Membership has its Co-benefits

Last week marked the first “informal ministerial consultations” in the run up to the UN climate talks in December. The objective of these informal meetings before The Meeting is to provide the opportunity to find common ground and an organizing framework for what the UN Climate Chief is ominously calling the “last chance for a meaningful agreement”.

Two decades worth of efforts to broker a binding global climate change treaty from the top down have largely failed. But hope springs eternal, and there is belief that a more “bottom up” approach which allows countries to define their own contributions will break the impasse.  Each country has been asked to submit a self-determined national “intention” to curb its greenhouse gas emissions. These pledges will provide the foundation for any climate deal reached in Paris.


Every time a country registers a pledge, this UNFCC tree sprouts a new green leaf  linked to the plan. As of today, 20 pledges have been submitted.

The prospects for free-riding make this process particularly daunting.  As Gernot Wagner and Marty Weitzman note in their recent book:

“Why act, if your actions cost you more than they benefit you personally? Total benefits of your actions may outweigh costs. Yet the benefits get spread across seven billion others, while you incur the full costs. The same logic holds for everybody else. Too few are going to do what is in the common interest. Everyone else free-rides.”

In other words, why would a country voluntarily commit to making significant, costly reductions in domestic greenhouse gas emissions (GHGs)?

Last week, I saw a paper presented by Ian Parry at a summer gathering of environmental economists that suggests this free riding problem might not be quite as dire as it appears. The jumping off point is that reducing the use of GHG producing fuels can generate domestic co-benefits for the countries that undertake them (such as improvements in local air quality or reduced traffic congestion). These authors set out to quantify these environmental co-benefits by country and evaluate the possible implications for GHG emissions. In my mind, the paper’s findings strengthen the pragmatic case for the new “bottom up” approach to global climate change mitigation.

Climate change policy from the bottom up

At the heart of the new approach to climate change negotiations is a new climate change acronym:  Intended Nationally Determined Contributions (INDCs). These national pledges describe steps that countries intend to take to reduce their GHG emissions. Countries have tremendous flexibility in drafting these plans.  They can pledge to cut emissions by a lot – or a little. Commitments can be binding – or voluntary.  The idea is to let countries decide what they are willing and able to contribute to this global effort and take it from there.

The figure below, taken from a special IEA report on climate change, projects the emissions associated with the climate pledges countries have already declared. This INDC trajectory (in blue) is contrasted with a “450 scenario” (in green) that would achieve the widely accepted target of limiting global warming to 2 degrees Celsius, or 450 parts per million of CO2 equivalent in the atmosphere.


The gap between the blue and green line is downright depressing. Although the INDC scenario could improve as more countries sign on, even Al Gore acknowledges that these INDC pledges will fall short of the critical target.

But the glass half full (or at least not empty) view is that these initial pledges are the sign of meaningful global cooperation taking hold.  Ultimately, the success of an approach that relies on voluntary contributions will depend on whether countries find the political will to engage in this global effort and pursue significant GHG emissions reductions. Co-benefits could provide a leg up in this regard.

Membership (in the global climate change mitigation club) has its co-benefits

When a country takes steps to reduce GHG emissions benefits beyond climate change mitigation often result.  Domestic “co-benefits” of GHG reductions include, for example, reductions in the number of deaths caused by air pollutants that are emitted along with CO2 when fossil fuels are burned, and reductions in congestion, accidents, and other externalities from motor vehicle use.

The paper I saw last week calculates the domestic co-benefits of pricing of carbon dioxide emissions for the top twenty emitting countries that are responsible for about 80 percent of global CO2 emissions.  Using country-level estimates of (non-CO2) environmental damages by fossil fuel from this study, together with fuel price data and fuel tax/subsidy information,  the researchers derive efficient CO2 prices that reflect domestic (non-internalized) environmental benefits and costs:


The figure summarizes the nationally efficient carbon prices that reflect domestic co-benefits excluding climate benefits.  The average (emissions weighted) price is remarkably high: $57/ton CO2. This exceeds the current, mid-range estimates of global climate-change damages per ton CO2.

The graph also shows how prices vary dramatically across countries. Some of this variation reflects differences in the extent to which fuels are subsidized/taxed across countries.  The extremely high prices in Saudi Arabia and Iran, for example, are largely due to large subsidies on transportation fuels and natural gas. The analysis assumes that these subsidies remain, and that the carbon tax works to offset the subsidy. The negative tax in Brazil reflects the fact that the existing fuel taxes exceed the author’s calculations of non-carbon external costs per unit of fuel use.

To put these tax estimates in perspective, the authors ask: How would CO2 emissions change if these nationally efficient carbon taxes were implemented?  The figure below summarizes emissions reductions estimates for each country (relative to the 2010 emissions that were actually observed).


Across all 20 countries, the authors estimate a 14 percent reduction in 2010 emissions. The majority of emissions reductions come from reduction in coal consumption. In  countries where the tax effectively offsets transportation fuel subsidies, reductions in diesel and gasoline play a larger role.

These numerical results are, of course, sensitive to some of the underlying (and sometimes uncertain) assumptions that are documented in the paper.  Qualitatively, the take away is that domestic co-benefits from climate change mitigation appear to be significant on average and highly variable across countries.

Boosting local motivation for global cooperation

The mere existence of co-benefits need not imply that countries will implement climate policies to pursue them. There are political constraints, distributional concerns, and other considerations that help explain why most countries have neglected to address domestic externality problems and other distortions in the first place.  These constraints will presumably limit a government’s ability to reduce greenhouse gas emissions via a carbon tax or other means

But large co-benefits can make it easier for countries to drum up support for pursuing reductions in domestic GHGs among a wide range of domestic actors, not all of whom are motivated by the spirit of global cooperation or the will to lead. Here at home, President Obama introduced the proposed Clean Power Plan, the centerpiece of his Climate Action Plan, in the asthma ward of a Children’s hospital.  Health co-benefits from reductions in local air pollution, including avoided asthma attacks, were estimated to yield approximately 60 percent of the gross benefits under the proposed Clean Power Plan. China offers another example of a country where concerns about air pollution are accelerating action on climate change (and vice versa).

Many economists will cringe at the thought of using climate change policies to address other unpriced externality problems. This is not the ideal, first-best path forward. Climate change policy is an indirect tool for addressing related but different problems of air pollution, traffic congestion, etc.  However, until efficient corrective policies are implemented, countries can and should consider these co-benefits in the design and implementation of climate change policy. This will help to mitigate domestic damages associated with the burning of fossil fuels at home while greasing the wheels of the global response to climate change in Paris and beyond.

Posted in Uncategorized | Tagged , , | 2 Comments

Are Clean Energy Tax Credits Equitable?

A new Energy Institute working paper finds that income tax credits for weatherization, solar panels, hybrids, and electric cars go predominantly to higher-income households.

Over the last decade, U.S. households have received more than $18 billion in federal income tax credits for weatherizing their homes, installing solar panels, buying electric vehicles, and other clean energy investments. In a new EI@Haas working paper, available here, Severin Borenstein and I use tax return data from the IRS to examine the socioeconomic characteristics of filers who receive these credits.


We first examine a set of income tax credits aimed at residential investments in energy-efficiency and renewables. Between 2006 and 2012 the largest categories of investments were energy-efficient windows ($4.0 billion), qualified furnaces ($2.4 billion), qualified air conditioners and water heaters ($2.4 billion), ceiling and wall insulation ($2.0 billion), and solar photovoltaic systems ($1.8 billion).

The figure below shows how use of these credits varies across income levels. We divided tax filers into six categories based on their Adjusted Gross Income (AGI). The first five categories are approximately quintiles, and then the last category ($200,000+) includes about 3% of returns.

