Building Blocks to a World of Flat Emissions

Unsurprisingly, global emissions have flattened for a third year in a row. The fortunate outcome is composed of the following building blocks: the peak of emissions in the OECD in 2007;  the termination of a twenty-five year phase of heavy industrialization in China and its ability to convert coal growth to zero; and, the accelerated deployment of wind and solar.

Peering closely at the year-over-year changes in new electricity generation helps us understand better how global economic expansion can be pressing forward, without pushing emissions higher. In 2015, for example, nearly all of the marginal growth in global power generation was provided by 187 TWh of new wind and solar. In the years prior, combined wind and solar took minority shares of new generation, at 158 TWh in 2013 and 121 TWh in 2014—but the majority of new electricity ex-wind+solar was provided by natural gas. This trend, where natural gas replaces retiring coal, is set to continue through the end of the decade. But combined wind and solar will eventually go to work also, on natural gas: in 2016 and this year, 2017, they are already expected to take half of new power generation.

Climate scientists will remind you, however, that emissions actually need to enter outright decline. When might that happen? Well, emissions have to flatten first before they can decline. One way to answer this question is with another: when will 100% of marginal growth in global power generation be taken up by wind and solar (or other low carbon additions, like nuclear)?  If you look ahead to the rough schedule of global coal retirements, natural gas additions, and the shift in marginal growth in wind and solar—in China, the US, and especially India—a reasonable guess is sometime after the year 2020.

But that still leaves the problem of emissions growth from oil.

As natural gas kills coal, it creates an emissions-saving gain. But those emissions-saving gains will eventually tail off. An actual decline in emissions after 2020 will be reliant not only on wind and solar dominance of powergrid growth, but a concurrent marginal transition of transportation growth, typically oil-based, towards electricity.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

The Oil Demand Downshift

The International Energy Agency in Paris (IEA) released its oil outlook this week, Oil 2017, warning that demand may begin to outpace supply by 2020 if the global oil industry doesn’t start investing again.

But what if the oil industry is right to not invest? The most important data point in the IEA’s report was their forecast for oil demand growth: 1.2%, on average, over the next 5 years. If that forecast turns out to be accurate—and I think it will—then the next five years will see the same sluggish demand growth as the past five. And on a longer timeline, it represents a major downshift from the kind of demand growth the industry used to enjoy. | see: Compound Annual Growth Rate of Global Oil Demand During Five Year Periods 1985-2015 | Forecast 2016-2020.

The transition from an era when oil demand growth tended towards 2.00% to a time when it tends towards 1.00% is the driver behind the most lively argument in the oil sector today: peak oil demand. Just about every oil-focused energy analyst is trying to model when policy, urban demographic shifts, and the rollout of electric vehicles will take oil demand growth down to zero.

So what are the risks? Right now, you would have to locate the greatest upside risk to global oil demand in Asia—in particular, overall economic growth in China and India. While it would not be possible for China or India to replicate the 20th C experience with the automobile seen in the West, oil adoption is still underway in Asia. But downside risks to oil demand are real too. Indeed, if we bundle the lower operating costs of EV with their higher price, the discount that currently supports ICE vehicle sales is fading rapidly. Risk is growing that EV sales will take off, accelerate, and start to dominate many markets in a disruptive trajectory.

But there’s a larger point to be made in these equations. Oil demand growth at this lower bound has already put the price of oil, and the industry, in a vulnerable position. Remember what happened to the coal industry. After supersizing itself to the greatest heights of demand ever, between 2000 and 2013, global demand growth for coal turned flat. And now the coal industry in tatters.

The global oil industry doesn’t need to reach peak demand growth to run into severe problems. Once demand growth falls below 1.00% and stays there, the price of oil is going to find a new lower band as well, and it’s likely $10 away from current prices to the downside.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Car Clarities Begin to Appear: Bring on the EV

Many predictions are fielded, but few are chosen. With 2017 fully underway, it’s now possible to sober up a little about the future direction of cars.

