The Congressional Budget Office is back in the news, after director Doug Elmendorf testified before Congress about the economic impacts of clean energy legislation. Opponents of that legislation rushed to hype a few of his comments out of context, and succeeded, as usual, in getting their voices amplified in the Washington Post.

Elmendorf didn’t say anything that wasn’t already in the CBO’s analysis of the Waxman-Markey bill, which was released in June and discussed to death at the time. Despite what you’d think from hysterical headlines, the top-line conclusion of that report was that the bill would reduce the federal deficit. The lowest-income quintile would gain in purchasing power; the middle class would pay modest costs. Those costs are, by CBO’s own admission, almost certainly lower than the analysis projects, since it doesn’t count the benefits of the aggressive energy efficiency provisions (or the economic benefits of avoided climate change). Overall, the CBO paints an incredibly optimistic picture of the economic costs of clean energy legislation, even given the flaws and conservative biases of its analysis.

Anyway, much ado about nothing, again. Just so we walk away from this latest hubbub having learned something, though, let’s take a look at some serious issues around the CBO, specifically its budget scoring and how it undercounts the cheapest, fastest source of clean energy: efficiency.

To acknowledge up front: it’s not easy to talk about CBO scoring without putting people to sleep. It’s not easy to talk about energy efficiency without putting people to sleep. CBO scoring of energy efficiency? That’s x-treme sleepy. We’re going to need puppies.

Grist thanks its sponsors. Become one.


Despite its obscurity, CBO budget scoring is actually quite important. It shapes legislation in subtle but pervasive ways. In particular, undercounting the potential for efficiency systematically distorts energy policy in ways that favor delay, compromise, and defensiveness. The good news is that properly accounting for efficiency opens the door to clean energy policy that’s fiscally responsible and environmentally accountable.

To understand why CBO accounts poorly for efficiency, some basic background on CBO scoring is necessary. Don’t worry. This guy will help you through it.


Grist thanks its sponsors. Become one.

The t-word

First, we’re talking about budget scoring here, not broader economic scoring (which CBO also does, as does the EPA, as do any number of think tanks and consultancies). Budget scoring is a narrow assessment of legislation’s projected effect on the federal budget, and only CBO does it. There are big problems with the broader analyses, too, but it’s budget scoring that has the most tangible effect on the legislative process.

Budget scoring matters because Obama and congressional Democrats promised that the climate bill will be deficit neutral. It “won’t add a dime to the deficit,” as Sen. John Kerry (D-Mass.) has put it. Who decides if the bill is deficit neutral? That’s the CBO.

The most important aspect of climate bill scoring is that the CBO treats pollution allowances as federal revenue, and allocating those allowances — handing them out — as spending. It effectively treats cap-and-trade as an indirect tax (like excise taxes, or customs duties). This has all sorts of implications.

The 25% haircut


For reasons we won’t get into, the CBO assumes a fixed nominal GDP — that is, they assume that legislation will not change the overall size of the economy, by either spurring or slowing growth. Practically speaking, that means any tax revenue that comes out of the private economy must, by definition, reduce private spending and thereby reduce taxable income, to balance the scales. In this case, the requirement to buy pollution allowances will lower the taxable income of the businesses and individuals who pay for them. That in turn will lower federal revenue from direct (income) taxes. So you get a new revenue stream, but at the cost of losing a little bit of the old revenue stream.

How big is the collateral damage to direct tax revenue? Rather than try to calculate that percentage for every piece of legislation and every set of taxed entities, the CBO (along with the Joint Committee on Taxation and the Treasury Department’s Office of Tax Analysis) has settled on a standard number, which it applies across the board: 25%. So for every buck that’s raised via an indirect tax, a quarter is lost in direct taxes and only $0.75 can be slated for new spending; $0.25 has to be set aside to balance the budget (assuming a balanced budget is the goal). This revenue offset is colloquially known, by the tiny number of people who have reason to know such a thing colloquially, as the “25% CBO haircut.”

Because the CBO treats pollution allowances as tax revenue, they take the haircut. Climate policy watchers will recall that the Kerry-Boxer climate bill introduced late last month was virtually silent on allowance allocation. The only thing it did specify is that 25% of the allowances would be set aside to reduce the deficit. However the allowances are allocated, whomever they’re given to, as long as 25% are set aside the result is guaranteed to be scored deficit neutral. That provision is what justified Kerry’s “won’t add a dime” boast.