Income Tax Credits for Clean Energy Residential Investments
Average Credit per Tax Return, By Income Level


The figure shows a strong positive correlation with income. Filers with less than $40,000 in AGI receive less than $10 in credits on average per tax return. The average credit amount more than doubles for filers with $40,000-$75,000 and then doubles again for filers with $75,000-$200,000 in AGI. Finally credits reach $80 per return for filers with AGI above $200,000. The figure above also plots 95 percent confidence intervals, though they are barely visible except for in the highest income category.

Another significant tax credit is the Alternative Motor Vehicle Credit, which provided a credit for hybrid vehicles until 2010 and continues to provide credits for natural gas, hydrogen, and fuel cell vehicles. As the figure below shows, this credit exhibits the same strong positive correlation with income. The bottom three income quintiles receive about 10% of all credits, while the fourth and fifth quintiles receive about 30% and 60%, respectively.

Alternative Motor Vehicle Credit
Average Credit per Tax Return, By Income Level


Finally, we looked at the Qualified Plug-in Electric Drive Motor Vehicle Credit, an income tax credit for electric and plug-in hybrid vehicles. The size of this credit ranges from $2,500 to $7,500 depending on the battery capacity of the vehicle. For example, the Toyota Prius plug-in hybrid qualifies for a $2,500 credit whereas the Chevrolet Volt qualifies for a $7,500 credit.

We find that this credit is considerably more concentrated in the highest income categories. As shown in the figure below, filers with less than $75,000 in AGI rarely claim the electric vehicle credit. The average credit amount jumps considerably in the $75,000-$200,000 category and then soars in the top AGI category ($200,000+).

Electric Vehicle Credit
Average Credit per Tax Return, By Income Level


Thus overall, we find that filers with AGI in excess of $75,000 receive about 60% of the tax credits aimed at energy-efficiency, residential solar, and hybrid vehicles, and about 90% of the tax credit aimed at electric cars.

We find that tax credits are less attractive on distributional grounds than pricing GHGs directly. Previous studies (here and here) have examined how a carbon tax or cap-and-trade program would impact households with different income levels. Whereas tax credits go disproportionately to high-income households, a carbon tax would be paid disproportionately by high-income households. It would seem difficult, therefore, to argue for tax credits on distributional grounds.

Our data come from individual income tax returns, so they miss tax credits received for electric vehicles and solar panels that are leased. Leasing has grown more common in both markets, though especially in the solar market with the well-documented move toward third-party ownership. However, previous research finds that the decision to lease is uncorrelated or even slightly positively correlated with income, so leasing is unlikely to undo the pronounced positive correlation between credits and income.


Why are these tax credits so concentrated among the higher income categories?

Part of the explanation is that all of these credits are non-refundable. You can use these credits to offset your tax bill, but you cannot go negative and receive a net payment from the IRS like you can with the Earned Income Tax Credit and many other tax credits. This is a significant distinction because a large fraction of filers do not have positive tax liability. In 2012, for example, more than one-third of U.S. tax returns had zero tax liability. These filers without tax liability tend to be lower-income, so this helps explain the low take-up in lower income categories. Making these credits non-refundable doesn’t make much sense. After all, what is the real difference between a filer who owes $0 in tax and another who owes $1000?  Both reduce carbon emissions when they install an energy-efficient window. Both stimulate innovation when they purchase an electric vehicle. So it seems odd to treat these filers so differently in our tax code.

Another issue is that renters are ineligible for most of these credits. Over 40 million American households are renters, and thus cannot take advantage of any of the credits aimed at weatherization, energy-efficiency, or solar PV. Addressing renters is challenging because of imperfect information and split incentives, but excluding this sector altogether misses a large share of the housing stock. The proportion of households that own a home increases steadily across income quintiles from 0.49 to 0.91 (here),  so excluding renters disproportionally impacts lower-income households.

With the electric vehicle credit there are also a couple of additional potential explanations. It may simply be that, for the moment, electric vehicles are only affordable for relatively high-income households. Even after the credit, electric and plug-in electric vehicles are expensive compared to equivalently-sized gasoline-powered vehicles. Finally, another possible explanation is that in California, electric vehicles owners are allowed to drive in high-occupancy vehicle lanes. The value of time is highly correlated with income so this could help explain why this credit is so highly concentrated in the highest income categories.


So what? Should we scrap these tax credits? Should we expand them to include more Americans? Ultimately, in evaluating tax credits or any public policy it makes sense to think about both equity and efficiency. Our new paper is mostly about equity, and the results imply that it probably does not make sense to argue for these tax credits on distributional grounds.

What about efficiency?  Although tax credits may initially seem like a good idea, they are actually quite a poor substitute for first-best policies like a carbon tax or cap-and-trade program.  Probably the single biggest limitation of a tax credit is that it cannot achieve the efficient level of usage. Take energy-efficient windows as an example.  A tax credit can encourage households to install better windows, but it cannot get households to use less heating and air-conditioning. A carbon tax, in contrast, would encourage households both to install better windows and to use less heating and air-conditioning.

Tax credits are also extremely coarse instruments. The social benefits from clean energy investments vary enormously by geography. For example, a new paper by Stephen Holland, Erin Mansur, Nick Muller and Andrew Yates finds that the environmental benefits from electric cars varies from $3000 in California (where most electricity comes from natural gas and renewables) to -$4700 in North Dakota (where most electricity comes from coal). Tax credits ignore this heterogeneity completely, whereas price-based policies would incorporate these differences.

In the end, it is hard not to be somewhat disappointed. The more we have studied these tax credits, the more we realize their limitations.  There are large potential social benefits from clean energy investments, but income tax credits are an inefficient instrument for realizing changes in behavior. Moreover, the distributional impacts are a real concern. Through several key features of the tax code, we have set up these credits in a way which excludes millions of Americans from participating, and higher-income households receive the lion’s share of total credit dollars.

Posted in Uncategorized | Tagged , , | 15 Comments

Hawai’i – The Next Frontier

hawaiian sunset

Aloha, dear readers. It’s quiet at the Energy Institute as most of us are out in the field. I just got to spend some time with a number of the world’s smartest economists on Oahu and some vacation time on Maui. Hawai’i is an awesome place. Not only because of its pristine beaches, balmy waters and glorious sunsets, but because of the energy challenges and opportunities it faces.

If you have ever flown to Hawai’i, you know it is far away from anywhere and does not have any significant local energy resources. Hence most inputs to electricity production are imported. This means mostly oil. Hawaiian Electric Industries Inc. (known as HECO) is the largest supplier of electricity, counting over 95% of Hawai’i’s population as its customers, with subsidiaries on all the major islands, with the exception of beautiful Kauai, which is served by a cooperative. And all the islands are separate grids; no transmission lines between them.

While I have pondered before what a world without coal would look like, Hawai’i provides an interesting case study.  HECO serves 300,000 customers just on Oahu, where most of the population of Hawai’i lives. Coal accounts for 9% of generating capacity, rooftop solar for 10% and oil for 65%. On Maui, Moloka’i and Lana’i, HECO serves 70,000 customers with zero coal, 29% of renewable generating capacity and the remainder coming from oil. On the big island, it serves 82,000 customers with a bigger share of wind and geothermal which results in 48% of renewable generating capacity.

Generating electricity with oil is expensive, which is why Hawai’i is leading the scoreboard for most expensive electricity in the country. The EIA quotes an average price per kWh of 31 cents! That is almost thrice the price of California’s fancy average kWh sold.