First, Silicon Valley is not going to manufacture automobiles. Not EV, and not AV either. That the Valley would not enter the competitive, low-margin business of auto manufacturing was never a difficult forecast to make. That said, this excellent wrap-up of the current scene from John McElroy at Autoblog, The race for autonomous cars is over. Silicon Valley lost., succinctly makes a case for the the most rational, plausible outcome: the Valley will dominate the data around cars, leaving their production to those who know best. I’d only add an additional point of support to this conclusion: the example of Tesla is already serving as a teaching-story to the Valley. Musk’s success clarifies, interestingly, that there’s no need for technologists to follow or duplicate his auto-manufacturing pathway. This is especially true if you understand that Tesla is becoming a battery and energy storage company.

Second, it’s finally more widely understood that chatting about a future of Autonomous Vehicles is effortless, but creating that future will take time. Lots of time. Chris Mims over at the WSJ wrote a good update on the issue last September, Self-Driving Hype Doesn’t Reflect Reality. We are at the beginning, not the end point, of fully autonomous vehicle engineering. There is still a great deal of material to be worked out in mapping, vehicle negotiation of unmapped obstacles and hazards, and the short-cycle updating required to fully pair an AV to the built environment. And we will still have the Trolly Problem to confront when fully autonomous AV are ready to deploy. 

Finally, the EV are coming. Norway, (yes, a small country), is looking to ensure through tax incentives that EV will constitute 100% of new car sales by 2025. But Norway is already getting its EV on. In 2015, over 20% of new sales were EV. Meanwhile, in the large country category, China just announced that new entrants to Beijing’s taxi fleet would only be open to EV, starting this year. Also, Tesla has now done its first 1$ billion in sales, confirming China will be a major market for Musk’s company. Most important is that China is now consuming nearly half the total global sales of EV. For a country that still consumes half the world’s coal, that’s encouraging.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Not the Road Ahead

A broad claim made in favor of autonomous vehicles (AV) is that their eventual deployment will usher in a massive reduction in national vehicle fleets. It’s an appealing image: our cities emptied out of excess vehicle inventory, and lots of real estate and automobile infrastructure converted to human-centered uses.

In September of last year, the co-founder of Lyft, John Zimmer, wrote a long piece on Medium in which he laid out this very vision. The Road Ahead has the familiar appeal which prosaic, silicon valley forecasts are very good at producing. Autonomous vehicles will quickly become widespread, and account for the majority of Lyft rides in five years. By 2025, private car ownership will all but end in major cities. There’s no question that AV are coming (though not nearly as quickly as many forecast). And I agree that both EV and AV will find their high adoption rates first, in global cities. But something about Zimmer’s precision struck me as too confident. And in particular, his accounting for the number of citizens that would shed car ownership. The idea that cheap and available AV would only usher in car-shedding struck me as one-sided.

This week, in a detailed report on the rise of TNC (Transport Network Companies) in New York City, we have a better picture of the neglected side of the AV rollout equation. According to Unsustainable, a detailed ridership and impact study conducted by Schaller Consulting, the success of Uber, Lyft, Via, Gett, and Juno in the five boroughs has uncovered demand not from riders looking to shed car ownership, but from transit users looking to take a cheap car ride. According to Schaller, the marginal growth among competing transport options in New York City began to tilt slowly towards TNC in 2014, and then really accelerated in 2015. So much so, that marginal growth in bus and subway ridership slowed considerably. And by 2016, TNC were trouncing, at the margin, both buses and subways.

| see: Figure 12. Changes in ridership by mode, 2015 to 2016.


I agree that AV penetration will impact the inventory of existing cars. But readily available and affordable AV will also impact the inventory of users who don’t own cars. You could argue that New York City contains the largest population of car-less users in the country. But you would also have to allow that latent demand for car services can be found in all cities; and if prices for TNC fall—regardless if fleets are AV, EV, or still have drivers—the impact of those falling prices will uncover growth in this latent demand. It’s already happening, in New York City.

