The (Estimated) Cost of Everything and the Value of Nothing: The Green New Deal

It has been almost a year since the Green New Deal was introduced as a Congressional Resolution. It didn’t take long for critics to start piping up. Within a few weeks Douglas Holtz-Eakin (DHE), former director of the Congressional Budget Office and now President of the American Action Forum, released an estimate of what the GND would cost. I didn’t have time to look at his estimate until recently, and while neither the GND or DHE’s estimates are in the news as much as they were a year ago, I think it’s still worth taking a closer look.

Economists and other cynics are said to know the cost of everything and the value of nothing. It’s one of those clichés that persist because it is so often true. And if we are ever going to improve people’s perceptions of economists, we’re going to have to do better than this. I’m not trying to argue that the Green New Deal is a good idea or a bad one. Instead, I’m trying to point out that these cost estimates are so unrealistic that they can’t possibly help anyone decide.

There is a lot wrong with DHE’s work on this, but first let me say this: I don’t know him personally, but I’ve been familiar with Holtz-Eakin’s work for a long time. I’ve often disagreed with him, but I’ve never had a reason to think he was a partisan hack. I’d like to think I was right, but he doesn’t make it easy.

The Value of Principals and the Principal of Value:

First, the Green New Deal is a set of broad principals that touch on huge swaths of the economy. It is in no way a policy proposal. It lists goals like achieving net zero carbon emissions, creating millions of jobs, and providing everyone with high quality health care and economic security. It’s a 14 page document. There are no details of what policies we would need to meet these goals. It defies policy analysis because it’s not a policy. Because it has no proposed actions, it defies efforts to cost them. Or at least it should.

Secondly, these estimates only look at costs. There is nothing in there at all about the economic benefits of the GND, DHE makes no attempt whatsoever to determine whether it would be worth whatever cost he came up with. It implicitly assumes that achieving the GND goals has no value whatsoever. 

After taking a closer look at DHE’s numbers, I can only conclude that they were meant to scare rather than enlighten, which is disappointing.

DHE’s analysis presents a low and high-end estimate of the costs of the GND, ranging from $52-$93 Trillion over 10 years. Importantly, the vast majority of the costs have nothing to do with the Green-ness of the New Deal. According to DHE, two parts of the GND account for between 82% and 87% of its costs: The universal health care and universal employment guarantees. Whether or not DHE is even in the right ballpark rests on his analysis of these two parts alone, so I took a closer look.

Universal Health Care: Don’t be Gross.

Like other parts of DHE’s “analysis,” the universal health care estimate is based on analysis performed elsewhere. In this case he relies on a study published by the Center for Health and the Economy. That analysis looks at the cost of a Medicare-for-All proposal modeled after Bernie Sanders’ proposal from the 2016 Campaign and comes up with a cost of $36 Trillion.

Critically, this analysis is based on a policy that would eliminate all private health care expenditures:

“All health care costs would be covered by the federal government. Premiums, deductibles, copays and any other out-of-pocket costs would be completely paid for through federal funds.”

The Medicare-for-All proposal is pretty clear: scrap the entire healthcare system and replace it with (you guessed it) Medicare for all. Since Medicare-for-All would pay for “all health care costs,” if you want to know what the impact would be on people’s budgets, you have to look at the net, not the gross cost. How much money would it cost us less how much money would it save us. If you move from one apartment to another, the impact on your budget is the difference in the cost between them. Subtract one, add the other. It’s common sense, at least to everyone except Douglas Holtz-Eakin. He portrays the costs of Medicare for All as an addition to healthcare costs without accounting for the fact that all other costs would go away.

Why would he do that? There’s a really good reason (or a really bad one, depending on your point of view):

The Centers for Medicaid and Medicare Services (CMS) produces annual projected health care costs for the following ten years. The most recent estimate of those costs covers 2018-2027 and projects total costs of roughly $47 Trillion. Updating the forecast window to cover the years 2020-2029 (matching DHE’s window) and assuming that costs increase in 2028 and 2029 at the 10-year average rate produces a 10-year cost estimate of $52.5 Trillion.

This means that the net cost of switching to universal healthcare, according to CMS and Holtz-Eakin himself would be a net savings of over $16 Trillion over 10 years.

This savings alone is enough to cover all of the assumed costs of the GND under DHE’s low-end assumptions achieving all its goals and leaving a savings of $500 billion. It cuts the high-end estimate by more than half.

The bottom line is that for Holtz-Eakin’s estimates to be correct, people would have to continue to pay insurance premiums on non-existing policies and to buy medicines that they already have. This is absurd of course, but as the next section shows, at least he’s consistent. 

Guaranteed Employment: A World without Walmart?