So why do I get all giddy when thinking about Hawai’i? Yes, Mai-Tai’s on the beach at sunset; but even more importantly, this is a set of islands, each of which has ample sunshine, plentiful wind, and potentially significant geothermal resources. Each island has a significant share of commercial (think hotels and restaurants) and residential customers. And electricity is already expensive. What we have here ladies and gentlemen is a unique opportunity to study smart integration of renewables on the supply side and demand side programs that go hand in hand with the rapidly growing share of renewables.

At the Energy Institute we have an impressive array of demand side studies underway in collaboration with the investor-owned and municipal utilities in California, which are integrated into the Western grid. While Jerry Brown is looking for collaboration with China, I would like to see us pay attention to what is happening half-way to Beijing!

In California we have some of the most innovative utilities in the country (I am looking at you SMUD!). The islands of Hawai’i provide us with a setting that would allow us to push our understanding of renewables integration and pricing further in a field setting. Plus, the thought of field work in Hawai’i is an appealing idea!

Posted in Uncategorized | Tagged , | 12 Comments

Do Residential Energy Efficiency Investments Deliver?

Today’s post is co-authored by Michael Greenstone (University of Chicago) and Catherine Wolfram

We recently released a paper presenting the findings of a first-of-its kind, randomized controlled evaluation of the returns to some common residential energy efficiency investments. The study’ s context is the nation’s largest residential energy efficiency program, the Weatherization Assistance Program (WAP). You can read media coverage of the paper here, here, here, and here.

For those who haven’t read about the paper, between 2011 and 2014, we administered a randomized controlled trial (RCT)—considered the gold standard in evidence—on a sample of more than 30,000 WAP-eligible households in the state of Michigan in order to shed some light on a critical question: Do investments in important residential energy efficiency measures (improved insulation, air sealing, weather-stripping, window replacement, furnace replacement, etc.) deliver the energy savings they promise?

The research revealed five main results: (1) The energy efficiency measures undertaken by households in the study reduced their energy consumption by between 10 and 20 percent on average; (2) However, these savings were just 39 percent of the average savings predicted by engineering models; (3) There is no evidence that the shortfall in savings is the result of rebound—households did not turn up their thermostats after the investments were made; (4) While the investments cost roughly $4,580 on average, our best estimate of the energy savings was about half of these costs[1]; and (5) The costs also greatly exceeded the benefits when the monetary value of pollution reductions are added to the energy savings to calculate benefits. While the WAP program has a number of goals, when measured by the energy savings and emissions benefits, these efficiency upgrades were not a good investment.


The urgency of the climate challenge means that it is critical to identify cost-effective strategies that will deliver real greenhouse gas emissions reductions. Energy efficiency is a crucial component of most climate change mitigation plans, underscoring the importance of developing a body of credible evidence on the real-world—versus projected—returns on energy efficiency investments in the residential sector and beyond.

Such a process will undoubtedly uncover some gems, but in some instances it will also be necessary to update our beliefs. When seemingly inconvenient evidence comes to light that challenges our beliefs—as we have uncovered with this analysis—that data should not be undermined and ignored.  Instead, it should be used to inform our strategy to confront climate change. The magnitude of the climate challenge requires that we ruthlessly pursue the most cost effective mitigation options.

Our paper has generated some strong reactions and important questions, some the result of misconceptions about what exactly we evaluated and how the evaluation was conducted. In the remainder of this blog, we respond to the most common criticisms of our study and its findings.

*              *              *

Reaction 1: This is just one study and scores of other studies have opposite findings.

Some critics have cited prior evaluations showing that residential energy efficiency programs are good investments and that our study is an anomaly. Many of these evaluations, however, are based on savings projections that- as we found- can significantly overestimate the savings when applied in the real world. Other studies use real- world data, but analyze these data using methods that can confuse the effects of energy efficiency improvements with other factors that drive changes in energy consumption.

Our study is different. It represents a first-of-its-kind evaluation using a randomized controlled trial, the gold standard for rigorous evaluation. Society routinely relies on this methodology to assess the efficacy of new drugs, treatments, and other interventions. This approach is increasingly used in the social sciences, including criminology, education, development economics, and energy economics. In many instances, the application of randomized control trials has changed the conventional wisdom. Our application of this approach to residential energy efficiency measures is therefore an important departure from, and improvement upon, previous analyses.

Reaction 2: The study unfairly paints WAP as an ineffective program.

WAP has multiple goals and improving the living standards of its recipients is clearly a central and worthy one.  Our study does not claim to provide a comprehensive evaluation of WAP, nor would it be appropriate to do so.

Rather, the study’s purpose is to measure the real-world energy savings resulting from WAP-funded energy efficiency improvements.  We then compare them to both the investment costs and the projected energy savings generated from detailed energy audits.

In interpreting the results, it is important to bear in mind that for a measure to be implemented under WAP, federal regulations require that it pass a cost-benefit analysis—that is, the projected cumulative energy savings must be greater than the investment costs. This cost-benefit analysis is based on an in-home energy audit conducted using an engineering model, in this case the National Energy Audit Tool (NEAT).

For the households we studied, NEAT-driven audits projected that the WAP measures would reduce annual energy consumption by 43.7 million British thermal units (MMBtu). Yet, when we observed the energy bills of households that received WAP measures, the actual energy savings were just 17.2 MMBtu. In other words, the model systematically over-predicted energy savings by a factor of 2.5.

This is an important finding. The investments in efficiency in our study underperformed relative to projected values and in a way that the program was expressly designed to avoid. Homeowners, program managers, and taxpayers only received 39 percent of the projected savings. According to the Department of Energy, the NEAT model is used by approximately 700 state and local Weatherization Assistance Program subgrantees in more than 30 states.

Broader program objectives notwithstanding, WAP is a compelling setting to learn about the returns to energy efficiency investments. WAP is the nation’s largest residential energy efficiency program. According to the Department of Energy, which administers the program, more than 7 million homes have participated in the program since its inception in 1976. If one is attempting to assess the performance of commonplace residential energy efficiency investments on a large scale, there may be no better option.

Reaction 3: The study’s calculations of costs and benefits are inaccurate.

Here again, it is important to note that we recognize WAP has benefits beyond saving energy. But, the intent of our study was focused solely on evaluating the energy-related (and associated emissions) costs and benefits. We never claim to evaluate the other benefits of these upgrades, as that is beyond the scope of our study. It is also important to note that, no matter how one decides to evaluate monetary costs and benefits, a central finding of our analysis remains unaffected: efficiency upgrades delivered just 39 percent of the energy savings they promised. It is therefore challenging to find a set of assumptions (e.g., about lifespans and discount rates) that would cause these efficiency investments to have energy savings and emissions benefits that exceed their costs.

To drill down a bit more, here are some of the criticisms of our calculation of the costs and benefits and our responses:


Some have argued that it is inappropriate to factor in costs that don’t directly lead to energy savings.  As anyone who has done home repairs knows, once you start down the path to do something like lay new insulation, additional costs are necessarily incurred. For example, weatherization can reduce indoor air quality by tightly sealing a house, so additional costs may be required to maintain indoor air quality. Separating what’s required to lay the insulation from what’s completely separate is not easy. The average household in our sample received approximately $4,600 in energy efficiency upgrades, which includes roughly $800 in costs required to make installation of the weatherization measures safe and functional, such as wiring upgrades. Our judgment is that the most reasonable assumption is to include all of these installation and materials costs. It is worth noting, however, that if we take the polar opposite view and exclude all costs that do not directly result in energy savings, the average cost per household still significantly exceeds our central estimate energy savings.

Moreover, there are other costs associated with these retrofits that are not reflected in our cost-benefit comparison. For example, we do not include any program overhead or administrative costs. Nor do we account for the hassle and effort that households expend to implement a weatherization retrofit, even one with zero out of pocket costs. An earlier blog makes the point that these process costs can be large (we found that it cost $1,050 per weatherized household to encourage take up of these measures). Accounting for these additional expenses would of course widen the gap between costs and savings.


We measure benefits by calculating the net present value of annual energy savings using a range of discount rates (3, 6, and 10 percent) and investment lifespans (10, 16, and 20 years). Our estimate of the benefits also includes an estimated upper bound on the benefits households derive from increased warmth (based on our analysis of “rebound” in demand for heat in the winter). In no case does the present value of energy savings reach parity with actual costs, even if we ignore the indirect efficiency-related improvements.

In calculating program benefits, we used real 2013 residential energy prices for electricity and natural gas in Michigan and assumed that these figures would increase at the rate of inflation over the lifetime of the investments. While some have criticized this as too conservative, it is standard to  use current energy prices as a predictor of future energy prices.

Reaction 4: The results cannot be generalized because they only relate to one part of Michigan, to one program, and to one subset of the population.

We study a subset of low income households in Michigan undertaking a particular set of residential efficiency measures recommended by NEAT. However, minimizing the significance of our findings on account of this context ignores the ubiquity of the measures we analyze and of the reliance on audit tools like NEAT.

As noted above, the households in the sample we studied were subjected to the same measurement tool that is used by residential weatherization programs throughout the country to gauge which upgrades are the most cost effective; and all implemented measures had to pass the same cost-benefit analysis. The types of upgrades installed at the WAP households in our sample (e.g., furnace upgrades, improved insulation, and weather stripping) are commonplace for home retrofits for all income groups.

Drawing implications from a study is not an all-or-nothing proposition.   For example, the results of a randomized controlled trial studying the effectiveness of a given drug or treatment on middle-aged men will in some instances tell us everything we need to know about its effectiveness on young women.  Of course, in other instances, less decisive conclusions are warranted until further research is conducted.

Our study tells us that a common set of efficiency measures installed in the low-income households we studied in Michigan did not deliver the expected energy savings, and that investment costs significantly exceeded these savings. Given similarities between the setting we evaluate and other efficiency applications, these findings likely generalize to a broader set of residential efficiency investments. There is logic behind this implication, while also acknowledging the need for  further experiments on the returns to energy efficiency investments  in other contexts  (Indeed, we have already begun to do them and, in at least one case, our preliminary results are qualitatively similar).

Reaction 5: The study period covered a time when the program experienced a significant increase in funding that led to poor results (e.g., inexperienced contractors).

The time period we studied included an unprecedented number of weatherizations because the American Recovery and Reinvestment Act (ARRA) increased the amount of money allocated to the program dramatically. As a result, some say new, inexperienced contractors were called in to do weatherizations and their work may not represent the norm.

To investigate this possibility, we compared savings at homes where the contractors were experienced to homes where the contractors weren’t experienced and found no difference in the average energy savings. Consequently, we find no evidence that inexperience during the time period played a role in explaining the lower-than-expected savings.

[1] This blog focuses on a subset of the numbers and results reported in the paper. Here we emphasize our preferred estimates from the randomized controlled trial that estimates average impacts for the subset of households whose participation in weatherization was the result of random assignment to our experimental intervention.  These households are associated with somewhat lower average costs ($4,580) as compared to the larger sample of recipient households from whom we collected data for our quasi-experimental analysis ($5,150).

Posted in Uncategorized | Tagged , | 24 Comments

Winners and Losers from Flattening Tiered Electricity Prices

The California Public Utilities Commission is moving closer to major changes in the steeply increasing-block residential electricity rates that the state has had since the 2000-01 California electricity crisis.  This Friday the Commissioners may decide to significantly flatten the tiered rate structure.  In a blog post last fall, I discussed the inefficiency of charging tiered prices that don’t reflect cost and the unfairness of charging different customers different prices for the same good – a kilowatt-hour (kWh) of electricity.

In that blog, I also addressed the three standard arguments that defenders of such steeply tiered pricing commonly put forward.  The first is that increasing-block pricing yields conservation.  While in theory this could happen, the best empirical work on this subject, by (my former student) Professor Koichiro Ito shows that it is likely to have about zero effect on overall consumption.  It does encourage high-consumers to consume less, but it also encourages lower-using households to consume more.  Professor Ito shows that the net effect is no reduction in overall consumption.

The second argument is that supplying electricity to high-use households is more expensive per kWh on average, because they consume more at peak times.  My own research has shown the difference is so small that it would justify less than a one-cent differential in price between high-use and low-use customers.

The third argument carries the most weight, that higher-use customers are on average higher-income customers.  That’s true, as I showed in research published in 2012, using household consumption data from 2006.  However, like most states, California has a separate tariff for the lowest-income customers: households up to 200% of poverty-level income are eligible for the CARE program and pay much lower rates.  With the CARE program now covering about 30% of all residential customers, is tiered pricing in the standard residential rate an effective way to help lower-income households?

My 2012 paper just showed the average bill change for households in each income bracket, not the distribution of changes within each bracket.  There is another issue of equity if a program designed to transfer wealth from high-income to low-income households actually does a poor job of targeting either group, harming many low-income and/or benefitting many high-income customers.  This week I went back to the usage data from 2006 to see how great of a concern that should be.

Using PG&E data, I applied my earlier work on usage and household income to current electricity rates and the flatter tariff that a CPUC administrative law judge has proposed.  I estimated the range of impacts the proposed change would have on households across the income spectrum.


The figure above shows Pacific Gas & Electric’s current residential rate and an alternative rate that would raise the same revenue, but would have only two tiers with a 20% increase between them, as the ALJ has proposed.[1]  The first thing to note is that the new rate would be higher for consumption out to what is currently the third tier, the point where the two lines cross at 130% of baseline quantity.  In order to be a “structural” winner with the flatter rate – that is, paying less without changing consumption at all — the household would have to be consuming out beyond the crossing point in order to offset the higher marginal prices for the lower kWhs.

In this case, the breakeven point is at 216% of baseline quantity.  The median household consumes about 130% of baseline quantity, so that means most households would pay more.  By my calculation about 21% of households would save, while about 79% would see a bill increase if no one changed consumption.   This reflects the fact that since the electricity crisis the great majority of price increases have been placed on the highest-use customers, resulting in the steep tiers.

Using census data and applying a statistical matching method I developed in the 2012 research, I estimated into which of 5 income bracket each household falls.   I focused on the customers who are not on CARE, because CARE households are on a separate tariff.   The 5 income brackets are based on census categories that are roughly equal parts of the entire population.  As the table shows, however, in this analysis the bottom two brackets are smaller due to the substantial participation in CARE.

Income Bracket 1 2 3 4 5
Income ($2015) Under $27,400 $27,400-$54,800 $54,800-$82,200 $82,200-$137,000 Over $137,000
Share of non-CARE customers 2% 16% 23% 31% 28%
Average change in Monthly Bill with 2-tier tariff $5.84 $4.59 $2.40 $0.70 -$5.78
Percent Structural Winners 5% 12% 18% 22% 32%

Next, I calculated how much the average monthly bill of each household would change if the rate were changed to the two-tiered structure in the figure above (and the household did not change its consumption).  The third row of the table above shows the average bill change for households in each income bracket.  Not surprisingly, because lower income households consume less on average, they are more likely to see their bills go up with the change.  But even in the highest income bracket more than two-thirds of customers would see bill increases.

What was particularly surprising to me is the figure below, which shows the distribution of the change in bills for each of the income brackets.  The impacts across brackets are more similar than I expected.     In every income bracket two-thirds or more of the customers see their bills rise between $0 and $20 per month, even in the highest-income bracket.  Are these high-income households that don’t use much electricity super-conservers?  Maybe, but I bet many of them are households with only one or two people, who work and travel a lot, and don’t spend much time at home.


About 4% of households are the biggest winners with bills dropping by at least $50 per month under the proposed tariff.  Are these “energy hogs”? Maybe, but I bet some of them are big families and people who are home all day, because they are retired or have small children.   Among these biggest winners, I estimate that slightly less than half are in the highest-income bracket.

Undoing the steeply tiered rates that were created during California’s electricity crisis will on average cause lower-income households to pay more.  If there were no other consequences of the steep tiering, I could see keeping it on that basis.  But there are other impacts on both efficiency and fairness, not the least of which is the monthly harm to higher-consuming, middle- and lower-income households that is caused by a rate structure that has no basis in costs.

Some opponents of the two-tiered rate proposal have presented it as a simple shift of payments from poor to rich.  This analysis shows that the story is not that simple.  Both winners and losers are present at every income level.  The two-tier proposal makes bills more cost-based and more proportional to usage, as they were before 2000.  And as they are in nearly all other states, and in the parts of California served by municipal utilities.

I’m still tweeting energy news articles and new research papers @BorensteinS 

[1] All of these calculations assume no consumption response to the tariff change.  As I show in my 2012 paper, accounting for the small elasticity that has been estimated for response to a change in increasing-block pricing makes very little difference to these calculations.

Posted in Uncategorized | Tagged | 18 Comments

How Should Distributed Generation be Distributed?

Growth in the residential solar market continues apace. In the United States, residential solar PV installations last quarter were up 11 percent over the previous quarter:



The figure  illustrates this impressive growth rate (in dark blue). However, this is growth on a very small base. By my crude calculations, less than half a percent of American households currently have solar panels on their roof.[1]

In those states where residential solar is starting to take hold, there are mounting concerns that rate structures currently in place to support residential PV will result in adopters bearing less than their fair share of system costs. If increasing levels of distributed solar generation puts additional pressure on grid equipment and aging infrastructure, these concerns loom even larger.

A new EI working paper takes a close look at how increasing levels of distributed solar generation can impact power system costs. For me, this paper raises a timely question: should we be paying more attention to where distributed generation gets distributed?

Bill savings for the adopters

Before diving into the details, let’s first review the basic issues.

If you have invested in putting solar panels on your roof, chances are the solar electricity you generate is valued at retail rates one way or another. This is thanks to net metering policies adopted in most states. If you consume the electricity you generate, you avoid purchasing electricity from your utility. If you do not use all the electricity that your solar panel is generating (it is estimated that almost half of electricity generated by net metering customers in California is exported), you can export the extra power to the grid and count it against consumption within the same billing period.

As we’ve discussed before on this blog,  the marginal retail price can significantly exceed the direct energy costs of producing a kilowatt-hour centrally. On the one hand, many of the costs that are reflected in retail prices (e.g. metering, billing, and infrastructure costs) are not avoided when you put solar panels on your roof. On the other hand, solar PV generates benefits that are not fully reflected in market prices.

So what is the right price for distributed solar generation?  Past blog posts have touched on some elements of this value-of-solar calculation that fall squarely in the purview of economists. But there are other important elements that push outside the boundaries of economics.  This week, we venture into the engineering-meets-economics world of distribution system costs.

I am no an engineer, but I am fortunately married to one, who is a co-author of the new EI working paper. Much to the chagrin of my kids (who’d rather be talking Frozen or fire trucks), I have been steering the family dinner table conversation towards this paper which looks at how distributed solar affects the electricity distribution system. The findings should be of interest to energy economists and engineers alike (but not so much three and five year olds it turns out).

The distribution system meets distributed generation

A quick summary of what I’ve learned at my dinner table.

If you install solar panels on your roof, this will impact how power flows through the distribution system that delivers power from high voltage transmission networks to the people in your neighborhood. The cartoon below helps to fix ideas.



Some of these impacts can reduce costs. For example, less electricity flowing into your neighborhood during peak times can reduce pressure on aging infrastructure (e.g. distribution lines, service transformers) and defer the need to invest in distribution system upgrades. That’s good. But increased PV penetration can also increase the need for investment in hardware such as voltage regulation because distribution systems are not designed to handle power flowing from customers back to the substation.  Not good.

The cartoon above shows a single “feeder” (a collection of distribution lines that carry power from the high voltage transmission system to customers). Using detailed data on all 3,000 feeders operated by the largest utility in California, PG&E, Duncan and coauthors simulate how increased solar PV penetration on each feeder would impact the need for system capacity upgrades, expenditures on voltage management, etc.

On average, they find that the levelized value of deferred investment in distribution system upgrades (avoided costs) is small: around half a cent per kWh. This is in line with rough estimates found in other reports that use highly aggregated data.

But the advantage of disaggregated data is that they can look beyond the average. It turns out that this economically insignificant average value obscures tremendous variation in feeder-specific capacity values. Capacity values are zero across a large majority of feeders where no capacity upgrades are anticipated over the next ten years. But for approximately ten percent of feeders, the picture looks quite different.

The figure below focuses on the 298 feeders where capacity upgrades are anticipated  in the next ten years under a business-as-usual scenario. This represents about 20 percent of the total capacity, or approximately 1 million customers.


The figure shows that estimated capacity values exceed $60/kW-year (or $33/MWh using their discounting and electricity production assumptions) for approximately 30 feeders (this assumes a solar PV penetration rate of 7.5 percent).  This is almost on par with the energy value.  The median capacity value in this select group exceeds $20/kW-year ($11 per MWh).

Although there has been much hand wringing over the potential for voltage regulation problems,  the authors find that these problems are actually relatively small.  Using PG&E’s current budget for repairing voltage regulating equipment, they estimated that even in extremely aggressive scenarios for PV deployment, the total costs to ratepayers would be less than half a million dollars a year.

Distribution matters

On average, this study finds that the average (distribution system related) net benefits of distributed solar PV are not very significant.  However, looking beyond the average, the value of deferred investments in distribution system infrastructure associated with a given level of distributed generation depends significantly on how these resources are distributed on the system. In other words, the net costs of future distributed generation could be significantly reduced if these resources are targeted to areas where they can generate the largest benefits.

There is precedent for targeting energy efficiency to defer investments in transmission and distribution system upgrades… why not solar PV?  If the next generation of distributed solar incentive programs and resource planning protocols reflect  the impacts that these resources could have on different parts of the distribution system, the next generation of distributed resources can be more efficiently distributed.


[1] To estimate the number of systems, I divide GTM research estimates of installed residential PV capacity in 2014 by the state-specific average residential system size (as reported in Tracking the Sun, 2014). I then divide this by the US Census estimate of the number of US households in 2014. Thanks to Naim Dargouth and Snuller Price for the solar PV numbers!

Posted in Uncategorized | Tagged | 11 Comments

Is the U.S. Investing Enough in Electricity Grid Reliability?

We had a 2-hour power outage at our house last week, together with 45,000 other customers in the East Bay. The lights flickered off just after 8PM and didn’t come back on until after 10PM. Nothing like going without something that you take for granted to make you realize just how valuable it is.

The East Bay outage was reportedly caused by a squirrel

The East Bay outage was reportedly caused by a squirrel

My son and I had fun gathering our candles and figuring out that our hand-crank radio played Mariachi music, but that only lasted for about half an hour. As the minutes ticked by without WiFi, the economist in me started thinking about just how much I would be willing to pay to get the electricity back. I had a meeting the next day to prepare for, and it was my turn to take a pass through the slide deck. I couldn’t even get good enough cell service to download the presentation to my phone, perhaps because local cell towers were also affected by the outage.

The beauty of the free market is that it allocates resources to the sectors of the economy where they are most valued. (Yes, I’m beating the economics drum, but this is econ 101 – we ALL agree on this one, even the two-handed economists.) If enough customers value a good highly and it’s inexpensive to produce, an innovative entrepreneur can make money by figuring out how to sell that good to consumers.

So, most goods and services that people value more highly than it costs to provide them exist, and things that aren’t valued don’t exist. The market supplies frozen pizzas and smart phones, but not condos in space, because they’re super expensive and not, currently, in high demand.

frozen pizzaThings are different with electricity. Given that the majority of the world’s citizens get electricity from some kind of regulated or state-owned monopoly, we’ve basically given up on using the market to figure out how much people value electricity reliability. So, regulators and the regulated companies are left guessing how much customers are willing to endure higher prices to cover a more robust system.

My personal hypothesis is that we have gotten this wrong in the U.S. I suspect we’re underproviding reliability and spending too little on making the grid more secure.

Even in areas of the U.S. that have restructured (or, what we used to call “deregulated”) their electricity industries, the distribution system remains regulated. Most outages are caused by failures at the distribution system level. Further, in most restructured wholesale markets, generation reliability is impacted by regulatory decisions on things like reserve margins.

Yes, there are many parts of the developing world where (only!) 2 hours without power is not a good day but an extraordinary day. But, there’s another side to the spectrum. Germany and other parts of Europe have much more reliable electricity systems than the U.S.

I first heard this anecdotally from a friend who grew up in Germany and said he could remember one outage throughout his entire childhood. The table below shows that his anecdote is true generally.


Source: Galvin Electricity Initiative report, Table 1.

Being on top of this list isn’t good. Larger values of SAIDI (System Average Interruption Duration Index) and SAIFI (System Average Interruption Frequency Index) indicate less reliable power. Roughly, SAIDI reflects the average number of minutes per year that customers are without electricity and SAIFI reflects the average number of outages customers experience per year. Americans endure 10 times as many minutes of outages compared to Germans.

stormRecent work from Lawrence Berkeley National Labs (LBNL) suggests that, if anything, reliability has been getting worse in the U.S. over time.

If the regulators in both Germany and the U.S. were doing a good job approximating market outcomes, these vast differences in the amount of reliability would suggest that either the German utilities can provide reliability at a much lower cost or that German customers have much higher demands for reliability. My guess is that neither of these things is true. The electricity systems are very similar, so I don’t think Germans are using a radically different technology to drive their costs down. Maybe Americans live in areas that are more exposed to storms, but 10 times more exposed seems implausible.

Why do I think the U.S. is spending too little on reliability and not that Germany is spending too much? At a very macro level, estimates of the annual economic losses from electricity outages are very high, ranging from $20 billion to $150 billion annually. This seems like a lot of lost productivity and I would hope there are relatively inexpensive investments we can make in the grid to avoid these losses. Also, as I have blogged about earlier, to the extent we can back out how much regulators think customers value reliability, the estimates seem low.

Is Elon Musk going to solve this for us? In the post-Powerwall world, people who value reliability highly can vote with their pocketbooks and spend $3,500 to get a battery backup that will deliver 10 kWh each time there’s an outage. From what I’ve read, they’ll spend another $3,500 on installation and the ancillary equipment, like a smart inverter. Someone I spoke to recently who didn’t like outages was looking forward to installing a Powerwall, although he is a senior employee of a large tech company and probably thinks about $7,000 investments the way most of us think about spending $50.

Let’s run some quick numbers on the Powerwall. Let’s say it costs $7,000 for a 10kWh battery, which I assume you use for four 2-hour outages per year. According to the table above, the U.S. average is 240 minutes of outages across 1.5 events, but let’s think about people who are experiencing many more outages than average. The Powerwall is supposed to last for 15 years, so at a 5% real interest rate, the rental cost of capital is about $675 per year to get 10 kWh 4 times per year. This amounts to almost $17 per kWh. Given that average U.S. customer pays 12 cents per kWh, that’s a SUPER expensive backup.

powerwallFinally, it’s not clear to me that having a Powerwall at your house will deliver the kind of reliability we really want. In our highly networked world, it’s possible the outage will disable other services. If the battery backups on the local cell towers run out, it could be hard to make calls.

In short, while the Powerwall might satisfy the demand for reliability for a handful of very wealthy or very outage averse U.S. customers, I suspect it will leave a lot of unmet demand. Plus, if we’re just talking about backup electricity, it’s not even clear that the Powerwall fills a niche that a diesel generator didn’t already fill, though it does look sleek.

We have a lot more to learn about reliability. This post makes some assertions that I would love to see substantiated with hard evidence! But, as the LBNL folks point out, we currently don’t even collect very good data.

The good news is that new technologies seem poised to deliver better information on reliability and to give us new ways to enhance the electric grid. But, whether utility companies and regulators have the right incentives to use this information to ensure that systems are delivering the correct amount of reliability is an open question.

Posted in Uncategorized | Tagged | 37 Comments

A Deeper Look into the Fragmented Residential Solar Market

(Today’s post is co-authored with Alex Chun, who just received his MBA at Haas and is an alumnus of our Cleantech to Market class.  Alex is the Business Intelligence Manager at Sungevity.)

Who sells residential solar photovoltaic systems (PV) in California? How many companies operate in this market?  What fraction of the market is controlled by the largest companies? How is this changing over time? In today’s post we look at data from the California Solar Initiative (CSI) to better understand the structure of this market.

Since 2007, 1.9GW of residential solar PV capacity have been installed under the California Solar Initiative. One of the truly laudable features of the CSI program is that data from the program has always been made publicly available. This has given market participants an unprecedented view into the market and spurred innovative research like the paper Severin Borenstein blogged about two weeks ago. For today’s post, we used the CSI data to look at the spectrum of companies working in this market.

Market Share by Company, Top 30 Companies 2013top30

We first looked at market share by company in 2013. The first thing that becomes apparent when looking closer at the CSI data is the high degree of fragmentation in this market. The largest firm, SolarCity, sold almost 18% of all installations in 2013, but after that market share falls precipitously. The top five companies together account for only about 40% of the market and beyond Solar City and Verengo at 9%, no other company comes close to having a double-digit market share.

Also notable is the long right tail. This is apparent among the top-30 sellers and can also be seen in the figure below which shows market share for all companies including those outside the top 30. The rainbow of tiny vertical slices shows that beyond the top five, this is a market made up of small companies, each with only a tiny market share.

Market Share by Company, All Companies 2013all

Before we go any deeper, an important caveat is that the CSI data include only those installations for which the household received a CSI subsidy. If you installed solar PV and received the Federal tax credit, but not the state subsidy, then you are not in the CSI data. This was not a large category a couple of years ago because most solar PV customers in PG&E, SDG&E, and SCE opted to participate, but the subsidy amounts have decreased over time and in 2013 the CSI covered only 22% of all new California residential solar installations. Data from Greentech Media suggest that the CSI data are approximately representative of the broader market, for example with similar market share for the top-five companies. Moreover, in 2012 the CSI covered 48% of all new California installations, and results (here) are very similar. Nevertheless, you have to be very careful interpreting these data, particularly because the fraction of the market covered by CSI has decreased so much over time.

