How carbon emissions will actually be reduced

Matthew Yglesias presents his take on how we can actually reduce carbon emissions:

Long story short -- futuristic electric cars? Good. Futuristic clean electricity sources and smart grids? Also good. But the killer ap is still reduced consumption. We have proven, longstanding technology that drastically reduces carbon emissions. To wit -- walking, biking, bus, trolley, light rail, metro, commuter rail, high-speed intercity rail. Unlike plug-in vehicles, there's nothing speculative about this technology -- we know that it works, it just costs money. And there's considerable reason to believe that investments in non-highway transportation infrastructure combined with a regulatory structure designed to encourage high-density development near key nodes would enhance economic growth rather than detract from it.

Not quite. There are two fundamental limiting factors in the shift to an economy with lower carbon emissions: technology and time. Yglesias is right to point out that futuristic technology will take time to research and implement, and that we have many existing technologies that could theoretically reduce our emissions right now. But he fails to grasp how long it will take for the infrastructure shifts he suggests to become a reality, and how even an unprecedented pace of change would fail to bring about substantial drops in emissions.

Right now, rail accounts for less than 1% of passenger transportation nationwide. As I've discussed, it's unlikely that intercity rail will become viable outside a few corridors anytime soon; most US cities are simply too far apart. Meanwhile, expansions in urban rail will require decades to start making a real dent in transportation. Consider my home city, Portland, which has pursued alternative transportation more aggressively than any other midsized urban area in the country. Since 1986, it has developed a model light rail system, the MAX, currently the fifth most ridden system in the nation. The net effect? The percentage of Portland-area commuters taking mass transit to work has actually fallen, from 9.8% to 7.6%. Yes, the low price of gasoline over the past few decades hasn't provided a large incentive to take the train, but low usage isn't really the issue; the MAX lines that do exist are quite popular. The real problem is that it takes a long time to create a light rail infrastructure that reaches a significant fraction of the population. Even incremental improvements demand years of planning and construction.

For the sake of argument, let's say that my pessimism is misplaced, and that passenger rail increases to 10% of all travel over the next 20 years. This is utterly unrealistic, demanding an improvement by a factor of 12, but it's still useful as an upper bound. With the existing mix of electricity generation, rail emits about 40% less CO2 per passenger mile than average passenger cars. The 9% increase in passenger rail as a proportion of travel, therefore, will mean a drop of less than 4% in CO2 emissions per passenger mile. Since transportation accounts for about a third of total CO2 emissions, we end up with a 1% decline in total emissions. And remember: this assumes an almost unfathomable increase in the prevalence of rail.

What could change this analysis? Well, it's possible that electricity generation will become significantly less CO2 heavy, but 100% carbon-free electricity (quite unlikely) will only enlarge our calculated decline in emissions to 4%. Intercity rail could displace some air flight, but emissions per passenger mile from air travel aren't too different from those of passenger cars, and our figures barely budge. However you slice the numbers, even an unprecedented rise in rail transit will make very little difference.

In fact, the benefits are probably even slimmer. Since the average passenger car lasts less than a decade, the automobile fleet can change far more rapidly than infrastructure does. With the skyrocketing price of oil, it's quite plausible that in 20 years, the majority of cars purchased will have the mileage of a Toyota Prius. And the CO2 emissions per passenger mile of rail travel are actually higher than those of a Prius.

Significant changes to the character of a city, like the denser development Yglesias endorses, proceed even more slowly. It's possible that in 40 or 50 years, we'll have cities noticeably denser than the ones we have now. But this will only start to make a difference after many of the "futuristic" technologies that he downplays are in force. Combine this with the fact that transportation only accounts for about one third of carbon emissions, and it's clear that denser cities will not play a large role in confronting global warming anytime soon.

And buses? Trolley buses using overhead electricity emit less CO2 than cars per passenger mile, but motor buses (on average) actually emit more. Unfortunately, rapid expansion is much more plausible for motor bus service, where it appears to be environmentally counterproductive. I do think that electric buses may play an increased role over the next few decades (more than rail); as with trains, however, the emergence of more fuel-efficient cars will cut into and even eliminate electric buses' CO2 advantage.

How will we really reduce emissions? In the near term, I think it'll be through a combination of conservation and more fuel-efficient vehicles. These don't require slow, long-term changes in infrastructure. People can drive less, organize carpools or relocate closer to work almost immediately, as evinced by the sudden 4.3% year-over-year decrease in miles driven this March. New motor vehicles rapidly slide in and out of usage, making it easy for improvements in technology to saturate the market in just a decade or two. More energy-efficient air-conditioners, refrigerators, and light bulbs will also become commonplace more quickly than we can change our electricity generation mix, or radically alter the infrastructure in our cities.

Within 15 to 25 years, I think that the necessary shift in electricity generation will start to become a reality. Solar electricity has grown exponentially, and we're seeing promising developments in PV cells and thermal generation, but its current role is so small that even relatively dramatic expansion won't start making a difference for at least a decade or two.

In 30 to 50 years, solar will become dominant in electricity production. Relative to other emissions-free sources, its cost decline is easily the most compelling; if anything, nuclear is becoming more expensive. Why? Solar power is essentially a high-tech manufacturing problem, while the main cost of nuclear power is the massive construction expense. In our economy, high-tech manufacturing of all kinds is rapidly becoming cheaper and more advanced, while construction costs are relatively static. Although both kinds of power are capital-intensive, the capital costs are heading in very different directions.

By this point, plug-in hybrids (or even fully electric vehicles) should also be almost universal. The economics of solar power will steadily become more favorable, and at $400 a barrel -- or whatever ungodly price oil achieves within several decades -- consumption of gasoline will become harder and harder to sustain. Petroleum usage may be mostly limited to areas where no clear alternative is available, like jet fuel.

All in all, it's likely that we will be able to make a sizable dent in our carbon output in the next 50 years. As I hope I've made clear, however, the vast majority of this decrease is going to come from technological improvements, not any radical change in our urban or transportation infrastructure. While it might seem that "existing" technologies are quicker to implement, fundamental lifestyle and infrastructure changes are actually slower than technology to advance. And in the end, that's a critical difference.