The advent of AV and EV will offer significant gains to society and well being. But the claim that urban congestion will be solved by AV in revolutionary fashion needs to account for the phenomenon of induced demand, and cannot rely on one-sided arguments that concentrate only the size of existing vehicle fleets.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Troubles Ahead for the Global Oil Sector

Global production of crude oil reached a new, all time high in November of 81.81 mbpd. That’s not good news. Especially at a time when the rate of demand growth for oil has downshifted from its historical 2% rate to something closer to 1%. More chilling still is that the price crash, which started in mid 2014 and took prices from a seemingly stable $100 level down to today, did not produce a supply response. No, instead, global oil production is up 4 mbpd (using the monthly data) since the start of 2014. That’s just grim for producers everywhere. Or rather, producers outside of low-cost OPEC.

Equally astonishing: according to the latest data, global crude oil production in 2016 will likely average right around 80.4 mbpd, a statistical match to 2015’s 80.5 mbpd. So even in 2016, the global oil complex was still making no response whatsoever to lower prices. Why haven’t lower prices driven supply lower?

The explanation is fairly straightforward: the rate of global oil demand growth is now slow enough that it can be adequately covered by the rate of newly developed oil brought to market. This equilibrium relieves both the futures market, and the oil extraction industry, of pressure.

In the latest issue of the TerraJoule newsletter, I discuss briefly The Hunt for Peak Oil Demand. Get the latest free, bi-monthly by signing up here: newsletter.

–Gregor Macdonald

Wind and Solar Fly to New Highs in the United States

Full year 2016 data just released from EIA shows that combined wind and solar reached 7% of total US electricity generation. In a system that produced a total of 4096 TWh, combined wind and solar provided 283 TWh. Breaking that down a bit further, wind provided 226.5 TWh; utility scale solar (both thermal and PV) provided 36.75 TWh; and small scale, distributed solar provided 19.5 TWh. That last figure is the one to watch. In a system that’s been plodding along without growth for a decade, producing 4000 TWh year after year, small scale solar has become an important force at the margin.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Photo: Occidental College solar array, Los Angeles, by Gregor Macdonald

Slow Growth in China Still Driving the Shape of Global Emissions

Slow growth in China continues to drive global emissions growth towards a flatline, leaving the gap open for renewables to exploit. According to the China Electricity Council (CEC), electricity generation grew 5% in 2016 to 5920 TWh with the combined contribution of new capacity in hydro, wind, and solar eclipsing the contribution of new capacity in coal.

However, for China’s electricity generation to have grown by a full 5% last year to 5920 TWh, it will require a downward revision in 2015’s generation total from 5811 TWh (the latest according to BP Statistical Review) to 5640 TWh. Yes, China’s energy data has become, in recent years, notoriously volatile. But if we combine the existing data set from BP with the latest CEC data, we find that China’s electricity generation grew a more reasonable 1.9% last year—a figure more in line with overall, economic readings.

For China to have achieved 5% growth in power generation last year, the downward revision to 2015’s generation will be large enough to record a significant decline in growth, from the previous year (2014). The implication is pretty clear: volatility in the baseline at this juncture does not change the overall picture, that China’s economy—which runs mainly on electricity—is moving along at a conventional rate.

But there’s a final data point buried in the recent data release from CEC, that strengthens the slow(er) growth thesis in China. The utilization rate of China’s thermal power generation capacity (coal, natural gas, and oil) actually fell, to the lowest level since 1964. This confirms a broader, more important trend that we’ve known about China for some years now: it’s powergrid system is a sprawling, stuffed with coal plants it no longer needs, and has already reached overcapacity.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Coal Jobs and the Power of a False Premise

Employment in America’s coal industry collapsed not five years ago, but fifty years ago. Since that time, President Nixon’s establishment of the EPA, the steady market share gains of natural gas in the US power generation mix, and the shift to machine-based, open pit coal extraction in the West have helped to ensure that growth in the US coal job market has been nil. The long-term, historical picture is captured well in a US Department of Labor MHSA series that starts in the year 1900:

There have been fluctuations, however. China’s choice of coal singlehandedly resurrected the entire global coal industry from its chronic, moribund state—for about 15 years, between 1995 and 2010. That blip in time revived, very briefly, US coal mining employment by a scant 17,000 jobs from a low in 2003 of 70,000 to a high of 87,000 in 2011. This shorter time frame is best seen starting in 1985, from the US Department of Labor’s CES series:

Depending on which data series you use, 2016 coal mining employment was running at 81,000 jobs (MHSA) or 50,000 (BLS/CES). But it doesn’t really matter. On the longer timeline, US coal employment has literally been decimated, falling from a high of over 800,000 jobs in the 1920’s,** to today.

