In the face of economic catastrophe, yesterday’s controversial assertion has become today’s conventional economic wisdom. That lack of regulation is one root of the current depression is not only the view of liberals and moderates, but also of sensible conservatives. And the need for public investment to fight the depression is no longer in doubt either. There are really only two tools in the conventional economic toolbox to fight a depression: lower interest rates, and public investment. Given that real interest rates are close to zero, that doesn’t leave a whole lot of alternatives.

Most of the economists who predicted the crash, Nobel laureates Joseph Stiglitz, Robert Solow, and George Akerlof, not to mention Roubini, now support [PDF] really large scale public investment — $300 to 400 billion annually. Dean Baker and Paul Krugman suggest the right number may be more like $500 to 600 billion.

At same time, Joe Romm wrote an open letter to Jim Hansen indicating how much technology we have to deploy very quickly to meet the 350 ppm target Hansen has set. Below the fold I describe in detail how we could deploy existing technology to meet these goals, how much we’d have to spend on grants, and how much we’d have to spend on loans. All figures are taken from the spreadsheet Jon Rynn and I have put together on green scenarios.

The bottom line is that we have about $275 billion a year we could productively invest in a green transformation, rising over the course of 20 years to $475 billion, and then dropping down to $265 billion for a decade after the transformation was complete, to pay off the last of the “green debt.” Those subsidies would make up for any difference in cost between green energy and dirty energy.

Social benefits from such a transformation (aside from global warming reduction) would result in net economic growth compared to the scenario in which we continue to use dirty energy.

This scenario (unrealistically) assumes zero technical breakthroughs that lower costs. In practice, we might never need to exceed the $265 billion base investment.

In the spreadsheet mentioned, Jon and I outline various scenarios, with varying degrees of aggressiveness in efficiency investment and varying degrees of technical improvements (zero, modest, and significant).

We have technology now to reduce emissions by 95 percent or more. Properly designed public investment and regulation could drive deployment of most of this technology.

Improvements in buildings

Improvements in buildings could be driven by regulation, along with an “efficiency” utility structured somewhat along the lines of British “rates” to make compliance painless.

For residential buildings, the regulations would specify emissions per square foot and per resident. For commercial buildings, the regulations would specify emissions per square foot and per full-time employee equivalent; retail and industrial emissions would be dealt with differently (including buildings).

What needs to be taken from the idea of “rates” is that they are paid by building occupants, not building owners — unlike property taxes, and unlike conventional loans. Thus a local efficiency utility could, for example, finance installing insulation in a home or office — with an agreement from both the current owner and the current occupant (who might or might not be the same person) — that whoever occupied the building would make payments for that insulation. Such payments would be suspended when not occupied.

It removes tenant disincentives for agreeing to such arrangements because tenants only have liability as long they occupy the building. When they move, it is the next tenant’s problem. Similarly landlords would know the liability is their tenant’s, not their own. Even an owner-occupier would know that, unlike an addition to their mortgage, they don’t have to worry about covering this payment in the sale price if they sell their home or office. For various reasons, it is much easier for tenants or owner-occupiers alike to incur a new utility bill that replaces part of an old utility bill and results in a net savings rather than to borrow money to achieve the same savings. Jon Rynn has written about an arrangement not unlike this used to finance solar energy in Berkeley, Calif.

Lastly these utilities could be subsidized in two ways. One would be for the federal government to use its ability to borrow to secure capital for them and then lend them out on a cost basis (lending the money for slightly higher rate than it paid in borrowing to cover transaction costs and defaults). Where necessary we could further subsidize such utilities by requiring partial rather than full repayments.

There is no physical reason we can’t upgrade every existing residence with appropriate weather and duct sealing, roof, floor, window, duct, and pipe insulation. Added to this would be leak repairs, faucet aerators, kick-pedal sink controls, and water-saving shower heads. Lighting, and large-scale water and electrical appliances could be replaced with more efficient ones over the course of 15 or 20 years as they wear out (or sooner if savings justified doing so). As a final step along the way, we could replace existing space heaters and air conditioners with either solar equipment or ground source heat pumps. The total cost of would be about $2 trillion. Repaying this at 5 percent interest would be approx.$131 billion a year for 30 years or approx. $161 billion a year for 20 years. About two-thirds of this cost could be paid back by residents at a savings on their current utility bills. That would leave about $44 billion annually over 30 years or $54 billion over 20 years, which would not be paid back.

New buildings would not require subsidies, since they would be required to meet standards from the beginning, and the costs of making a new building efficient are trivial compared to the costs of retrofitting an existing one.

Like residential buildings, office buildings would have to meet standards per square foot and per full-time employee equivalent. Greening an office building pays back its costs in a number of ways. Unlike residential improvements, no subsidy other than the use of government borrowing power would be needed.

Greening small office buildings is similar to greening residences. In large office buildings, office spaces can be greened with window tints that let in light, but exclude glare, and cut down heat in the summer and then let in light, but exclude glare, and keep in heat in the winter. Lighting tubes can bring sunlight into interior offices without windows. Various types of lighting improvements, including better lights and the use of reflectors, as well as putting lighting intensity directly under the control of workers who have to live with its consequences. Like homes, offices can use more efficient appliances. These include computers, printers, and copiers. More and more document management systems are reducing the amount of printing done in offices, and the paper and energy use associated with printing. (Years after the myth of the paperless office was exploded, thanks to the “less paper” office, paper use in offices in the U.S. is actually dropping.) In addition, modern computerized control systems can spot waste in climate control systems, which are literally too complex for humans to keep adequate track of. Greening office buildings will cost $1.3 trillion over the course of 20 years, essentially all private. However increased [PDF] productivity [PDF] alone will pay back that cost to private industry, even before energy and maintenance savings are considered.


Often overlooked for fast energy-reduction potential, freight can be shipped less expensively per ton-mile now by rail than by truck. Consider an 80 percent, or better, energy savings, which factors the need to still use trucks to get freight to and from freight yards and the less direct routes by which freight travels over rail compared to highways. But there are three major limitations which keep us from shifting large amounts of freight from truck to rail. First, our current rail infrastructure is being used at close to its limit. Without major improvements in our rail system, we cannot even double what we currently carry let alone increase use by many times. Second, shipping freight by rail is slower than shipping it by truck. Third, our current rail system is less reliable than our trucking system — even with all the reliability problems trucking companies have.

All of this could be solved in by investing about $450 billion to upgrade our freight infrastructure. This would include double-tracking heavily used routes, installing a certain amount of new lines, or reviving closed lines. Most importantly it would involve electrifying the most heavily used routes. Although electrification doubles — at the very least — the energy efficiency of rail compared to electric diesel, that is actually the least important reason to electrify freight rail. Most importantly an electric freight car can travel faster than a diesel one. Electrification is the key to allowing rail to greatly increase ton-miles moved while maintaining both speed and reliability comparable to that of the trucking industry. Because this could be done comparatively quickly, I would say this should be done as straight grant over the course of six to 10 years. A lot of the cost of such electrification would be running to grid to places it currently does not exist. It has been suggested that since we will ultimately need more long-distance transmission, this could be combined with running a national grid over rail rights-of-way.

There are about $500 billion worth of local unfunded transit projects — rail, bike paths, sidewalks, bus systems — in the U.S. These won’t replace cars, but they will nibble around the edges of and perhaps at least slow the rate of growth of automobile use. Again this should be done as a straight grant over the course of 10 to 20 years.

Although various means may reduce automobile use, I doubt we will see the automobile disappear for a long time, if ever. Given how unsustainable biofuels for automobiles have proven to date, and how expensive hydrogen and hydrogen cars still are, that leaves battery electric cars and plug-in electric vehicles as the only alternatives available today. Highway-speed electric cars run from $20,000 two-seaters to $120,000 roadsters and are still produced in small craft-car quantities that cannot take full advantage of the economies of scale in mass production. It would be reasonable to guess that mass production would let us make five-passenger electric cars at the same price we produce two-person cars for in craft quantities.

Assuming we still have any major U.S. owned car companies a few months from now, one way to produce such vehicles would be a public/private partnership similar to the one that created the Volkswagen. We could follow Amory Lovins’ suggestion and have the government purchase the initial output both for federal fleets and for leases. Purchasing large numbers of electric vehicles would not only help lower their costs in general, but it also could help break the chicken-egg deadlock that keeps electric-car-quality battery prices high. I’m suggesting $500 billion total in investments and purchases, about half of which would be recovered in fleet gasoline savings, and lease payments, for a net cost of $250 billion.

For transport regulation, I’d suggest that as we invest in electric cars that we also require that all new automobiles meet efficiency standards both in terms of fuel and electricity consumption per vehicle mile. At least one-quarter of vehicles sold should be required to be all-electric, and all vehicles should initially be required to be able to plug in to a wall and travel at least the first 20 miles on electricity. This standard should have escalator clauses so that within a short period of time all new vehicles have a 50 mile pure electric range and then a 70 mile pure electric range.

For freight I’d suggest rules on overall consumption allowed per ton (not ton-mile shipped) so standards could be met either by shipping freight efficiently, or by manufacturing goods closer to use point and shipping them shorter distances. I would also require that all the broadband suppliers, who have received valuable public rights-of-way in return for not much, extend net-neutral, 5 gigabit or better service to all Americans, with a basic rate of $25 monthly or less. If they can’t make a profit offering this, we should use the federal government’s powers of eminent domain to buy their assets at a fair price and supply such broadband ourselves. Universal broadband could, among other things, encourage much more telecommuting, and allow more office workers to work at home at least some days a week.