The Long Right Tail

We next constructed a figure to look at installation volume by company. In 2013, there were a total of 840 companies who installed systems that received CSI rebates and the median volume installed was 21KW per company.  A typical system is about 4KW, so the median company installed only about 5 systems during the entire year. Even the third quartile company installed only 79KW.

Distribution of Installer 2013 KW Volumevolume

Solar City, in contrast, installed 28,000KW, approximately 1000 times the median volume. Solar City, Verengo Solar, and REC Solar look like outliers when compared with the rest of the market. The line at the bottom represents the median, 25th, and 75th percentiles. With this scale, three-quarters of the companies in the market are nearly imperceptible at the bottom of the figure, underscoring the fact that while there were 840 companies competing in the market, most of them were operating at a tiny fraction of the scale of the biggest players.

Market Consolidation Since 2010

While the market remains highly fragmented with large numbers of small installers, the CSI data show that over the past three years there has been a considerable amount of consolidation.

Installer Count Over Timeovertime

Between 2007 and 2010, the number of companies in the market increased by about 200 per year. It then peaked at 1,050 in 2010. Since 2010, however, the number of companies has decreased every year with an overall ~25% drop from 2010-2013.  While it is possible that this pattern is driven by the decreasing coverage of the CSI data, the evidence is suggestive of a consolidating market.

We also looked at the breakdown of new entrants versus surviving companies. The green line below shows the fraction of companies each year who were not in the market during the previous year. From 2008-2010, nearly 50% of the companies in the market were not in the market during the previous year. This is a remarkably high level of entry and reflects that this is a rapidly growing market. Since 2011, this “churn” has decreased somewhat with only about 30% new entrants each year.

New Entrants vs Previous Year Survivorssurvivor

The figure also shows the number of new and surviving companies in the market each year. Starting in 2010, the number of survivors (companies who were in the market the previous year) started to flatten out at around 600. Meanwhile, the number of new companies entering the market decreased significantly in 2011 and appears to have further decreased in 2013.

Long-Term Market Structure

It is too soon to say, but we may be headed toward a more consolidated solar PV market with fewer companies controlling larger market shares. In his Business Strategy class, Haas Professor Ned Augenblick discusses the five factors that lead to a sustainable strategic advantage. One of us took this class recently and can still remember all five:

  • Network Effects
  • Switching Costs
  • Restricted Access to Resources
  • Economies of Scale
  • Economies of Scope

Companies in the residential solar market do not benefit from any of the first three factors. Consumers do not receive an additional benefit from going solar from the same company as other people in their network. Sales in the residential solar market are a one-time sale, thus preventing any switching costs, and there is no restricted resource that would give companies in this market a sustained strategic advantage. To the contrary, the large number of smaller companies in this sector suggests that barriers to entry are minimal.

In the absence of these sustainable strategic advantages, the solar market does not have the winner-take-all dynamics that will drive it naturally to an oligopoly market in the same way as, for example, the cell phone OS market. Thus, consolidation will be driven solely by economies of scale and economies of scope. Where potentially there could be economies of scale is in the “soft” costs like marketing, customer acquisition, permitting, and inspections. Companies are hoping that the larger they become, the more they will be able to spread these costs across customers. Moreover, economies of scope could become important as companies take advantage of their growing customer bases to sell related products and services.

For market observers, understanding these underlying economics and market trends is important because it provides an idea of where the market is going and what the long-term market structure will be. For companies operating in the space, however, it’s even more critical.  Understanding the scale necessary to capitalize on economies of scale and scope and hitting this volume can be the difference between surviving and exiting the market. The CSI data doesn’t have all the answers, but it does suggest that the market has been consolidating over the last couple of years. If economies of scale and scope end up being important, then we would expect to continue to see smaller companies exit the market and the remaining companies increase their share of the pie.

Posted in Uncategorized | Tagged , | 3 Comments

Exiting Coal?

On March 11, 2011 I was sitting in a coffee shop in Berlin, dressed appropriately in a black turtleneck and leather jacket, reading about the terrible Fukushima Daiichi Nuclear disaster. The next day I read that the German government was pushing for “Atomausstieg,” which is German for “let’s retire all nuclear generating capacity.” 80% of Germans surveyed were in favor of this move. The nine remaining German nukes are being phased out and the last one will shut its doors by 2022.

The Energiewende Law, which was proposed only months before the Fukushima disaster, was enthusiastically approved in 2011 and has led to rapid growth in the penetration of solar PV and wind power across Germany, as the advertising below indicates.


While there is no way to establish causality here, no one can argue with the fact that the installed cost of PV has come down by 66% in a decade. And the creation of the German market could have had something to do with this. In 2012 Germany (1.1% of the world’s population) had 32% of installed solar capacity globally, according to government figures. And capacity continues to grow – 2014 installed capacity was 113% above that in 2010 suggesting a 21% growth rate p.a. This has come at a cost. While owners of PV installation have to pay for some of the cost of the solar panels privately, the average German household now pays about 260 Euro per year to subsidize renewables, which is nothing to sneeze at. But it’s also not the end of the world as some have suggested (about the equivalent of a Starbucks latte twice a week, which unlike the renewable subsidy, does not come with a green halo). The Energiewende enjoys less, but still strong, public support. So now the government is starting to contemplate what to do next to achieve its ambitious emissions reduction goal of 80% by 2050.

germany energy long

Sigmar Gabriel, who is Angela Merkel’s Energy minister, has started talking about something called “Kohleausstieg” (German for Coal Exit). When visitors from Germany to the Energy Institute lunch table mentioned this, I thought I misheard. But I did not. There is a slowly emerging vision of the German energy system, which will no longer have domestic baseload generation. Just say Nein to coal and nukes. This is fascinating. Let’s take a look at what estimated power supply looks like in May 2012 versus 2020:

german energy power demand

What we are seeing here is the huge variability in generation of renewables, which of course does not line up quite as beautifully with demand as has been pointed out elsewhere. This picture also shows nicely that by 2020 renewables are generating more power than is demanded (at least on the weekends). And if the installation trend continues, this will be true for most weekdays, too.

This means that we may not need the always-on baseload (coal and nuclear in most places). In one version of the world you use fossils that ramp up quickly to meet residual demand (e.g., gas from Russia). In a second version of the world you use clean hydro power from Northern Europe instead. In a third version, which is the one Elon Musk would like you to consider, you use a giant battery in your house, which stores renewable power at times when there is plenty of it to be had for cheap (requiring a pricing revolution).

I am a confessed hip-/techie. I like the last version of the future. But I have some questions.

  • Is Germany this bold since it can always buy cheap nuclear baseload from France if things go terribly wrong? What if you are a country like the US, where you do not have this type of backup at scale?
  • What about the political economy of a coal exit? Coal mining unions are very powerful and this would put a lot of people in poor areas out of jobs. And miners will not go into installing PV panels on people’s roofs, since the sunny rich areas are not usually where the coal mines are.
  • How much storage do we need to make this work? I can see a residential model, where Elon Musk sells me a battery or my car serves as storage. But what about BMW, Porsche, and Intel? Will we come full circle where firms will have their own fossil backup generation (which is the case for most manufacturers in China currently)?
  • What if the major players exit coal? That shift in demand, drives down price and leads to consumption elsewhere. In order to make this work you would have to exit coal and find a way to leave what you don’t consume in the ground.

While writing this blog post I was surprised by how similar California’s and Germany’s energy policies and challenges are. Both places are pushing hard for an almost fossil free future using a combination of market based policies and huge number of competing standards. Both places have political leadership proposing radical long range policy targets, which we do not necessarily know how to achieve. Both places are relatively wealthy. Both places have industries that have been at the forefront of technological innovation, especially in the STEM fields.