Now comes the Trump Administration promising a second resurrection of America’s coal industry; not by capturing natural demand, but by changing policy. Here, like so many claims from the new administration, the premise is false. Markets and technology, not policy, are what ails the US coal industry because the US coal industry like the US oil industry is a feature of a global market. And in today’s global market, solar, wind, and natural gas are trouncing coal for new, marginal additions in power generation in every important domain. Especially China.

Sadly, the Trump administration wants to sell their false premise to the small, remaining group of American workers connected to the coal industry. The humane, and rational policy would be to design the best ever job-retraining program for these workers, and to roll it out on their behalf. Such a program could be further supported by new infrastructure spending not only in the few states where coal extraction still takes place, but in those states where coal combustion, and coal fired power plants (many 40-50 years old), are retiring.

Instead, using coal miners as props, the administration has enacted a series of social signaling legislation, like the repeal of the Stream Protection Rule. This chimera has been covered well by Brad Plumer at Vox, and also by Jenna Johnson and Dave Weigel at The Washington Post. But the most basic point is not hard to understand: you can save maybe several hundred jobs, but, you cannot save 70,000 coal jobs through the adjustment of small scale regulation, as claimed by the coal industry, when only 50-80,000 jobs remain.

Where are all new the jobs today? In wind, solar, and natural gas. To illuminate the problem for coal further, see the relentless market share gains that natural gas has made, over the past 20 years, in US power generation. In a country that’s been generating roughly 4000 TWh per year of electricity, without much growth for a decade, natural gas has charged onto the scene and reached 1,400 TWh:

Coal remains a cheap and powerful energy source globally but, unfortunately for coal miners everywhere, extraction and combustion costs, construction timelines to erect new coal infrastructure, and the associated supply chain have made new coal increasingly uneconomic. Finally, path dependency—which at one time worked in coal’s favor—now works against it. Many coal plants in the US and the West came up for retirement just as natural gas, wind, and solar were getting cheap. Alas, coal has missed out on its market share opportunity for decades to come. Another reason to do the humane thing: offering robust job re-training to workers connected to America’s (long) dying coal industry.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

**Correction: data points in the first chart between 1930 and 1960 were corrected per the linked data set, to fix a data transposition error. The resulting chart does not change the story, but does indeed remove the spike/crash in the pre-corrected version, and offers a smoother portrait of the decline in coal jobs between 1930 and 1960.

The Next Emissions Story

If you’ve been following the emissions story the past few years, you will know that slower economic activity and the rise of renewables have started to seriously blunt the growth of global carbon. In the OECD, for example, CO2 output has surely peaked, with virtually no risk whatsoever that developed nations will return to the carbon output highs of last decade. Globally, emissions growth started to stall out 2-3 years ago. Because, even in the Non-OECD, China has begun to thwart marginal growth of fossil fuel consumption through a revolutionary buildout of wind and solar power. (see: Are We Finally Seeing Good Climate News from the US and China?)

The next emissions story therefore will not be about the growth of CO2, but rather, the difficulty in engineering a new phase of carbon output decline. Like the cyclist behind a belching bus (photo above), an encouraging adoption phase globally of wind, solar, energy efficiency, walkable urban space—and yes, cycling—is now running up against legacy, fossil fuel systems. If you are a climate scientist, the risk is now shifting, therefore, from the next phase of carbon growth to a long standoff between marginal growth from clean sources and fossil fuel dependency.

A better conversation about these risks can benefit, therefore, from a delineation between adoption and dependency. All too often, fluctuations in energy usage in dependent systems are mis-read as growth (or adoption). Oil use in the United States is a perfect example. Let me explain.

