Skip to content

Lessons in Regulatory Hubris

What the Solar Rooftop Standard has to tell us about our climate policies.

If you follow this blog, you are probably already aware that last Wednesday, the California Energy Commission, the state agency responsible for adopting energy-efficiency related building standards, added a new standard: starting in 2020 almost all new homes in California must install rooftop solar.

While I have already published an op-ed criticizing the move, I’m going to look on the bright side here. You see, this regulation provides what we in the University biz call a teaching moment. Many of us here have been working on climate and energy policy for a long time and realize that some core problems with regulatory approaches keep cropping up in policy after policy. Let’s call them the symptoms of regulatory groupthink.

So rather than picking exclusively on the new solar standard, we can use it as a case study to help illustrate a broader set of problems with many of our approaches to energy policy. As our climate goals get more ambitious, the stakes are getting much higher and we may not be able to continue to ignore these flaws. We are going to need to treat these symptoms or our climate policy could ride off the rails.

Symptom Number 1: If we like a technology, we mandate it!  If we don’t like a technology, we ban it!

One thing the press coverage of the standard seems to overlook is that there are lots of other policy options for promoting low-carbon energy and that we don’t necessarily have to do all of them. There are many pathways to zero-carbon electricity, and many of them don’t require solar PV on all houses. Standards, such as this new mandate from the CEC, are the strongest, and most extreme, tool in the toolkit of energy and environmental regulators. In general a regulatory standard eliminates choice and assumes that the mandated solution is the best one for everyone, everywhere, no matter what their circumstances. Contrary to the stereotype of economists, most believe that regulatory standards have a role to play, but as a kind of last resort. A rule of thumb should be, if everyone were fully informed and faced the right incentives (including the costs of their pollution), they would choose this option. Insulation in a new home meets that description.

In other words, standards should be limited to cases where they mandate an obvious, slam-dunk best option for solving a problem. This is where rooftop solar runs into trouble. In the case of rooftop solar, it should be not just the obvious best choice right now, but for much of the lifetime of a new home. That’s a tough criterion to meet. Consider what you thought to be the slam-dunk best technology 30 years ago.

My colleagues and I who have been critical of the CEC standard in the press have pointed to the fact that electricity from solar on residential rooftops is on average way more expensive than solar in other settings (warehouses, solar farms). I’ve gotten feedback from various commenters, friendly and not-so friendly, making arguments that point to the costs of high-voltage transmission, or concerns over large installations in sensitive desert eco-systems. Such factors should definitely be taken into consideration, and while they may narrow the gap between residential solar and other options, many believe they still don’t outweigh huge cost advantages enjoyed by large-scale installations.

But even having this debate misses a more important point. The burden of proof isn’t on me to show that there are better options for low-carbon energy than residential solar. I’m not the one requiring almost every new home install a nearly $10,000 technology. The burden of proof is on those pushing for a mandate that ignores any concept of a level playing field between possible solutions to show that their technology option is obviously the best choice amongst many others, not just now but also over the coming decades. I don’t think the CEC has come close to meeting that burden.

Symptom Number 2: We continue to treat infrastructure cost-shifts as savings.

This has been a recurring theme on the blog over the years. As Severin Borenstein has pointed out, one of the main reasons residential solar in California looks like a good deal, from the perspective of many households, is because its costs are compared to full retail rates. The problem with this comparison is that retail rates contain a large amount of fixed and sunk infrastructure costs. Those costs don’t go away when a home generates solar energy. Solar homes shift those costs onto non-solar homes.

Consider this situation: we spend $5 Billion to build a bridge across the Bay, and charge commuters $10 per round trip to pay for it.  It’s a good public policy decision because each commuter gains $20 in personal value from using the bridge. Right after the bridge is finished, the transportation commission determines that all commuters could own jet skis, and ski across the Bay at a cost of $8 a day. Two dollars less than having to pay the bridge toll! A win-win! It’s clear that for the best interest of those commuters, we should mandate they all buy jet skis. Except that those billions of dollars in jet skis won’t save us the cost of the bridge. Public policy evaluation needs to look at the full social costs of policies, and not treat cost shifts as free money.

Symptom Number 3: The infatuation with “net-zero <something bad>”.

The CEC solar decision is another step in the push for a net-zero energy standard for new homes. The solar mandate has been justified as one way to get the state closer, but still not all the way, to that goal. Rather than make the solar standard look better, this decision should draw more scrutiny upon the net-zero energy aspiration. This building standard target hasn’t received nearly enough attention or scrutiny, possibly because it hasn’t been taken seriously up till now. It’s a great example of the net-zero everything movement. Economists, or anyone who has spent any time thinking about comparative advantage, have long shaken their heads in disbelief at this stuff. Apparently we haven’t articulated the counter-argument clearly enough. Sure, net-zero energy for every new home could be way to reach our climate goal but it’s almost certainly not the best way. Requiring each California resident to sell enough goods and services to China to offset the imported goods they consume would be way to address our trade imbalance, but not the right way.