In the long run, smart development would lead to more transit-friendly cities and towns and goods manufactured closer to where they were build. But this would be mainly an area-by-area local initiative. Though many of the policies outlined would contribute to this, especially the transit funding, ultimately it is something that could be carried out by cities, towns, counties, municipalities, and states.


There are some areas where rule-based regulation of the type I described can lead to improvements in industrial efficiency. For example since the majority of industrial energy consumption is used to power industrial boilers, we could set rules mandating a minimum ratio of heat delivered to heat input. We could also mandate efficiency standards for pumps and motors which consume most industrial electricity. Also, standards that mandated reductions in paper use, packaging, water, and soil erosion could produce indirect energy savings. And in the short run, given the immediate recession, these probably are where we should concentrate our efforts.

In the long run, and specifically once the current depression is over, the only way I know outside of these and perhaps a few other very specific regulations to cut industrial energy use is via energy pricing. Efficiency or emissions-reduction regulations for something as broadly produced as greenhouse gases, cannot simply specify a reduction. They have to specify emissions per what. And in industry that “what” is very hard to define, varying per industry, per sub-sector of that industry, and per process. Ultimately the answer is emissions per unit of economic output, but then it is very hard to tell what the indirect inputs are for many industries such as banking which appears to have very low emissions until you consider the projects banks finance. While there is a role for regulations in areas, ultimately the only thing we can do is put a price on emissions and let the businesses sort out where to cut.

Electric power

A low-carbon economy will be, to the greatest extent practical, an electric economy. We can’t avoid using some fossil fuels and biofuels as a basis for industrial and chemical processes, for small amounts of transport, and for backups when renewable sources fail. But we need to use electricity from direct solar or ground source geothermal whenever possible. We can save the small amounts of fossil fuels and biofuels for uses for which they are necessary and reduce their use to the minimum that really is sustainable. In the U.S. that would be about five quads of natural gas, and three to perhaps 11 quads of biofuels. (Though I suspect seven quads is the most U.S. biofuel we can produce sustainably.)

If we really are going to power most remaining transport, heating, and industry needs electrically, than we need to implement the most aggressive efficiency improvements that appear to be possible, an 80 percent reduction in energy input per unit of GDP, and, after looking at non-electrical potential usable renewable inputs, we need to double our electrical production.

We have to replace all existing U.S. coal-powered plants with renewable electricity production — along with most use of natural gas for electricity. The bottom line: With plausible major breakthroughs in storage and renewable production, this could cost as little $4 trillion over the course of 20 years. With more moderate improvements, this could cost $7 trillion over 20 years. With zero technical improvements, (unlikely, but a limiting case) this could cost $14 trillion over 20 years. (Figures for renewables would be even higher if we don’t make really aggressive improvements in efficiency.) If the full cost was paid for by taxpayers, that would run $200 billion to $700 billion annually for renewables, depending on degree of technical improvement. (The high end, representing zero improvement is highly unlikely.) But of course taxpayers don’t pay our electric bills now. There is no reason they should pay 100 percent of them in the future. And one point of all the investments in efficiency in the previous section is that they allow us to pay more per unit of electricity and still end up with a lower electricity bill.

Thus the main way to drive renewable energy is with regulations. We need to require all power to be 95 percent emission-free — compared to today’s grid — by 2030, with incremental requirements every five years (25 percent by 2015, 50 percent by 2020 and so on). We also need to require utilities to pay feed-in tariffs and to buy emission-free power at a minimum rate. For this to be a reasonable thing to do, we would have assume some of the burden of grid management — creating smart grids, building a true national grid with long-distance transmission, and financing storage.

According to Al Gore, the cost of a real national grid, including both long-distance transmission and smart-grid management, would run about $400 billion. If financed with 5 percent, 20-year bonds, that would cost around $32 billion per year.

Similarly, one of the more expensive aspects of renewables is storage, which also helps stabilize a grid. The cost of this would be $1.7 trillion. Financed over 30 years, this would cost around $111 billion per year.