(Incidentally, for a more in-depth but still relatively clear explanation of the haircut, see the CBO’s immortal tract, “The Role of the 25 Percent Revenue Offset in Estimating the Budgetary Effects of Legislation” [PDF].)

puppies!Sidebar: What about the House? or Get a haircut, ya hippie!

You may be wondering: why didn’t the Waxman-Markey bill in the House reflect this same 25% haircut? How come it got away with setting aside under 10% (through 2025) of allowances for deficit reduction?

A partial answer is that the House operates within what used to be the standard Congressional budget window: 10 years. That is to say, for House bills, “deficit neutral” means “deficit neutral for the first 10 years.” The House can use various budgetary tricks, e.g. selling future years’ allowances to the near-term, to pad revenue during the crucial 10-year window. It’s a bit dodgy, but then so is the pretense that human beings are capable of predicting economic circumstances more than 10 years out with any precision.

The Senate, though, loooves deficit hawkery. Loves it! Just wants to kiss it. Senators compete to see who can be more Serious about the deficit; it’s like Deficit Idol on the Sunday cable talk shows. The Senate not only passed a budget point of order forbidding legislation that increases the deficit inside 10 years, it passed another that prohibited legislation “that would cause a net increase in deficits in excess of $5 billion in any of the four 10-year periods beginning in 2018 through 2057.” (What, are they joking? What about the 2058 budget?!?)

Because of this delightful and not-at-all stupid and shortsighted provision, to get through the Senate, legislation has to be scored deficit neutral effectively in perpetuity. That’s a lot harder than appearing deficit neutral for 10 years, especially in regards to a program that most economic forecasts (wrongly!) show rising steadily in price in the post-2020 years.

puppies!One more wrinkle: offsetting offsets

One important twist is that the revenue offset (haircut) can itself be offset by … are you ready for this? … an offsetting offset. (Yes, the CBO really calls it that.) In the haircut analogy, an offsetting offset would be Rogaine. Or maybe extensions.

Remember, the revenue offset is a reduction in direct taxes caused by an indirect tax. An offsetting offset is when tax revenue is spent in a way that raises other incomes, which thereby raises (re-raises?) direct tax revenue. It offsets the offset! Are we having fun yet?

What kind of spending counts as an offsetting offset, thus avoiding the haircut? Any revenue returned directly, via tax cuts or credits, to a taxable entity (a business or an individual) qualifies.

If, by contrast, revenue goes to, say, state energy efficiency block grants, federal RD&D, rebates to low-income households with no taxable income, transit assistance, jobs training programs, international reforestation, or any of a gazillion other worthy investments, it takes the haircut. It is assumed to be displacing spending in the private economy, and since the CBO assumes fixed GDP, that means someone’s taxable income must take a hit.

OMFG another wrinkle!

If revenue is returned directly to a taxable entity but with strings attached — mandates about how the money must be spent — then it does take the haircut. The CBO figures that’s tantamount to government spending through an intermediary, which like all government spending displaces private spending. This will turn out to be a crucial distinction.

Offsetting offsetting offsets

Just kidding. Here’s a puppy.


Luvagod tell us why any of this matters

OK! Right now, in the Kerry-Boxer bill, 25% of allowances are set aside for deficit reduction. That leaves 75% of the allowance money from cap-and-trade to be spent on buying votes various good causes.

Here’s the thing, though: before the final legislation is done, that 25% will go down. It will be partially offset, because some of the revenue will be spent on direct transfers to taxable entities, and direct transfers get no haircut. In fact, the more that’s spent on direct transfers, the more of the total pie there is to spend.

puppies!This creates a somewhat perverse incentive structure. Since Congress would obviously like to have 100% of the revenue to spend, that’s going to bias decisionmaking in favor of direct transfers: tax cuts and tax breaks, aka the same crappy tax-based energy policy the U.S. has been limping along with for 30 years. Any other spending gets that extra 25% cost tacked on.

Direct transfers aren’t necessarily bad policy: they include the good (rebates to consumers to shield them from energy price increases), the pretty good (subsidies for renewables), and the not-so-good (subsidies for polluters). But a great deal of what we’d like government to help with is included in all that spending that does get the haircut — most importantly, energy efficiency programs.