Or think of it this way, why not have net-zero energy standard for each room in a house? That would be crazy, you say? What if the roof above one room isn’t facing south and another is?  Well, what if one house has a roof amenable to large amounts of solar and another is a north-facing home with an odd-shaped roof shaded by trees?

Maybe, maybe, net-zero goals have a role as a  gimmick for spurring ideas in a demonstration project setting. When we start taking them too literally and applying things like this to every home, we’ve gone too far.

Symptom Number 4: Our renewable energy policies do not adequately account for the energy or capacity value of the resources we acquire.

In much of our policy a green kWh is considered better than a dirty kWh, but not all kWhs are the same. For example, California has been at the forefront of rapidly shifting its electricity production to renewable generation technologies. Utility-scale renewables have been driven by the renewable portfolio standard (RPS). Distributed renewable generation has enjoyed a raft of implicit and explicit subsidies, including net-metering and an extreme rate structure, which together make rooftop solar an attractive option for many California households. What all of these policies share is a complete indifference, incentive wise, to where and when the energy is produced.

Screen Shot 2018-05-13 at 12.28.14 PM
Annual Average Wholesale Prices and Solar Output in CAISO Market

Under the RPS everyone has had an incentive to install the technology that produces the most renewable KWh regardless of when they are produced, and, in California, that has turned out to be solar. Recent work by myself and Kevin Novan, also of UC Davis, quantified the impact of the solar binge in California on the wholesale prices of power. In the last 5 years, thanks to the surge of wholesale solar energy (right panel) the price profile has moved from the blue (solid) line above to the red (dashed) line (left panel). What this means is all the solar we have already installed has driven down the value of midday energy. The next KW of solar capacity we install will be far less valuable than the first. The 10,000th MW of solar capacity we install will generate only about half of the value that the 2000th MW did. But the RPS doesn’t care, it just wants its KWh.

Screen Shot 2018-05-14 at 7.21.14 AM
Estimated Value of Incremental Solar Capacity (Bushnell and Novan, 2018)

The new CEC rooftop solar mandate kind of cares, but in bizarre ways. The CEC uses something called the Time Dependent Valuation methodology, a forecast of the time value of energy and other infrastructure over the coming decades (good luck with that). Who knows, maybe the CEC’s consultants are exactly right and this is what the hourly marginal cost profile will look like, but it sure looks different than what I would have come up with.   Maybe reasonable people can disagree on this. This brings us back to point number 1, if reasonable people can disagree, then we shouldn’t be requiring almost everyone to install a technology justified by this one set of debatable assumptions.

As more aggressive and difficult carbon reduction goals loom for California, there seems to be an inclination to grasp at every policy we can think of that can add to the carbon reduction body count. It’s a spaghetti on the wall approach to carbon policy. However, it’s now more important than ever to focus on the efficient tools and policies that can push our carbon reductions in cost-effective ways. We could get away with inefficient policies like net-energy metering and zero-carbon schools when they were relatively small polices. From here on out the costs are going to start to matter.

Suggested citation: Bushnell, James. “Lessons in Regulatory Hubris.” Energy Institute Blog, UC Berkeley, May 14, 2018,

63 thoughts on “Lessons in Regulatory Hubris Leave a comment

  1. I completely agree with the post and its assumptions. It appears to be another really bad decision by a regulatory agency in California.

  2. Thoughtful, incisive post. The optimal role for the regulatory authorities is to create an environment in which everyone bears the full marginal costs, and derives the full marginal benefits, from every choice they make. Mandates should be reserved for the obvious stuff, to prevent egregious errors of judgment. Obviously, failure to install rooftop solar is not an error for many if not most people.

    The bridge analogy struck me at first as poorly chosen. The capital cost of the bridge is a sunk cost, which should be ignored when making decisions. If it’s truly possible for people to cross the bay at lower marginal cost by jet-ski than by bridge, then that’s how they should cross, and it’s how they will choose to cross if they bear the full marginal costs of their decisions. The remaining bridge-crossers will find their embedded-cost tolls rising, making the jet-ski alternative even more attractive. The capital cost will fall on the bondholders and stockholders of the entity that built the bridge, and will deter them from making such ill-advised investments in the future.

    But, given the bridge exists, it’s almost certain that the marginal cost of crossing the bridge is lower than alternative modes of crossing. Charging a toll based on the average embedded capital cost of the bridge may raise the price of crossing well above the marginal cost of alternative methods, particularly if they are subsidized. Unfortunately, borrowing constraints may mean that only the well-to-do can afford jet-skis, and as they cease paying tolls the embedded-cost tolls will only rise, creating a death spiral that places a disproportionate share of the capital-cost burden on the poor. Similarly, until the long-run marginal cost of residential solar falls below the short-run marginal cost of utility power, the solar mandate will raise both electricity rates and housing costs (including rents), making life even more difficult for California’s already-beleaguered working class.