The adoption phase of oil—the uninterrupted sequence of years in which the economy required ever higher volumes of oil to grow GDP—ended roughly 40 years ago in the United States. Oil consumption last year, of 35.89 quadrillion btu, was first reached 40 years ago when, in 1976, US oil consumption crossed above 35 quadrillion btu for the first time. In the last 40 years, the US population has grown from roughly 220 million to 330 million people. GDP (nominal) has grown roughly 10X, from 1,877 T to 18,567 T. Ergo, oil use per capita, and, per unit of economic growth, has been in steady decline for four decades.

But a new phase began in US oil consumption starting four decades ago: the dependency phase. Understanding this dependency phase is crucial to a broader framework-understanding of where we are today in global coal dependency, oil dependency, and natural gas dependency.

Consider, for example, the chronic discussion of peak car. Animated by data points on per capita auto ownership, miles driven, and gasoline consumption in the United States, the conversation appears to be teetering always to a question of whether the next new growth phase is coming in the American automobile complex. But growth in the US automobile complex is largely off the table. Dependency, and the failure of the US to invest more aggressively in public transport (like the rest of the world) is the real question, the real problem.

The qualities of the peak car discussion can be found also in the discussion about peak coal. Because wind and solar are doing so much work now at the margin, and because solar costs are falling so dramatically, the risk of a new increase in global coal consumption has crashed. Coal growth is over, full stop. And yet, many climate models still embed a risk of future coal growth, in part, because we are too close in time to coal’s resurrection (1995-2010) when coal and coal infrastructure was cheaper, and easier to adopt, than renewables. But that era is now behind us as both China, and now India, see that solar is better, cheaper—and most important of all—faster. (see: India Will Crush You with Solar, 13 February newsletter).

The probability is high, very high, that global emissions have now peaked. Indeed, if the world puts together a few more years of flat emissions, there will be a temptation to declare an initial, first-goal victory. And that’s fine. The next goal in global emissions—covering 100% of marginal energy consumption growth, while at the same time getting to work on existing fossil fuel dependency—will be harder, or perhaps more pointedly, slower to achieve.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

The Large Scale Force of Small Scale Solar

Soon, you will learn that US solar growth doubled last year, as the country added a gargantuan 14.6 GW of new capacity. The utility scale sector saw the highest growth rate in 2016, and that has caused some small consternation among analysts who observe a concurrent slowdown in the rate of new rooftop solar, in the residential market. But there is little to worry about in the performance of non-utility scale solar for just a single year. Why? Because small scale scale solar is now a major force, cumulatively, in US renewable generation.

According to EIA, small scale solar (typically referred to as distributed solar) generated 11.23 TWh compared to total solar generation of 28.92 TWh in 2014, or 38%. In 2015, small scale solar generated 14.94 TWh compared to total solar generation of 39.03 TWh, in 2015—again, a 38% share. In 2016, according to EIA (whose data tends to lag), small scale solar is on pace to generate about 19.94 TWh in estimated total solar generation of 56.85 TWh, a 35% share. And given the more updated data now emerging on last year’s installation of capacity, these generation totals may be revised higher.

A larger point is worth making here, however. The US electricity system has been running along for a decade now, generating roughly 4000 TWh per year, without any growth or decline. This data point alone is often marshaled to make the case that economic growth itself is on the wane in the US; and it’s certainly the case growth overall is running at a new normal, around 1%, in the developed world. But it’s also the case that efficiency is driving a large part of this phenomenon. Energy capture—the kind enabled by wind and solar—is alot more efficient than energy combustion. Meanwhile, the US power generation fleet continues to turnover as new plants with the latest turbines (especially in natural gas) replace old coal plants. Finally, a quiet revolution continues to take place in US building construction, driving further efficiencies. So yes it’s now true: US economic growth has quite effectively decoupled from “growth” in energy consumption.

–Gregor Macdonald

Get the free, bi-monthly newsletter.

Photo: Duxbury Beach, Massachusetts, by Gregor Macdonald