Eight Questions to Ask About CRA’s May 2009 Economic Analysis of the Waxman-Markey Cap and Trade Program
Posted July 16, 2009
This blog is structured from: http://switchboard.nrdc.org/blogs/ljohnson/seven_questions_to_ask_of_any.html
1. Top Line Question: How much larger (than 2008) does this model predict GDP will be in 2030 under an emissions cap (do not accept "how much less GDP is relative to no policy" answers)?
CRA does not provide absolute levels of GDP (and refused to make their complete set of results publicly available when I requested them). However, the appendix reports using for its baseline GDP EIA's Annual Energy Outlook levels which, when combined with the changes in GDP given in the report, allows one to calculate GDP levels under Waxman-Markey. Under the legislation, by 2030 GDP will be almost 70% higher than today (compared to approximately 72.3% without climate legislation), hardly the picture of economic collapse the report tries to convey.
2. CRA claims households will lose on average $830 dollars per year if we put a cap on carbon. How do they arrive at this conclusion?
Despite predicting decreasing household utility bills for the early years (during which regulated utilities are allocated free allowances whose value must be passed on to their customers), CRA still reports significantly larger overall costs for consumers than EPA or CBO (almost 3 times CBO's 2020 estimate of $175, and more than 8 times higher than EPA's average estimate through 2050). How? Energy efficient technologies are assumed to be expensive, due to no underutilized cost-effective efficiency opportunities being available in the reference case (see 3. below). In addition, renewables do not become cost competitive over time as their market share grows (see 6. below). These assumptions cause much more significant decreases in productivity (i.e. output per worker), wages, and employment relative to the no policy case (see 4. below) than the EPA analysis, which uses a similar model as CRA. As with any opposition-funded study, unrealistic assumptions generate unrealistic costs.
3. You should always ask how much consumers save in total from energy efficiency improvements, and at what cost.
CRA exaggerates costs by forcing reductions in fossil fuel energy use while simultaneously assuming that no alternative low cost sources of energy are available, such as energy efficiency, or cost competitive renewable energy (see 6. below).
Without seeing the full set of results and assumptions (which CRA refused to make public), it is impossible to infer how much energy efficiency is adopted and at what cost. It is also not possible to determine what assumptions are being made about how GDP responds to increased energy costs, other than to say that it declines relative to the reference case. Remarkably, CRA notes that assumptions behind GDP responses are critical inputs to the model, yet doesn't provide them (in technical jargon, we need to know assumptions about "elasticities of substitution" between different inputs to production; these determine how much flexibility producers have in changing their mix of inputs in response to energy price increases-the more flexibility, the less costly it is to adjust and the less GDP will decrease). Lacking any of these pieces of information, we have no way to compare CRA's results against other estimates, or against the assumptions contained in other studies. (Note: CRA does assume increasing energy efficiency over time based upon historical trends, but assumes it is the same for the policy case as for the reference case. Energy efficiency in response to a carbon price also does occur, but none of it is a result of energy inefficiency being present in the reference case. I elaborate on this below).
Just enough (barely) information is given to suggest that GDP declines more sharply in response to the carbon price than historical experience warrants: CRA tells us that consumers begin with an "optimal consumption bundle" prior to the imposed carbon cap. Translation: prior to imposing carbon restrictions, no unused energy efficiency opportunities exist at net savings. In contrast, EPA assumes that some energy efficiency is available at 3 cents per kWh (on average, it costs 7 to 8 cents per kWh to produce new electricity, so that these efficiency gains yield a net saving of 4 to 5 cents per kWh). As a result, in EPA's model, energy consumption levels that would be reached by 2015 without policy are not reached until 2040 under the cap. In contrast, CRA's total electricity generation increases under the policy case. However, CRA does not provide us with the exact amount of energy consumption (just bar graphs for electricity), so we cannot calculate how much energy is needed per unit of GDP. As with elasticities, the amount of energy needed per unit of GDP gives us an indication of the economy's ability to adapt to higher energy prices. With this basic output data, we could at least compare this ratio (GDP/unit of energy) against historically observed values.
Assuming away unutilized efficiency opportunities in the reference contradicts much of what we know about the economy's use of energy. Numerous studies have found an abundance of such opportunities. From better lighting, to more efficient appliances, new building shell efficiency, and better mileage automobiles, energy efficiency gains under a cap and trade program can create net economic gains. And energy savings will increase (as opposed to decrease) the amount of money available to spend on other goods and services, supporting new jobs to produce these goods and services. What's more, needing less energy per unit of GDP raises worker productivity, wages, and output, exactly the opposite results CRA's clients (see 8. below) are paying for.
4. A good follow up to any of the first three questions: How is it that this study finds net job losses, when it predicts substantial increases in GDP, when economic analyses of past environmental regulations on average show an increase in jobs from environmental regulation (http://epi.3cdn.net/83dfae8d6d0c6151e1_55m6id8x6.pdf) and when energy efficiency and clean energy create 3 times as many jobs as fossil fuel energy (http://www.peri.umass.edu/green_recovery and http://switchboard.nrdc.org/blogs/paltman/clean_energy_just_the_stimulat.html)?