    • stratdouglas
      You’ve hit the nail on the head: the utility shareholders have no skin in the game when investing in utility infrastructure and procuring generation. Decoupling revenue from profits and allowing full cost recovery of generation costs no matter how poorly the portfolio is managed are the instruments that have led to this situation in California. The single most important reason to have private ownership of network utilities is to transfer risk to a group that has higher risk tolerance (shareholders) than the general public at large. If we’ve transferred all the risk to ratepayers, why do we continue to have private ownership?

  3. Currently, electrical grids are primarily cyclo-dissipative AC Systems (transmitted external real power that is dissipated in load (Watts) exits from the system and internal reactive power (vars) oscillates between the source and the load. It seems that a change to this 100+ year old reactive power paradigm would result in less electrical losses and waste with a more circular economy approach where the majority of var oscillations are found locally nearest the loads and also between the loads and the POC with the grid. Now with IEEE 1547:2018 inverters and energy storage are required to have reactive power support between 95 percent lagging and 95 percent leading. Even some customers will be required to maintain a power factor between 95 percent lagging and 95 percent leading. Anything providing AC to the grid, inverters, energy storage or being supplied AC that is not operating at true unity power factor is hurting the AC grid system for all with losses and should be held accountable by mandating this energy efficient PF=1 approach which is an obvious, slam-dunk best option for solving this problem. Smart meters can measure PF and smart power electronics can control PF.

  4. The article referenced says that there are exemptions for cases where solar is not appropriate such as shady areas. It also says that standards will call for better insulation. It doesn’t say how much solar is mandated. From the estimated cost I’d guess about 3 kW. We are planning to add 5.5 kW for about $16k.

    This is not a lot of power. It will just balance the AC, refrigerator and freezer for homes in warmer areas. The article doesn’t say whether leased projects would qualify.

    This is good for California. Mandating solar will drive costs down even further. One possible benefit would be if this helps normalize rules for installation in each city.

    If there is so much residential solar available maybe it will incentivize utilities to build in storage at the neighborhood level.

  5. Right — mandating rooftop solar on all new houses is ill-advised. However the economists’ dream of a perfect market is laughable. “if everyone were fully informed and faced the right incentives (including the costs of their pollution), they would choose this option. Insulation in a new home meets that description.” First of all you know it doesn’t work, so you come up with the fully informed and right incentives ploy — which never exists.
    Right now putting in an AC and heating system (including the delivery system) that is proper for our climate would save much more than the solar panels. What is more it would not aggravate the “duck curve”. Upgrading our existing housing stock would be another “slam dunk”. But it will not happen within the real market — too many competing priorities for people. Meanwhile global warming marches on.
    You are right “net-zero” was the shiny new object. In the same time period the CEC and the CPUC agreed that California should move over to a climate appropriate HVAC system and that incremental improvements should be encouraged. Unfortunately rather than do that they allowed the utilities’ evaluators to redefine climate appropriate to something that was not climate appropriate.

  6. For those readers that require additional background regarding this debate, please take the time to at least read the executive summary of this informative report regarding another inherently intermittent power source, namely wind. The key finding is that with increased wind generation penetration, emissions actually increase above those of a fossil-generation power system that does NOT have wind generation. This analysis is equally appropriate for California solar photovoltaic (or thermal) generation. Per Cal-ISO, California has more than 11,500 MW (nameplate) of solar generation and more than 6,000 MW (nameplate) of wind generation. CGNP provided a careful analysis of California wind generation and solar generation during the half-year period ending on 31 January 2017. The result was that the “capacity factor,” or percentage ON time was about 20% – which means that fossil-fired generation backs up those intermittencies about 80% of the time. As noted above, California is already PAYING Arizona to take some of California’s un-curtailed solar power around solar noon, as California is Sun-synchronous throughout the state. Clearly, California’s solar energy regulations are causing unintended consequences that end up causing both environmental and economic harm – while only special interests are enriched.

    The pursuit of low-quality, non-dispatchable solar and wind generation should be subject to clear scientific, engineering, and economic analysis. Instead, it appears as if the scientific, engineering, and economic analysis of power systems in Germany and southern Australia as examples are just ignored by special interests obtaining financial benefits promoting these low-quality means of generation.

    To obtain some understanding of why high ramp rates (decreasing and increasing) for fossil-fired generation increase emissions – and wear and tear on generators – consider the difference in gasoline mileage when a vehicle is traveling using cruise control on a sparsely-traveled highway in comparison with alternately “flooring” the car and letting it coast to achieve the same average speed as the car using cruise control.