This is the other big-ticket item in DHE’s cost estimate. Like universal health care, the projected cost of guaranteed employment relies on estimates prepared by others. Once again, since he didn’t have the chance to manipulate the estimates themselves, DHE has to resort to some indefensible tactics to make the numbers look as scary as possible. In this case, he starts with a relatively small number and throws in a couple of assumptions that inflate the estimate by over 650%. Unfortunately for him, the assumptions he makes describe a future that is literally impossible to believe in.

DHE’s estimates of the Guaranteed Employment provision are based on estimates published by the Center on Budget and Policy Priorities. That estimate is based on a policy offering employment to unemployed job seekers and underemployed workers, offering them a minimum wage rate of $11.83 plus benefits. The original study estimates the costs of bringing the U-6 unemployment rate (currently 7.2%) to 1.5%, and finds that it would cost about $543 Billion per year. DHE updates this for inflation and moves the base year estimate to 2020 from 2019 arriving at a low-cost scenario of $569 Billion in 2020 and $6.8 Trillion over 10 years. 

So far so good. $6.8 Trillion is a lot of money, and you can argue whether or not you think it’s worth it. But why have an honest discussion when you can grossly inflate the numbers? Holtz-Eakin clearly wants this to look as expensive as he possibly can, so he does performs economics at its worst: make assumptions until you get the answer you like.

First, DHE assumes that the availability of guaranteed employment increases the labor force participation rate among prime-age workers to its most recent peak (reached in 2007). This increases the 10-year cost to $7.4 Trillion. 

Second, he increases that cost by a factor of 4 by further assuming that anyone who has a job that pays less than $11.83 quits and goes to work for the government instead, increasing costs to $31.8 Trillion.

Third, he adds the assumption that anyone earning less than the projected average wage under the program of $15.62 would also quit and take employment under the guarantee. This gets him to $44.6 Trillion over ten years.

The absurdity of these assumptions becomes clear with just a little bit of data and clear thinking.

First, the current effective minimum wage is already $11.80 once you factor in the existing federal rate as well as all of the workers living under higher state and local minimums. If the average low-wage worker can find a job for $11.80 an hour and people still stay out of the labor force despite that fact, it’s incumbent on DHE to provide some justification of why nearly a million such workers would choose to jump back in for $11.83. Keep in mind that the U-6 unemployment rate already includes the underemployed and discouraged job seekers. By definition, anyone still out of the labor force after the guarantee takes care of U-6 unemployment is out for a reason other than not being able to find a job.

Next, consider the fact that Walmart just raised its entry-level wage to $11 per hour. DHE is assuming that every single entry-level Walmart worker would switch jobs for an additional 83 cents per hour. And since their average wage is $14.25 per hour (not including benefits), many of their more experienced workers would also quit and take the guaranteed job. This seems unlikely, but its not categorically impossible.

But in order to believe DHE, we also have to believe that Walmart and every other private employer in the country would allow every employee earning less than $15.62 to quit and never replace them. Every Walmart, fast-food restaurant, clothing store, and other retail establishment would have no entry- to mid-level employees in them ever. Under DHE’s assumptions, nearly 60 million workers would leave their jobs and those positions would never be filled again at any wage. 40% of all private sector jobs in the U.S. would remain permanently vacant.

This is absurd, of course, and just demonstrates the lengths that GND opponents will go to come up with as high a cost estimate as they possibly can. In reality, a scenario like this would force employers like Walmart to compete on wages to attract workers, which is part of the point; GND supporters think the current federal minimum wage is too low. This would raise the floor.

These estimates are logically impossible and mathematically impossible as well. Because to add insult to injury, DHE assumed that the government would spend $10,000 on benefits for every worker under the program, including the cost of health care. That thing they are already getting under the universal healthcare provision. DHE has managed to achieve the rare feat of actually triple-counting some health care costs. It's a degree of difficulty that even the most dishonest economic consultants decline to attempt, at least in public.

Rather than twisting ourselves into logical knots of impossibility, if we instead assume that the universal job guarantee would work something like existing unemployment insurance programs, i.e. providing benefits only to workers who don’t currently have a job, who lose their existing jobs involuntarily, and actually wish to work, the costs are much closer to the original estimate of $6.2 Trillion rather than the $44.6 Trillion headline.

Summing up:

DHE’s assumptions are obviously intended to scare rather than enlighten. For them to be remotely accurate, one must be willing to ignore reality on the one hand and to believe in an impossible reality on the other.

Rolling back all these silly assumptions reduces the low-cost estimate to a net savings of about $2 trillion and the high-cost estimate to a cost of $2 trillion, less than the cost of the recent Trump tax cuts. All while delivering jobs and health care to every person who wants it, achieving a low-carbon electricity system, a net-zero carbon transportation system, and guaranteeing housing and food security to all Americans. If Trump really wanted to live up to his campaign promises of fixing health care and creating jobs for working class families, the GND might actually be a great place to start.