The productivity losses CRA engineers (see 3. above and 6. below) are used to create the job losses they report. Without energy efficiency and cost competitive alternative energies, reduced energy use (relative to the reference case) and higher priced goods translate into less output per worker (i.e. lower productivity), a decrease in demand for workers, and lower wages. In fact, in order to generate the employment losses, CRA does a short run calculation that contradicts the long run structure of their model, which actually does not produce employment losses. The mechanism behind this is somewhat complicated; readers who are interested in more detail should feel free to call me. Suffice to say here that the most important output determining employment effects are the productivity losses CRA engineers.
5. How do CRA's household costs compare to estimates by the EPA and CBO?
CRA projects costs that are approximately 3 to more than 8 times higher than CBO and EPA. A large part of these costs are due to the productivity losses CRA generates through its unrealistic assumptions about energy efficiency (see 3. above) and the cost of renewables (see 6. below).
Nevertheless, even if we accept CRA's household cost estimates, they should be put in perspective: along with healthy GDP growth is an unavoidable increase in household income and purchasing power, usually more than an order of magnitude higher than estimated household costs or changes in wages relative to the reference case. Thus, while climate protection isn't going to be free, it will still be much less than increases in household income, making climate protection very affordable. For example, for 2020 EPA projects a cost of about .10-.11% of household purchasing power, but a 15-19% increase in purchasing power overall from economic growth. The ratio of increased purchasing power to the estimated cost is at least 136 (15/.11)! Of course, we don't know how much purchasing power grows in CRA's analysis, because they don't report it and refuse to make it publicly available. Most likely, it is well above their cost estimates. If it is not, they need to explain why it is so different from analyses that are not paid for by someone whose financial interests conflict with climate legislation; see 8. below).
6. How much do the cost of renewables and advanced coal w/ CCS decline as their market share grows in response to the carbon price, and does the model permit a plausible amount of capacity growth?
CRA presents a graph showing how much new capacity there is for each of the technologies for a low and high cost case, but not its core case. The graphs themselves don't have numbers in them for each category, so you cannot get an exact amount, and CRA won't provide it.
More important, however, are the assumptions CRA makes about technology costs, and the obliqueness with which they pretend to transparently reveal them.
To begin, CRA appears to assume the same technology costs under policy as they do in their reference case, benchmarked to the Energy Information Administration's 2009 Annual Energy Outlook projections (except for nuclear and geothermal). This may sound honest and non-partial, but it is not.
EIA's model has costs declining as a function of how much is produced, to reflect the fact that as more is built of a given technology, economies of scale (i.e. mass production) and learning reduce costs. Because more renewables are produced under a carbon cap, relative to the no policy case, their costs will decline more than they do in the reference case. CRA does not seem to allow this.
Next, CRA artificially limits how much renewable energy can be built, justifying the limits based upon a "variety" of sources and models, including its own. CRA doesn't even provide its capacity limits for solar thermal or photovoltaic, arguing that only regional capacity limits are imposed instead of national. But it doesn't tell you what these regional limits are or what they imply in terms of national limits. For solar thermal costs, CRA says it uses an assumption from a 2006 EPA study. This may sound honest, but again it is questionable: CRA does not reveal the cost, nor indicate whether it chose a particular scenario within the study (for example, a "high cost" scenario) and, if so, what kind of assumptions were associated with the chosen scenario. For photovoltaic, CRA states that it made its own assumptions (which they don't provide).
Next, after 20 years, CRA assumes no more cost reductions will take place.
Then, after 2030, the amount of additional renewables that can be produced are capped, while non-renewable energy is allowed to expand.
Combined, this collection of assumptions prevent renewables from ever achieving real cost reductions. It's no surprise, then, that CRA is able to generate the productivity losses it needs to create exaggerated costs.
In terms of CCS, there is no CCS in the reference case, so costs cannot decline more in the policy case. CRA does say that the CCS subsidy it models brings down costs, but not how much relative to a case in which there is no subsidy.
7. Did the study model the costs of inaction, such as property lost to sea level rise, more intense hurricanes, forest fires, water shortages, national security threats (from millions of people losing fresh water supplies as glaciers evaporate and droughts intensify), increasing world food prices due to water shortages, and lost ecosystems and species?
8. Was the study paid for by someone whose financial interests conflict with climate legislation?
Yes. The study was paid for by the National Black Chamber of Commerce. Since 1998, the Chamber has received $275,000 from Exxon Mobil, and includes members from Chevron and Exxon. When Greenpeace identified climate change "denial" groups that ExxonMobil was still funding, even after the oil giant claimed that it had ceased doing so, the NBCC figured very prominently on the list.