Germany, specifically, has been at the forefront of pushing new distributed generation technologies and shouldering much of the cost of the global energy transition. This is laudable. California is along for the ride and doing its part. It looks like we might be the ones leading the charge on designing cost effective storage. Thanks Elon. While I don’t think a coal free Germany is necessarily an unrealistic idea, I want us to keep our eye on the prize. What we should shoot for are drastic global reductions in CO2. Germany and California are small. If what comes out of our policies is a way to drive coal and natural gas up the merit order in places like China and India, this would be the real success.

Posted in Uncategorized | Tagged , , | 41 Comments

What Put California at the Top of Residential Solar?

California leads the nation in residential solar photovoltaic installations.  In fact, nearly half of all systems installed have been in the Golden State.

So why is California the leader?  Sure, California has plenty of sunshine, but there are many other states that can compete on that dimension, including Florida, the Sunshine State.  It’s not the federal tax benefits, which are available to all US residents.  It’s not California’s Renewables Portfolio Standard, which effectively excludes residential solar.  Some point to the California Solar Initiative that gave rebates for new systems from 2007 to 2013, and that is surely part of it.

But another factor is the “solar friendly” residential electricity prices.  Not only do California’s two largest utilities have some of the country’s highest average residential electricity prices, the rates are also tiered, meaning that they increase for additional kilowatt-hours as the household consumes more over the month.  As a result, large users face rates for much of their power that can be three times higher than rates in many other states, including the Sunshine State.


Have the level and structure of retail rates been a major factor in California’s residential PV boom?  I’ve been wondering that for a while, so in the last few months I’ve been sizing up the various solar incentives for customers of Pacific Gas & Electric, the state’s largest utility, which has by far the most residential rooftop solar capacity in the country.  The result of this work is being released today in a new Energy Institute working paper, “The Private Net Benefits of Residential Solar PV: And Who Gets Them”.

Using data on PG&E households that installed solar from 2007 to 2013 (and for some data, into early 2014), I examine the collection of incentives that were available, whether the system was bought by the homeowner or owned by a solar company, known as third-party owners (TPOs).  TPOs can lease the panels to the homeowner or agree to sell the electricity the panels generate under a power purchase agreement that specifies the price per kilowatt-hour (kWh), usually for 20 years.   I then put all these incentives together with reported prices of the systems to calculate the net benefits.

The incentives include direct rebates and tax credits, as well as indirect incentives from the structure of retail tariffs and the credit for electricity grid injections from the panels under “net metering” policies, as I’ve discussed in an earlier blog.

To start with the easiest ones, the California Solar Initiative was offering $2.50 per watt rebates back at the beginning of this period – when the full systems cost around $10 per watt on average.  The CSI rebate stepped down over time, eventually hitting $0.20/watt in 2013 just before it disappeared.  In the first half of 2014, the average full system price was down to around $4.50/watt.


If you bought the system, you got the CSI rebate.  With a TPO, the company that owns the system got the rebate and — I hope — you got a lower price reflecting at least part of that savings. In either type of transaction, how much the price adjusted to pass through the savings to the homeowner, or how much the installer captured, is a point of strong dispute.  Different analyses have estimated 17%, 45% and 99% passthrough rates to homeowners.  Unfortunately, my study can’t unpack that even in the simple case of system purchases, let alone with much more complex lease or power purchase agreements.  I estimate the incentive the homeowner and seller jointly received, not how they divided it up.

At the same time as California had the CSI, the federal government was giving a 30% tax credit for solar, but only up to $2000 for the entire system if a homeowner bought it in 2007 or 2008.  TPOs got the full credit from the start.  Since 2009, homeowners have also had no cap on the tax credit.

If you think figuring out federal tax credits could get a bit tedious, imagine the thrill of analyzing the economics of accelerated depreciation.  I’ll spare you the details here (a phrase that may have been more welcome a couple paragraphs earlier), but the bottom line is that accelerated depreciation — which only TPOs can utilize — amounted to an additional 12%-15% incentive, about half the size of the 30% federal tax credit, and larger than the CSI since 2010.  The figure below shows my estimates of the size of these incentives, all per kW of installed capacity, from 2007-2013.


That brings us to the incentive from residential rates.  During the period I studied, the 5 tiers of PG&E’s rate structure averaged $0.13, $0.15, $0.28, $0.37, and $0.40 per kilowatt-hour (kWh).  The solar PV on your rooftop crowds out the most expensive kWh first by reducing the total kWh for which you get billed.    Over these years, the systems installed were on average displacing kWhs that would have cost the customer an average of about 26 cents.  Importantly, that is much higher than the 19 cents per kWh they would have saved if PG&E charged a single flat rate for electricity (that raised the same revenue).  If PV adopters expected the tiered prices to stay at those levels (adjusted for inflation), I show that PG&E’s tiering of rates created nearly as much additional incentive to install solar as did the 30% federal tax credit.

The savings are so large in part because of net energy metering (NEM), which means the household only pays for the net consumption — that is, total consumption minus the electricity the panels produce — even if some of the panel production gets injected into the grid (which happens any time that the household consumption is lower than production).  An alternative approach, used in other parts of the world, is to pay the household a lower price for grid injections than the retail price the household pays for receiving electricity.  Surprisingly — at least to me — moving from NEM to that alternative approach, but keeping the same tiered rates, would reduce the incentive for solar by only about half as much as moving from tiered to a flat electricity rate.  The steep tiers create a much larger incentive than NEM, though the combination creates a still larger incentive.

Important note: those steep tiers created strong incentives only if they were expected to last.  Maybe they were, but they didn’t.  Already, in 2015, the lowest tier prices have risen and the highest have fallen so much that the highest tier price is now about twice the lowest rather than three times.  Proposals now before the California Public Utilities Commission would change the spread to just 20% or 66% depending on which proposal is adopted.  This will further lower the average price of electricity that the solar panels replace, and lower the incentive for large users to install PV.

Beyond the size of these incentives, I also wondered who was going solar, particularly how much the recipients of incentives tilt towards high-income households.  Using very granular census data, I estimated household incomes for each PG&E customer who installed solar.  Not surprisingly, they are heavily skewed to the wealthy with 35%-40% of systems going to households in the top 20% of earners.  But that has been changing since 2011, with the measure of inequality among adopters declining by nearly one-fifth from 2010 to 2014.  In the first few months of 2014, households in the highest of the five income brackets were still 82% more likely to adopt solar than households in the middle bracket, but that’s down from 116% in 2010.

Estimating incomes of solar adopters also give some insight into how the private benefits vary among those who do install PV systems.  As you would expect, the lower income adopters tend to consume less electricity and put in smaller systems, but they actually put in larger systems relative to their consumption.  That means they start lower down on the tiered rate structure and they crowd out a larger share of kWh, which are kWh that wouldn’t have cost that much anyway.  Systems on the roofs of the highest income bracket households crowded out electricity that would have cost them 27 cents per kWh on average, while the systems on middle income households displaced 25 cent power, and the households in the lowest bracket displaced 21 cent electricity on average.  Among those who installed solar in 2007-14, the wealthiest customers were likely to get the largest savings.

As I wrote a few weeks ago, we need a careful analysis of the societal costs and benefits of deploying renewable power at grid scale versus distributed generation.  At the same time, we also need a careful analysis of the incentives that have been created for generating energy from all sources.  Regardless of one’s views on solar, distributed generation, or renewables generally, understanding the size of the financial incentives from direct and indirect factors is critical to evaluating which programs are likely to have the greatest effect on adoption and which customers are likely to get the greatest benefits.

Posted in Uncategorized | Tagged , , | 9 Comments