So, in budget terms, tax credits cost what they cost; energy efficiency programs cost what they cost plus 25%.

What about all those kickbacks to utilities?

Good question.

As noted the other day, utilities get lots and lots of allowance value (about 40%) in Waxman-Markey. Do those handouts take the haircut? The utilities (LDCs) are expected to use the revenue for “consumer benefit,” but what does that mean in budget terms?

Remember, direct transfers to taxable entities don’t take the haircut, unless there are strings attached. So is allowance value given to LDCs a form of direct transfer to businesses and utilities? That wouldn’t take the haircut. Or is it a direct transfer to LDCs, with strings attached on how it’s to be used? That would take the haircut.

What a fun wonky question! Let’s savor it with a wagon full of puppies.


A CBO letter to Waxman (PDF) clarifies the treatment of LDC allocations … a little. But some questions remain.

The letter seems to say that LDC allocations won’t take the haircut. For one thing, the bill doesn’t yet mandate that the money be used for consumer benefit, it just strongly suggests it (see this post for more). As far as what counts as consumer benefit, the letter explicitly cites reductions in electricity rates or fixed-dollar rebates. But that distinction matters; if utilities just lower electricity rates, the electricity price signal will be muted and other forms of carbon use will become proportionately more expensive. A fixed-dollar rebate returned separately would keep consumers whole while still preserving the motivating force of the price signal. Could the bill mandate that “consumer benefit” be in the form of fixed-dollar rebates? It’s unclear.

puppies!What if the bill mandates that some percentage of utility allowance money go to energy-efficiency programs (rather than being returned directly to ratepayers)? Bam: haircut. Here’s what it says in the letter:

One might argue that this example [spending on utility efficiency programs] is similar to the previous example in which the LDCs kept consumers’ electricity prices from rising (and in which federal revenues did not fall) because both involve giving resources back to consumers. However, this example differs from the earlier one due to the increase in electricity prices that, by the assumption of fixed nominal GDP, must be offset by lower spending elsewhere in the economy.

Now, this would make sense if spending on efficiency were like any other spending. But it isn’t! Yes there would still be an “increase in electricity prices,” but there would be a corresponding decrease in energy bills. That’s the whole point of efficiency improvements: they produce ongoing savings, functionally equivalent to rebates, in perpetuity. Every dollar not spent on energy is money that can be used for other purposes. It doesn’t lead to “lower spending elsewhere in the economy.” (It’s worth noting that utility efficiency programs are incredibly cost-effective means of reducing emissions — about the most cost-effective way currently available.)

It’s a little crazy that handing money to electricity ratepayers doesn’t take the haircut, but handing energy savings to ratepayers does. If anything, the transfer of value via efficiency is economically more advantageous; every efficiency investment effectively establishes a grant fund that produces savings year in year out.

The CBO should revisit this issue. The utilities should be required to spend a good chunk of the allowance value on efficiency programs, and policymakers shouldn’t be punished with a 25% cost hit for making that decision.

puppies!State efficiency programs too?

Yes. Allowance value granted to states to fund energy efficiency programs should not suffer the haircut. As long as there’s good monitoring and verification, state efficiency programs save taxpayers money, thus increasing their take-home income. It’s an offsetting offset by god!

Beyond budgeting

The bias against efficiency — counting it as a cost but refusing to take account of its benefits — extends beyond budget scoring to broader economic analysis as well. Economic models assume that cost-effective efficiency investments would already have been made by the private sector (no one leaves money lying on the street, etc.). So government efficiency investments must displace more efficient private sector spending.

In sharp contrast, research from the McKinsey & Co., based on ground-up research, shows that the 2020 target in Waxman-Markey can be met at negative cost. If Kerry and Boxer made the improvements to the bill recommended by the American Council for an Energy-Efficiency Economy, the legislation would not only create more jobs but offer households $283 annual savings by 2020. (If, instead, Kerry and Boxer integrate the energy bill that passed through the Energy Committee — the American Clean Energy Leadership Act — it would leave enormous potential savings on the table.)

If it was more widely understood that U.S. taxpayers could save billions of dollars by lowering greenhouse gas emissions, the political landscape would look very different. The vote count would look very different. The federal budget would look different too.

Aw, now look what I did: