individually rational—choices.

Early evidence from California suggests that because of these drawbacks, relying on tiny decisions is no silver bullet to a green energy future—and it could be a big mistake.

Many of California's rooftop solar panels are clustered in areas that receive less sunlight than other areas of the state.

How else to describe the state’s decision to partner with homeowners to install 1,923 residential rooftop solar installations in San Francisco, which the National Oceanic and Atmospheric Administration characterizes as having persistent clouds and fog during the summer? Despite the weather, San Francisco is still among the top ten cities in California with the most distributed solar panels per household. Solar panels in San Francisco may generate enough energy over their 20-25-year life spans to repay homeowners for their investments. But it’s clear those green energy dollars could be put to far better use in desert communities, like Apple Valley, where 22% more solar energy lands each year (according to the National Renewable Energy Laboratory), and yet only one in 90 homes harvest it.

If San Francisco’s residential solar panels were relocated to Apple Valley, they would produce another 2.1 million kilowatt-hours (kWh) of electricity each year—enough to power 320 average California homes. That would add as much green energy to the state’s electrical grid as installing another 415 average solar systems in San Francisco. At a $0.12 rate for electricity and a $0.03 environmental value for avoided carbon-based generation, the typical San Francisco solar homeowner could afford to pay an Apple Valley homeowner $167 per year to rent his rooftop and be no worse off.

If less efficient residential solar panels throughout California were moved to the sunniest cities in the state, the entire stock of residential solar would generate 15% more green energy. An optimized rooftop solar program would generate an additional $12.5 million in benefits per year and could power an additional 13,000 homes. The optimization would generate as much additional green energy as adding 7,800 more average residential solar installations to the state’s homes.

In a world where energy conservation programs are heralded as successes if they yield 1-2% demand reductions, a 15% efficiency improvement would be huge. Imagine if the build-out required to meet Gov. Brown’s planned 12 gigawatts by 2020 continues apace. A 15% increase in energy production on 12 gigawatts of installed capacity fuels a half-million homes.

If solar investment decisions were the product of a rational cost-benefit analysis, then solar should be far more prevalent in inland communities that lack the coastal fog and marine layer. After all, if it is individually rational for a San Francisco homeowner to install solar, then it must also be individually rational for just about every other homeowner to follow suit. And yet they don’t.

The solar boom along the fog-laden California coast may be spurred by environmental preferences and even conspicuous conservation effects that yield additional, non-pecuniary dividends to homeowners in coastal green communities. But solar is also well-represented in places you would expect (and hope) to find it. San Diego, for instance, has 2,238 residential solar installations, the most in the state. And Central Valley communities with more sun and more conservative political views, like Fresno, Clovis, and Bakersfield, also rank among the state’s top-ten solar cities.

It is perhaps reassuring that political party affiliation is a poor predictor of solar panel density in California. But it can be of little consolation that NREL’s solar energy potential is only a weakly-significant predictor of solar density, even after controlling for housing stock characteristics, population and housing growth, and household income.

Free-market economists embrace the power of many individual decisions to guide markets and direct resources to their highest value. Yet the burgeoning field of behavioral economics has documented many instances in which consumers make mistakes. Consumer errors are much more likely with large investment decisions that require complex calculations involving discounting and expectations of future prices. Even if consumers execute the discounted cost-benefit calculus correctly, their objectives may be different from those of society, causing efficiency losses for policies that rely on their utility maximizing behavior.

So if homeowners can’t be trusted to make social welfare maximizing investments, and if industrial scale solar projects can’t clear legal hurdles, what should be done?

One option would be to permit the solar swapping that Dubner proposed. But this only works if a homeowner in the desert who won’t invest his own money in a solar array would be willing to rent his roof for less than $200 on average. He may not find that inducement sufficient to overcome transaction costs, not to mention the potential nuisance of housing someone else’s green investment.

A better solution may be for the state to coordinate investment by individual residents in larger, community-based solar installations that could be situated in the sunniest cities—perhaps on rooftops of public or commercial buildings or above parking lots in inland Southern California. Homeowners would receive payments for the energy produced from their share of the projects, much as they do from the projects on their own roofs. Only the payments would be bigger because of their better location. These larger projects would lower transactions costs and capitalize on scale economies, which a consultancy for the state’s energy commission reported to be critical to generating positive net benefits for society. And, as Dubner suggests, the residential investors could get big signs to put on their roofs to replace the “green” signal their solar panels would have sent.

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. Joshua Northey says:

    “It also places infrastructure investment decisions in the hands of homeowners, who, as this space has suggested, may not make socially optimal—or even individually rational—choices.”

    Yep because governments and corporations have such a good track record of doing the same?

    Lets take your little caption and twist it:
    Many of California’s SUVs are clustered in neighborhoods where the need for such vehicles is minimal .
    Many of California’s tax dollars are clustered in military spending that helps no one and encourages destructive foreign misadventures.
    Many of California’s tax dollars are clustered in tax breaks/credits/exemptions for individuals and industries who manifestly don’t need the extra money.
    Many of California’s educational resources are clustered in communities where they make the least impact.
    Many of California’s policing resources are clustered in communities where there is little crime.
    Many of California’s social service resources are to people/communities beyond help.
    Many of California’s businesses/entrepreneurs are focused on lines of business which not only do not help the general well being, but are actively detrimental to it.
    Many of California’s citizens spend much of their lives paying off debts that were horrible decisions and the government should make it illegal to offer.

    and so on…people whether individually, or collectively in the form of governments or corporations, do not spend money particularly well/efficiently.

    We have enough economic resources to provide everyone in the world with a decent standard of living, and create institutions that would eliminate war, and most of our other problems. But we don’t, because we are selfish little animals. We do things like build solar panels at OUR house instead of building them 800 miles away where they will do more good. This is not news, it is human nature, and it is that way for a lot of good/understandable reasons.

    I am really not trying to make the “as long as there is one problem bigger than the problem you are discussing, discussing it is pointless argument”, but this is a pretty small issue in the grand scheme of inefficiently spent public (or private dollars).

    Well-loved. Like or Dislike: Thumb up 25 Thumb down 6
  2. Allen says:

    Could tax incentives for solar panels be tied to cities and towns with the greatest amount of sunny days?

    Well-loved. Like or Dislike: Thumb up 8 Thumb down 0
  3. nobody.really says:

    I agree with much of Sexton’s analysis. Yes, individuals have a variety of incentives that are unrelated to optimizing social benefit. A different program might better align the benefits with the subsidies.

    But Sexton’s analysis overlooks some of the beneficial aspects of distributed generation. We can’t simply consider where the sun shines; we need to also consider where the power is needed, and how to get it there. A solar cell in the middle of the desert will generate more power than one on a San Francisco rooftop. But once you consider the cost of the transmission facilities required to move the power from the point of generation to the point of consumption, the benefits of the desert location may, well, evaporate. In contrast, to paraphrase an old saying, “What’s generated in San Francisco stays in San Francisco ” — no transmission facilities required.

    (Similarly, we can’t merely consider where power is needed; we need to consider WHEN power is needed. Compare the relative benefits of solar and wind power: Yes, wind seems more efficient by many measures. Yet the wind tends to blow in the mornings and evenings – not times of peak demand. In contrast, the sun tends to shine throughout the day – precisely when demand is most acute, and the cost of substitute power is high. Any failure to consider these distinctions will distort your analysis.)

    In short, there may be more things in power systems optimization than are dreamt of in Sexton’s philosophy.

    Well-loved. Like or Dislike: Thumb up 19 Thumb down 1
    • Chris says:

      you’re ignoring the point of the analysis- San Francisco does not have optimal sunny days. On days of peak demand, San Francisco may very well be fogged in with temperatures in the 60s and not producing any solar power. I would agree that location needs to be factored in to the analysis, but you can’t ignore that San Francisco is an extremely poor location to invest state funds into distributed solar power. Even if San Francisco is blanketed with solar, on a foggy day that does nothing to reduce the reliance on utility scale power plants and you’re basically wasting money…..

      The second thing to consider is around the amount of distribution upgrades that will be needed in areas with significant amount of rooftop solar. Those costs are not drops in the bucket by any means. San Diego is cited as an example of putting in solar where it’s most beneficial- what’s not noted is the costs that SDG&E will incur to allow two-way flow of electricity (distribution grid is designed as a one way street, so to speak), as well handling the voltage disruptions from rooftop solar at the transformer level. Asking PG&E to invest in upgrading the distribution grid in San Francisco is on completely different level altogether.

