A second look at CBO's employment study: Climate legislation and a healthy economy can go hand in hand

The Congressional Budget Office (CBO) recently released a study examining the employment effects of putting a cap on carbon emissions. It concluded that climate legislation would have modest, if not negligible, impacts: “[Climate legislation] would probably have only a small [slightly negative] effect on total employment over the next few decades…most laid-off workers would find employment in other industries.” These small impacts would fall mostly on workers in energy-intensive industries and related sectors.

It’s a fairly benign conclusion, but still off the mark: climate change legislation contains many provisions designed to prevent and moderate even these small impacts, as well as speed up the transition to a clean energy economy. To the extent that workers are negatively impacted, the solution is not preventing climate legislation from being enacted. Rather, climate legislation must contain effective safeguards to re-train these workers, provide alternative sources of employment to them, and restore their health care benefits should they lose them. With so few workers affected, only a very small portion of the proceeds from selling pollution allowances would be needed to accomplish this.  

The problems with CBO’s analysis

CBO relied upon three studies that make its conclusions problematic. Two of the studies (click here and here), conducted by non-partisan research institutions, Resources for the Future (RFF) and the Brookings Institution (BI), only model a cap on carbon. None of the complementary policies and incentives in climate legislation, designed to allow a smooth transition to a cleaner economy, are taken into account. The third study, conducted by Charles River Associates (CRA) on behalf of industries opposed to climate legislation, relies upon a set of unrealistic and false assumptions that generate large job losses. Partially adjusting for only some of these problems, I estimated that CBO’s economy-wide projections change from negative to positive: from -.25% in 2015 and -.27% in 2025, to +.16% and +.22%, respectively. This analysis is presented further below.

Generally, there are six overarching problems with the studies CBO relies upon.

1. The RFF and BI studies do not model the allowance allocations in climate and energy legislation to manufacturers to offset their energy price increases; these allocations have been found to fully offset these costs—and in some instances more than offset them (click here for a federal interagency study and here for a study from Stanford University). CRA models these provisions, but constructs their model to make them ineffective.
2. None of the studies model provisions in climate legislation that direct billions of dollars from the proceeds of selling pollution allowances toward promoting more efficient industrial processes, or assistance to manufacturers to help them retool, retrain workers, and lower their energy use.
3. As CBO describes, the studies use models that cannot effectively capture employment gains in sectors expected to grow under climate and energy legislation: “[The] industry classifications used in the studies coincide better with industries that would lose employment than with industries that would be expected to gain.” These include production of more fuel-efficient vehicles, and increases in construction “resulting from consumers seeking to improve the energy efficiency of their homes” and from “producers of electric power [building] new wind or solar generation facilities.” Such shifts in production are critical for job creation: shifting expenditures away from energy toward the rest of the economy is a driving force behind job growth projected from climate legislation, as is switching from fossil fuel energy to renewable energy. Every dollar spent on energy efficiency or renewables is estimated to create on average 3 times the number of jobs as every dollar spent on fossil fuel energy (click here  and here for more examples of job creation from energy efficiency).
4. RFF and BI do not model efficiency standards in climate legislation that will save firms and households billions of dollars in energy expenditures (click here, and here). CRA does model them but, contradicting evidence documenting the opposite, starts with an assumption that the economy is fully energy efficient. Perversely, this has the effect of making the economy less efficient, rather than the reverse, artificially increasing CRA’s job loss projections.
5. None of the studies model break-through innovations in clean energy (note that climate legislation also has incentives promoting innovation in new and emerging technologies), for the simple reason that it is not possible to model technologies that don’t yet exist. But how likely is it that there won’t be any unforeseen cost-reducing innovations in low carbon energy over the course of forty years? Just think about cell phones and the internet over only the last decade. The profit motive is a very powerful thing indeed, and putting a price on carbon creates just that for clean energy.
6. None of the models account for damages from climate change (click here and here) that could result in job losses from not taking action.

Two other miscellaneous but important points. CBO writes that real wages will decline. However, this is relative to a baseline of increasing income (i.e., incomes are projected to increase in all models, just slightly less so under climate policy). Similarly, most of the job impacts are projections relative to a baseline of increasing jobs. In most instances, no actual workers will be laid off; employment opportunities just shift to different parts of the economy.

In contrast to the CBO report, studies that more completely model provisions in climate legislation (click here, here, and here), demonstrate the numerous employment opportunities a low carbon economy offers. Consistent with the historical relationship between environmental regulation and job growth (click here and here and here) over the last four decades, climate and energy legislation is likely to lead to net job growth rather than the reverse.[1] Given the magnitude of the structural changes that will occur in the transition to a clean energy economy, job growth resulting from climate and energy legislation is likely to be much larger than that created from past environmental regulations. In the last section, I elaborate on how incentives and complementary policies in climate legislation can generate economic growth and significant employment opportunities.

First I’d like to turn to CBO’s actual estimates: see what they look like in the aggregate—i.e., relative to the economy as a whole, and what they might look like if we correct for some of these errors. We will see that adjusting CBO’s estimates for only the first error and partially the third turns CBO’s economy-wide job loss estimates positive. It is not possible to make any adjustments to account for errors 2, 4, 5 and 6 with CBO’s data.

CBO’s estimates in the aggregate: a relative perspective

CBO provides job impact estimates for several sectors, but no economy-wide number aggregating them, nor any metric with which one could compare it relative to other economic shifts. So our first task is to see what the sectoral impacts look like relative to the economy as a whole.

Using data from CBO’s analysis,[2] I estimated that CBO’s combined sectoral impacts sum to roughly a -.25% change in the number of jobs in 2015, and -.27% in 2025. To put this in perspective, from 2000 to 2007, manufacturing sector employment fell 2.9% per year[3]; year-to-year variations in unemployment rate forecasts fluctuate +/- .16% per year on average.[4] Consistent with the tenor of CBO’s overall assessment, their economy-wide employment impacts are statistically insignificant.

What CBO’s estimates look like after making adjustments for some of the errors

As small as CBO’s estimates are, adjusting for some of the errors discussed above turns their economy-wide employment impact from negative to positive. The table below compares CBO’s economy-wide impacts to alternative estimates that sequentially remove losses legislation is designed to prevent, or losses from sectors where actual gains in employment might occur:

Removing just CRA’s service sector losses changes CBO’s economy-wide change to essentially zero:  -.03% in 2015 and +.01% in 2025 (from -.25% and -.27%, respectively). If we then also remove losses in non-fossil-fuel sectors, the change is +.13% in 2015 and +.18% in 2025. Finally, if you remove BI’s (small) projected losses in the electricity sector you get +.16% and +.22%, respectively.

The rationale for removing these losses:

• CBO notes that the service sector is expected to grow the most under climate legislation: it is less carbon-intensive than the rest of the economy, so production and consumption will shift to it. I therefore removed losses CRA projects in that sector, which CBO notes results from an arbitrary assumption it regards as questionable.
• Allowance allocations compensating manufacturers’ for increased energy prices fully offset these costs (per point 1 above); employment in industry and sectors that use industrial products as inputs should not be affected. In fact, increased industrial efficiency and new industries might result in net job increases (see next section).
• Should energy efficiency or increased electricity prices reduce electricity consumption and production (and therefore employment in that sector), shifting energy expenditures to other sectors of the economy generates job creation elsewhere. Additionally, some fossil-fuel based electricity will be replaced with low carbon electricity. Both energy efficiency measures and clean energy are more labor-intensive and use more domestic content for inputs than fossil-based energy sources, so these shifts would result in a net increase in jobs (see point 3 above). There is already an extensive domestic supply chain supporting a cleaner economy and clean energy production, which will only grow in size.

Although I did not take out the small projected losses in the fossil-fuel sector (i.e. in coal, oil and gas extraction, and refining), for reasons outlined in the next section concerning enhanced oil recovery (EOR) and carbon capture and sequestration (CCS), these could be smaller than CBO projects.

These revised numbers tell a very different (and more positive) story, but still significantly understate job gains, because industry classification codes in the studies do not effectively capture sectors that will grow under climate legislation (point 3 above, highlighted by CBO), because the studies do not capture energy efficiency savings (points 2 and 4 above) stimulated by climate legislation, and because unforeseen cost-reducing innovations cannot be modeled.

The 21st century clean energy economy

CBO’s analysis is clearly limited when it comes to modeling job creation in the new energy economy. So how might this economy evolve, and what are the possibilities?

There are two important factors. The first is how quickly the cap on emissions shrinks, and the second how fast the economy can reduce emissions.

In regards to the first, caps on emissions are phased in very gradually over time. Having the lowest cost of emissions reductions, the electric sector is the first covered sector. Under proposed legislation, it reduces emissions to 4.5% below 2005 levels by 2013, with that percentage increasing gradually over time. In 2008, before the recession began, emissions were already 2% below 2005 levels. If the electric sector reduced emissions 17% below 2005 levels by 2020, the economy-wide target for covered sectors, that would imply a reduction of 1.7% per year between 2013 and 2020, or 1.2% if starting from 2010, very reasonable rates of decline that would not be disruptive. Industrial sources do not have to begin emissions reductions until 2016, with the cap also phased in. There will be plenty of time to adjust.

In regards to the second factor, the economy is already well underway in lowering its carbon emissions. On May 6th, the Energy Information Administration reported that in 2009 total energy consumption fell across all end-use sectors by an unprecedented 4.8%, attributable in part to the recession but also to a dramatic improvement in the rate at which the economy becomes more energy efficient over time.[5] From 2000 to 2008, efficiency improved on average 2% per year. In 2009 this went up to 2.4%, a 20 percent improvement. And there is ample room for more.

According to a 2009 McKinsey & Company report, the manufacturing sector could cut primary energy consumption with a positive payback by 18 percent by 2020 (click here for another study on industrial efficiency). Climate legislation can help industry obtain these improvements. Allowances will go toward promoting more efficient production processes, R&D, low-cost loans, and other assistance to help manufacturers retool, retrain workers, and lower their energy bills. Increased efficiency will improve competitiveness, not only creating jobs in the process, but also keeping the ones we already have.

Indeed, some of the very same sectors needed to run the fossil fuel economy, including glass, steel, cement, chemicals, automobiles, and special minerals (click here, here, here and here for some examples of greening and growing these industries) will be needed to drive the clean energy economy. Retooling and retraining programs will help automakers build cleaner cars (click here for a recent study on net job creation in the auto sector), help glass companies convert to making high-efficiency windows, or help steel manufacturers make components for wind turbines. And transporting these products will require investments in our energy infrastructure, such as high speed rail, cleaner buses and trucks and the roads and bridges that support them, modernization of our grid, and pipelines, also generating employment.

Finally, climate legislation will provide billions of dollars to develop and deploy carbon capture and sequestration (CCS), which could moderate impacts on coal miners. CCS is already a proven technology employed at industrial facilities. If it can be scaled up in the utility sector it would bring a surprising additional benefit: billions of barrels of new oil could be captured from existing stranded oil fields through a process called enhanced oil recovery (EOR), which uses CO2 to increase the amount of oil that can be captured from wells, creating new domestic jobs in this sector.[6] (Click here, here, here, and here to learn more about EOR). EOR is already being used to extract oil, but has been constrained by limited supplies of CO2 (currently it uses CO2 from natural sources). EOR could yield more than ten times the oil the American Petroleum Institute projects is likely to come from new offshore exploration.

Conclusion

Taken together, provisions in climate legislation that are ignored in the CBO study will help American manufacturers become more efficient, cleaner, and more competitive. They will create new markets and demand for American-made advanced materials and products at the core of our 21st century clean energy economy, and they will advance clean energy production. As a result, climate legislation will both preserve jobs and open up numerous new employment opportunities across America.[7]

[1] A net increase in jobs is possible due to the presence of unemployment, i.e. there is a pool of workers available from which to draw labor. Historically, the U.S. economy has never operated at full employment, making clean energy investments a source for net increases in jobs.

[2] Estimated from reading the second graph in Figure 1 of the report and EIA’s employment forecasts (Table 20).

[3] CBO notes that from 1979 to 2007 manufacturing employment fell from 20 million jobs to 14 million jobs, with 3.5 of the 6 million loss occurring between 2000 and 2007. Thus, in 2000 there were roughly 17.5 million jobs (20 million minus the 2.5 million jobs lost between 1979 and 2000). The loss of 3.5 million from 17.5 represents a 20% decline between 2000 and 2007. Averaging over 7 years gives an annual decrease of 2.9%.

[4] Calculated from year-to-year changes in projected unemployment rates from the Energy Information Administration’s Annual Energy Outlook 2010 between 2012 and 2035.

[5] This is measured by “energy intensity,” the amount of energy used per dollar of GDP. Energy intensity declines “naturally” (i.e. without climate policy) over time as older, less efficient appliances and machines get replaced by more efficient models.

[6] Note that even if demand for oil decreases in the U.S., world demand for oil will continue to grow, providing a market for U.S. produced oil.

[7] Though not the focus of this blog, we should also add that climate and energy legislation will increase national security by substantially cutting our addiction to oil (click here, here, and here).