In terms of a subsidy, we could pass a modified RTC that offers 1.9 cents per kWh for any source with emissions less than or equal to wind, that does not expire for 20 years, one that is a refundable tax credit, that is not limited to passive income, and that is available to all entities — for-profit, non-profit, and individual — that can produce power. If we ultimately reached 95 percent carbon-free energy, this could cost as much as $200 billion in the 20th year. This is another area where a price on emissions would encourage emissions-free development


Making the unrealistic assumption of zero technical breakthroughs in efficiency or renewable technology, the total cost of a complete transition to 95 percent (or better) emissions-free energy in the U.S. would be about $1.7 trillion annually, if financed at 5 percent over 30 years. (There is no reason a 20-year build-out could not be financed over 30 years. It would provide paybacks for at least that long, and many aspects such as transmission lines and railroads would last 40 years or longer.) From a social standpoint, total paybacks would be $600 billion a year more than this, meaning in the 20th year, the economy woul` grow $600 billion more per year net than without such investments (not counting global warming reductions, but only immediate and short-term social returns.) Energy costs only, not counting possible fossil-fuel price increases or any social costs, would be about 31 percent higher in this scenario than under a business-as usual-scenario. However the particular subsidies I projected start at around $275 billion annually, average to $365 billion a year for the first 20 years, and peak at $475 billion annually in the 20th year. They drop back to $275 billion a year in the 21st year, as the renewable industries mature and can get by without further subsidy. The other $275 billion continues for another 10 years to pay back the green debt. These not only overcome most of the bottlenecks to phasing out fossil fuels, but they also compensate for what would otherwise be increases in energy costs.

Last points

1) I’m considering population growth, but not economic growth per capita in this outline. That is because it does not pertain to the argument. Economic growth will require more energy, thus more investments in renewables and efficiency than in this scenario, but obviously will have more places to use this energy, and more social costs if this energy comes from fossil fuels. So total payback compared to investment is the same, both in saved fossil fuels and in paybacks in social costs. The ratio remains constant. You just get large numerators and denominators, if you recalculate

2) Because this post is about public investment and regulation, I concentrated mainly on this topic. But putting a price on emissions is not optional. To the extent public investment is more palatable than such pricing, it does allow it to be delayed. We can, if we have to completely eliminate emissions in the building and power generation sectors without such pricing, eliminate most emissions in transportation, and a significant percent even in industry. But there is no way, except via an emissions price, to capture most possible savings in industry. There are just too many efficiency means we don’t know about in advance. Similarly, even in transport it is really hard to see how to reduce emissions in shipping and air travel without an auctioned permit system.

There are technologies that work at the margins. Airplanes can substitute sustainable biofuels (to the extent they really are sustainable) for kerosene, and they can also fly lower and slower. But at minimum, even given all technology available to day, or likely to be developed in the next 20 years, we are probably going to have to reduce total air miles. The best way to do this is to stop expanding and subsidizing airports, to auction permits for total emissions from aviation, and to let the air industry sort out to what extent they can cut emissions while maintaining flights and to what extent they have to reduce service.

Similarly, ships can add filters to existing engines to reduce black carbon, use flying sails to save an average of ten percent of their fuel use, and run their ships off electricity rather than fuel while in port. To the extent that truly sustainable biofuel is available, they use that. They can also buy very expensive new ships that used advanced engines, hulls, and propellers to reduce emissions more drastically. Since a cargo ship is not a fast means of transport in any case, they can also travel a little slower to reduce emissions per ton-mile. And of course it is possible that more goods will be made locally, reducing the total need for shipping. Again, the best means of driving all this is to auction permits to the shipping industry and to let them and their customers sort the various trade-offs. Making sure port electrical service is available — on the scale cargo ships and passengers liners would need to turn off their engines in port — would be a good additional and relatively low-cost public investment.

3) Mostly I’ve been talking about new regulations and new public investments. But there are also a great deal of existing subsidies and regulations that could be modified or eliminated. For example, one form of emission-free electricity is to take waste heat from industrial processes and use it to produce kWh with no additional emissions. One big barrier to this is that anytime you ship power over a public line you are considered a public utility and regulated as such. Allowing combined heat and power (and for that matter small scale generation of renewable electricity) to cross public rights of way under simpler regulations would remove a major barrier to such development. (You can’t have zero regulation because you still want to make sure they don’t block existing utility lines or roadways, For those obstacles would make future maintenance of more difficult. A regulatory and approval process that deals only with issues like these could be structured so as not to be a barrier to such development.)

We give huge subsidies to the fossil-fuel and nuclear industries that could be eliminated or redirected to efficiency and clean energy. And we have huge agricultural subsidies that should be redirected to benefit sustainable practices that build soil instead of erode it and produce near zero run-off, zero toxic run-off, and low nutrient run-off.


We can phase out greenhouse gas emissions over the course of 20 years (paying for it over the course of 30) at a profit. We can gain increased economic growth that far exceeds additional energy expenditures. We can also publicly invest money that will overcome most barriers to this transformation and recoup the price difference between clean and dirty energy so that overall energy prices don’t rise as compared to a business-as usual-scenario. All of this can occur at a cost much less than the total measurable social gains from this green deal.

[Update] Dean Baker made a point to me that all this should only be considered a stimulus until the recession/depression ends, hopefully by 2011. So I guess after this is should just be considered a sound public investment rather than a stimulus.