      On the rate design issue- let’s remember that the majority of customers in San Francisco never pay more than $.13 kwh, thanks to the tiered rate structure and use of climate zones to set baselines (there’s little air conditioner use in San Francisco, and people tend to live in apartments and not houses). customers who live in hotter areas of the state will pay more because their usage tends to be higher (even accounting for the change in baselines) due to larger houses and air conditioning (among other things). What this means is the customer in San Francisco is only avoiding $.12/kwh, whereas the customer in the hotter area could be avoiding up to $.30-.40/kwh. again, if you do avoided cost, there is simply no way to recover the costs of a rooftop solar facility in San Francisco over a meaningful time period. Even accounting for the increase in transmission facilities, i still fail to see how it makes sense to throw money at a generation asset that’s unusable for a good portion of the year (coupled with the increased distribution upgrade costs). As someone below said, what makes more sense is to base the subsidy on location, hopefully to drive investment in areas where it makes more sense.

      Thumb up 7 Thumb down 3
      • Raffi says:

        I agree with you regarding the distribution costs/upgrades, however, how does this address the issue of transmitting power from the sunnier areas to San Francisco-level area of fog? While a place like San Francisco is less efficient, a home can still produce a good fraction of its usage, depending on the exact situation (with the homeowner having more incentives to only install if it’s producing a good sized fraction).

        Thumb up 2 Thumb down 0
    • James says:

      A good point re cost of transmission. Also, I noticed that the map of solar installations vs available solar energy is rather misleading. It makes a lot more sense if you mentally overlay a map of California population density, ’cause the solar installations are – surprise! – where the people are. You may get more solar out in the Mojave Desert, but (except for that cluster around Palm Springs) there aren’t a lot of people living there. No people, no roofs, no rooftop solar.

      Well-loved. Like or Dislike: Thumb up 13 Thumb down 0
    • Rick says:

      You have a good point about transmission facilities. West Texas is a prime area for wind generation and there are farms of wind generators. When the wind is optimal for production a significant amount of them are inactivated because there is not enough transmission capacity to get the power to the cities in central and east Texas that can use it. It’s in the works but there are significant political barriers that are slowing it down.

      Thumb up 1 Thumb down 0
  4. Enter your name says:

    It wouldn’t be difficult to scale the subsidy to match the solar potential of the area. If Apple Valley receives 22% more solar energy than San Francisco’s, then make their subsidy 22% bigger than the subsidy for San Francisco.

    As for why people don’t make this investment even when it’s clearly sensible: Solar panels cost money, a scarce resources. You not only have to have enough money to pay for them, but you also must not have any higher priorities for that money, such as a higher standard of living, sending your children to private schools or college, saving for retirement, etc. The difference in uptake between communities says a lot about the relative wealth of the communities.

    Well-loved. Like or Dislike: Thumb up 11 Thumb down 0
    • James says:

      Exactly. It’s not really the cost of solar or the payback time that’s driving current installations, it’s the cost as a fraction of one’s disposable income. That’s a lot lower for your typical resident of say Atherton than Apple Valley.

      However, should the goal at this point be to maximize efficiency of installed panels, or rather to increase production & so drive down the cost, so that eventually it’ll be economic to install solar in what are now marginal areas?

      Well-loved. Like or Dislike: Thumb up 9 Thumb down 0
      • Owen says:

        Lots of good points in the comments but James is the most important. It seems like Jerry Brown is stealing the rhetoric of the moment to justify the program. Whether or not his public reasoning is the true reason remains to be seen. It’s impossible for us to predict the added benefit to the economy by increasing demand for solar panels and we certainly can’t do it by using the economics we are familiar with today.

        Thumb up 0 Thumb down 0
  5. Recent Vegas Visitor says:

    I was amazed during a recent trip to Vegas that as I stood on the observation deck of the Stratosphere, I could only see one building with solar panels. Seems like an obvious place to have them, given the amount of sun and need for electricity.

    Well-loved. Like or Dislike: Thumb up 13 Thumb down 1
  6. Clancy says:

    I would have a difficult time writing an article like this without using the phrase “stick it where the sun don’t shine.”

    Thumb up 5 Thumb down 1
    • Sam says:

      Another factor to consider is that PG&E has a tiered rate system for electricity consumption. Many of the houses in San Francisco may actually be paying an incremental cost of 30 or 40 cents per kWh which makes their solar projects cost effective even if they are getting less sun.

      Well-loved. Like or Dislike: Thumb up 6 Thumb down 0
  7. Andy says:

    Having put solar on my house in the bay area, I can say the following from an economic perspective.

    East of the San Francisco peninsula, we get more sun and a big consumer of electricity in the warm summer months is the air conditioner. Mine draws 9 KWatts when its running, which works out to about 37 Amps at 240 Volts.

    San Francisco proper has less of this, in fact even in the city there are many microclimates, you can grow tomatoes in the Mission, but not in Twin Peaks.

    As an investor in solar, it is a hedge on energy price increases. With the California rebate, the Federal tax credit and other incentives from our local electric company, the investment returns about 11% a year, paying for itself in 8 years. The remaining 17 years of its warrantied 25 year lifespan is profit. These projections assume a 3% increase in electrical costs, which I believe is conservative based on historical increases.

    A net metering system that scales the costs and therefore the buy-back rate for electricity on the time of use is what we use. We buy in the off peak hours at $0.12 a KW hour and sell in peak summer hours at $0.49 a KWh.

    Our system produces approximately 70 to 80% of our annual electrical needs (7000 KW hours annual, but dropping from conservation measures light LED lights), but it pays for 100% of the electrical bill.

    If the electrical company would pay market rates on the electricity I produce in the form of a check every year on the balance, I’d put in more panels, but until then its only worth getting electricity credit.

    Although I do agree with some of the points and previous posts that part of what might drive a PV solar installation in parts of the Bay area may include: being ecologically driven, the perception from your neighbors, economic access to the available cash, not needing to move for several years or even the desire to be more self sufficient. I have some of those tendencies myself, but when I can almost guarantee an 11% annual ROI in these stable economic times ;-), the selfish economic whore tendency in me wins.

    Much like other areas, its hard to generalize about individuals in their specific cases. As they say, “your mileage may vary.”

    Well-loved. Like or Dislike: Thumb up 7 Thumb down 2
    • Andy says:

      Also, the CA solar subsidy application form does have a efficiency factor built into it which accounts for the azimuth, tilt and location of the array to adjust for less than optimal installations. I don’t remember if shadowing or solar obstructions were taken into account, but I don’t think so.

      Based on your lat. and long. you can find historical average solar radiation numbers across the year, which is what the state uses to estimate effective efficiency.

      So my theoretical 2.5 KW maximum system is adjusted by the 0.96 efficiency factor to the 2.4 KW system that was the number used for the CA rebate. Since the panels point 240 degrees (60 degrees west of 180 South), and are on an 18 degree tilt (a 4:12 roof), the estimated a slight drop from perfect to 0.96. We have a pretty solar friendly site though.

      So if you do put in panels that face north in a foggy area, your efficiency rebate is lowered accordingly.

      Thumb up 4 Thumb down 0
      • Charlie says:

        Yes, the California incentive calculator includes deductions for obstructions and shadowing. It also includes the specific weather at the location, including fog and other impacts to daily (actually hourly) solar energy.

        Thumb up 0 Thumb down 0
    • Charlie says:

      I am sorry to point it out, but you don’t get that big benefit for selling your power at the high peak rates. While it is true that you would pay at those rates if you didn’t have solar, it isn’t true that you get reimbursed at that same rate. The reason is that PG&E (the utility in San Francisco) uses a rather odd calculation to determine the baseline rate for Time of Use customers. This calculation has the result of artificially boosting the level of the baseline power. The more you make in the peak periods, the higher the baseline is set, and the less likely it is that you can produce power above this new baseline. In fact, the formula is designed to allow the baseline to approach infinity if need be to prevent you from getting paid at higher tiers that would credit at a higher rate. There is an exclusion in the calculation that specifically avoids a possible divide by zero error as the baseline approaches infinity!

      I think this is highly unfair, circumvents the intent of the law that says you will be credited at the rate you would have paid without the solar, but that is how it is. I contacted the CPUC to see if they were aware of this little trick and managed to talk to a lady who said she is they one that created the formula. She said that her intent was to prevent people from getting paid back at the rates of the higher tiers because she thought that would be unfair to the utility. She was successful in that effort.

      Thumb up 0 Thumb down 0
  8. Michael says:

    This article appears to be based on a single premise: Solar panels don’t work on cloudy days, so installing a solar panel in an area with more clouds is a losing game.

    No references to back this contention up – it’s taken as a given.

    Yes, Apple Valley has more sunlight. But what of those homes in San Francisco? The article (cleverly) doesn’t mention whether their installation of solar has been successful or not – as in producing a positive ROI or an effect on that very demand reduction percentage you say is so important – which is the single most important thing an article in an Economics blog should tell us.

    Essentially, you’ve taken a stance and produced only the evidence that favors your stance. And that’s not economics at all – it’s politics.

    Oh, and nice move singling out the most cloudy area of California. This proposal would also put panels in Apple Valley. Oops!

    Well-loved. Like or Dislike: Thumb up 9 Thumb down 4