Paul Krugman has been on the warpath lately regarding climate change economics. He has devoted his last two NYT columns (here and here) to the subject, as well as back-to-back blog posts (here and here). True to form, Krugman accuses those who disagree with him of abject stupidity and evil intent; for Krugman it is impossible that any decent economist who cares about human beings could actually think the costs of cap-and-trade legislation will be high. But as we’ll see, Krugman’s own figures don’t jibe with the narrative he’s pushing.
In his September 27 blog post, Krugman takes up his familiar theme of denouncing those who dare to say that Waxman-Markey carries a large price tag. After using a diagram to explain the textbook distinction between the compliance costs of a new tax (or mandate), versus the “deadweight loss,” Krugman excoriates economist Martin Feldstein for allegedly spreading lies:
[Feldstein] took the CBO’s estimate of “compliance costs”, which was $1600 per household in an early report (it’s now down to $900, but who’s counting?), and implied that this was the economic cost of the legislation. But “compliance costs” are basically the sum of the blue rectangle and the red triangle; the true economic costs are just the triangle, and are much smaller.
OK now, this is quite simply hilarious, if you can follow me through the argument. I really don’t think Krugman realizes just how much his pants are down on this one.
First off, Krugman is correct that there really is a distinction between the impact of a new tax in terms of paying extra revenues, versus the overall loss to the economy because of distorted incentives. But when the public wants to know “how much will cap-and-trade cost?”, it is quite reasonable for them to wonder, “How much will my electricity bill, or gasoline prices, go up because of this?” Most people do not realize that Krugman & Co. are netting out the gains to the recipients of free allowances and government expenditures when computing the “net burden on U.S. households.”
For an analogy, consider the debate over health care reform. Using Krugman’s approach to cap-and-trade, the proponents of reform could say, “We suggest imposing an extra tax of $1 trillion on upper-income earners, in order to provide insurance to all Americans. But the cost to the economy would be quite minimal. After all, every penny of that trillion dollars would be spent on hospitals, doctors, pharamaceuticals, etc., thus recycling that money right back into the pockets of U.S. households. The true economic cost of the tax hike would only be in the form of reduced incentives to work among the super-wealthy, which I estimate at around $75 billion over a ten-year period, which works out to a Slurpee per day for the average U.S. household.”
Would anybody buy the above argument? Of course not. A $1 trillion tax hike is just that, a $1 trillion tax hike. Of course that money “gets back into” the economy when the government spends it, but so what? The reason people dislike taxes is that it’s no fun to pay them, and they have a lot less control over how politicians spend their money once it’s confiscated. And yet, the above argument that I invented for health care is very analogous to how Krugman is saying we should score the “economic cost” of cap-and-trade.
So what’s the true number, Professor Krugman? But now we’re ready for the really fun part. Krugman is saying that because Feldstein focused on the gross costs of compliance–rather than the net “deadweight loss” in terms of altered incentives–he frightened the American people with the scary number of $1,600 per household. So what is the more accurate number, according to Krugman? In the same blog post he tells us:
Now, the cost to the economy of this limit is the benefit the private sector would have gotten by emitting more than is allowed under the cap. It’s shown in the figure as the red triangle labeled “deadweight loss”. CBO puts these losses under Waxman-Markey at 0.2-0.7 percent of GDP in 2020, 1.1 to 3.4 percent in 2050. These costs have to be set against the environmental benefits.
Well hold on just a second! What if we convert that statistic back into a dollar amount per household? Krugman would have you believe it would be a lot lower than $1,600, right? And yet simple arithmetic shows us that he’s way way off.
Let’s convert the range into a single point estimate: We’ll say that Waxman-Markey will cost 2.2% of GDP by 2050 (a little less than the midway point of the range). If we turned that percentage into a gross dollar amount with current GDP levels (year 2008), it works out to about $315 billion. (Current GDP is $14.3 trillion, so 2.2% of that is $315 billion.) Dividing by 117 million U.S. households, that works out to almost $2,700 per household. So already, we’re way above the allegedly dishonest figure put out by Feldstein.
But wait, it gets much worse. The CBO estimate refers to the cost of Waxman-Markey in the year 2050, when GDP will be much higher. As Krugman explains in one of his recent NYT columns: “[T]he budget office also predicts that real G.D.P. will be about two-and-a-half times larger in 2050 than it is today, so that G.D.P. per person will rise by about 80 percent.” (The interested reader can see the CBO’s actual GDP forecasts in the last tab of this Excel file.)
So assuming that the number of people per household stays roughly constant between now and 2050, we need to increase our figure of $2,700 by 80%, allowing us to conclude: Krugman’s own preferred mid-range estimate of the cost of Waxman-Markey works out to $4,800 per household by the year 2050.
Conclusion. By focusing on the “deadweight loss,” rather than the gross compliance costs, Krugman is minimizing the huge spikes in energy prices that Waxman-Markey would impose on consumers. If such an analysis were used to score a simple tax hike, everyone would recognize the chicanery involved–and this is why the politicians are pushing cap-and-trade, rather than an explicit carbon tax.
However, even on his own terms, Krugman’s own statistics yield a mid-range cost estimate of Waxman-Markey of $4,800 per household by the year 2050. So when Martin Feldstein and others throw around figures like $1,600 or even $3,100, they’re not scaring the public enough! When the caps under Waxman-Markey are squeezed as tightly as they will go, the CBO itself admits that the “deadweight loss”–the measuring rod favored by Krugman–will be in a range with a mid-point value of $4,800 per household.
One final point: It would be misleading to conclude, “Oh, so the CBO says the cost of Waxman-Markey’s cap-and-trade program will be $4,800 per household.” That figure applies to the year 2050, when households will be much richer. This is one of the reasons the CBO and other groups express costs (whether from carbon restrictions or from climate change) in the far-distant future as a fraction of GDP, rather than as a dollar amount. However, even if we looked at the mid-point 2.2% of GDP estimate using 2008 numbers, the per-household cost is $2,700, much higher than Feldstein’s figure of $1,600. What Krugman and other proponents of Waxman-Markey are doing is reporting the dollar-cost per household figures for early years (like 2020), when the emission caps are relatively lenient. But then in the later years (like 2050), when the caps are very tight, Krugman and others report the figure as a % of GDP, which means nothing to most Americans.