Kerry-Lieberman is Looking Like a Nuclear Energy Jobs Bill

Trevor Houser and his colleagues at the Peterson Institute have a sharp economic analysis of the Kerry-Lieberman energy and climate bill out this morning. Take a look here. (There’s a lot more in it – on emissions, oil consumption, and energy prices – than I discuss here.) This is the first analysis of any climate bill I’ve seen that actually tells a plausible story of the “green jobs” front. Their conclusion is that the bill would add modestly to job growth during the next decade, while depressing it slightly in the decade after, but still leaving a (tiny) net gain over the two-decade period. This is a big deal if it holds up.

I’ve been skeptical of green jobs arguments in the past, and still have some big questions about the sensitivity of the authors’ conclusions to a few important assumptions. But the authors make the first serious quantitative case I’ve seen for net job creation from a climate bill. Moreover, they tell a new sort of green jobs story (which will make many environmentalists uncomfortable): the bill, by their estimates, would create 165,000 jobs in the nuclear industry (averaged annually over the 2011-2020 period) but only 19,000 jobs in renewable power.

I’ve been quite critical of most “green jobs” analyses for missing the big picture: they normally count jobs created in clean energy and jobs lost in fossil fuels but ignore the negative impacts of higher energy prices and of macroeconomic distortions. They thus tend to inflate job projections, tipping those from negative to positive. Typical economic models, in contrast, incorporate these effects. Those models conclude, I still think correctly, that for an economy at or near full employment, carbon pricing will probably send employment numbers down a smidge.

But the problem with those models is that we aren’t anywhere near full employment – and the EIA doesn’t think we’ll be back there until the end of this decade. As Houser and his colleagues argue, this provides an opportunity. If we use carbon pricing to spur replacement of old power plants with new (low-carbon) power generation in the near term, it will create jobs building those plants.  Since there is a lot of slack in labor markets, those won’t necessarily come at the expense of other jobs. The same is true, they argue, for capital: there is a lot of money on the sidelines, and there will be for a while, minimizing the capital market distortions that would normally arise from a big push for new investment in one sector.

These arguments have been made at a qualitative level before. But they always needed quantitative analysis to see whether there would be more jobs created than destroyed. Houser and his colleagues conclude that, for 2011-2020, the answer is yes, to the tune of 203,000 on average for any year. For 2021-2030, when the economy is back to normal, the answer reverses (as intuition suggests it should), with an average of 190,000 shaved back off (a number which presumably would increase after 2030). (The authors don’t actually give a number for 2021-2030; they give a net average annual gain of 6,300 for 2011-2030, from which one can infer the 2021-2030 number.)

I have one big worry about the analysis: its results appear to be sensitive to some very important assumptions about nuclear power. (I have similar concerns, for roughly the same reasons, about its treatment of carbon capture and sequestration.) Page 12 of the report has an extraordinary plot that shows how jobs are created and destroyed. Here it is:

More than half of the clean energy jobs created over 2011-2020 are in nuclear power – the model projects 165,000 average annual jobs added in that sector. (Another 31% are in CCS; CCS and nuclear together yield 85% of the clean energy jobs.) This has a special and potentially big impact on the analysis.

Nuclear plants are expensive and take a long time to build. (Ditto for CCS.) The model that the authors use (which is the EIA’s own model) assumes that utilities aren’t allowed to pass on the costs of power plant construction to their customers until the plants are in operation. This means that while construction jobs are created in the coming decade, while the plants are being built, the economy doesn’t feel the hit from higher energy prices until after 2020, when they go into operation. That juices the jobs numbers for the 2011-2020 time period. I can imagine two ways, though, that this assumption might turn out to be wrong.

The first is if nuclear construction turns out not to be feasible at the scale and pace that the authors predict. (Same, again, for CCS.) In that case, the bill would presumably need to spur more renewable energy investment (and greater use of offsets). Wind and other renewable facilities are quicker to build. Their costs would thus be passed on to consumers sooner; the net result would be greater near-term job destruction through higher energy prices, and, perhaps, a tipping of the net jobs numbers into negative territory (though one would need to rerun the model to find out).

The second way that the nuclear math could prove wrong is if regulators allow utilities to pass on some of their construction costs to ratepayers before the new plants go online. (Yes, once again, this applies to CCS too.) This practice (which is allowed in some states) has been the subject of much debate in the regulatory community. It is also high on the wish list of the nuclear energy lobby. To the extent that it is allowed, higher energy prices would come sooner, just as would be the case with renewable energy investment. The net result would be the same too: bigger numbers in the negative jobs column. Again, it’s impossible to know what the net jobs outcome would be; the authors would need to rerun their model with different assumptions to find out.

The biggest take away for me, though, is that you can actually do analysis that acknowledges the investment-pull benefits of carbon pricing in an economy that’s well away from full capacity while respecting the negative effects of higher energy prices, as well as the macroeconomic limits that real labor and capital markets impose. If this holds up to scrutiny by serious economists (a club that I’m not a member of) then it’s an important advance in how we model climate legislation. The EIA, EPA, and CBO need to seriously study this analysis, and make any necessary modifications to their own models, before they pronounce their own judgments on this and future bills.