    Here’s the reference:

    Click to access BENTEK-How-Less-Became-More.pdf

    How Less Became More…. Wind, Power and Unintended Consequences in the Colorado Energy Market, Bentek Energy, LLC, 16 April 2010
    Wind energy promises a clean, renewable resource that uses no fossil fuel and generates zero emissions. Careful examination of the data suggests that the numbers do not add up as expected.
    The “must take” provisions of Colorado’s Renewable Portfolio Standard require that other sources of generation, such as coal plants, must be “cycled” to accommodate wind power. This cycling makes coal generating units operate much less efficiently…so inefficiently, that these units produce significantly greater emissions.
    This study reviews the data that supports this conclusion, outlines mitigation measures which can be used to realize the full potential of wind generation, and provides recommendations for policy makers.

  7. As with Lucas Davis’ post ( I’m disappointed with lack of looking at the complete picture in the analysis. I’ll walk through each of the listed issues:

    Symptom Number 1: The mandate overcomes a strong bias in the building community to NOT offer cost-saving measures with higher upfront costs. The CEC found the expected savings on a $10,000 investment to be $16,000. I don’t know if the CEC’s analysis was done correctly, but you’re ignoring that finding in your discussion. I won’t walk through again the reasons why Davis’ analysis was incorrect–you can read my comments there. And until the investor-owned utilities start “leveling” the playing field by ending postage stamp transmission rates and hiding ratemaking at FERC, the argument about level playing fields is empty.

    Symptom Number 2: There is only cost shifting if you accept the utilities’ disingenuous claim that shareholders should bear no risks and be able to pass on all “approved” costs without regard to how they manage their systems. This presumption has been the key to the very problem we now face in our utility system. Why should I prudently manage costs if I’m guaranteed a 14% return on every dollar that I invest? And further, the utility system is not a static system. Parts are being added and replaced constantly. A more dynamic analysis recognizes how distributed energy resources eventually lowers everyone’s costs.

    Symptom Number 3: The mandate will require some trade offs where solar isn’t appropriate,but the point is that communities should be aiming to be “net zero.” And perhaps the best single reason is that buildings are infrastructure that will last decades if not longer than a century. What we build today will be here long the future. So we can’t pass up opportunities to impose requirements that we might not really need for many years. For example, microgrids are not yet economic, but likely will be in the coming decade. Yet it costs less than two thousand dollars today to install the microgrid-ready infrastructure in a new house, but to retrofit a house likely will cost $20,000 (yes, I have two different data sets to back this up.) Why be penny wise / pound foolish now? And again, we are addressing a market failure in which the available solutions are biased toward large scale players who hide their real costs. Further, this mandate is likely to spur development of the distributed storage market. This in turn will help solve the real reliability problem which is in our distribution system (these outages are 15 times more likely in California than transmission level outages.)

    Symptom Number 4: Hourly prices are no longer a relevant measure to value in California’s electricity market. The CAISO DM & RTM markets are used solely for load balancing and not to serve incremental loads. A review of the CAISO’s Annual Market Performance Reports since 2001 ( shows that the CAISO markets have NEVER produced enough revenue to justify investment in new generation. Based on the cancellation of several natural gas projects over the last year, I (and many other knowledgeable participants) believe that there won’t be another CT built in California ever again. We almost certainly will rely on storage and other DERs going forward. These as yet to be defined resources represent the real marginal costs or market values in the future.

  8. Hi James – your points are thoughtful and well taken. As a member of the regulatory community, however, it’d be really awesome if you put as much thought into framing as you do the suggestions. The tone of this piece is likely to make many feel defensive, and any lessons you can share ultimately unheard.

    It is clear there are things you think the policy community is doing wrong. The next question, then, is, Why are these things happening? “The other side doesn’t understand” or “they’re not smart enough” is too simplistic for a person of your intelligence. Why are they happening? What do they see that economists, perhaps, cannot? What is the alternative? How can we meet in the middle, and come out with a better outcome for all?

    This sort of dialogue is how we miss opportunities for collaboration — and collectively suffer as a result.

    • I realize your question is for Jim but this article might shed a little light:

      The author seems to conclude that doing something – anything – is better than doing nothing. Perhaps. But I suspect several other possible answers to your question including a) opinions from experts like Jim, Severin Borenstein and Frank Wolak were solicited and ignored, or worse, b) their opinions weren’t solicited at all, c) the CEC chose to ignore the impacts Jim discusses in this post, d) regulators have forgotten about the problem of “stranded costs” and how they impact customers who choose not to install solar for one reason or another.

      A more likely reason is political pressure from the coalition that backed the mandate.

      I’d be interested in a reference to whatever analysis the CEC used to derive its estimates, because I suspect they’re based on average conditions that do not account for long winter nights or switching from gas to electricity or what to do with summer surpluses, all of which are details that matter.

%d bloggers like this: